<%BANNER%>

Wildlife and Habitat Responses to Prescribed Burning, Roller Chopping, and Grazing of Florida Rangeland and Pasture.

Permanent Link: http://ufdc.ufl.edu/UFE0041538/00001

Material Information

Title: Wildlife and Habitat Responses to Prescribed Burning, Roller Chopping, and Grazing of Florida Rangeland and Pasture.
Physical Description: 1 online resource (218 p.)
Language: english
Creator: Willcox, Emma
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: abundance, arthropods, avifauna, chopping, fire, pasture, rangeland, richness, season, vegetation
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In Florida, prescribed burning and roller chopping are management practices widely promoted under assistance-based programs to improve rangeland livestock forage and wildlife habitat conditions. In addition, many assistance-based programs in Florida also promote management activities that intend to improve monoculture and mixed pasture habitats for the benefit of wildlife. However, there is a lack of information concerning the impacts of these treatments and little is known of the role grazing lands play in providing habitat for wildlife, impeding management and conservation efforts. This study was designed to fill knowledge gaps regarding the response of native rangeland (pine flatwoods) vegetative, avian, and invertebrate communities to prescribed burning and roller chopping. It also examined vegetation and bird community responses to grazing of monoculture and mixed pastures to provide information of use in tailoring management programs to sites where avian conservation is a priority. Roller chopping, particularly during the growing season, was effective at reducing shrub cover, height, and density in pine flatwoods. However, burning and roller chopping practices frequently resulted in decreases in herbaceous vegetation. Burning and roller chopping treatments led to reductions in arthropod familial richness. Total arthropod abundance was lower on all but growing season roller chop sites. Depending on treatment and season of application, reductions in familial richness and abundance were observed in a variety of individual arthropod orders. Dormant season burning resulted in decreases in non-breeding, overwintering avian species richness and abundance. However, growing season burning resulted in increases in species richness and abundance for this avian group. Growing season roller chopping resulted in increases in abundance of permanent resident and breeding avian species. On monoculture pasture, an increase in grazing intensity led to declines in total avian species richness and short-distance migrant, neo-tropical migrant, and permanent resident avian species richness and abundance. Declines in total avian species richness and neo-tropical migrant avian species richness and abundance were observed on mixed pastures subject to increasing grazing intensity. However, short-distance migrant and urban avian species richness and grassland avian abundance increased on this pasture type in the presence of grazing.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Emma Willcox.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Giuliano, William M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041538:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041538/00001

Material Information

Title: Wildlife and Habitat Responses to Prescribed Burning, Roller Chopping, and Grazing of Florida Rangeland and Pasture.
Physical Description: 1 online resource (218 p.)
Language: english
Creator: Willcox, Emma
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: abundance, arthropods, avifauna, chopping, fire, pasture, rangeland, richness, season, vegetation
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre: Wildlife Ecology and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: In Florida, prescribed burning and roller chopping are management practices widely promoted under assistance-based programs to improve rangeland livestock forage and wildlife habitat conditions. In addition, many assistance-based programs in Florida also promote management activities that intend to improve monoculture and mixed pasture habitats for the benefit of wildlife. However, there is a lack of information concerning the impacts of these treatments and little is known of the role grazing lands play in providing habitat for wildlife, impeding management and conservation efforts. This study was designed to fill knowledge gaps regarding the response of native rangeland (pine flatwoods) vegetative, avian, and invertebrate communities to prescribed burning and roller chopping. It also examined vegetation and bird community responses to grazing of monoculture and mixed pastures to provide information of use in tailoring management programs to sites where avian conservation is a priority. Roller chopping, particularly during the growing season, was effective at reducing shrub cover, height, and density in pine flatwoods. However, burning and roller chopping practices frequently resulted in decreases in herbaceous vegetation. Burning and roller chopping treatments led to reductions in arthropod familial richness. Total arthropod abundance was lower on all but growing season roller chop sites. Depending on treatment and season of application, reductions in familial richness and abundance were observed in a variety of individual arthropod orders. Dormant season burning resulted in decreases in non-breeding, overwintering avian species richness and abundance. However, growing season burning resulted in increases in species richness and abundance for this avian group. Growing season roller chopping resulted in increases in abundance of permanent resident and breeding avian species. On monoculture pasture, an increase in grazing intensity led to declines in total avian species richness and short-distance migrant, neo-tropical migrant, and permanent resident avian species richness and abundance. Declines in total avian species richness and neo-tropical migrant avian species richness and abundance were observed on mixed pastures subject to increasing grazing intensity. However, short-distance migrant and urban avian species richness and grassland avian abundance increased on this pasture type in the presence of grazing.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Emma Willcox.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Giuliano, William M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041538:00001


This item has the following downloads:


Full Text

PAGE 1

1 WILDLIFE AND HABITAT RESPONSES TO PRESCRIBED BURNING, ROLLER CHOPPING AND GRAZING OF F LORIDA RANGELAND AND PASTURE By EMMA V. WILLCOX A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIA L FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Emma V. Willcox

PAGE 3

3 To my Mum and Dad

PAGE 4

4 ACKNOWLEDGMENTS I express my deepest appreciation to my commit tee chair and mentor Dr William Giuliano. His encoura gement, guidance, and persistent help made this dissertation possible. I am also grateful to Dr Jaret Daniels, James Selph, Dr George Tanner and Nicholas Wiley for guiding my research over the past s everal years and helping me develop a strong background in my chosen field. In addition, I thank Dr. John Hayes and Dr. Mel Sunquist for their assistance w hen completing my degree. I am extremely appreciative for the assist ance of my research technicians: Dixie Cline, Mary Hobby, Courtney Hooker, Spencer Ingley, Karen Ridener, and Christine Sciarrino. Their hard work, friendship, and support were invaluable. To the landowners and managers who allowed me access to their lands to conduct research I am indeb ted. Their participation was critical to the completion of this study. For the love and support of my parents, I am ever thankful. They instilled in me many important qualities and taught me the importance of hard work self respect persistence, and independence. It is their faith in my abilities that allowed me to achieve this important milestone in my life. My thanks go also to my parents in law, Joanne and Dick, who have welcomed me into their family and provided me with unwavering encouragement and support. Last, but my no means least, I would like to thank my husband and best friend Adam Willcox. His love, patience, and unwavering belief in me have provided the inspiration I needed to complete this journey.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 11 ABSTRACT ................................................................................................................... 12 CHAPTER 1 INTRODUCTION .................................................................................................... 14 2 WILDLIFE HABITAT EFFECTS OF PRESCRIBED BURNING AND ROLLER CHOPPING IN PINE FLA TWOODS ....................................................................... 19 Introduction ............................................................................................................. 19 Methods .................................................................................................................. 21 Study Sites ....................................................................................................... 22 Treatment Types .............................................................................................. 23 Habitat Sampling .............................................................................................. 23 Analyses ........................................................................................................... 25 Results .................................................................................................................... 26 Dormant Season Burn ...................................................................................... 26 Growing Season Burn ...................................................................................... 28 Dormant Season Roller Chop ........................................................................... 30 Growing Season Roller Chop ........................................................................... 32 Roller Chop/Burn .............................................................................................. 33 Treatment Type Comparisons .......................................................................... 35 Discussion .............................................................................................................. 36 Management Implications ....................................................................................... 44 3 SEASONAL EFFECTS OF PRESCRIBED BURNING AND ROLLER CHOPPING ON SAW PALMETTO IN FLATWOODS ............................................. 57 Introduction ............................................................................................................. 57 Methods .................................................................................................................. 60 Study Sites ....................................................................................................... 60 Treatment Types .............................................................................................. 61 Saw Palmetto Sampling ................................................................................... 61 Analyses ........................................................................................................... 62 Results .................................................................................................................... 63 Dormant Season Burn ...................................................................................... 63 Growing Season Burn ...................................................................................... 64 Dormant Season Roller Chop ........................................................................... 64

PAGE 6

6 Growing Season Roller Chop ........................................................................... 64 Roller Chop/Burn .............................................................................................. 65 Treatment Type Comparisons .......................................................................... 65 Discussion .............................................................................................................. 66 Management Implications ....................................................................................... 69 4 INFLUENCE OF ROLLER CHOPPING AND BURNING ON ARTHROPOD CO MMUNITIES OF FLORIDA RANGELANDS ...................................................... 73 Introduction ............................................................................................................. 73 Methods .................................................................................................................. 75 St udy Sites ....................................................................................................... 75 Treatment Types .............................................................................................. 76 Arthropod Sampling .......................................................................................... 76 Habitat Sa mpling .............................................................................................. 78 Analyses ........................................................................................................... 79 Results .................................................................................................................... 81 Dormant Season Roller Chop ........................................................................... 81 Growing Season Roller Chop ........................................................................... 83 Dormant Season Burn ...................................................................................... 84 Growing Seas on Burn ...................................................................................... 85 Roller Chop/Burn .............................................................................................. 86 Treatment Type Comparisons .......................................................................... 87 Art hropodHabitat Relationships ....................................................................... 89 Discussion .............................................................................................................. 90 Managment Implications ......................................................................................... 94 5 EFFECTS OF PRESCRIBED BURNING AND ROLLER CHOPPING ON AVIAN COMMUNITIES AND THEIR HABITAT ASSOCIATIONS IN FLATWOODS ........ 105 Introduction ........................................................................................................... 105 Methods ................................................................................................................ 108 Study Sites ..................................................................................................... 108 Treatment Types ............................................................................................ 109 Bird Surveys ................................................................................................... 109 Habitat Sampling ............................................................................................ 111 Analyses ......................................................................................................... 114 Results .................................................................................................................. 115 Dormant Season Burn .................................................................................... 115 Growing Season Burn .................................................................................... 116 Dormant Season Roller Chop ......................................................................... 117 Growing Season Roller Chop ......................................................................... 118 Roller Chop/Burn ............................................................................................ 119 Treatment Type Comparisons ........................................................................ 120 Avian Habitat Relationships ........................................................................... 121 Discussion ............................................................................................................ 122 Management Implications ..................................................................................... 127

PAGE 7

7 6 DIURNAL LEPIDOPTERAN RESPONSE TO PRESCRIBED BURNING AND ROLLER CHOPPING IN FLORIDA FLATWOODS ............................................... 143 Introduction ........................................................................................................... 143 Methods ................................................................................................................ 146 Study Sites ..................................................................................................... 146 Tr eatment Types ............................................................................................ 147 Lepidopteran Surveys .................................................................................... 147 Nectar Producing Plant Sampling ................................................................... 148 Habitat Sampling ............................................................................................ 148 Analyses ......................................................................................................... 149 Results .................................................................................................................. 151 Lepidopteran Species Richness and Abundance ........................................... 151 Flowering Forb and Shrub Species Richness and Abundance ....................... 151 LepidopteraHabitat Rel ationships ................................................................. 152 Discussion ............................................................................................................ 153 Management Implications ..................................................................................... 155 7 AVIAN COMMUNITY RESPONSE TO GRAZING INTENSITY ON MONOCULTURE AND MIXED PASTURES ......................................................... 158 Introduction ........................................................................................................... 158 Methods ................................................................................................................ 160 Study Sites ..................................................................................................... 160 Vegetation Sampling ...................................................................................... 162 Bird Surveys ................................................................................................... 163 Analyses ......................................................................................................... 164 Results .................................................................................................................. 166 Vegetation ...................................................................................................... 166 Avian Abundance and Species Richness ....................................................... 168 Avian Habitat Relationships ........................................................................... 172 Discussion ............................................................................................................ 173 Vegetation ...................................................................................................... 173 Avian Abundance and Species Richness ....................................................... 176 Management Implications ..................................................................................... 180 8 CONCLUSIONS ................................................................................................... 199 LIST OF REFERENCES ............................................................................................. 201 BIOGRAPHICAL SKETCH .......................................................................................... 218

PAGE 8

8 LIST OF TABLES Table page 2 1 Effects of treatment on habitat characteristics of flatwoods in Florida, 2007 2008 ................................................................................................................... 46 2 2 Effects of treatment time interactions on habitat characteristics of flatwoods in Florida, 2007 2008 ......................................................................................... 47 2 3 Effects of treatment grazing interactions on habit at characteristics of flatwoods in Florida, 2007 2008 ......................................................................... 50 2 4 Effects of treatment season interactions on habitat characteristics of flatwoods in Florida, 2007 2008 ......................................................................... 52 2 5 Comparison of the effects of treatment type on habitat characteristics of flatwoods in Florida, 2007 2008 ......................................................................... 53 2 6 Comparison of the effects of trea tment type time interactions on habitat characteristics of flatwoods in Florida, 2007 2008 ............................................. 54 2 7 Comparison of the effects of treatment type grazing interactions on habitat characteristic s of flatwoods in Florida, 2007 2008 ............................................. 55 2 8 Comparison of the effects of treatment type season interactions on habitat characteristics of flatwoods in Florida, 2007 2008 ............................................. 56 3 1 Effects of treatment on saw palmetto height, cover, and density in Florida flatwoods, 2007 2008 ........................................................................................ 71 3 2 Effects of treatment time interacti ons on saw palmetto height, cover, and density in Florida flatwoods, 2007 2008 ............................................................ 71 3 3 Comparisons of the effects of treatment and treatment time interactions on saw palmetto height, cover, and density in Florida flatwoods, 2007 2008 ......... 72 4 1 Arthropod abundance ( no of individuals) within orders and families collected on prescribed burned, roller chopped, and control sites in Flori da flatwoods, 2007 2008 ......................................................................................................... 96 4 2 Effects of treatment on arthropod familial richness and abundance in Florida flatwoods, 2007 2008 ........................................................................................ 98 4 3 Effects of treatment time interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 2008 ...................................................... 99

PAGE 9

9 4 4 Effects of treatment season interactions on arthropod fam ilial richness and abundance in Florida flatwoods, 2007 2008 .................................................... 101 4 5 Effects of treatment grazing interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 2008 .................................................... 102 4 6 Comparison of the effects of treatment time interactions on arthropod familial richness and abundance in Florida flatw oods, 2007 2008 ................... 103 4 7 Comparison of the effects treatment season interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 2008 ................... 104 5 1 Avian migratory and breeding category composition and abundance in prescribed burned and roller chopped Florida flatwoods, 2007 2008 .............. 129 5 2 Effects of treatment on avian richness and abundance in Florida flatwoods, 200 7 2008 ....................................................................................................... 132 5 3 Effects of treatment season interactions on avian richness and abundance in Florida flatwoods, 2007 2008 ....................................................................... 133 5 4 Effects of treatment grazing interactions on avian richness and abundance in Florida flatwoods, 2007 2008 ....................................................................... 134 5 5 Effects of treatment time interactions on avian richness and abundance i n Florida flatwoods, 2007 2008 ........................................................................... 134 5 6 Comparison of the effects of treatment on avian richness and abundance in Florid a flatwoods, 2007 2008 ........................................................................... 135 5 7 Comparison of the effects of treatment season interactions on avian richness and abundance in Florida flatwoods, 2007 2008 ............................... 135 5 8 Comparison of the effects of treatmen t time interactions on avian richness and abund ance in Florida flatwoods, 2007 2008 ............................................. 136 5 9 Comparison of the effects of treatment grazing interactions on avian richness and abundance in F lorida flatwoods, 2007 2008 ............................... 137 5 10 Habitat characteristics that best predict total, category, and guildspecific avian abundance and rich ness in Florida flatwoods, 2007 2008 ..................... 138 6 1 Lepidopteran abundance (no. of individuals) in prescribed burned and roller chopped Florida flatwoods, 2008 ...................................................................... 156 6 2 Comparison of the e ffects of treatment on lepidopteran, flowering forb, and flowering shub species richness and abundance in Florida flatwoods, 2007 ....................................................................................................... 157

PAGE 10

10 7 1 Avian guild composition and seasonal abundance on monoculture and mixed pastures at MacArthur AgroEcology Research Station, Highl ands County, Florida, 1999 2003 ........................................................................................... 182 7 2 Effects of grazing intensity on vegetation attributes of monoculture and mixed pastures at MacArthur AgroEcology Research Station, Highl ands County, Florida, 1999 2003 ........................................................................................... 185 7 3 Effects of a grazing intensity season interaction on vegetation attributes of monocu lture and mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. .............................................. 186 7 4 Effects of a grazing intensity time interaction on vegetation attributes of monoculture and mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. .............................................. 188

PAGE 11

11 LIST OF FIGURES Figure page 7 1 Effects of grazing intensity on avian abundance by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ........................................................................................... 189 7 2 Effects of grazing intensity on total avian species richness and avian species richness by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ............................... 190 7 3 Effects of a grazing intensity season interaction on avian abundance and species richness by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ................. 191 7 4 Effects of a grazing intensity time interaction on avian abundance by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ............................................................ 193 7 5 Effects of grazing intensity on total avian abundance and avian abundance by guild in mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ............................................................ 194 7 6 Effects of grazing intensity on avian species richness by guild in mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ........................................................................................... 195 7 7 Effects of a grazing intensity season interaction on total avian species richness and avian abundance and species richness by guild in mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ........................................................................................... 196 7 8 Effects of a grazing intensity time interaction on avian species richness by guild on mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003 ............................................................ 198

PAGE 12

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 WILDLIFE AND HABITAT RESPONSES TO PRESCRI BED BURNING, ROLLER CHOPPING AND GRAZING OF F LORIDA RANGELAND AND PASTURE By Emma V. Willcox May 2010 Chair: William M. Giuliano Major: Wildlife Ecology and Conservation In Florida, prescribed burning and roller chopping are management practices w idely promoted under assistancebased programs to improve rangeland live sto ck forage and wildlif e habitat conditions In addition, m any assistance based programs in Florida also promote management activities that intend to improve monoculture and mixed pasture habitats for the benefit o f wildlife. However, there is a lack of information concerning the impact s of these treatments and l ittle is known of the role grazing lands play in providing habitat for wildlife, impeding management and conservation efforts. This study was designed to fill knowledge gaps regarding the response of native rangeland (pi ne flatwoods) vegetative, avian, and invertebrate communities to prescribed burning and roller chopping. It also examined vegetation and bird community responses to grazing of monoculture and mixed pastures to provide information of use in tailoring manag ement programs to sites where avian conservation is a priority Roller chopping, particularly during the growing season was effective at reducing shrub cover, height and density i n pine flatwoods. However, bur n ing and roller chopping practices frequent ly result ed in decreases in herbaceous vegetation. Burning and roller chopping treatments led to reductions in arthropod familial richness.

PAGE 13

13 Total arthropod abundance was lower on all but g rowing season roller chop sites Depending on treatment and seaso n of application, reductions in familial richnes s and abundance were observed in a variety of individual arthropod orders Dormant season burning resulted in decreases in non breeding, overwintering av i a n species richness and abundance. However, growing season bur n ing resulted in incr e ases in species richness and abundanc e for this av i a n group. Growing season roller c hopping resulted in increases in abundance of permanent resident and breeding av i a n species On monoculture pasture, a n increase in grazing intensity led to declines in total avian species richness and short distanc e migrant, neotropical migrant, and permanent resident avian species richness and abundance. Declines in total av i a n species richness and neotropical migrant av i a n species rich ness and abundance were observed on mixed pastures subject to increasing grazing intensity. However, short distance migrant and urban av i a n species richness and grassland avia n abundance increased on this pasture ty pe in the presence of grazing.

PAGE 14

14 CHAPTER 1 INTRODUCTION Florida is an ecologically diverse state with a climate ranging from temperat e to subt ropical and is home to a wide diversity of flora and fauna, including 425 bird, 184 reptile and amphibian, 75 mammal 126 fish, and 3,500 vascular plant s pecies (Ewell 1990, Cox et al. 1994). The state ranks fourth in the contiguous United States in terms of overall species richness (Noss and Peters 1995), having more species than any other state east of the Mississippi River (Ewell 1990). Unfortunately, r apid population growth poses a pervasive threat to the future of Floridas ecosystems and wildlife. In 2000, the state was ranked fourth in the United States in population size, being home to 15,982,378 people (United States Census Bureau [USCB] 2000) and its population density was approximately double that of the most populous state, California ( USCB 2000). In 2004, Floridas human population had exceeded 17,000, 000 (USCB 2000), an i ncrease of over 15,600,000 from 1936 (Florida Biodiversity Task Force 1993). By 2030 it is projected the states population will have undergone a furt her 80% increase totaling approximately 28.7 million (USCB 2000). One consequence of past and predicted increases in population size has been residential, commercial, and ind ustrial development. This development has resulted in the loss, degradation, and fragmentation of wildlife habitats, causing the disruption of ecological connectivity and disturbance to local and landscapelevel ecological functions (e.g., wildlife moveme nts, fire regime s, hydrology and sediment movement; Florida Fish and Wildlife Conservation Commission [FWC] 2005). The actions of Floridas residents and landowners (e.g. road constr uction, alterations in fire regimes changes in grazing and other land management practices, introduction of exotic plants and animals, and

PAGE 15

15 degradation of water resources) place additional pressures on the states ecosystems ( e.g., Melaluca [ Melaleuca brevifolia Turcz] Japanese climbing fern [ Lygodium japonicu m (Thunb.) Sw.] Brazillian pepper [ Schinus terebinthifolius Raddi ] ; Myers and Ewel 1990, FWC 2005, Gordon et al. 2005). One hundred and eighteen animal species, many of which are associated with habitats undergoing considerable declines in quantity and condition because of the stresses described above, are state listed as endangered, threatened, or of special concern by FWC (2005) Fifty seven of these animals are also federally listed as endangered or thre atened (United States Fish and Wildlife Service [ US FWS] 2006). In an assessment of ecosystem risks in the United States based on development pressures, endangered habitats, and imperiled species, Florida was ranked first in the extreme risk category out of 10 states (Noss and Peters 1995). Like other terrestrial habitats in t he state, Floridas rangelands and pastures are under threat of destruction, degradation, and fragmentation with several of high conservation priority because of declines in quantity and quality (FWC 2005, Gordon et al. 20 05). These declines can partly be attributed to suburban, urban, and industrial development (FWC 2005, Gordon et al. 2005). In recent years, the demand for suitable development lands has risen dramatically, with th e highest burden being placed on coastal and upland habitats (FWC 2005), sites usually considered most desirable or appropriate for construction. In addition to development, conversion to intensive agricult ure (e.g., monoculture pasture, sugar cane planti ngs citrus groves, pine plantations, and row crops), changes in land management techniques (e.g., alteration of natural fire regimes and increases in grazing pressure), and establishment of exotic

PAGE 16

16 plants (Myers and Ewel 1990, FWC 2005, Gordon et al. 2005) pose additional threats to the persistence and integrity of remaining rangeland and pasture areas. Although rangeland and pasture problems are complex, because these areas potentially occupy more than 8 million hectares of the state, their maintenance and improvement is key to the persistence of more than 140 associated wildlife species considered to be of greatest conservation need (Millsap et al. 1990, FWC 2005, Gordon et al. 2005). The majority of Floridas land area is characterized as nonfederal rural lands (United States Department of Agriculture Natural Resource Conservation Service [USDA NRCS] 1997), with a large proportion privately owned (FWC 2005). The F lorida Fish and W ildlife C onservation Commission believe s the key to conserving many of Flori das native species involves maintaining or enhancing habitats that currently exist on private lands. They hope this can be partly achieved through the implementation of nonregulatory, assistance based activities. In conjunction with other state and federal agencies such as the USDA FWC is utilizing assistance based programs to encourage private landowners to employ land management practices of potential benefit to wildlife and their habitats. Many assistance based programs provide technical and financ ial assistance for the implementation of such management activities. These include USDA Farm Bill programs such as the Environmental Quality Incentives Program and Wildlife Habitat Incentives Program. In Florida, assistance based programs are being us ed t o facilitate the maintenance and improvement of a variety of native rangeland habitats. H owever, few quantitative data exist detailing plant, invertebrate, and vertebrate population and

PAGE 17

17 community response to the management practices promoted by such progr ams (Berkland et al. 2005, FWC 2005, Gray et al. 2005). In Florida, a l ack of knowledge as to how management activities endorsed by assistance based programs impact rangeland vegetation, livest ock forage, wildlife, and habitat may hinder implementation. If landowners are to successfully utilize these management practices and their promotion under assistance based programs is to be fully achieved, we need a clearer understanding of how they affect rangeland vegetative communities and wildlife. In this dis sertation, I quantified the effects a number of land management practices have on central and south Floridas native rangeland habitats and associated w ildlife. In particular, I examined the response of native rangeland (pine flatwoods) vegetative, avian, and invertebrate communities t o prescribed burning and roller chopping, 2 management activities promoted through assistance base d programs across the state. In addition, I investigated wildlife habitat associations on monoculture and mixed (also known as improved and semi improved) pasture. In order to meet livestock production demands, larg e areas of native rangeland have been converted to monoculture or mixed pasture. Many assistance based programs promote management activities that intend to improve t hese altered rangeland habitats to the benefit of wildlife (e.g., brush management, prescribed burning, prescribed grazing, brush control, and fencerow maintenance). However, we know little of the role these grazing lands play in providing habitat for wil dlife in Florida, hindering management and conservation efforts. If we are to apply management practices of benefit to Floridas native species to these lands and subsidize the most appropriate management activities under assistance based programs, we nee d to identify features of these nonnative

PAGE 18

18 rangeland habitats important to wildlife. I examine d vegetation and avian community response and avianhabitat associations on monoculture and mixed rangelands subject to various livestock grazing densities in an attempt to identify habitat characteristics important to avian conservation and of use in tailoring management programs.

PAGE 19

19 CHAPTER 2 WILDLIFE HABITAT EFF ECTS OF PRESCRIBED BURNING AND ROLLER CHOPPING IN PINE FLA TWOODS Introduction Pine flatwoods occur thr oughout the southeastern coastal plain of the United States and formerly cover ed approximately 50% of the land area of Florida (Abrahamson and Hartnett 1990). Unfortunately, large areas of this pine savanna habitat currently exist in a highly degraded state (Means 1996). Historically, flatwoods habitats were maintained by frequent, low intensity, lightni ng ignited fires during the May July thunderstorm season (Komarek 1968, Abrahamson and Hartnett 1990, Pyne et al. 1996). However, over the past 50 years on much of Floridas pine flat woods fire suppression, reductions in fire frequency, or a shift in fire season, commonly a result of human intervention, have resulted in excessive shrub growth and proliferation of species such as saw palmetto ( Serenoa repe ns [Bartram] Small ) gallberry ( Illex glabra [L.] A. Gray ), wax myrtle ( Morella cerifera [L.] Small ), and fetterbush ( Ly oni a lucida [Lam] K. Koch ) These increases in shrub dominance have resulted in the loss of many grass and forb species and declines in the otherwise species rich herbaceous ground layer of many pine flatwoods habitats ( Wade et al. 1980, Huffman and Blanchard 1991, Glitzenstein et al. 1995 ). Such changes threaten the integrity of remaining pine flatwoods and their suitability as habitat for many wildlife species of conservation concern (F lorida Fish and Wildlife Conservation Commission [F WC ] 2005). These changes have also resulted in declines in forage quantity and quality, potentially restricting the use of these areas for livestock production (Moore 1974). Today, most landowners and managers of pine flatwoods are faced with the challenge of habitat maintenance and improvement In many instances, this is being

PAGE 20

20 achieved through the reintroduction of prescribed burning and application of mechanical treatments (e.g., roller chop ping). Many landowners and managers in Florida are cattle ranchers, whose goal is to reduce shrubs and increase the growth and production of more palatable grasses and forbs as livestock forage (Yarlett 1965, Moore 1974, Kalmbacher and Martin 1984, Tanner et al. 1988). Concomitantly, the majority of wildlife species that occupy pine flatwoods habitats benefit from increases in grass and forb cover as they provide diverse food, cover, and other habitat resources (Huber and Steuter 1984, Madden et al. 1999, FWC 2005). However, limited numbers of understory trees and shrubs are also an integral component of wildlife habitat and their maintenance at low levels is typically desirable. The United States Department of Agr iculture in collaboration with FWC, is currently using the Environmental Quality Incentives Program (EQIP) and Wildlife Habitat Incentives Program, among others, to encourage private landowners to maintain and restore wildlife habitat. These programs are providing financial and technical assistance for landowners to implement management practices that reduce shrub and understory hardwood cover in pine flatwoods with the intention of increasing herbaceous plant growth. Practices promoted under assistance programs include prescribed burning and roller chopping during dormant (November March) and growing ( April October) seasons. Prescribed burning and roller chopping can reduce shrubby vegetation and promote growth of herbaceous groundcover species in southeastern rangeland habitats (Wade et al. 1980, Kalmbacher and Martin 1984, Tanner et al. 1988, Robbins and Myers 1992, Glitzenstein et al. 1995, Watts and Tanner 2003, Watts et al. 2006),

PAGE 21

21 potentially improving their quality for wildlife. However, the seaso nal effects and benefits of these treatments are unclear. Studies detailing and comparing the habitat effects of these practices, particularly dormant and growing season roller chop ping are lacking Research conducted has typically been extremely locali zed, being confined to a single study area. If we are to make general recommendations on the use of these practices to individuals managing pine flatwoods habitat across the state, there is a need for detailed research that evaluates vegetation response t o prescribed burning and roller chopping practices over a larger area. In addition, we know little regarding how livestock management, a dominant landuse in many pine flatwoods, alters the effects of prescribed burning and roller chopping on vegetation and wildlife habitat. A lack of understanding of how season of prescribed burning and roller chop ping impact s pine flatwoods vegetation, livestock forage, and wildlife habitat in grazed and nongrazed areas could hinder practice implementation and may limit landowner participation in assistance based management programs. Therefore, detailed research that considers vegetative community and wildlife habitat response to prescribed burning and roller chop ping practices in pine flatwoods in varying seasons is re quired The objectives of my study were to 1) fill recognized gaps in our understanding of how prescribed burning, roller chopping, and combinations of the 2 practices, effect native pine flatwoods vegetation structure and composition and wildlife habitat, and 2) explore whether grazing interacts with prescribed burning and roller chop ping to effect vegetation and habitat.

PAGE 22

22 Methods Study Sites I conducted m y study in privately and publicly owned pine flatwoods across a 6county area (Desoto, Highlands, Lee, Manatee, Osceola, and Sarasota) of central and south Florida. In these flatwoods, I established 50paired treatment and control sites with varying management (i.e., prescribed burning and roller chop ping) histories and grazing regimes. When grazed, both the treatment and paired control study sites were subject to similar grazing pressures at similar times. During the study, local landowners and managers prescribed burned and roller chopped these pine flatwoods sites using varying, individual protocol s. Floridas pine flatwoods are characterized as having a pure or combined overstory stand of scattered slash ( Pinus elliotti Engelm. ) and longleaf ( P. palustris Mill. ) pine. The understory and shrub layer includes saw palmetto, gallberry, wax myrtle sta ggerbush ( Lyonia fruticosa [ Michx.] G.S. Torr.), dwarf huckleberry ( Gaylussacia dumosa [Andrews] Torr. & A. Gray ), dwarf live oak ( Quercus mimima [Sarg.] Small ), and tarflower ( Befaria racemosa Vent ). When the shrub and understory layer is relatively open, an often diverse herbaceous layer exists. This layer contains a wide variety of grasses ( e.g., Agrostis spp. Andropogon spp. Aristida spp. Eragrostis spp. Panicum spp. and Paspalum spp ). Common forbs include legumes ( e.g., Cassia spp. Crotalari a spp. Galactia spp. Tephrosia spp.), milkweeds ( Asclepias spp.), milkworts ( Polygala spp.), and a wide variety of composites ( e.g., Aster spp. Chrysopsis spp. Eupatorium spp. Liatris spp. and Solidago spp ; Abrahamson and Hartnett 1990, United S tate s Fish and Wildlife Service 1999).

PAGE 23

23 Treatment Types Treatment types incl uded dormant season (November March ) burn, gro wing season (April October ) burn, dormant season roller chop, growing season roller chop, and a roller chop/burn combination. The roller chop/burn combination (hereafter referred to as roller chop/burn) involved roller chop ping in the dormant season followed by burning within 6 months. I established a total of 11 dormant season burn, 9 growing season burn, 9 dormant season roller chop, 12 growing season roller chop and 9 roller chop/burn sites, each paired with an adjacent untreated control. Habitat Sampling I assessed the effects of management treatments (i.e. prescribed burning, roller chop ping, and combinations of the two) on pine flat woods habitat characteristics using a pairedsample approach, where plant community composition and structure and ground layer variables were compared between sampling point s randomly located within paired treated (e.g., burned ) and untreated sites. Sampling point s in untreated sites were randomly located adjacent to treated sites, and were of similar current and past management (e.g., grazing intensity), surrounding landuse, plant community (e.g., overstory cover), and soil conditions, and were located in the same pasture or managem ent unit. Within each site, 1 randomly selected treatment or control sampling point was established. Sampling point s that occurred within 50 m of the edge of a site were rejected and randomly relocated to minimize edge effec ts. Sites, within which treatment and control sampling point s were located, ranged from 2 20 ha in siz e. I assessed habitat characteristics at each sampling point once in winter (January March), spri ng (April May), and summer (June September ), during each of 2 years (2007 2008) following treatment At each point plant community composition and

PAGE 24

24 structure, li t ter and soil variables, and vertical obstruction were examined in several strata (i.e., ground, herbaceous, shrub, understory, and overstory levels), using a 0.03 ha nested circular plot design similar to that described by Dueser and Shugart (1978) and Higgins et al. (2005). Ground layer I determined litter c over (%; ocular estimate) within 4 1m2 sub sample plots, 1 randomly locat ed in each quadrant of the 0.03ha circular plot, along with soil density (g/cm3), moisture (%), and pH. Litter cover was rec orded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 99%, and 7 = 100% (Hays et al. 1981, Higgins et al. 2005). Soil density was measured as the dry weight dens ity (g/cm3) of a 5 cm diameter 10cm deep soil core sample after oven drying at 45C for 48 hours (Donhaue et al. 1971) I used a Kelway soil tester to measure soil pH and moisture (Rodewald and Yahner 2001). Herbaceous layer I measured s pecies richness ( no. of species) cover (%; ocular estimate), and maximum height (cm) of forbs and graminoids with in the 1m2 subsample plots. Forb and graminoid cover were rec orded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25% 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 99%, and 7 = 100% (Hays et al.1981, Bullock 1996, Krebs 1999, Higgins et al. 2005). Vertical o bstruction. Vertical obstruction (%) from 0 2 m above ground was measured using a cover pole (Griffin and Yo utie 1988) centered on the 0.03ha circular plot at a distance of 5 and 10 m. Shrub layer I counted and measured the height of all shrubs (woody vegetation <2.0 m in height) in 2 perpendicular 20m2 quadrats centered on the 0.03ha plot to estimate species richn ess (no. of species), density (no./m2), and maximum height (cm)

PAGE 25

25 for individual species and all combined (Hays et al. 1981, Krebs 1999, Higgins et al. 2005). Shrub cover (%) was assessed along 2 perpendicular 20m transects centered on the 0.03ha circular plot using the line intercept method (Hays et al. 1981, Higgins et al. 2005). Understory I counted, identified t he species, and measured the diameter at breast height (dbh) of a ll understory (woody vegetation <7.5 cm dbh, height) plants within the 0.03ha circular plots to estimate species richness, density (no./ha), and basal area (cm2/ha) for individual species and all combined (Krebs 1999). Understory canopy cover ( %) was estimated from 41 evenly spaced, vertical ocular tube sightings taken at a height of 0.75 m along 2 perpendicular 20m transects centered on the 0.03ha circular plot (James and Shugart 1971). Overstory All overstory (woody vegetation dbh ) plants were also counted, species identified, and dbh measured within the 0.03ha circular plot s to estimate species richness, density (no./ha), and basal area (cm2/ha) for individual species and all combined (Krebs 1999). I estimated overstory canopy co ver (%) from 41 evenly spaced, vertical ocular tube sightings taken at a height of 1.5 m along 2 perpendicular 20m transects centered on the 0.03ha circular plot (James and Shugart 1971). I recorded the presence of livestock on study sites for incorporat ion into analyses. Analyses I used repeated measures mixed model regressions, followed by Fishers Protected LSD tests, to examine differences in habitat characteristics (e.g., understory cover, forb richness, graminoid cover) between untreated (control) a nd treated sites, both within (e.g., dormant season burn) and among (i.e., dormant season burn, growing

PAGE 26

26 season burn, dormant season roller chop, growing season roller chop, and roller chop/burn) treatment types. Repeated measures were time since treatment (time ) and season Study site pair was included as a blocking factor and presence of grazing as an additional influential independent variable. I n my results and discussion, I focus ed on treatment rather than repeated m easures or grazing effects. Two, 3 and 4way treatment effect interactions were noted in the results, if they occurred. Due to the timing of data collection, it was not possible to test for 3way interactions for growing season burning and roller chopping treatments As 3 and 4way treatment interaction ef fects are difficult to reliably interpret, they were not discussed further (Zar 1999). In the case of twoway treatment interactions, when differences in linear combinations of groups or biologically meaningless comparisons (e.g., shrub cover in growing burn sites in year 1 versus shrub cover in control sites in year 2) arose, I stated that post hoc comparisons revealed no differences based on treatment and the interacting factor. I rank transformed all data sets prior to analyses due to violations of normality and homogeneity of v ariance assumptions (Conover 1998, Zar 1999, SYSTAT 2007). Statistical significance was concluded at P I used this value rather than the more common P ity of making a Type II error ( Mapstone 1995, Zar 1999). All statistical tests were performed using SYSTAT (2007) statistical software. Results Dorman t Season Burn Ground layer Variance of litter cover was affected by dormant season burning (Table 2 1). Mean litter cover and mean and variance of litter depth were affected by a dormant season burning time interaction (Table 22 ). Dormant season burning alone

PAGE 27

27 and in all combinations with time, season, and grazing had no effect on maximum, mean, and var iance of soil pH, maximum, mean, and variance of soil moisture, and mean and variance of soil density ( P Herbaceous layer A dormant season burning time interaction affected variance of forb height and graminoid cover and mean and variance of graminoid height. Mean graminoid cover was also affected by a dormant season burning time interaction. However, no differences in this habitat characteristic based on burning and time were observed from post hoc comparisons (Table 2 2). Mean graminoid cover and variance of graminoid height were affected by a dormant season burning grazing interaction. A dormant season burning grazing interaction also affected mean forb species richness and mean and variance of forb height. However, post hoc comparisons revealed no differences in these habitat characteristics because of burning and grazing (Table 2 3). Mean graminoid height was affected by a dormant season burning season interaction (Table 2 4). Variance of forb species richness and mean forb c over and height were affected by a dormant season burning time season interaction ( P and variance of graminoid species richness a burning time season grazing interaction ( P = 0.018). Variance of forb cover and maximum and mean graminoid richness were unaffected by dormant season burning alone and in all combinations with t ime, season, and grazing ( P Vertical obstruction. Vertical obstruction from 10 m was affected by dormant season burning (Table 2 1). Vertical obstruction at 5 m was affected by a dormant season burning time interaction (Table 2 2).

PAGE 28

28 Shrub laye r Shrub cover was affected by a dormant season burning time interaction (Table 2 2) and dormant season burning grazing interaction (Table 2 3). However, post hoc comparisons found no differences in shrub cover as a result of burning and grazing. Shrub height was affected by a dormant season burning time season interaction ( P = 0.021). Dormant season burning alone and in all combinations with time, season, and grazing had no effect on shrub density ( P Understory A dormant season burning time interaction affected understory species richness (Table 2 2). Understory species richness and understory density were affected by a dormant season burning grazing interaction (Table 2 3). Understor y cover was unaffected by dormant season burning alone and in all combinations with time, season, and grazing ( P Overstory Overstory canopy cover was affected by dormant season burning (Table 2 1). Dormant season burning alone and in all combi nations with time, season, and grazing had no effect on overstory richness, density, and basal area ( P Growing Season Burn Ground layer Mean litter cover and soil pH and mean and variance of litter depth were affected by growing season burning (Table 2 1). A growing season burning grazing interaction affected variance of soil pH (Table 2 3). Variance of soil moisture was affected by a growing season burning time, mean soil density by a burning grazing, and variance of soil pH by a burning season interaction. However, post hoc comparisons found no differences in any of these habitat characteristics as a result of burning and, as appropriate, time, grazing, or season interactions (Tables 2 2, 2 3, and 2 4, respectively). Variance of litt er cover was affected by a growing season burning season grazing interaction ( P = 0.072). Maximum and mean soil moisture and

PAGE 29

29 variance of soil density were unaffected by growing season burning alone and in all combinations with time, season, and grazing ( P Herbaceous layer Growing season burning affected mean graminoid species richness and forb cover and variance of forb height (Table 2 1). Mean graminoid cover and height were affected by a growing season burning time interaction. A growing season burning time interaction also affected variance of forb cover. However, post hoc comparisons found no differences in this habitat characteristic based on burning and time (Table 2 2). M ean forb height and variance of graminoid height were affected by a growing season burning grazing interaction. However, again no differences in these habitat characteristics based on burning and grazing were observed from post hoc comparisons (Table 2 3). Variance of graminoid cover was affected by a growing sea son burning time season grazing interaction ( P = 0.066 ). Growing season burning alone and in all combinations with time, season, and grazing had no effect on maximum, mean, and variance of forb richness, and variance of graminoid richness ( P 0.280). Vertical obstruction. Vertical obstruction from 10 m was affected by growing season burning (Table 2 1). Visual obstruction from 5 m was affected by a growing season burning time interaction (Table 2 2). Shrub layer Shrub height and cover were aff ected by a growing season burning time interaction (Table 2 2). Shrub species richness was affected by a growing season burning grazing interaction, but differences in this habitat characteristic based on burning and grazing were not observed from pos t hoc comparisons (Table 2 3).

PAGE 30

30 Shrub density was unaffected by gr owing season burning alone and in all combinations with time, season, and grazing ( P Understory All understory habitat characteristics were unaffected by growing season burning alone and in all combinations with time, season, and grazing ( P 0.220). Overstory All overstory habitat characteristics were unaffected by growing season burning alone and in all combinations with time, season, and grazing ( P 0.500). Dormant Season Ro ller Chop Ground layer Mean and variance of soil moisture were affected by dormant season roller chopping (Table 2 1). A dormant season roller chopping time interaction affected mean and variance of litter depth (Table 2 2). Variance of litter depth and mean soil pH were affected by a dormant season roller c hopping grazing interaction. However, post hoc comparisons revealed no differences in shrub density based on roller chopping and grazing (Table 2 3). Dormant season roller chopping alone and in all combinations with time, season, and grazing had no effect on mean and variance of litter cover, variance of soil pH, and maximum, mean, and variance of soil density ( P 0.200). Herbaceous layer Mean forb species richness, cover, and height and mean graminoid cover were affected by a dormant season roller chopping time interaction. However, for none of these habitat characteristics did post hoc comparisons reveal differences based on roller chopping and time (Table 2 2). A dormant season roller chopping grazing interaction affected variance of graminoid cover, but post hoc comparisons revealed no differences in this habitat characteristic as a result of roller

PAGE 31

31 chopping and grazing (Table 2 3). A dormant season roller chopping season interacti on affected mean graminoid height (Table 2 4). Variance of forb cover and graminoid species richness were affected by a dormant season roller chopping season grazing interaction ( P choppi ng time season grazing interaction ( P = 0.080). Variance of forb richness, maximum and mean graminoid richness, and variance of graminoid height were unaffected by dormant season roller chopping alone and in all combinations with time, season, and g razing ( P Vertical obstruction. A dormant season roller chopping time interaction affected vertical obstruction from 5 m and 10 m (Table 2 2). Visual obstruction from 5 m was also affected by a dormant season roller chopping grazing interaction. How ever, post hoc comparisons revealed no differences in this habitat characteristic based on roller chopping and grazing (Table 2 3). Shrub layer Shrub height was affected by dormant season roller chopping alone (Table 2 1). A dormant season roller choppi ng time interaction affected shrub density and cover. However, post hoc comparisons revealed no differences in shrub density based on roller chopping and time (Table 2 2). Shrub species richness was affected by a dormant season roller chopping grazin g interaction, but no differences in this habitat variable based on roller chopping and grazing were observed from post hoc comparisons (Table 2 3). Understory All understory habitat characteristics were unaffected by dormant season roller chopping alone and in all combination with time, season, and grazing ( P 0.316).

PAGE 32

32 Overstory All overstory habitat characteristics were unaffected by dormant season roller chopping alone and in all combinations with time, season, and grazing ( P Growing Season Roller Chop Ground layer Mean and variance of litter d epth were affected by growing season roller chopping (Table 2 1). A growing season roller chopping time interaction affected variance of soil pH and mean soil density. Post hoc comparisons found no differences in mean soil density based on roller chopping and time (Table 2 2). A growing season roller chopping grazing interaction affected mean litter cover (Table 2 3). Mean litter cover was affected by a combination of growing season roller chopping and season, but no differences in this habitat attr ibute based on roller chopping and season were observed from post hoc comparisons (Table 2 4). Variance of soil density was affected by a growing season roller chopping time season grazing interaction ( P = 0.023). Growing season roller chopping alone and in all combinations with time, season, and grazing had no effect on variance of litter cover, maximum and mean soil pH, and maximum, mean, and variance of soil moisture ( P Herbaceous layer Mean forb height and graminoid species richness, cover, and height were affected by a growing season roller chopping time interaction (Table 2 1). A growing season roller chopping grazing interaction affected variance of graminoid height, but post hoc comparisons revealed no differences in this habitat characteristic as a result of roller chopping and grazing (Table 2 2). Mean graminoid height was affected by a growing season burning season interaction (Table 2 3). Var iance of graminoid species richness was affected by a growing season roller chopping time season grazing interaction ( P = 0.050). Maximum, mean, and variance of forb

PAGE 33

33 richness, maximum, mean and variance of forb cover, and variance of forb height, gr aminoid cover, and graminoid height were unaffected by growing season roller chopping alone and in all combinations with time, grazing, and season ( P Vertical obstruction. Vertical obstruction from 5 m and 10 m were affected by a growing season roller chopping time interaction (Table 2 2). Shrub layer Growing season roller chopping affected shrub species richness (Table 2 1). Shrub density and cover were affected by a growing season roller chopping grazing interaction. A growing season roller chopping grazing interaction also affected shrub height. However, no differences in this habitat characteristic based on roller chopping and grazing were observed from post hoc comparisons (Table 2 2) Understory All understory habitat characteri stics were unaffected by growing season roller chopping alone and in all combinations with time, season, and grazing ( P Overstory All overstory habitat characteristics were unaffected by growing season roller chopping alone and in all combinati ons with time, season, and grazing ( P Roller Chop/Burn Ground layer Variance of litter depth was affected by roller chopping/burning (Table 2 1). Mean and variance of litter cover and mean litter depth were affected by a roller chopping/burning time interaction (Table 2 2). A roller chopping/burning grazing interaction affected variance of litter cover mean litter depth, and soil density, but post hoc comparisons revealed no differences in these habitat characteristics based on roller chopping/burning and grazing (Table 2 3). Roller chopping/burning alone and

PAGE 34

34 in all combinations with time, season, and grazing had no effect on maximum, mean, and variance of soil pH and maximum and mean soil moisture (P Herbaceous layer A roller chopping/burning time interaction affected mean and variance of forb height. No differences in graminoid species richness based on roller chopping/burning and time were observed from post hoc comparisons (Table 2 2 ). A roller chopping/burning grazing interaction affected mean forb species richness, cover, and height and mean graminoid cover and height. However, post hoc comparisons revealed no differences in any of these habitat characteristics because of roller chopping/burning and grazing (Table 2 3). Mean forb height and mean graminoid cover and height were affected by a roller chopping/burning season interaction. Post hoc comparisons found no differences in mean forb height based on roller chopping and se ason (Table 2 4). Variance of graminoid height was effected by a roller chopping/burning time season interaction ( P = 0.024) and variance of graminoid cover by a roller chopping/burning time season grazing interaction ( P = 0.046). Maximum and v ariance of forb richness, variance of forb cover, and maximum and variance of graminoid richness were unaffected by roller chopping/burning alone and in all combinations with time, season, and grazing ( P Vertical obstruction. Visual obstruction from 10 m was affected by roller chopping/burning (Table 2 1). A roller chopping/burning time interaction affected visual obstruction from 5 m (Table 2 2). Shrubs Shrub cover was affected by roller chopping/burning (Table 2 1). A roller chopping/burning time interaction affected shrub species richness. However, post hoc comparisons revealed no differences in this habitat characteristic based on roller

PAGE 35

35 chopping/burning and time (Table 2 2). Shrub height was affected by a roller chopping/burning grazing interaction (Table 2 3). Shrub density was unaffected by roller chopping/burning alone and in all combinations with time, season, and grazing ( P = 0.330). Understory Roller chopping/burning aff ected understory species richness and density (Table 2 1). Understory basal area and cover were unaffected by roller chopping/burning alone and in all combinations with time, season, and grazing ( P 0.150). Overstory All overstory habitat characteristi cs were unaffected by roller chopping/burning alone and in all combinations with time, season, and grazing ( P 0.363). Treatment Type Comparisons Ground layer Treatment type affected variance of litter depth (Table 2 5). A treatment type time interac tion affected mean litter cover and litter depth (Table 2 6). A treatment type grazing interaction affected mean litter cover (Table 2 7). Mean soil moisture was affected by a treatment type season grazing interaction ( P = 0.087) and variance of li tter cover by a treatment type time season grazing interaction ( P = 0.020). Mean and variance of soil pH, variance of soil moisture, and mean and variance of soil density were una ffected by treatment type alone and in all combinations with time, sea son, and grazing ( P Herbaceous layer Mean forb species richness and cover, variance of forb height, and mean graminoid species richness and height were affected by a treatment type time interaction (Table 2 6). A treatment type grazing int eraction affected mean forb richness and cover, variance of forb height, and variance of graminoid richness and

PAGE 36

36 height. However, post hoc comparisons revealed no differences in variance of graminoid height based on treatment type and grazing (Table 2 7). Mean graminoid height was affected by a treatment type season interaction (Table 2 8). Mean forb height ( P = 0.030) was affected by a combination of treatment type time season and mean graminoid cover by a treatment type grazing season interact ion ( P = 0.062). Treatment type alone and in all combinations with time, season, and grazing had no effect on vari ance of forb richness and cover and variance of graminoid cover ( P 0.127). Vertical obstruction. Vertical obstruction from 5 m and 10 m were affected by a treatment type time (Table 2 6) and a treatment type grazing (Table 2 7). Shrubs A treatment type time interaction affected shrub height and cover (Table 2 6). Shrub species richness, density, and height were affected by a treatme nt type grazing interaction (Table 2 7 ). Understory Understory species richness, density, and cover were affected by treatment type (Table 2 5). Understory basal area was una ffected by treatment type alone and in all combinations with time, season, and grazing ( P Overstory Overstory species richness, density, basal area, and cover were affected by a treatment type grazing interaction (Table 2 7) Discussion I found that improvements to wildlife habitat in flatwoods, through reductions in shrub cover and potential increases in herbaceous growth, can be achieved through the application of prescribed burning and roller chopping. However, these treatments vary in the length of time they exhibit an effect on vegetati on structure and composition.

PAGE 37

37 The reduction in litter cover and depth that occurred on dormant and growing season burn sites into the second year post treatment is likely beneficial for plant development. The combustion of litter during a fire can result in nutrients from ash, part icularly nitrogen and phosphorus, important in the growth of many plant species being returned to the soil (Ahlgren 1960, Hendricks et al. 2002, Christensen 2004). In addition, the presence of litter creates a physical barrier for seedling and sprout emergence and to seeds reaching the soil. As a result, litter removal can result in increase d plant growth (Ahlgren 1960, Facelli and Pickett 1991, Hendrick s et al. 2002, Christensen 2004). While dormant and growing season roller chopping also resulted in r eductions in litter cover and depth, considerable litter remained on the soil surface. This litter may have resulted in some suppression of herbaceous regrowth (Ahlgren 1960, Facelli and Pickett 200 1, Hendrick s et al. 2002, Christensen 2004). Dormant and growing season burning caused reductions in many herbaceous characteristics related to forb and graminoid richness, height and cover. On dormant season burn sites, affected herbaceous habitat characteristics typically returned to, but not above, preburn levels later in the study However, effects on herbaceous characteristics were often more prolonged on growing season burn sites. Increases in herbaceous species biomass or abundance, which might be reflected in increases in cover and height, have been observed on dormant burn compared to control sites (Fitzgerald 1990, Robbins and Myers 1992) often when there has been a reduction in litter cover and depth (Ahlgren 1960, Facelli and Pickett 199 1 Hendricks et al. 2002, Christense n 2004 ). Growing season burning is frequently promoted as a means to restore grass and forb cover in areas invaded by hardwoods and shrubs (Glitzenstein et

PAGE 38

38 al. 1995). Many studies examining the response of herbaceous pine flatwoods vegetation to growing season burning have focu sed on flowering and seed production. Wiregrass ( Aristida stricta Michx.) a common flatwoods species, flowers profusely and will only produce viable seeds following growing season burning (Outcalt 1994). Cutthroat grass (Panicum abscissum Swallen ) has also been found to flower abundantly when burned between midApril and mid August, but rarely flowers if burned at other times (Myers and Boettcher 1987). Little bluestem ( Schizachyrium rhizomatum [Swallen] Gould) and other bluestem grasses ( Andropogon spp .) have exhibited similar patterns (Platt et al. 1988). However, few studies have examined changes in herbaceous cover following fires in different seasons and most information for Florida is anecdotal. Studies of tallgrass prairie suggest that growing season fires may increase grass production but decrease forb production relative to dormant season burns (Towne and Owensby 1984). In Oklahoma pine flatwoods communities, growing season burning increased herbaceous species richness and diversity and forb abundance (Sparks et al. 1998). That increases in herbaceous cover above pre burn levels did not occur on dormant or growing season burn sites during this study may be a result of its short duration. However, as understory and shrub density and cover wer e not significantly reduced, competition for resources may have remained, restricting forb and graminoid growth. A number of forb and graminoid characteristics only exhibited reductions following dormant and growing season burning when grazed. Certain nat ive pine flatwoods grasses, including creeping ( Schizachyrium scoparium [Michx.] Nash var. stoloniferum [Nash] Wipff ) and chalky bluestem ( Andropogon capillipes Nash ) and wiregrass have

PAGE 39

39 been found to decline when grazed immediately following burning (Whit e and Terry 1979, Sievers 1985). The attraction of livestock to new herbaceous growth produced following a fire ( Hilmon and Hughes 1965) may have been responsible for a lowering of graminoid characteristics during this study. G razing deferment following burning may benefit some grass species commonly found on pine flatwoods (White and Terry 1979, Sievers 1985, Orr and Paton 1997). While dormant season roller chopping had little effect on herbaceous habitat characteristics, growing season roller chopping and roller chop/bur n ing, like growing season burning, caused prolonged reductions in forbs and graminoids. Again, grazing deferment is recommended to maintain the herbaceous plant community as reductions in many forb and graminoid characteristics were ass ociated with grazing. Few studies have examined the response of herbaceous vegetation to roller chop ping or roller chopping/burning However, those that have suggest an increase in herbaceous characteristics following treatment application with certain forbs flower ing more profusely (Huffman and Werner 2000). Increases in certain grasses, including wiregrass, creeping bluestem broomsedge ( Andropogon virgnicus L. ), and lopsided Indiangrasss ( Sorghastrum secundum [Elliot] Nash ) have been reported followi ng dormant season roller chop ping ( Yarlett 1965, Yarlett and Roush 1970, Kalmbacher and Martin 1984). However, season of roller chop ping comparisons have found no differences in yield of bluestems, other grasses, or forbs (Kalmbacher and Martin 1984). A study of Texas range found that after 2 years herbaceous cover was up to 135% greater on roller chopped than untreated sites (Bozzo et al. 1992).

PAGE 40

40 The use of a single dormant season or growing season burn is not likely to be sufficient to reduce shrubs in pine flatwoods habitats, particularly when they occur at high densities and provide excessive cover as is often the case after a period of fire exclusion. These treatments had no effect on shrub density and only minimal effect on shrub height and cover, w hich exhibited swift regrowth post burn and returned to pretreatment levels within a year Dormant season burning may have minimal medium to long term effects on many shrubs, particularly saw palmetto, with cover typically returning to preburn levels as soon as 1 year post burn (Abrahamson 1984a, b; Fitzgeral d 1990; Glitzenstein et al. 1995; Robbins and Myers 1992). Saw palmetto a dominant shrub in pine flatwoods, is fire adapted with rhizomes that serve as large carbohydrate reserves and meristematic tissues that are protected from the flames during a burn. It is these characteristics that allow this species to resprout rapidly following a burn, regaining up to 80% of its preburn canopy cover after just 1 growing se ason (Hilmon 1968). Increases in s hrub growth are often not as quick following growing season burning as resprouting tends to be lower than for dormant season burning (Lewis and Harshbarger 1976, Fitzgerald 1990, Huffman and Blanchard 1991, Olson and Platt 1995, Drewa et al. 2002) Shrub density tended to be lower on growing season burn than control sites when grazed. Shrubs, primarily saw palmetto, can comprise up to 20% of the diet of cattle grazing pine flatwoods in winter when grasses and f orbs are scarce (Kalmbacher et al. 1984). Th ere was a reduction in the cover of grasses and forbs following growing season roller chopping and livestock may have been forced to graze remaining shrubs, causing a reduction in their density. The effects of roller chopping/burning have not

PAGE 41

41 been well st udied. During this study, the majority of sites were burned after roller chopping, but it has been suggested that burning prior to roller chopping may remove some of the above ground shrub biomass making subsequent roller chop ping more effective at shrub reduction (Kalmbacher and Martin 1984). O n pine flatwoods, in situations where significant, rapid shrub reduction is a primary goal, the use of alternative management techniques instead of or in addition to dormant season burning may be preferred. However regular dormant season burning with a 2 3 year return interval, where fuel loads permit, will keep saw palmetto crown size small and can slow the spread of this species in situations where it is not dominant (Hilmon and Lewis 1962, Hilmon and Hughes 1965) A fire return interval of 4 years or more will typically result in cover increases of 2 3% per year (Hough 1968, Moore 1974). While not a suitable management activity in areas of high shrub growth, prescribed burning treatments provides a useful and c ost effective means of preventing shrub spread in situations where low shrub levels are to be maintained. Roller chopping may be the preferred technique for significant, rapid shrub reduction and the best method for initial treatment and control of shrubs in areas where they have high density and cover It also has potential for use in controlling shrub increases in sites near urban developments and recreational areas where smoke management issues may aris e during and following burning. Roller chopping and roller chopping/burning treatments resulted in prolonged lowering of s hrub height and cover, and, in the case of growing season roller chopping, shrub density. The prolonged effects of these 2 treatments are largely what separated them from the practice s of dormant and growing season burning. Similarly, other studies have found roller

PAGE 42

42 chop ping to be useful for improving poor condition flatwoods sites with dense shrub cover. In such situations, this treatment has resulted in considerable decreases in sh rub abundance and cover; being particularly useful as a form of saw palmetto control (Moore 1974, Tanner et al. 1988). Roller chopping can result in 83% saw palmetto kill and shrub cover reduction of up to 3% on treated compared to control sites up to 2 y ears post treatment (Moore 1974). These reductions often persist in the long term, with 70% kill and crown cover reduced to less than 12% up to 5 years after roller chop p ing (Hilmon 1968, Lewis 1970). This study found no difference in the effects of dor mant and growing season roller chopping on shrub height and cover. However, i t has been suggested that chopping during periods of high soil moisture, i.e., during the growing season, may result in greater shrub reduction due to the greater penetration of the equipments blades and deeper severing of the plant s roots that occur at this time of year (Tanner et al. 1988). Prolonged reductions in a number of understory characteristics, including richness, density, and cover, on dormant season burn sites contr adicts the findings of other studies, which suggest increases in hardwood stems following this treatment. These increases are attributed to resprouting of topkilled plants (Langdon 1981, Waldrop et al. 1987). In other studies while increases in underst ory density were not observed, dormant season burning had no effect on density or cover of species due to insufficient top kill (Robbins and Myers 1992). Decreases in understory density, basal area, and cover because of dormant season burning were often only observed on non grazed sites. One possible explanation is that livestock, following a burn, may focus their attention on consumption of newly resprouting herbaceous species ( Hilmon and Hughes

PAGE 43

43 1965) rather than the consumption of less palatable understory browse, which may be consumed when there is reduced herbaceous forage availability (Kalmbacher et al. 1984). I found understory ri chness, density, basal area, and cover to be largely unaffected by growing season burning and dormant season roller chopping It is not possible to report on the effect growing season roller chopping and roller chop/ burning had on understory characteristics as both control and treatment sites were largely free of understory trees. U nderstory stem density has been found to be lower on growing season burn compared to dormant season burn sites due to greater root kill from hotter summer fires (Grelan 1975, Langdon 1981, Waldrop et al. 1987, Fitzgerald 1990, Robbins and Myers 1992). In other situations, understory density has been found to be similar on both dormant and growing season burn sites (Lotti 1956, Robbins and Myers 1992). L ower overstory canopy cover on dormant season burn than control sites suggests there may have been damage to tree crowns. Maintenance of the pine overstory is a typical goal of pine flatwoods management. Crown scorch, which can cause damage to developing buds and threaten growth, is often of particular concern following growing season burning (Robbins and Myers 1992). In the dormant season, buds ar e not present in the canopy and the chances of damage are lower (Robins and Myers 1992). Throughout the 2 years of the study there were no changes in overst ory pine density or basal area on dormant or growing season burn sites suggesting scorching had no significant effect on overstory density or basal area.

PAGE 44

44 Neither dormant nor growing season roller chopping affected overstory characteristics. There are concerns that roller chop ping may cause damage to pine roots, leading to reduced growth or mortality (Willcox personal observation). However, no evidence of pine damage following roller chop ping or roller chop/ burning was observed during this study. M anagement Implications Dormant and growing season burning would be best utilized to suppress understory and shrub growth in pine flatwood s habitats w here cover and height is not excessive and shrub species are not dominant. Where shrub levels are high, the single application of either of these treatments are unlikely to cause prolonged reductions in shrub density, height, or cover In situations where shrub proliferation has not occurred, a single dormant or growing season burn will likely prove a cost effective means of maintaining desirable understory, shrub, and herbaceous conditions although care should be taken to protect the overstory canopy. Alternative management practices such as roller chopping may be needed in combination with, or instead of, dormant and growing season burning if there is a need to reduce excessive understory and shrub density and cover, and maintain or restore herbaceous plants following fire exclusion. As decreases in herbaceous characteristics were often associated with grazing on dormant burn sites, initial deferment from grazing following the use of dormant season or growing season burning, as for all practices discussed, is likely to be necessary to permit re establishment and growth of forbs and graminoids. C onsideration should be given to t he use of growing over dormant season burning as a means of reducing shrub cover for up to 1 year post burn. Although during this study, growing season burning resulted in decreases in herbaceous cover and height, it may in the longer term

PAGE 45

45 increase flowering and seeding of important flatwoods plants, maintaining and restoring areas prev iously dominated by shrubs. In situations where shrubs have become a nuisance, a rapid reduction in their height and cover in the longer term ( application of dormant or growing season roller chop ping or roller chop/burning If reductions in shrub density are required, growing season roller chopping may be the only suitable treatment.

PAGE 46

46 Table 2 1 Effec ts of treatment on habitat characteristics of flatwoods in Florida, 2007 2008. Habitat Characteristics by T reatmenta Treatment ( SE) P Control Treated Dormant season burn Visual obstruction (% from 10 m distance) 40.9 2.3 28 1.7 Overstory canopy cover (%) 11.9 1.7 9.1 1.3 0.070 Variance of litter cover (%) 0.6 0.2 1.1 0.2 0.040 Growing season burn Mean forb cover (%) 1.8 0.2 1.2 0.1 0.028 Variance of forb height (cm) 3.5.3 44.9 129. 5 29.5 0.024 Mean graminoid richness (n o. of spp.) 3.3 0.2 3.0 0.2 0.016 Visual obstruction (% from 10 m distance) 37.5 2.6 26.8 1.9 0.015 Mean litter cover (%) 5.4 0.1 3.0 0.3 Mean litter depth (cm) 10.3 1.0 2.5 0.7 Variance of litter depth (cm) 38.9 7.7 10.0 7.2 Mean soil pH 5.7 0.1 5.9 0.1 0.036 Dormant season roller chop Shrub height (cm) 101.8 3.0 77.3 2.8 Mean soil moisture (%) 8.0 2.5 13.42 2.8 0.017 Variance of soil moisture (%) 21.4 6.0 82.9 20.4 0.008 Growing season roller chop Shrub richness (n o. of spp.) 5.3 0.3 4.3 0.4 0.033 Mean litter depth (cm) 10.1 0.9 6.0 0.8 0.055 Variance of litter depth (cm) 26.7 5.0 32.0 23.2 Roller chop/burn Shrub cover (%) 64.4 4.4 30.2 3.5 Visual obstruction (% from 10 m distance) 37.3 3.0 18.8 1.8 Understory richness (n o. of spp.) 0.2 0.1 0.0 0.0 0.024 Understory density (no./ha) 153.8 64.5 0.0 0.0 0.025 Variance of litter depth (cm) 29.6 8.3 10.7 5.0 a Only habitat characteristics significantly affected by treatment presented ( P

PAGE 47

47 Table 2 2 Effects of treatment time interactions on habitat characteristics of flatwoods in Florida, 2007 2008. Habitat Characteristics by T reatmenta Timeb Treatment ( SE) c, d P Control Treated Dorman season burn Shrub cover (%) 1 64.2 4.6 Aa 39.1 4.6 Aa 0.085 2 67.5 3.4Aa 67.9 3.2Ab Variance of forb height (cm) 1 317.1 47.7 Aa 176.2 42.5 Ba 0.053 2 273.4 45.6 Ab 253.0 48.0 Aa Mean graminoid cover (%) 1 2.5 0.2 Aa 2.0 0.1 Aa 0.035 2 2.2 0.1 Aa 2.4 0.1 Aa Variance of graminoid cover (%) 1 0.5 0.1 Aa 0.3 0.1 Ba 0.006 2 0.5 0.2 Aa 0.4 0.1 Aa Mean graminoid height (cm) 1 59.1 4.2 Aa 39.5 3.4 Ba 2 56.8 3.4 Aa 56.5 2.6 Aa Variance of graminoid height (cm) 1 390.3 77.3 Aa 134.4 27.7 Ba 0.026 2 338.1 68.0Aa 2 22.9 53.3Aa Visual obstruction (% from 5 m distance) 1 25.1 2.9 Aa 15.2 3.2 Ba 0.024 2 28.2 2.2 Aa 26.3 1.8 Ab Understory richness (no. of spp.) 1 0.4 0.2Aa 0.2 0.1Aa 0.027 2 0.6 0.2 Ab 0.2 0.1 Aa Mean litter cover (%) 1 5.5 0.1 Aa 3.4 0.3 Ba 0.016 2 5.3 0.2 Aa 4.5 0.2 Ba Mean litter depth (cm) 1 8.2 1.1Aa 4.8 2.9Ba 0.005 2 9.4 1.1 Aa 6.3 0.7 Aa Variance of litter depth (cm) 1 27.6 11.8 Aa 4.2 2.2 Aa 0.035 2 24.7 7.0Aa 38.3 23.5Ab Gr owing season burn Shrub height (cm) 1 135.9 20.8 Aa 51.2 5.7 Ba 0.021 2 133.3 7.6Aa 92.2 3.4Ab Shrub cover (%) 1 67.3 6.6 Aa 16.2 5.9 Aa 0.019 2 73.8 4.7 Aa 76.6 4.3 Ab Variance of forb cover (%) 1 0.5 0.1 Aa 0.2 0. 1 Aa 0.031 2 6.8 6.7Aa 0.3 0.0Aa Mean graminoid cover (%) 1 2.7 0.2 Aa 1.6 0.2 Ba 0.035 2 2.2 0.1 Aa 2.7 0.4 Aa Mean graminoid height (cm) 1 73.1 4.2Aa 31.6 5.4Ba 0.012 2 56.3 3.2 Aa 52.6 3.1 Ab Visual obstruction (% from 5 m distance) 1 26.1 3.8 Aa 8.6 2.0 Ba 0.021 2 28.2 2.7 Aa 21.5 1.6 Ab Variance of soil moisture (%) 1 29.9 11.4 Aa 36.6 13.1 Aa 0.070 2 56.3 17.8 Aa 22.2 8.2 Aa Dormant season roller chop Shrub density (n o./m 2 ) 1 3.7 0.3 Aa 3.2 0.7 Aa 0.074 2 3.3 0.3 Aa 3.5 0.5 Aa Shrub cover (%) 1 58.3 5.7 Aa 19.4 4.1 Ba 0.072 2 57.6 3.1Aa 35.5 4.0Ba

PAGE 48

48 Table 2 2 Continued Habitat C h aracteristics by T reatment a Time b Treatment ( SE) c, d P Control T reated Mean forb richness (n o. of spp.) 1 4.5 0.3 Aa 3.9 0.3 Aa 0.067 2 4.4 0.3 Aa 5.0 0.3 Aa Mean forb cover (%) a 1 1.6 0.1 Aa 1.5 0.1 Aa 0.018 2 1.6 0.1 Aa 1.9 0.1 Aa Mean forb height (cm) 1 29.3 1.5 Aa 24.8 2.2 Aa 0.00 2 2 30.3 1.5 Aa 38.8 2.6 Aa Mean graminoid cover (%) 1 2.8 0.2Aa 2.0 0.1Aa 0.064 2 2.5 0.1 Aa 2.3 0.1 Aa Visual obstruction (% from 5 m distance) 1 23.2 2.1 Aa 6.7 1.2 Ba 2 20.0 2.0Aa 14.6 1.6Ab Visual obstruction (% from 10 m distance) 1 31.6 2.0 Aa 11.6 1.3 Ba 0.005 2 30.5 1.9 Aa 20.5 1.5 Ab Mean litter depth (cm) 1 8.7 1.0 Aa 6.2 0.6 Ba 0.045 2 7.7 0.6Aa 3.7 0.4Ba Variance of litter depth (cm) 1 35.5 10.3 Aa 11.8 3.5 Ba 0.064 2 30.1 5.2 Aa 6.3 1.5 Ba Growing season roller chop Mean forb height (cm) 1 33.5 2.4 Aa 17.6 3.7 Ba 0.004 2 29.8 1.9 Aa 30.7 2.2 Ab Mean graminoid richness (n o. of spp.) 1 4.3 0.4 Aa 3.3 0.4 Ba 0.028 2 4.0 0.2 Aa 4.6 0.3 Ab Mean graminoid cover (%) 1 2.6 0.2 Aa 1.9 0.2 Ba 0.099 2 3.1 0.7 Aa 2.3 0.1 Aa Mean graminoid height (cm) 1 57.5 2.4Aa 36.6 3.6Ba 0.005 2 58.4 2.3 Aa 49.6 2.5 Aa V isual obstruction (% from 5 m distance) 1 24.4 4.7 Aa 4.6 1.0 Ba 0.020 2 17.8 1.2Aa 7.8 1.0Ba Visual obstruction (% from 10 m distance) 1 29.3 4.1 Aa 6.9 1.5 Ba 0.009 2 23.0 1.4 Aa 10.8 1.2 Aa Variance of soil pH 1 0.1 0.0 Aa 0 .2 0.0 Aa 0.034 2 0.3 0.1 Aa 0.1 0.0 Bb Mean soil density (g/cm 3 ) 1 1.2 0.0 Aa 1.2 0.0 Aa 0.087 2 1.2 0.0 Aa 1.2 0.0 Aa Roller chop/burn Shrub richness (n o. of spp.) 1 5.1 0.3 Aa 4.8 0.5 Aa 0.084 2 4.8 0.2 Aa 6.0 0.4 Aa Mean forb height (cm) 1 35.2 2.6Aa 28.1 4.4Aa 0.048 2 37.6 1.9 Aa 38.7 2.8 Ab Variance of forb height (cm) 1 302.0 74.2 Aa 101.5 26.2 Ba 0.091 2 233.4 37.6Aa 286.5 63.0Aa Mean g raminoid richness (n o. of spp.) 1 3.5 0. 3 Aa 3.4 0.4 Aa 0.084 2 3.2 0.2 Aa 4.0 0.4 Aa Mean graminoid cover (%) 1 2.4 0.2 Aa 1.7 0.2 Ba 0.089 2 2.4 0.2Aa 2.0 0.1Aa

PAGE 49

49 Table 2 2 Cont inued Habitat Characteristics by T reatment a Time b Treatment ( SE) c, d P Control Treated Mean graminoid height (cm) 1 61.4 4.0 Aa 37.1 4.5 Ba 0.055 2 62.1 4.4 Aa 43.9 3.8 Ab Visual obstruction (% from 5 m distance) 1 28.5 3.4 Aa 8.2 2.1 Ba 0.085 2 26.8 2.7 Aa 17.5 2.2 Aa Mean litter cover (%) 1 5.5 0 .1 Aa 2.6 0.3 Ba 0.029 2 5.0 0.2 Aa 3.5 0.3 Aa Variance of litter cover (%) 1 0.6 0.4 Aa 1.5 0.4 Ba 0.033 2 1.2 0.3 Aa 1.4 0.3 Aa Mean litter depth (cm) 1 10.4 1.2 Aa 4.4 2.8 Ba 0.016 2 8.8 0.9 Aa 3.9 0.7 Ba a Only habitat c haracteristics significantly affected by a treatment time interaction presented ( P b Time since treatment application (years). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1). d Means in a column follo wed by the same lowercase letter not significantly different ( P > 0.1).

PAGE 50

50 Table 2 3 Effects of treatment grazing inte ractions on habitat characteristics of flatwoods in Florida, 2007 2008. Habitat characteristics by treatmenta Grazing Treatment ( SE) b P Control Treated Dormant season burn Shrub cover (%) Nongrazed 88.5 7.7 A 54.9 7.5 A 0.030 Grazed 61.2 2.6A 53.5 3.8A Mean forb richness (n o. of spp.) Nongrazed 3.2 0.2 A 2. 3 0.4 A 0.028 Grazed 3.6 0.3 A 4.4 0.3 A Mean forb cover (%) Nongrazed 1.9 0.2 A 1.1 0.1 B 0 .001 Grazed 1.4 0.1 A 1.7 0.1 B Mean forb height (cm) Nongrazed 36.3 2.9 A 27.0 3.9 A 0.088 Grazed 34.7 2.2 A 34.7 2.7 A Vari ance of forb height (cm) Nongrazed 163.5 51.9 A 375.9 91.3 A 0.003 Grazed 322.6 37.0 A 176.8 31.5 A Mean graminoid cover (%) Nongrazed 1.8 0.2 A 2.2 0.2 B 0.061 Grazed 2.5 0.1A 2.2 0.1A Understory richness (n o. of spp.) Nongraz ed 1.1 0.3 A 0.1 0.1 B 0.009 Grazed 0.4 1.2 A 0.2 0.1 A Understory density (no./ha) Nongrazed 225.9 74.5A 6.0 6.0B 0.023 Grazed 51.1 17.5 A 33.3 11.4 A Growing season burn Shrub richness (n o of spp.) Nongrazed 7.1 0.3 A 5.0 0.6 A 0.069 Grazed 5.3 0.5A 4.8 0.6A Mean forb height (cm) Nongrazed 38.9 2.8 A 19.6 2.6 A 0.017 Grazed 25.8 5.6 A 26.2 5.3 A Variance of graminoid height (cm) Nongrazed 489.5 121.6A 121.6 27.2A 0.006 Grazed 218.1 77. 4 A 299.9 64.6 A Variance of soil pH Nongrazed 0.1 0.0 A 0.1 0.0 B 0.047 Grazed 0.1 0.0A 0.0 0.0A Mean soil density (g/cm 3 ) Nongrazed 1.2 0.0 A 1.2 0.0 A 0.064 Grazed 1.3 0.0 A 0.3 0.0 A Dormant season roller chop Sh rub richness (no. of spp.) Nongrazed 5.9 0.4A 4.2 0.3A 0.067 Grazed 4.2 0.2 A 3.7 0.3 A Variance of graminoid cover (%) Nongrazed 0.5 0.1 A 1.5 1.0 A 0.064 Grazed 0.4 0.1A 0.2 0.0A Visual obstruction (% from 5 m distance) Nong razed 19.2 1.7 A 12.3 2.0 A 0.082 Grazed 23.4 2.2 A 9.4 1.2 A Variance of litter depth (cm) Nongrazed 26.4 6.6 A 10.4 3.3 B 0.046 Grazed 38.1 8.7 A 7.6 2.0 B Mean soil pH Nongrazed 6.0 0.1 A 5.9 0.1 A 0.052 Grazed 5.8 0.1 A 6 .0 0.1 A Growing season roller chop Shrub density (n o./m 2 ) Nongrazed 3.4 0.4 A 3.0 0.6 A 0.080 Grazed 3.7 0.8 A 1.4 0.3 B

PAGE 51

51 Table 2 3 Continued Habitat characteristics by treatment a Grazing Treatment ( SE) b P Control Treated Shrub height (cm) Nongrazed 106.6 5.3 A 57.5 3.4 A 0.014 Grazed 89.1 5.1 A 65.9 3.4 A Shrub cover (%) Nongrazed 43.8 4.7 A 17.4 4.2 A 0.031 Grazed 45.5 3.3 B 8.1 2.0 B Variance of graminoid height (cm) Nongrazed 1 92.4 47.8 A 236.1 60.5 A 0.097 Grazed 220.3 91.0 A 95.6 18.0 A Mean litter cover (%) Nongrazed 5.4 0.2A 4.6 0.2B 0.098 Grazed 4.8 0.2 A 4.4 0.4 B Roller chop/burn Shrub height (cm) Nongrazed 128.4 7.8A 73.0 4.6B 0.054 G razed 111.4 6.7 A 82.1 4.6 A Mean forb richness (n o. of spp.) Nongrazed 5.1 0.6 A 4.4 0.5 A 0.068 Grazed 4.5 0.3 A 5.6 0.5 A Mean forb cover (%) Nongrazed 2.5 0.7A 1.5 0.1A 0.006 Grazed 1.5 0.1 A 2.0 0.1 A Mean forb hei ght (cm) Nongrazed 36.8 2.3 A 28.9 2.6 A 0.081 Grazed 36.5 2.0A 42.0 4.2A Mean graminoid cover (%) Nongrazed 2.0 0.2 A 1.9 0.2 A 0.014 Grazed 2.9 0.2 A 2.0 0.2 A Mean graminoid height (cm) Nongrazed 54.0 3.3 A 40.0 3.5 A 0.090 Grazed 72.1 4.8 A 43.2 5.0 A Variance of litter cover (%) Nongrazed 1.3 0.3 A 1.3 0.3 A 0.011 Grazed 0.6 0.2 A 1.6 0.4 A Mean litter depth (cm) Nongrazed 8.1 0.8A 2.9 0.7A 0.087 Grazed 11.2 1.1 A 3.2 0.8 A Mean soil density (g/cm 3 ) Nongrazed 1.3 0.0 A 1.3 0.0 A 0.042 Grazed 1.2 0.0A 1.4 0.2A a Only habitat characteristics significantly affected by a treatment grazing interaction presented ( P b Means in a row followed by the same uppercase letter no t significantly different ( P > 0.1)

PAGE 52

52 Table 2 4 Effects of treatment season interactions on habitat characteristics of flatwoods in Florida, 2007 2008. Habitat characteristics by treatmenta Season Treatment ( SE) b P C ontrol Treated Dormant season burn Mean graminoid height (cm) Winter 65.3 .5.2 A 41.6 4.9 B Spring 51.8 4.2A 44.9 3.9A Summer 58.0 4.5 A 57.0 3.2 A Growing season burn Variance of soil pH Winter 0.1 0.0 A 0.1 0.0 A 0.042 Spring 0.1 0.0 A 0.1 0.0 A Summer 0.1 0.0 A 0.1 0.0 A Dormant season roller chopping Mean graminoid height (cm) Winter 57.6 5.2 A 40.2 3.1 B 0.051 Spring 44.2 2.2 A 39.7 1.8 A Summer 49.7 2.1 A 49.6 2.5 A Growing season roller chop Mean graminoid height (cm) Winter 56.1 5.8 A 43.4 3.9 B 0.014 Spring 58.8 2.5 A 4 4.1 2.3 B Summer 58.9 1.8A 49.0 4.0A Mean litter cover (%) Winter 4.9 0.3 A 5.0 0.3 A 0.049 Spring 4.9 0.4 A 4.4 0.4 A Summer 5.4 0.2 A 4.4 0.3 A Roller chop/burn Mean forb height (cm) Winter 35.9 2.1 A 30.6 3.9 A 0. 061 Spring 37.7 2.8 A 29.1 4.3 A Summer 36.2 0.7A 41.7 3.9A Mean forb height (cm) Winter 35.9 2.1 A 30.6 3.9 A 0.061 Spring 37.7 2.8 A 29.1 4.3 A Summer 36.2 0.7A 41.7 3.9A Mean forb height (cm) Winter 35.9 2.1 A 30.6 3.9 A 0.061 Spring 37.7 2.8 A 29.1 4.3 A Summer 36.2 0.7 A 41.7 3.9 A Mean graminoid cover (%) Winter 2.9 0.2A 1.8 0.2B 0.044 Spring 2.4 0.2 A 35.0 4.1 A Summer 2.2 0.2 A 2.1 0.2 A Mean graminoid height (cm) Winter 71. 9 8.0A 30.9 7.1B 0.020 Spring 56.3 5.4 A 35.0 4.1 B Summer 61.1 3.4 A 52.5 3.5 A a Only habitat characteristics significantly affected by a treatment season interaction presented (P b Means in a row followed by the same uppercase l etter not significantly different ( P > 0.1)

PAGE 53

53 Table 2 5 Comparison of the effects of treatment type on habitat characteristic s of flatwoods in Florida, 2007 2008. Habitat characteristics a Treatment type ( SE) b P Control Dormant burn Gr owing burn Dormant roller chop Growing roller chop Roller chop/burn Understory richness (no. of spp.) 0.3 0.0A 0.2 0.1B 0.4 0.1A 0.1 0.0AB 0.0 0.0AB 0.0 0.0AB 0.019 Understory density (n o./ha) 57.7 14.5 A 28.1 9.3 B 20.4 6.5 AB 3.8 1.5 B 1.7 1.7 AB 0.0 0.0 AB 0.010 Understory cover (%) 0.2 0.1 A 0.0 0.0 B 0.3 0.2 AB 0.0 0.0 A 0.0 0.0 AB 0.0 0.0 A 0.020 Variance of litter depth (cm) 30.4 3.0 A 21.6 12.1 AB 10.1 7.2 AB 9.0 1.9 B 32.1 23.2 AB 10.0 5.0 B 0.097 a Only habitat characteristics significantly affected b y treatment type presented ( P b Means in a row followed by the same letter not significantly different ( P > 0.1)

PAGE 54

54 Table 2 6 Comparison of the effects of treatment type time interactions on habitat characteristic s of flatwoods in Florida, 2007 20 08. Habitat characteristics a Time b Treatment type ( SE) c P Control Dormant burn Growing burn Dormant roller chop Growing roller chop Roller chop/burn Visual obstruction (% from 5 m distance) 1 25.5 1.4 A 15.2 3.2 B 8.6 2.0 B 6.7 1.2 B 4.6 1.0 B 8.2 2.1 B 2 24.0 1.0 A 26.3 1.8 AB 21.5 1.5 AC 14.6 1.6 BD 7.8 1.0 D 17.5 2.2 BCD Visual obstruction (% at 10 m distance) 1 35.6 1.7 A 21.6 2.4 B 18.9 4.6 B 11.6 1.3 B 6.9 1.5 B 12.6 2.3 B 2 33.9 1.3A 34.2 1.8AB 29.5 1.7AB 20.5 1.5B 10.8 1.2B 22.7 2.2AB Shrub height (cm) 1 115.2 3.9 A 95.6 7.1 B 51.2 5.7 B 69.2 4.1 B 59.4 6.5 B 63.8 2.0 AB 0.013 2 118.7 3.0 A 125.3 3.7 A 92.2 3.4 AB 84.8 3.4 B 60.9 2.7 B 85.2 3.6 AB Shrub cover (%) 1 60.1 2.7A 39.1 4.6B 16.2 5.9B 19.4 4.0B 9.4 5.1B 20.1 4.7B 2 61.5 2.1 A 67.9 3.2 A 76.6 4.3 A 35.5 3.6 B 15.7 3.4 B 36.5 4.6 A Mean forb richness (n o. of spp.) 1 3.9 0.2 A 3.8 0.5 A 1.2 0.4 A 3.9 0.3 A 3.2 0.6 B 4.0 0.4 A 0.042 2 4.0 0.2 A 4.2 0.4 B 3.0 0.4 A 5.0 0.3 AB 4.4 0.3 AB 5. 5 0.5 AB Mean forb cover (%) 1 1.6 0.1AC 1.5 0.1C 0.8 0.2B 1.5 0.1A 1.4 0.2B 1.6 0.2C 0.042 2 1.8 0.2 A 1.8 0.1 B 1.3 0.1 A 2.0 0.1 AB 1.9 0.2 AC 1.7 0.1 BC Variance of forb height (cm) 1 251.4 25.5 A 176.2 42.5 B 44.1 21.0 A 164.2 57.0 A 121.5 42.7 AB 101.5 26.2 B 0.024 2 234.0 19.3A 253.0 48.0B 157.9 37.3AB 142.5 21.0AB 172.6 38.3AB 286.5 63.0AB Mean graminoid richness (no. of spp.) 1 3.6 0.1 AB 3.1 0.2 AB 2.2 0.4 A 3.4 0.1 B 3.3 0.4 AB 37.1 4.5 AB 0. 005 2 3.7 0.1 A 4.1 0.2 A 3.2 0.2 B 4.2 0.3 A 4.6 0.3 A 3.9 0.4 A Mean graminoid height (cm) 1 57.3 1.9 A 39.5 3.4 B 31.6 5.4 B 40.4 1.8 B 36.6 3.6 B 37.1 4.5 AB 2 57.4 1.5 A 56.5 2.6 B 52.7 3.1 BC 45.9 2.4 C 49.6 2.5 BC 43.9 3.8 ABC Mean litter cover (%) 1 5.6 0.3 A 3.4 0.3 B 2.7 0.6 C 5.3 0.3 B 4.9 0.3 AB 2.6 0.3 BC 0.052 2 5.1 0.1 A 4.5 0.2 B 3.1 0.3 C 4.5 0.2 B 4.4 0.2 B 3.5 0.3 B Mean litter depth (cm) 1 9.9 0.6A 4.8 2.9B 1.2 0.4B 6.2 0.6C 8.5 2.1ABC 1.7 0.5BC 0.001 2 8.8 0.4 A 6.3 0.7 B 2.9 1.0 C 3.7 0.4 D 5.1 0.8 BD 3.9 0.7 BD a Only habitat characteristics significantly affected by a treatment type time interaction presented ( P b Time since treatment application (years). c Means in a row followed by the same letter not significantly different ( P > 0.1).

PAGE 55

55 Table 2 7 Comparison of the effects of treatment type grazing interactions on habitat characteristics of flatwood s in Florida, 2007 2008. Habitat characteristics a Grazing Treatment type ( SE) b P Control Dormant burn Growing burn Dormant roller chop Growing roller chop Roller chop/burn Visual obstruction (% from 5 m distance) Grazed 26.0 1.3 A 20.2 2.3 A 25.5 2.6 A 9.4 1.2 BC 5.8 0.9 C 14.1 2.1 AB Nongrazed 23.1 1.1 A 23.5 3.3 A 16.2 1.7 AB 12.3 2.0 AB 7.6 1.2 B 13.8 2.6 AB Visual obstruction (% at 10 m distance) Grazed 35.5 1.5 A 26.4 1.7 AB 35.0 3.6 A 14.2 1.6 B 8.9 1.5 B 19.4 2.5 AB 0.020 Nongrazed 33.5 1.5A 34.5 4.6A 24.5 2.0AB 18.5 1.8AB 10.3 1.3B 18.3 2.5AB Overstory richness (n o. of spp.) Grazed 0.5 0.0 AB 0.8 0.1 A 0.5 0.2 AB 0.0 0.0 B 0.0 0.0 AB 0.8 0.1 A Nongrazed 0.5 0.1 A 0.5 0.2 A 0.4 0.1 A 0.6 0.1 A 0.6 0.2 A 0.4 0.1 A Over story density (No./ha) Grazed 32.2 4.7AB 56.5 7.7A 33.3 12.6AB 0.0 0.0B 0.0 0.0AB 26.3 3.2A 0.002 Nongrazed 28.5 4.5 A 33.3 10.1 A 22.6 6.2 A 26.4 6.3 A 44.0 13.4 A 12.0 3.3 A Overstory basal area (cm2/ha) Grazed 2.6 0.9 AB 3.0 0.3 A 0.6 0.2 AB 0.0 0.0 B 0.0 0.0 AB 0.9 0.2 A 0.005 Nongrazed 1.0 0.2 A 2.2 0.7 A 0.9 0.3 A 1.4 0.4 A 0.9 0.3 A 0.5 0.2 A Overstory cover (%) Grazed 6.4 0.9AB 10.3 1.5A 3.0 1.4AB 0.0 0.0B 0.0 0.0AB 7.7 1.3A Nongrazed 7.6 1.1 A 3.8 1.6 A 4.4 1.2 A 7.5 2.0 A 6.2 2.2 A 1.9 1.0 A Shrub richness (n o. of species) Grazed 5.3 0.2 ABC 6.2 0.3 C 4.8 0.6 AB 3.7 0.3 B 3.9 0.5 AC 6.1 0.5 C 0.019 Nongrazed 5.8 0.2A 4.6 0.4AB 5.0 0.6B 4.2 0.3AB 4.6 0.5AB 5.1 0.4A Shrub density (No./m2.) Grazed 4.0 0.2 A 5.3 0.6 B 3.4 0.5 ABC 4.1 0.8 AB 1.4 0.3 C 3.8 0.5 AB 0.015 Nongrazed 3.7 0.2 A 3.8 0.5 A 5.6 0.5 A 2.5 0.3 A 3.0 0.6 A 2.9 0.4 A Shrub height (cm) Grazed 111.8 3 .0 A 104.2 4.6 A 92.5 8.7 AB 71.9 3.2 B 65.9 3.4 B 82.1 4.6 AB 0.010 Nongrazed 123.0 3.6 A 117 8.3 A 79.0 4.6 B 85.6 4.6 B 57.5 3.4 B 73.0 4.6 B Mean forb richness (n o. of spp.) Grazed 3.8 0.2 A 4.4 0.3 B 2.7 0.8 A 4.6 0.4 AB 3.8 0.4A B 5.6 0.5 AB Nongrazed 4.0 0.2 A 2.3 0.4 A 2.5 0.3 A 4.3 0.3 A 4.3 0.4 A 4.4 0.5 A Mean forb cover (%) Grazed 1.5 0.0A 1.7 0.1B 1.1 0.2A 1.7 0.1ABC 1.4 0.1AC 1.9 0.1BC Nongrazed 2.0 0.2 A 1.1 0.1 A 1.2 0.1 A 1.7 0 .1 A 2.0 0.3 A 1.5 0.1 A Variance of forb height (cm) Grazed 258.3 21.2 A 176.8 31.5 B 210.4 68.0 AB 111.8 21.2 AB 111.2 37.4 AB 191.7 65.6 AB 0.017 Nongrazed 222.8 22.3A 376.0 91.3A 106.4 32.0A 200.9 57.4A 186.6 42.2A 232.8 55.4A Variance of graminoid richness (n o. of spp.) Grazed 1.1 0.1 AC 1.4 0.2 ABC 0.7 0.2 C 1.4 0.4 AB 1.1 0.3 ABC 1.7 0.3 B 0.048 Nongrazed 1.4 0.2 A 0.8 0.3 A 1.4 0.6 A 1.9 0.4 B 1.3 0.2 A 1.0 0.2 A Variance of graminoid height (cm) Grazed 301.5 39.5 A 204.1 36.8 A 299.9 64.6 A 177.3 40.4 A 95.6 18.1 A 248.8 86.0 A 0.007 Nongrazed 347.7 43.8 A 76.3 23.0 A 121.6 27.2 A 331.4 98.0 A 236.1 60.5 A 339.7 124.5 A

PAGE 56

56 Table 2 7 Continued Habitat characteristics a Grazing Treatment typ e ( SE) b P Control Dormant burn Growing burn Dormant roller chop Growing roller chop Roller chop/burn Mean litter cover (%) Grazed 5.3 0.2A 4.1 0.2B 1.6 0.3C 4.7 0.6BC 4.4 0.4BC 3.4 0.4B 0.014 Nongrazed 5.2 0.1 A 3.2 0.3 B 3.4 0 .3 B 5.1 0.1 B 4.6 0.2 B 3.0 0.2 B a Only habitat characteristics significantly affected by a treatment type grazing interaction presented (P b Means in a row followed by the same letter not significantly different (P > 0.1). Table 2 8 Comp arison of the effects of treatment type season interactions on habitat characteristics of flatwoods in Florida, 2007 2008. Habitat characteristics a Season Treatment type ( SE) b P Control Dormant burn Growing burn Dormant roller chop Growing roller chop Roller chop/burn Mean graminoid height (cm) Winter 60.4 2.8 A 41.6 4.8 B 52.2 7.2 A 40.2 3.1 B 43.4 4.0 B 31.0 7.1 B 0.004 Spring 52.6 2.0 A 44.9 4.0 AB 45.3 5.0 AB 40.0 2.0 B 44.1 2.3 B 35.0 4.1 AB Summer 59.0 1.5 A 57.0 3.2 AB 46.0 4.4 AB 65.1 5.0 B 49.0 4.0 AB 52.5 3.5 AB a Only habitat characteristics significantly affected by a treatment type season interaction presented ( P b Means in a row followed by the same letter not significantly different ( P > 0.1)

PAGE 57

57 CHAPTER 3 SEASONAL EFFECTS OF PRESCRIBED BURNING A ND ROLLER CHOPPING O N SAW PALMETTO IN FLAT WOODS Introduction In rangeland habitats across the United States (U S ), including the southwestern arid and subarid grasslands, central and northern tall and mi xed grass prairies, and southeastern pine savannas, a reduction in disturbance, most often fire, has resulted in shrub encroachment (Collins and Gibson 1990, Van Auken 2000, Lett and Knapp 2003). This encroachment has resulted in the degradation of many o f these disturbance maintained rangelands systems and reduced their value for native wildlife species, particularly birds associated with more open herbaceous dominated habitats (Lloyd et al. 1998, Madden et al. 1999, Hunter et al. 2001). Saw palmetto ( Ser enoa repens [Bartr.] Small), a low growing, branched, fan palm, has become a dominant shrub species in many pine savanna habitats of the southeastern Coastal Plain. This includes the pine flatwoods, a rangeland community with an open pine overstory and an ofte n rich herbaceous layer (Hilmon 1968). Pine flatwoods cover approximately 50% of the land area of Florida (Abrahamson and Hartnett 1990). Unfortunately, due to shrub invasions, particularly of saw palmetto, large areas of this pine savanna habitat are in poor condition and currently exist in a highly degraded state (Means 1996, Florida Fish and Wildli fe Conservation Commission 2005). Historically, flatwoods habitats were maintained by frequent, low intensity, lightning ignited fires during the May July thunderstorm season. These fires prevented encroachment by saw palmetto, a species that is able to spread prolifically in the absence of this disturbance (Hilmon 1968, Komarek 1968, Abrahamson and Hartnett 1990, Pyne et al. 1996). During the past 50 years, fire suppression, reductions in fire

PAGE 58

58 frequency, or a shift in fire season, commonly a result of human intervention, have resulted in excessive saw palmetto growth on many pine flatwoods sites across Florida. On Floridas pine flatwoods, increases i n saw palmetto height, cover, and density have become a concern, potentially resulting in the loss of many grass and forb species and declines in the species rich herbaceous ground layer (Wade et al. 1980, Huffman and Blanchard 1991, Robbins and Myers 1992 Olson and Platt 1995). Such changes threaten the integrity of pine flatwoods and their suitability for many wildlife species of conservation concern. These include a variety of mammals, birds, amphibians, and reptiles such as Shermans fox squirrel ( S ciurus niger shermanii L. ), Florida black bear ( Ursus americanus floridanus Merriam ), red cockaded woodpecker ( Picoides borealis Vieillot ), Bachmans sparrow ( Aimophila aestivalis Lichtenstein), flatwoods salamander ( Ambystoma cingulatum Cope), gopher frog ( Rana capito LeConte), and gopher tortoise ( Gopherus polyphemus Daudin ; FWC 2005). They have also resulted in declines in forage quantity and quality, potentially reducing the value of these areas for livestock production (Hilmon 1968, Moore 1974, Tanner et al. 1988). In recent years, a common goal among managers of pine flatwoods has been to reduce the proliferation of shrubs, particularly saw palmetto. For most Florida landowners, many of whom are cattle ranchers, the objective of shrub reduction is to increase the growth and production of more palatable grasses and forbs as food for livestock (Yarlett 1965, Moore 1974, Kalmbacher and Martin 1984, Tanner et al. 1988). Concomitantly, most wildlife species that occupy pine flatwoods habitats benefit from increases in groundcover of grasses and forbs, as they provide diverse food and cover resources (Huber and Steuter 1984, Madden et al. 1999).

PAGE 59

59 T he United States Department of Agricultures Environmental Quality Incentives Program and Wildlife Habitat Incentives Program are currently promoting management activities thought to reduce saw palmetto and other shrub species, and maintain areas of pine flatwoods in Florida. These programs provide landowners financial and technical assistance to implement manageme nt activities, including prescribed burning and roll er chopping during dormant ( Novembe r March) and growing ( April September) seasons. Prescribed burning and roller chopping can reduce shrubby vegetation in southeastern rang eland habitats (Wade et al. 1980, Kalmbacher and Martin 1984, Tanner et al. 1988, Glit zenstein et al. 1995, Watts and Tanner 2003). A number of studies have compared the effects of season of burning on shrub regeneration and growth, suggesting growing season burning results in greater r eductions than dormant season burning (Robbins and Myers 1992). In addition, roller chopping has been shown to cause considerable declines in shrub cover. However, the results of some of these studies are contradictory (Lewis 1970, Moore 1974, Tanner et al 1988 Watts and Tanner 2003, Watts et al. 2006). S tudies that compare shrub responses to prescribed burning and roller chopping are few, as are those that specifically examine effects on saw palmetto, a problem species in many habitats (Watts and Tan ner 2003, Watts et al. 2006). In addition, s tudies that have been conducted are extremely localized, typically being confined to a single study area. If we are to make general recommendations on the use of these practices to individuals managing pine flatwoods across the state, we need detailed research that compares saw palmetto response to prescribed burning and roller chopping practices over a larger area. Therefore, the objective of my study was to

PAGE 60

60 fill recognized gaps in our understanding of how prescribed burning, roller chopping, and combinations of the two, applied during different seasons, affect the height, cover, and density of saw palmetto over a broad range of pine flatwoods habitats. The study was unique in that it examined saw palmetto res ponse to these practices on pine flatwoods sites over a broad geographic area. The intention was to obtain information representative of the real world where landowners use a range of burning and roller chopping techniques to control different levels o f saw palmetto under a variety of conditions. Methods Study Sites I conducted research on 50 privately and publicly owned, paired treatment and control sites across 6 counties (Desoto, Highlands, Lee, Manatee, Osceola, and Sarasota) in central and south F lorida. Study sites consisted of pine flatwoods habitat s with varying management histories and grazing regimes that were being prescribed burned and roller chopped by local landowners and land managers using varying, individual protocols. Floridas pine flatwoods are characterized as having an overstory of scattered slash ( Pinus elliotti Engelm.) and longleaf ( P. palustris Mill. ) pine, either in pure stands or in combination. The understory and shrub layer includes saw palmetto, wax myrtle ( Morella cerif era [L.] Small), gallberry ( Ilex glabra [Pursh] Chapm.), fetterbush ( Lyonia lucida [Lam.] K. Koch), staggerbush ( Lyonia fruticosa [Michx.], G. S. Torr), dwarf huckleberry ( Gaylussacia dumosa [Andrews] Torr. & A. Gray), dwarf live oak ( Quercus mimima [Sarg. ] Small), and tarflower ( Bejaria racemosa Vent.). An appreciable herbaceous layer exists when the shrub layer is relativel y open. This layer contains a wide variety of grasses ( e.g., Agrostis Andropogon Aristida Eragrostis

PAGE 61

61 Panicum, and Paspalum spp. ) Common forbs include legumes ( e.g., Cassia Crotalaria Galactia Tephrosia spp.), milkweeds ( Asclepias spp.), milkworts ( Polygala spp.), and a wide variety of composites ( e.g., Aster, Chrysopsis, Eupatorium, Liatris, and Solidago sp p. ; Abrahamson and H artnett 1990, U S Fish and Wildlife Service 1999). Treatment Types Treatment types included dormant season ( November March) burn, growing season ( April October ) burn, dormant season roller chop, growing season roller chop, and a roller chop/burn combinati on treatment. The roller chop/burn combination treatment (hereafter referred to as roller chop/burn) involved roller chopping in the dormant season followed by burning within 6 months. A total of 11 dormant season burn, 9 growing season burn, 9 dormant s eason roller chop, 12 growing season roller chop, and 9 roller chop/burn sites were established, each paired with and adjacent untreated control Saw Palmetto Sampling I used a pairedsample approach to assess the effects of management treatments (i.e., pr escribed burning, roller chopping, and combinations of the two) on saw palmetto height, cover, and density. These 3 saw palmetto variables were compared between sampling points randomly located in paired treated (e.g., dormant season burned) and untreated (control) flatwoods sites. I randomly located control sampling points in untreated sites adjacent to those in treated sites. These control sites were of similar current and past management (e.g., grazing intensity), surrounding landuse, plant community (e.g., overstory cover), and soil conditions, to treated sites and were located in the same pasture or management unit. Within each site, I established 1 randomly

PAGE 62

62 selected treatment or control sampling point. I rejected and randomly relocated sampling p oints that occurred within 50 m of the edge of a site to minimize edge effects Sites, within which treatment and control sampling poi nts were located, ranged from 2 20 ha in size. Saw palmetto height, cover, and density at each sampling point were assess ed following treatment, once in winter (February March), spring (April May), and summer (July August), during each of 2 years (2007 2008). I counted all saw palmetto st ems within 2 perpendicular 20m2 quadrats centered on the sampling point to determine d ensity (No./m2; Hays et al. 1981, Bullock 1996, Higgins et al. 2005). Within these quadrats, I also obtained a single measurement of maximum saw palmetto height (cm). I assessed saw palmetto cover (%) along 2 perpendicular 2 0 m transects centered on the sampling point using the line intercept method (Stephenson and Buell 1965 Hays et al. 1981, Higgins et al. 2005). Analyses I performed repeated measures analyses using mixed model regressions, with season and time since treatment (time) as repeated measures and study site pair as a blocking factor, followed by Fishers Protected LSD tests, to examine differences in saw palmetto height, cover, and density between untreat ed (control) and treated sites. Difference were examined within (e.g., dormant season burn) and among (i.e., dormant season burn, growing season burn, dormant season roller chop, growing season roller chop, and roller chop/burn) treatme nt types. In my results and discussion, I focus on treatment rather than repeated measures effects. I pr esent res ults for treatment alone and 2and 3way treatment interactions in the text (Tables 3 1, 3 2, and 33 ). Due to the timing of data collection, it was not possible to test for 3way interactions for growing

PAGE 63

63 season burning and roller chopping treat ments. When differences in linear combinations of groups or biologically meaningless comparisons (e.g., cover in dormant burn site in year 1 versus cover in control site in year 2) arose, I stated that post hoc comparisons revealed no differences based on treatment and t he interacting factor. As 3 way interactions are difficult to reliably interpret, they were not discussed further (Zar 1999). All data sets were rank transformed prior to analyses due to violations of normality and homogeneity of variance assumptions (Conover 1998, Zar 1999 SYSTAT 2007). I concluded s tatistical significance at P than the more common P to minimize the probability of making a Type II error ( Mapstone 1995, Zar 1999) All statistical tests were performed using SYSTAT (2007) statistical software. Results Dorma nt Season Burn Saw palmetto height was affected by a dormant season burning time interaction (Table 3 2). Differences in height were not observed between dormant season burn and control sites the first or second year following treatment. However, saw palmetto height increased by 20% on burn sites from the first to the second year of the study. A dormant season burning time interaction also affected saw palmetto cover (Table 3 2), which was 46% lower on burn than control sites the first year following treatment. No differences in saw palmetto cover between dormant season burn and control sites were observed the second year following treatment, with cover on burn sites increasing by 51% from the first to the second year of the study. Dormant season burning alone and

PAGE 64

64 burning time, burning season, and burning season time interactions had no impact on saw palmetto density ( P Growing Season Burn Growing season burning alone affected saw palmetto height (Table 3 1), which was 23% lower on burn than control sites. Saw palmetto cover was affected by a growing season burning year interaction (Table 3 2), being 79% low er on growing season burn than control sites the first year following treatment. No differences in saw palmetto cover between growing season burn and control sites were observed the second year following treatment, with cover on burn sites increasing by 80% from the first to the second year of the study. Growing season burning alone and burning time, burning season interactions had no impact on saw palmetto density ( P Dormant Season Roller Chop Dormant season roller chopping alone affected saw palmetto height (Table 3 1), which was 28% lower on dormant season roller chop than control sites. Saw palmetto cover was also affected by dormant season roller choppi ng alone (Table 3 1), being 56% lower on dormant season roller chop than control sites. Saw palmetto density was affected by a dormant season roller chopping year interaction (Table 3 2). However, examination of post hoc comparisons revealed no differences in density based on roller chopping and year. Growing Season Roller Chop Growing season roller chopping alone affected saw palmetto height and cover (Table 3 1). Saw palmetto height was 40% lower on roller chop than control sites, and cover 70% lower on roller chop than control sites. Saw palmetto density was also

PAGE 65

65 affected by growingseason roller chopping alone (Table 3 1), being 27% lower on roller chop than control sites. Roller Chop/Burn A roller chopping /burning year interaction affected saw p almetto height (Table 3 2), which was 38% lower on roller chop/burn than control sites the first year following treatment. The second year following treatment, saw palmetto height was 23% lower on roller chop/burn than control sites. Saw palmetto cover w as affected by a roller chopping/burning season year interaction ( P = 0.030). R oller chopping /burning alone, and roller chopping /burning time, roller chopping /burning season, and roller chopping /burning time season interactions had no effect on saw palmetto density ( P Treatment Type Comparisons A treatment year interaction affected saw palmetto height (Table 3 3 ). The first year following treatment, height was lower on dormant and growing season burn, dormant and growi ng season roller chop, and roller chop/burn than control sites, but the effects of the 5 active treatments were similar. The second year following treatment, dormant season burning had no effect on saw palmetto height. However, height was lower on growing season burn, dormant and growi ng seaso n roller chop, and roller chop/burn than control si tes, with growing season roller chopping causing the greatest reduction. Saw palmetto height increased on dormant and growing season burn and roller chop/burn sites between the first and second year of the study. Saw palmetto cover was also affected by a treatment year interaction (Table 3 3). The first year following treatment, cover was lower on dormant and growing season burn, dormant and growi ng season roller chop, and roller chop /burn than control sites,

PAGE 66

66 with similar effects of the 5 active treatments. The second year following treatment, dormant s eason burning had no effect on saw palmetto cover. However, cover was greater on growing season burn compared to control sites, and lower on dormant a nd growing season roller chop and roller chop/burn than control si tes, with growing season roller chopping causing the greatest reduction. Saw palmetto cover increased on dormant and growing season burn and dormant season roller chop sites between the fir st and second year of the study. Treatment type alone affected saw palmetto density (Table 3 3). Density was greater on dormant and growing season burn compared to control sites, but lower on dormant and growing season roller chop and roller chop /burn com pared to control sites. The greatest reduction in saw palmetto density was observed on sites s ubject to growing season roller chopping. Discussion This study suggests a single dormant season burn is unlikely to be sufficient to reduce saw palmetto in pine flatwoods habitats, particularly when growth is dense and the cover excessive. Dormant season burning had no effect on saw palmetto density or height, and while cover was lower following dormant season burning, reductions were short lived. Other studies suggest dormant season burning has minimal medium to long term effects on many shrubs, with cover returning to preburn levels within 12 months post burn (Abrahamson 1984a, b; Fitzgerald 1990; Robbins and Myers 1992; Glitz enstein et al. 1995). As a fire adapted species with rhizomes serving as large carbohydrate reserves and meristematic tissues that are protected f rom the flames during a burn, saw palmetto is able to resprout vigorously after fire and regain up to 80% of its preburn canopy cover after j ust 1 growing season (Hilmon 1968). Typically,

PAGE 67

67 on pine flatwoods, in situations where rapid, significant shrub reduction is a primary goal, alternative management techniques should be utilized in place of or in addition to dormant season burning. However regular dormant season burning with a 2 3 year return interval, where fuel loads permit, has been found sufficient to keep saw palmetto crow n size small (Hilmon and Hughes 1965) and prevent saw palmetto proliferation. Therefore, this treatment is likely to be beneficial in reducing growth in situations where the shrub has not proliferated and maintenance at current levels is a goal. A fire return interval of palmetto cover (Hough 1968). Growing season burning also had no effect on saw palmetto density. However, it appeared to cause greater reductions in saw palmetto height and cover in the short term than did dormant season burning. Overall, treatment type comparisons suggest potentially greater and extended effects of growing season burning on shrub height and cover than dormant season burning. Studies examining shrub communities indicate regrowth is lower after growing than dormant season burning because of reduced shrub resprouti ng (Lewis and Harshbarger 1976, Fitzgerald 1990, Huffman and Blanchard 1991, Olson and Platt 1995, Drewa et al. 2002). The effects of growing season burning on saw palmetto, although slightly greater than for dormant season burning, were still relatively temporary. However, the effects of both dormant and growing season roller chopping were greater and more prolonged than for either burn treatment. Dormant and growing season roller chopping resulted in lower saw palmetto height and cover than on control sites. However, saw palmetto density was only lower on growing season burn than control sites. In general, the effects of these treatments persisted through the duration of this study, although

PAGE 68

68 comparisons suggest saw palmetto height and cover were lower on growing than dormant season roller chop sites. The longer term effects of these 2 treatments are largely what separated them from the practice of growing season burning. R oller chopping can result in considerable decreases in shrub abundance and cover, and is useful on poor conditi on pine f latwoods with dense shrub cover, being particularly useful as a form of saw palmetto control (Moore 1974, Tanner et al. 1988). An 83% kill and reduction in crown cover to 3% of pretreatment levels has been reported 2 years after practice impleme ntation (Moore 1974), and a 70% kill and reduction in crown cover to 1 2% of pretreatment levels 5 year s after treatment (Hilmon 1968, Lewis 1970). However, many studies do not specifically examine effects of roller choppi ng on saw palmetto density. My results contradict those of a study which found saw palmetto density was lower on dry prairie roller chopped in the dormant versus the growing season (Watts and Tanner 2003, Watts et al. 2006) Chopping during periods of high soil moisture, i.e., during t he growing season, may result in greater shrub reduction due to the greater penetration of the equipments blades and deeper severing of plant roots (Moore 1974, Tanner et al. 1988). The additional stress placed on saw palmetto when roller chopped during the growing season, a period when it is actively producing new roots and stems, may result in greater reductions when compared to roller chopping conducted during the dormant season. Roller chopping, in either the dormant or growing season, is likely to be the best method for initial, rapid treatment and control of saw palmetto in high density and cover areas. It also has potential for use controlling saw palmetto on sites near urban developments and recreational areas, where smoke management issues may aris e during and following burning.

PAGE 69

69 Similar to dormant and growing season roller chopping, roller chopping /burning resulted in prolonged lowering of saw palmetto height compared to control sites. However, this treatment had no effect on saw palmetto densit y and its effect on saw palmetto cover was unclear due to an interaction with season and year. The effects of roller chopping /burning on plant communities have not been well studied. Other studies have found a reduction in saw palmetto density on burned and roller chopped plots (Watts and Tanner 2003, Watts et al. 2006). Burning prior to roller chopping may remove some of the above ground shrub biomass making subsequent roller chopping more effective in terms of shrub r eduction (Kalmbacher and Martin 198 4). During this study, combination treatment sites were burned after roller chopping and this may be why decreases in density were not observed, as fire can stimulate saw palmetto sprouting (Hilmon 1968). Further study of roller chopping /burning combinat ion treatments is recommended to determine if there are differences based on timing of burning and roller chopping practices. Management Implications The single application of a dormant or growing season prescribed burn will likely be insufficient to reduc e saw palmetto growth in areas where it has proliferated. In such situations, burn effects on saw palmetto height and cover are likely to be minor and short lived, with no observable effect on density. Alternative management practices such as roller chopping will likely be needed in combination with, or in place of, dormant and growing season burning if there is a need to quickly and significantly reduce excessive saw palmetto height, cover, and density. If rapid, more prolonged control and greater reduct ions in saw palmetto height, cover, and density are desired, dormant or growing season roller chopping should be

PAGE 70

70 conducted. These practices result in decreases in saw palmetto height and cover for at least 2 years post treatment. If reductions in shrub d ensity are desired, growing seaso n roller chopping may be the only suitable treatment.

PAGE 71

71 Table 3 1 Effects of treatment on saw palmetto height, cover, and density in Florida flatwoods, 2007 2008. Saw Palmetto Variables by Treatment a Treatment ( SE) P Control Treated Growing season burning Cover (%) 60.5 2.8 46.7 4.3 0.001 Dormant season roller chop Height (cm) 101.8 3.0 73.4 2.5 Cover (%) 45.2 1.7 20.0 2.0 Growing season roller chop Height (cm) 95.8 3.8 57.4 2.7 Cover (% ) 34.6 2.2 10.1 1.6 Density (n o./m 2 ) 1.5 0.0 1.1 0.1 0.032 R oller chop /burn Height (cm) 105.2 3.7 74.8 3.2 Cover (%) 50.0 3.1 20.2 2.1 a Only saw palmetto variables sign ificantly affected by treatment presented ( P 0.1). Table 3 2 Effects of treatment time interactions on saw palmetto height, cover, and density in Florida fla twoods, 2007 2008. Saw Palmetto Variables by Treatment a Time b Treatment ( SE) c d P Control Treated Dorman season burn Height (cm) 1 96.6 7.2 Aa 84.4 4.8 Aa 0.073 2 114.2 4.4 Aa 118.2 4.2 Ab Cover (%) 1 45.2 4.7 Aa 24.4 3.8 Ba 0.021 2 51.2 4.4 Aa 50.2 3.9 Ab Growing season burn Cover (%) 1 57.5 5.4 Aa 12.1 3.6 Ba 0.005 2 61.5 3.3 Aa 58.2 3.3 Ab Dormant season roller chop Density (n o./m 2 ) 1 1.7 0.1 Aa 1.2 0.1 Aa 0.023 2 1.4 0.1 Aa 1.4 0.1 Aa R oller chop/burn Height (cm) 1 100.8 7.9 Aa 62.1 4.8 Ba 0.041 2 108.0 3.5 Aa 82.8 3.5 Ba a Only saw palmetto variables significantly affected by a treatment time interaction presented ( P 0.1). b Time since treatment application (years) c Means in a row foll owed by the same uppercase letter not significantly different ( P > 0.1). d Means in a column followed by the same lowercase letter not significantly different ( P > 0.1).

PAGE 72

72 Table 3 3 C omparisons of the effects of treatment and treatment time interactions on saw palmetto height, cover, and density in Florida flatwoods, 2007 2008. Saw Palmetto Characteristics by Treatment and Interaction a Time b Treatment Type ( SE) b,c P Control Dormant Season Burn Growing Burn Dormant Season Roller Chop Growing Seas on Roller Chop Roller Chop/Burn Treatment Density (No./m 2 ) N/A 1.6 0.0A 1.7 0.1B 2.2 0.1B 1.3 0.1C 1.1 0.1D 1.5 0.1C Treatment Year Height (cm) 1 98.3 3.3 Aa 84.4 4.8 Ba 51.2 5.7 Ba 118.2 4.2 Ba 59. 1 6.4 Ba 62.1 4.8 Ba 2 106.1 1.9 Aa 118.2 4.2 ABb 92.2 3.4 BCb 67.9 3.7 Ca 56.8 2.9 Da 82.8 3.5 Cb Cover (%) 1 45.3 2.2Aa 24.4 3.8Ba 12.1 3.6Ba 12.6 1.8Ba 6.9 3.0Ba 11.5 2.0Bb 2 48.7 1.6 Aa 50.2 3.9 ABb 58.2 3.3 Bb 26.7 2.8 C b 11.2 1.9 Da 25.7 2.7 Cc a Only saw palmetto variables significantly affected by treatment or a treatment time interaction presented ( P 0.1). b Time since treatment application (years) c Means in a row followed by the same letter not significantly different ( P > 0.1). d Means in a column followed by the same lowercase letter not significantly different ( P > 0.1).

PAGE 73

73 CHAPTER 4 INFLUENCE OF ROLLER CHOPPING AND BURNING ON ARTHROPO D COMMUNITIES OF FLORIDA RANGELANDS Introduction Frequent f ires are considered essential to maintain the structure and diverse herbaceous groundcover of many southeastern rangel and habitats (Christensen 1981, Abr ahamson and Hartnett 1990, Platt 1998). However, in many situations, f ire exclusion, reductions in fir e frequency, and/or a shift in fire season have result ed in excessive shrub growth and declines in the species rich herbaceous ground layer of these habi tats (Wade et al. 1980, Platt et al. 1988, Huffman and Blanchard 1991, Glitzenstein et al. 1999 ), poten tially reducing their value to livestock and certain wildlife species. Depending on season of application roller chopping and prescribed burning have been shown to improve southeastern rangeland condition by reducing the cover of shrubs such as saw palmetto ( Serenoa repens [bartr.] Small ), gallberry ( Ilex glabra [Pursh] Chapm. ), and wax myrtle ( Morella cerifera [L.] Small ) and promoting the growth and seeding of herbaceous groundcover species ( Chapter 2; Wade et al. 1980, Kalmbacher and Martin 1984, Tann er et al. 1988, Glitzenstein et al. 1995, Watts and Tanner 2003). In Florida, the use of these practices in a variety of r angeland habitats is promoted as a mea n s to maintain or enhance wildlife habitat increase livestock forage quantity and quality, and reduce fuel build up and wildfire risk. Arthropods are a critical component of rangeland systems and their management and conservation should be considered when implementing roller chopping and prescribed burning activities. They are major contributor s t o biodiversity and play an important role in ecosystem processes as pollinators and predators, providing benefits to agricultural and livestock producers (Warren et al 1987, Triplehorn and Johnson

PAGE 74

74 1995). In addition, they provide an important food source for rangelandassociated wildlife. In Florida, this includes numerous avian species considered of conservation priority e.g., northern bobwhite ( Colinus virginianus L. ), Bachmans sparrow ( Aim ophila aestivalis Lichtenstein), common grounddove ( Columbina passerina L. ), and grasshopper sparrow ( Ammodramus savannarum Gmelin ; Vickery 1996, Brennan 1999, Bowman 2002, Dunning 2006) In contrast, some arthropods cause extensive damage to grasslands and crops. Nymphal, larval, and adult arthropods can cause considerable injury to the leaves, stems, roots, and reproductive structures of plants (Warren et al. 1987) while some phytophagous species transmit plant diseases (Hardwood and James 1979). In addition, many arthropods are parasites of humans, livestock, and wildlife (Hardwood and James 1979). Despite their importance, the effects roller chopping and prescribed burning have on the arthropod communities of Floridas rangelands have not been extensively studied (Robbins and Myers 1992, Hanula and Wade 2003) In other rangeland systems across the United States arthropod response to fire has been shown to be highly variable and influenced by a variety of factors including order, family or species examined, mobility, life stage at time of burning, burn freque ncy, degree of flame exposure, and reaction to changes in community composition and habitat (Lussenhop 1976; Seastedt 1984; Warren et al. 1987; Anderson et al. 1989; S iemann et al. 1997; Swengel 1996, 1998; Hanula and Wade 2003). Many rangeland, particularly grassland and savanna, arthropods are fire adapted (Evans 1984, Anderson et al. 1989, Siemann et al. 1997). However, the application of prescribed burning to Floridas rangeland habitats may cause alterations

PAGE 75

75 to associated arthropod communities as a r esult of changes in vegetative structure and composition (Herman et al 1998) The response of the vegetative community to roller chopping is also likely to result in alterations to associated arthropod communities. A more comprehensive understanding of how Florida rangeland arthropod communities respond t o roller chopping and prescribed burning is needed to ensure they are used appropriately in situations where arthropod management or conservation are a consideration. The objectives of my study were to 1) c ompare composition (e.g., abundance and richness) of arthropod communities on treated (management activities implemented) and untreated (no management activi ties implemented) pine flatwoods sites during d orman t (November March) and growing ( April Octob er) seasons and 2) e xamine the effects local pine flatwoods habitat characteristics (e.g., shrub density forb cover, and graminoid height) have on the composition (e.g., abundance and richness) of arthropod communities Methods Study Sites I conducted res earch on 50 paired treatment and control pine flatwoods sites with varying management (i.e., roller chopping and prescribed burning) histories and grazing regimes. When grazed, both the treatment and paired control study sites were subject to similar graz ing pressures at similar times. F latwoods sites were located on privately and publically owned lands across a 6 county area (Desoto, Highlands, Lee, Manatee, Osceola, and Sarasota) in central and south Florida, and were being managed by landowners and ma nagers using a variety of prescribed burning and roller chopping protocols Floridas pine flatwoods are rangelands characterized as having a pure or combined overstory stand of scattered longleaf ( Pinus palustris Mill. ) and slash ( P.

PAGE 76

76 elliotti Engelm ) pi ne and an often diverse herbaceous layer. This herbaceous layer is comprised of a wide variety of grasses ( e.g., Agrostis spp. Andropogon spp. Aristida spp. Eragrostis spp. Panicum spp. and Paspalum spp. ). Common forbs include legumes ( e.g., Cassia spp. Crotalaria spp. Galactia spp. Tephrosia spp.), milkweeds ( Asclepias spp.), milkworts ( Polygala spp.), and a range of composites ( e.g., Aster spp. Chrysopsis spp. Eupatorium spp. Liatris spp. and Solidago spp. ). The understory and shrub layer i ncludes saw palmetto, gallberry, wax myrtle, fetterbush ( Lyonia lucida [Lam.] K. Koch), staggerbush ( Lyonia fruticosa [Michx] G. S. Torr ), dwarf huckleberry ( Gaylussacia dumosa [Andrews] Torr. & A. gray ), dwarf live oak ( Quercus mimima [Sarg] Small ), and t arflower ( Befaria racemosa Vent ; Abrahamson and Hartnett 1990, United States Fish and Wildlife Service 1999). Treatment Types Treated sites were subject to 1 of 5 treatment types: dormant season (November March) roller chop, growing season (April October) roller chop, dormant season burn, growing season bur n, or a roller chop/burn combinat ion treatment. The roller chop/ burn combination treatment (hereafter referred to as roller chop/burn) involved roller chopping in the dormant seaso n followed by burning within 6 months. I established a total of 9 dormant season roller chop, 12 growing season roller chop, 11 dormant season burn, 9 growing season burn, and 9 roller chop/burn and control pairs. Arthropod Sampling I used a paired sampling approach to assess the effects of treatment type (i.e., roller chopping, prescribed burning, and roller chopping/burning) on arthropod familial richness and abundance. Richness and abundance were compared between sampling point s randomly located in paired treated (e.g., dor mant season roller chopped) and

PAGE 77

77 untreated (control) flatwoods sites. Paired treatment and control sampling point s were adjacent, being located in the same pasture or management unit, and of similar current and past management (e.g., grazing intensity), su rrounding landuse, plant community (e.g., overstory cover), and soil conditions. Within each treatment and control site, I established 1 randomly selected sampling point To minimize edge effects, sampling point s that fell within 50 m of the edge of a t reatment or control site were rejected and randomly relocated. Sites within which treatment and control sampling points were located ranged from 220 ha. I collected arthropods at each sampling point once in winter (February March), spring (April May), an d summer (July August), during each of 2 years (2007 2008) following treatment. Sub samples of arthropods occupying vegetation less than 30 cm above the ground were taken from within 4 1m2 plots, randomly loc ated in each quadrant of a 0.03ha nested circ ular plot centered on the sampling point ( Dueser and Shugart 1978, Higgins et al. 2005) Arthropods were sampled using a suction s ampler (Wright and Stewart 1993, Ausden 1996). Within each 1 m2 plot, the suction sampler was turned on and systematically m oved around the subsample area, no more than 30 cm above the ground, for a 3minute period collecting arthropods. Suction sampling was not conducted if vegetation was damp or had been flattened by wind, rain, or trampling (Ausden 1996). I separated arthropods collected in each suction subsample from coarse vegetation and combined them in a vial containing a preservation agent of 75% ethanol and 25% distilled water (Schauff 1986). I collected sub samples of mobile arthropods and arthropods occupying vegetation more tha n 30 cm above the ground along 2 perpendicular 2 0 m transects centered on

PAGE 78

78 the sampling point (Dueser and Shugart 1978 Higgins et al. 2005). Arthropods were sampled using a sweep net (Ausden 1996). I made 50 sweeps (1 sweep comprising a fo rward and backward stroke of the sweep net) along each of the 20 m transects, ensuring the sweep net did not pass within 30 cm of the ground (Schauff 1986). Arthropods collected in each sweep net subsample were combined and preserved as described for those collected using suction sampling. In the laboratory, I identified arthropods contained in each suction and sweep net sample to the family level using a microscope and appropriate identification keys (Triplehorn and Johnson 2005, Ubick et al. 2005). Hab itat Sampling Habitat sampling was conducted once in winter (February March), spring (April May), and summer (July August), during each of 2 years (2007 2008) following trea tment. Within each of the 0.03ha nested circular plots established at arthropod sampling point s, I examined a variety of habitat characteristics including herbaceous and shrub community composition and structure and ground layer variables (Dueser and Shugar t 1978, Higgins et al. 2005). Ground layer I assessed litter c over (%; ocular estimate) within 4 1 m2 sub sample plots, 1 randomly locat ed in each quadrant of the 0.03ha circular plot, along with soil density (g/cm3), moisture (%), and pH. Litter cover was recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 99%, and 7 = 100% (Donhaue et al. 1971, Hays et al. 1981, Higgins et al. 2005). I recorded soil density as the dry weight density (g/cm3) of a 5 cm diameter 10cm deep soil core sample after oven drying at 45C for 48 hours. Soil pH and moisture were measured using a Kelway soil tester (Rodewald and Yahner 2001).

PAGE 79

79 Herbaceous layer I measured s pecies richness ( no. of species) cover (%; ocular estimate), and maximum height (cm) of forbs and graminoids with in the 1m2 subsample plots. Forb and graminoid cover were recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 9 9%, and 7 = 100% (Hays et al.1981, Krebs 1999, Higgins et al. 2005). Shrub layer I counted and determined the height of all shrubs ( woody vegetation <2.0 m i n height) in 2 perpendicular 20m2 quad rats centered on the 0.03ha plot to estimate species richness (no. of species), density (no./m2), and maximum height (cm) for individual species and all combined (Hays et al. 1981, Krebs 1999, Higgins et al. 2005). Shrub cover (%) was assessed along 2 perpendicular 20 m transects centered on the 0.03ha circular plot using the line int ercept method (Hays et al. 1981, Higgins et al. 2005). I recorded the presence of livestock on study sites for incorporation into analyses. Analyses Prior to analyses, I combined data collected from suction and sweepnet samples. I analyzed differences in total arthropod familial ri chness and abundance (all families combined) and familial richness and abundanc e by order (orders with abundance 500 individuals) between treated and untreated (control) sites. Differences were examined both within (e.g., dormant season roller chop) and among (i.e., dormant season roller chop, growing season roller chop, dormant se ason burn, growing season burn, and roller chop/burn) treatment types using repeated measures mixed model regressions. Repeated measures were season and time since treatment (time). Study site pair was included as a blocking factor and presence of grazing as an additional influential independent variable. I used Fishers Protected LSD tests to make post hoc

PAGE 80

80 comparisons. In my results and discussion, I focus ed on treatment rather than repeated measures or grazing effects. I presented results for 2, 3 and 4way treatment interactions. Due to the timing of data collection, it was not possible to test for 3way interactions for growing season burning and roller chopping treatments. As 3and 4way treatment interaction effects are difficult to reliabl y interpret, they were not discussed further (Zar 1999). If, when examining twoway treatment interactions, differences in linear combinations of groups or biologically meaningless compari sons (e.g., arthropod abundance in growing season roller chop sites in year 1 versus arthropod abundance in control sites in year 2) arose, I stated that post hoc comparisons revealed no differences based on treatment and the interacting factor. Multiple linear regression was used to examine which combination of habitat characteristics best described changes in arthropod family richness and abundance. I reduce d multicollinearity problems by subjecting all predictor variables involved in pair wis e correlations with r to a univariate, oneway analysis of variance (ANOVA) with each dependent variable. For each pair of highly correlated predictor variables, I retained the one with the greatest F value (Noon 1981, McGarigal et al. 2000). All r egression models were fit using a backward stepwise procedure with Tolerance = 0.001, F to enter = 0.15, and F to remove = 0.15. These values are considered appropriate for predictor variables that are relatively independent (SYSTAT 2007). I considered r egression models statistically and biologically significant at P and R2 0.2 Only m odels considered significant were presented. The relative importance of each variable in the best model was assessed by examining standardized regression

PAGE 81

81 coefficients (SC; i.e., variables with higher coefficients made greater individual contributions to the explanatory power of t he model). I rank transformed all data prior to analyses due to violations of normality and homogeneity of variance assumptions (Conover 1998, Zar 1999, SYSTAT 2007). To minimize the probability of making a Type II error, I concluded statistical significance for all tests at P P ). All statistical tests were performed using SYSTA T (2007) statistical software. Results During the study I collected arthropods from 24 orders and 162 families (Table 4 1). Of these orders, Orthopt era, Coleoptera, Hemiptera, Diptera, Araneida, Hymenoptera, and Blattodea had an abundance 500 individuals and were subject to further analyses. Dormant Season Roller Chop Total arthropod familial richness was affected by dormant season roller chopping al one (Table 4 2), being 15% lower on roller chop than control sites. Dormant season roller chopping also affected Araneida familial richness, which was 18% lower on roller chop than control sites. Hemiptera, Diptera, and Hymenoptera familial richness were affected by a dormant season roller chopping time interaction (Table 4 3). However, post hoc comparisons revealed no differences in richness between roller chop and control sites for these arthropod orders based on time. A dormant season roller choppi ng season interaction also affected Hemiptera, Diptera, and Hymenoptera familial richness (Table 4 4). In winter, Hemiptera familial richness was 38% lower on roller chop than control sites. Diptera familial richness was 14% greater on roller chop than control sites in spring and 9% lower on roller chop than control sites in winter. In

PAGE 82

82 spring, Hymenoptera familial richness was 67% lower on roller chop than control sites. Blattodea familial richness was affected by a dormant season roller chopping gr azing interaction (Table 4 5). However, no differences in Blattodea familial richness between roller chop and control sites based on grazing were observed from post hoc comparisons. Dormant season roller chopping alone and in all combinations with time, season, and grazing had no effect on Coleoptera familial richness ( P 0.260). Dormant season roller chopping alone affected total arthropod abundance (Table 4 2), which was 20% lower on roller chop than control sites. Orthoptera, Hemiptera, and Araneida abundance were also affected by dormant season roller chopping alone. Orthoptera abundance was 16% lower, Hemiptera abundance 22% lower, and Araneida abundance 35% lower on roller chop than control sites. Diptera and Hymenoptera abundance were affected by a dormant season roller chopping time interaction (Table 4 3). H owever, post hoc comparisons revealed no differences in abundance between roller chop and control sites for these arthropod orders based on time. Diptera abundance was also affected by a dormant season roller chopping season interaction, being 18% lower on roller chop than control sites in spring and 28% greater on roller chop than control sites in winter. A dormant season roller chopping grazing interaction affected Diptera and Blattodea abundance (Table 4 5). Diptera abundance was 17% higher on nongrazed dormant season roller chop than control sites but 46% lower on grazed dormant season roller chop than control sites. Examination of post hoc comparisons revealed no differences in Blattodea abundance between roller chop and control sites based on grazing. Coleoptera abundance was unaffected by dormant

PAGE 83

83 season roller chopping alone and in all combinations with time, season, and grazing ( P 0.327). Growing Season Roller Chop Total arthropod familial richness was affected by a growing season roller chopping time interaction (Table 4 3), being 38% lower on roller chop than control sites the first year following treatment. However, no differences in total arthropod richness were observed the second year following treatment. Growing season roller ch opping alone affected Blattodea familial richness, which was 57% lower on roller chop than control sites (Table 4 2 ). A growing season roller chopping time interaction also affected Orthoptera familial richness, but post hoc comparisons revealed no diff erences in richness between roller chop and control sites based on time. Hemiptera familial richness was affected by a growing season roller chopping grazing interaction (Table 4 5). However, no differences in Hemiptera familial richness between roller chop and control sites based on grazing were observed from post hoc comparisons. Dormant season roller chopping alone and in all combinations with time, season, and grazing had no effect on Coleoptera familial richness ( P 0.338). Diptera and Blattodea abundance were affected by growing season roller chopping (Table 4 2), being 32% and 50% lower on roller chop than control sites, respectively. A growing season roller chopping time interaction affected Orthoptera abundance (Table 4 3), which was 55% lower on roller chop than control sites the first year following treatment. However, no differences in Orthoptera abundance were observed between roller chop and control sites the second year following treatment. Total arthr opod abundance and Coleoptera abundance were unaffected by growing

PAGE 84

84 season roller chopping alone and in all combinations with time, season, or grazing ( P 0.254). Dormant Season Burn Total arthropod familial richness was affected by a dormant season burning time interaction (Table 4 3), being 49% lower on burn than control sites the first year following treatment. However, no differences in total arthropod familial richness were observed between burn and control sites the second year following treatment. Total arthropod familial richness was also affected by a dormant season burning grazing interaction (Table 45). However, post hoc comparisons revealed no differences in total arthropod familial richness between burn and control sites based on grazing A dormant season burning time interaction also affected Orthoptera, Hemiptera, Diptera, and Hymenoptera familial richness. Orthoptera familial rich ness was 42% lower, Hemiptera richness 62% lower, Hemiptera familial richness 64% lower, and Hymenoptera familial richness 45% lower on burn than control sites the first year following treatment. However, there were no differences in richness between burn and control sites for these arthropod orders the second year following treatment. Hymenoptera familial richness were affected by a dormant season burning grazing interaction (Table 4 5). However, post hoc comparisons revealed no differences in richnes s for this order between burn and control sites based on grazing. Dormant season burning alone and in all combinations with time, season, and grazing had no effect on Coleoptera familial richness ( P 0.147). A dormant season burning time interaction af fected total arthropod abundance, which was 64% lower on burn than control sites the first year following treatment (Table 4 3). The second year following treatment there were no differences in total arthropod

PAGE 85

85 abundance between burn and control sites. O rthoptera, Hemiptera, Diptera, Araneida, and Hymenoptera abundance were also affected by a dormant season burning time interaction. Orthoptera abundance was 68% lower, Hemiptera abundance 74% lower, and Hymenoptera abundance 73% lower on burn than contr ol sites the first year following treatment. However, there were no differences in abundance between burn and control sites for these orders the second year following treatment. Post hoc comparisons revealed no differences in Diptera or Araneida abundanc e between burn and control sites based on time. A dormant season burning grazing interaction affected Coleoptera and Blattodea abundance (Table 4 5). Coleoptera abundance was 69% greater and Blattodea abundance 125% greater on Nongrazed burn than contr ol sites. There was no difference in abundance between grazed burn and control sites for these arthropod orders. Growing Season Burn A growing season burning time interaction affected total arthropod familial richness (Table 4 3), which was 56% lower on burn than control sites the first year following treatment. There were no differences in total arthropod familial richness between burn and control sites the second year following treatment. Hemiptera familial richness was affected by growing season bur ning alone (Table 4 2), being 23% lower on burn than control sites. Araneida familial richness was affected by a growing season burning time interaction, being 57% lower on burn than control sites the first year following treatment. The second year fol lowing treatment there was no differences in Araneida familial abundance between burn and control sites. Blattodea familial richness was affected by a growing season burning season interaction (Table 4 4). In winter, Blattodea familial richness was 89% lower on burn than control sites. Coleoptera

PAGE 86

86 familial rich n e ss was affected by a growing season burning season grazing interaction ( P = 0.063). Orthoptera, Diptera, and Hymenoptera familial richness were unaffected by growing season burning alone and in all combinations with time, season, and grazing ( P Total arthropod abundance was affected by growing season burning alone (Table 4 2), being 40% lower on burn than control sites. Hemiptera and Araneida abundance were also affected by growing season burning. Hemiptera abundance was 40% lower and Araneida abundance 50% low er on burn than control sites. A growing season burning season interaction affected Blattodea abundance (Table 4 4). In winter, Blattodea abundance was 96% lower on burn than control sites. Orthoptera abundance was affected by a growing season burning grazing interaction (Table 4 5). However, post hoc comparisons revealed no differences in abundance for this order between burn and control sites based on grazing. Coleoptera and Diptera abundance were affected by a growing season burning time season grazing interaction ( P Dormant season burning alone and in all combinations with time, season, and grazing had no effect on Hymenoptera abundance ( P 0.363). Roller Chop/Burn Orthoptera and Blattodea familial r ichness were affected by a roller chopping/burning time interaction (Table 4 3). Blattodea familial richness was 88% lower on roller chop/burn sites the first year following treatment. However, there was no difference in Blattodea familial richness bet ween roller chop/burn sites the second year following treatment. Examinati on of post hoc comparisons revea l ed n o differences in Orthoptera familial richness between roller chop/burn and control sites based on time. Coleoptera familial richness was affect ed by a roller chopping /burning season

PAGE 87

87 interaction (Table 44 ). In spring, Coleoptera familial richness was 43% lower on roller chop/burn than control sites. A roller chopping/burning time season interaction affected Hemiptera and Araneida familial richness ( P Roller chopping/burning alone and in all combinations with time, season, and grazing had no effect on total arthropod familial richness and Diptera and Hymenoptera familial richness ( P Roller chopping/burning alone affected Diptera ab undance (Table 4 2), which was 32% lower on roller chop/burn than control sites. A roller chopping/burning time interaction affected Blattodea abundance, but examination of post hoc comparisons revealed no differences in abundance between roller chop/burn and control sites based on time. Coleoptera abundance was affected by a roller chopping/burning season interaction. In winter, Coleoptera abundance was 343% greater on roller chop/burn than control sites. Total arthropod abundance and Orthoptera, H emiptera, Araneida, and Hymenoptera abundance were affected by a roller chopping/burning time season interaction ( P ) Treatment Type Comparisons A treatment type time interaction affected total arthropod familial richness (Table 4 6). The fi rst year following treatment total arthropod familial richness was lower on growing season roller chop and dormant and growing season burn compared to control sites, but the effects of the 3 treatments were similar. There were no differences in total arth ropod familial richness between treatment and control sites during the second year of the study. Orthoptera, Diptera, Araneida, Hymenoptera, and Blattodea familial richness were also affected by a treatment type time interaction. Orthoptera familial ri chness was lower on growing season roller chop and dormant and growing season burn compared to control sites during the first year of the study. However, the 3

PAGE 88

88 treatments had similar effects. The second year following treatment, there were no differences in Orthoptera familial richness between treatment and control sites. Hymenoptera familial richness was lower on dormant season burn compared to control sites the first year following treatment and greater on dormant season burn compared to control sites the second year following treatment. Blattodea familial richness was lower on dormant and growing season burn and roller chop/burn sites during the first year of the study. There were no differences in Blattodea familial richness between treatment and control sites during the second year of the study. Post hoc comparisons revealed no differences in Diptera and Araneida familial richness based on treatment time. Orthoptera and Hymenoptera familial richness were affected by a treatment type season int eraction (Table 4 7). Orthoptera familial richness was greater on roller chop/burn sites in winter and lower on growing season roller chop sites in spring. Hymenoptera familial richness was also lower on dormant season roller cho p and roller chop/burn s ites in spring. Total arthropod abundance was affected by a treatment type time interaction (Table 4 6), being lower on dormant and growing season burn compared to control sites the first year following treatment. However, the effects of the 2 treatment s were similar. The second year following treatment there were no differences in total arthropod abundance based on treatment type. Orthoptera, Hemiptera, Diptera, Araneida, Hymenoptera, and Blattodea abundance were also affected by a treatment time in teraction. During the first year of the study, Orthoptera abundance was lower on growing season roller chop and dor mant and growing season burn than control sites, with similar effects of the 3 treatments. Hemiptera abundance was lower on dormant

PAGE 89

89 and growing season burn than control sites the first year following treatment. However, the effects of the 2 treatments were similar. Araneida abundance was lower on dormant season roller chop and growing season burn compared to control sites during the first y ear of the study, but there were no differences in the effects of these 2 treatments. The first year following treatment, Hymenoptera abundance was lower on dormant season burn sites. There was no difference in Ort hoptera, Hemiptera, Araneida, and Hymenoptera abundance based on treatment type during the second year of the study. Blattodea abundance was lower on dormant and growing season burn and roller chop/burn sites the first year following treatment, with similar effects of the 3 treatments. The sec ond year following treatment Blattodea abundance was lower on growing season roller chop compared to control sites. Examination of post hoc comparisons revealed no difference in Diptera abundance based on treatment type and year. Coleoptera and Araneida abundance were affected by a treatment type season interaction (Table 4 7). Coleoptera abundance was lower on roller chop/burn sites in winter and g reater on growing season roller chop compared to control sites in summer. In spring, Araneida abundance was lower on growing season burn, growing seas on roller chop, and roller chop/ burn than control sites. ArthropodHabitat Relationships Habitat characteristics that best predicted total arthropod familial richness were mean forb height (SC = 0.215), mean l itter depth (SC = 0.201), mean soil moisture (SC = 0.129), mean graminoid height (SC = 0.124 ), shrub density (SC = 0.120), variance of soil density (SC = 0.104), variance of forb cover (SC = 0.074), mean graminoid species richness (SC = 0.073), and mean s oil density (SC = 0.064; P R2 = 0.270). The combined effects of mean litter depth (SC = 0.203), mean soil moisture (SC = 0.177),

PAGE 90

90 mean forb cover (SC = 0.134), mean graminoid species richness (SC = 0.124), shrub height (SC = 0.121), mean graminoid height (SC = 0.117), vari ance of soil density (SC = 0.105), shrub species richness (SC = 0.101), and mean soil pH (SC = 0.076 ; P 0.001, R2 = 0.262) best explained Araneida richness Habitat characteristics that best predicted total arthropod abundance were mean litter depth (S C = 0.193), mean forb height (SC = 0.181), mean graminoid height (SC = 0.159), mean soil moisture (SC = 0.141), variance of soil density (SC = 0.135), shrub density (SC = 0.104), variance of litter depth (SC = 0.096), variance of forb cover (SC = 0.092), mean forb species richness (SC = 0.062; P R2 = 0.284). Shrub height (SC = 0.236), mean litter depth (SC = 0.233), mean forb richness (SC = 0.195), mean forb height (SC = 0.133), mean graminoid cover (SC = 0.129), shrub cover (SC = 0.107), mean soil pH (SC = 0.103), and variance of litt er cover (SC = 0.085; P R2 = 0.242) were the habitat characteristics that best described Orthoptera abundance. The combined effects of mean litter depth (SC = 0.210), mean soil moisture (SC = 0.180), mean graminoid height (SC = 0.173), mean forb height (SC = 0.152), variance of soil density (SC = 0.120), shrub species richness (S C = 0.096), mean soil pH (SC = 0. 082), shrub height (SC = 0.080), variance of litter cover (SC = 0.056; P R2 = 0.308) best explained Araneida abundance. Disc ussion Dorma nt season roller chopping led to a prolonged decrease in total arthropod familial richness and abundance that lasted throughout the 2 years of this study. Growing season roller chopping resulted in lower total arthropod familial richness, but the effect was short lived occurring the first year following treatment only. Based on research by Panzer (1988) in prairie remnants, Hall and Schweitzer (1992) suggested

PAGE 91

91 that, in southeastern flatwoods, many arthropods would survive and recolonize best a fter growing season burning, as this is when they are typically most mobile. This study suggests growing season roller chopping may also be pref erable to dormant season roller chopping to maintain arthropod familial richness, possibly because it occurs when the majority of arthropods are mobile and able to f lee roller chopped areas where conditions are unsuitable. During the dormant season, many species are in an inactive, nymphal stage below ground or beneath dead wood (Swengel 2001). Disturbance of so i l and logs by a passing roller chopper during this period may result in the loss of many of these dormant individuals, later reducing the abundance of more mobile life stages. Examination of arthropod orders indic ated that dormant season roller chopping ca used prolonged reductions in Araneida familial richness and Araneida and Hemiptera abundance that lasted throughout the 2 years of th e study. Growing season roller chopping caused prolonged reductions in Blattodea familial richness and Blattodea and Dipte ra abundance that also lasted throughout duration of the study. In addition, s hort term reductions in Orthoptera abundance that lasted throughout the first year of the study were observed following this treatment Studies examining the effects of roller chopping on arthropods are scarce. Roller chopping in both seasons causes short term reductions in litter depth and cover, graminoid and forb cover and height, and shru b cover and height (Chapter 2), potentially reducing food and cover availability for me mbers of these orders and in turn their familial richness and abundance. Reductions in graminoid and forb cover are likely to be a particular concern for many Orthoptera, the majority of which are herbivorous and rely on herbaceous plants as a food source

PAGE 92

92 (Warren et al 1987). Blattodea and certain Orthoptera are ground dwelling, living in the litter layer (Warren et al 1987). Removal of litter and disturbance of the soil during roller chopping may contribute to reductions in abundance within these orders both directly and indirectly. Dormant and growing season burning resulted in lower total arthropod familial richness and abundance. However, reductions in richness and abundance were more prolonged on growing season burn sites, lasting throughout the 2 years of the study. On dormant season burn sites, changes were relatively short lived, occurring only during the first year of the study. These findings do not support the suggestion that southeastern flatwoods arthropods survive and recolonize best fol lowing growing season burning, a period when they are most mobile (Hall and Schweitzer 1992). In contrast arthropods may be better able to survive dormant than growing season burning. D uring the dormant season, many arthropods are in an inactive, nymphal state below gr ound and beneath woody debris. This may offer them protection from passing flames during a burn (Swengal 2001). Examinati on of arthropod orders indicated th at dormant season burning led to short term reductions in Orthoptera, Hemiptera, and Hymenoptera familial richness and abundance. Growing season burning had a prolonged effect on Hemiptera familial richness and Hemiptera and Araneida abundance, which lasted throughout the 2 years of the study, and a short term affect on Araneida familial richness. The majority of Orthopterans are voracious plant feeders and can be negatively affected by fire due to a reduction in herbaceous plant foods (Evans 1984). Dormant and growing season burning have both been found to cause reductions in graminoid and forb height and

PAGE 93

93 cover on pine flatwood s sites for at least the first year post treatment ( Chapter 2). Reduced Orthopteran numbers following burning have also been attributed to reduced litter cover (Tester and Marshall 1961). Reductions in litter de pth and cover are considerable following prescribed burning of pine flatwoods habitats and can be prolonged, lasting for at least 2 years post treatment (Chapter 2) Accumulation of litter has been found to restore Orthoptera populations within 1 year of treatment (Tester and Marshall 1961). Other studies suggest that Orthoptera abundance may increase on burned sites in the first to second year post treatment (Nagel 1973, Reed 1997, Chambers and Samways 1998). Short term reductions in Hemiptera abundance have been observed following burning (Morris 1975, Anderson et al 1989). Like Orthoptera, many Hemiptera are reliant on herbaceous vegetation as a food source (Warren et al 1987) and burning significantly alters availability of graminoids and forbs ( Cha pter 2). Unlike Orthoptera, the majority of Hemiptera occupy aerial habitats (Warren et al 1987) and, as a result, the loss of litter following burning may not be such a concern. In other studies, Hemiptera abundance has been found to be greater on burned areas in the short and/ or intermediate term (Rice 1932, Cancelado and Yonke 1970, Hurst 1971). In the case of Hymenoptera, declines in abundance have been observed in the short term (Bulen and Barrett 1971 ). However, increases in predaceous species of ten occur, presumably because of greater numbers and vulnerability of prey on burned areas (Van Amburg et al. 1981). Other studies indicate an increase in Hymenoptera abundance on burned areas a few months following treatment (Hurst 1971, Nagel 1973). Ar aneida are only able to survive the combustion phase of a burn by seeking refuge in the soil or under nonflammable debris (Warren et al 1987). Immediately following a burn, their

PAGE 94

94 survival is likely to be largely dependent on prey density and diversity (W arren et al 1987), which the results of this study suggest may be reduced. Roller chopping/ burning had a short term effect on Blattodea familial richness and abundance which lasted for 1 year post treatment. As mentioned previously Blattodea occur largely in the litter layer (Warren et al. 1987), the cover and depth of which is severely reduced after roller chopping /burning ( Chapter 2). In addition, the passing of the roller choppers blades may cause soil disturbance that directly and indirectly leads t o reductions in the abundance of members of this order. Habitat characteristics that were most often identified as positively related to total arthropod familial richness and abundance, Araneida richness and abundance, and Orthoptera abundance were related to graminoid and forb cover and height and litter cover and depth. All of these characteristics are lower on dormant and growing season burn and roller chop than control sites some for 2 years post treatment ( Chapter 2). Therefore, the application of these treatments should be applied with caution in areas where arthropod conservation and management are a priority. Managment Implications Studies examining the response of arthropods to prescribed burning are often contradictory G enerally, total arthro pod abundance and richness in Florida flatwoods are lower following prescribed burning and roller chopping treatments, at least in the short term. However, greater variability in response to these treatments is observed when individual orders are examined. Certainly, prescribed burning and roller chopping may be viable management tools for the control of certain arthropod pests. However, further inv estigation of pest phenology and habit s is needed to ensure treatments applied are the most appropriate for control. When maintenance of overall arthropod

PAGE 95

95 biodiversity is a priority, care should be taken regarding the use of prescribed burning and roller chopping treatments over large areas. C onsideration should also be given to the season of application, wit h growing season burning limited to minimize negative effects on certain orders and families Until further research is conducted, arthropod biodiversity may be best promoted by applying a combination of the treatments examined in a mosaic across the landscape. In a ddition it may be necessary to implement techniques that help protect populations of immobile arthropods. Leaving untreated areas adjacent to treated areas to serve as refugia may be beneficial to these populations. Ultimately, t he applicati on of prescribed burning and roller chopping has the potential to harm beneficial arthropods, which may include pollinators, scaveng ers, and predators of parasites. Future research should take into account the relative value of all arthropods and how they are affected by prescribed burning and roller chopping activities to allow treatment application to be appropriately tailored to benefit desired orders or families.

PAGE 96

96 Table 4 1 Arthropod abundance ( no of individuals) within orders and f amilies collecte d on prescribed burned, roller chopped, and control s ites in Florida flatwoods, 2007 2008 Order a Family b Araneida (5298) Agelenidae (983) Araneidae (513) Corinnidae (5) Ctenidae (5), Ctenizidae (2), Clubionidae (5), Cyrtaucheniidae (1), Deinopidae (7) Dictynidae (2), Filistatidae (18), Gnaphosidae (815) Lycosidae (62) Nephilidae (287) Oonopidae (14) Oxyopidae (316) Pisauridae (6), Pholcidae (13) Salticidae (1853) Segestriidae (10) Tetragnathidae (14) Theraphosidae (4), Theridiidae (80) Th omisidae (220), Unknown (63) Hemiptera (2511) Acanaloniidae (5) Alydidae (6) Anthocoridae (7) Aphididae (63) Aradid i ae (3), Berytidae (64) Blissidae (1), Cercopidae (26) Cicadellidae (1385), Cicadidae (11), Cixiidae (15) Coreidae (8), Cydnidae (3), Delphacidae (20) Derbidae (172) Dictyopharidae (7), Flatidae (51), Hydrometridae (2), Issidae (171) Lygaeidae (18) Membracidae (77) Miridae (147) Nabidae (28), Pachygronthidae (72) Pachytroc t idae (2), Pentatomidae (33) Pyrrhocoridae (5), Reduvii dae (44), Rhopalidae (7), Scolytidae (1), Scutelleridae (6), Tin i gidae (26), Tropiduchidae (1), Unknown (24) Orthoptera (2465) Acrididae (749) Gryllidae (1415) Rhaphidophoridae (10) Tetrigidae (98) Tettigonidae (190), Unknown (3) Hymenoptera (1528) Apidae (34) Braconidae (2) Chalcididae (1) Cimbicidae (1) Crabronidae (2), Eulophidae (1), Formicidae (1038) Halyctidae (27), Ichneumonidae (75) Megachilidae (2), Mutillidae (3), Pteromalidae (1), Sphecidae (2), Vespidae (8), Unkn ow n (60) Diptera (1027) Bibionidae (138) Bombyliidae (1) Cecidomyiidae (4) Chamaemyiidae (2), Chironomidae (7), Chloropidae (12) Curtonotidae (1), Culicidae (150) Dolichopodidae (176) Drosophilidae (183) Empid id ae (1), Hippoboscidae (2), Lauxaniidae (1), Muscidae ( 17) Mycetophilidae (2), Otitidae (2), Phoridae (202) Pipunculidae (2), Platypezidae (1), Platystomatidae (6), Sciomyzidae (6), Simuliidae (1), Stratiomyi i dae (2), Syrphidae (1), Tabanidae (1), Tachinidae (11) Tephritidae (4) Tipulidae (5), Ulidiidae (6), Unknown (92)

PAGE 97

97 Table 4 1 Continued Order a Family b Coleoptera (942) Anthicidae (12) Cantharidae (11) Carabidae (9) Cerambycidae (2) Chrysomelidae (573) Cleridae (43), Cuc jidae (11) Curculionidae (106) Coccinellidae (61) Elateridae (13) E ndomychidae (1), Erotylidae (11), Eucnem idae (3), Euglenidae (1), Gyrinidae (1) Lampyridae (17) Latridiidae (7), Lyctidae (1), Meloidae (1), Mordellidae (4), Nitidu lidae (3), Pselaphidae (1) Scarabaeidae (9), Scolytidae (7), Silphidae (7), Staphylinid ae (2), Tenebrionidae (6), Unknown (17) Blattodea (552) Blattellidae (511), Blattidae (32). Unknown (9) Psocoptera (156) Pachytroc t idae (143) Unknown (13) Co llem b ola (150) Isotomidae (134) Poduridae (10) Sminthuridae (2) Entomobryidae (4) Pha smatodea (103) Heteronemiidae (12) Phasmatidae (40), Pseudophasmatidae (38) Unknown (13) Trichoptera (82) Hydroptilidae (65), Unknown (17) Acari (66) Ixodidae (2) Trombiculidae (64) Sco rpiones (58) Buthidae (1), Unknown (57) Lepidoptera (53) Ge ometridae (23) Hepialidae (1) Nymphalidae (1), Unknown (28) Par asitengona (47) Trombiculidae (47) Mantodea (35) Mantidae (23) Photinae (12) Thy sanoptera (25) Thripidae 24, Unknown (1) Neuroptera (19) Ascalaphidae (2) Chrysopidae (14) Corydal idae (1), Megcorydalidae (1), Unknown (3) Pse udoscorpionida (17) Chernetidae (17) Isoptera (15) Rhinotermitida e (15) Opiliones (5) Stylocellidae ( 1), Unknown (4)

PAGE 98

98 Table 4 1 Cont inued Order a Family b Spi robolida (3) Spirobolidae (1), Unknown (2) Diplura (3) Campodeidae (1), Unknown (2) Embiidina (1) Teratembiidae (1) a Abundance by order presented in parentheses. b Abundance by family presented in parentheses. Table 4 2 Effects of treatment on arthropod familial richness and abundance in Fl orida flatwoods, 2007 2008. Arthropod Richness and Abundance by Treatment a Treatment ( SE) P Control Treated Dormant season roller chop Richness (n o. of families) Total 12.3 0.6 10.4 0.7 0.017 Arane ida 3.7 0.2 3.0 0.2 0.020 Abundance (n o. of individuals) Total 37.0 3.0 29.5 3.1 0.032 Orthoptera 8.8 0.7 7.4 1.0 0.085 Hemiptera 6.3 0.7 4.9 0.8 0.042 Araneida 12.4 1.4 8.0 1.0 0.007 Growing season roller chop Richness (n o. of families) Blattodea 0.7 0.1 0.3 0.1 0.010 Abundance (n o. of individuals) Diptera 1.9 0.1 1.3 0.6 0.082 Blattodea 2.6 0.7 1.3 0.4 0.030 Growing season burn Richness (n o. of families) Hemiptera 2.2 0.2 1.7 0.2 0.021 Abundance (n o. of individuals) Total 40.0 4.2 24.0 2.8 0.001 Hemiptera 6.3 1.0 3.8 0.7 0.002 Araneida 14.9 1.7 7.5 1.0 Roller chop/burn Abundance (n o. of individuals) Diptera 3.7 1.0 2. 0 0.7 0.043 a Only arthropod families with richness or abundance significantly affected by treatment presented ( P 0.1)

PAGE 99

99 Table 4 3 Effects of t reatment time interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 2008. Arthropod Richness and Abundance by Treatment a Time b Treatment ( SE) c P Control Treated Dormant season roller chop Ric hness (n o. of families) Hemiptera 1 2.3 0.3 A 1.5 0.2 A 0.022 2 2.1 0.2 A 2.3 0.2 A Diptera 1 1.1 0.2 A 0.4 0.1 A 0.022 2 1.0 0.2 A 1.1 0.2 A Hymenoptera 1 1.2 0.2 A 0.6 0.1 A 0.007 2 0.8 0.2 A 0.8 0.1 A Abundance (n o. of individuals) Diptera 1 1.7 0.4 A 0.7 0.3 A 0.014 2 1.9 0.5 A 1.9 0.4 A Hymenoptera 1 3.6 0.9 A 3.4 1.3 A 0.052 2 1.8 0.4 A 2.4 0.7 A Growing season roller chop Rich ness (n o of families) Total 1 15.0 1.2 A 9.3 1.1 B 0.097 2 11.6 1.1 A 11.0 1.0 A Orthoptera 1 2.3 0.3 A 1.7 0.2 A 0.042 2 1.7 0.2 A 2.0 0.3 A Abundance (n o. of individuals) Orthoptera 1 7.5 1.7 A 3.4 0.7 B 0.070 2 4.7 0.9 A 5.1 1.0 A Dorman season burn Richness (n o. of families) Total 1 13.8 1.1 A 7.0 0.9 B 2 12.0 1.1 A 14.2 1.0 A Orthoptera 1 1.2 0.2 A 0.7 0.1 B 2 1.5 0.2 A 1.9 0.2 A Hemiptera 1 2.6 0.3 A 1.0 0.2 B 0.003 2 2.5 0.3 A 2.8 0.3 A Diptera 1 1.4 0.2 A 0.5 0.1 B 0.004 2 1.3 0.3 A 1.5 0.2 A Hymenoptera 1 1.1 0.1 A 0.6 0.1 B 0.009 2 0.8 0.1 A 1.0 0.2 A Abun dance (n o. of individuals) Total 1 51.4 9.0 A 18.3 3.8 B 2 36.2 5.6 A 43.8 5.2 A Orthoptera 1 4.0 0.7 A 1.3 0.4 B 2 4.2 0.8 A 6.0 1.0 A Hemiptera 1 8.1 1.8 A 2.1 0.6 B 0.001 2 6.3 1.1 A 7.0 1.2 A

PAGE 100

100 Table 4 .3 Continued Arthropod Richness and Abundance by Tr eatment a Time b Treatment ( SE) c P Control Treated Diptera 1 3.7 0.9 A 1.1 0.4 A 0.008 2 2.7 0.8 A 4.3 1.1 A Araneida 1 17.3 2.8 A 7.1 1.6 A 2 14.7 2.5 A 17.8 2.5 A Hymenoptera 1 8.3 3.1 A 2.2 0.7 B 0.028 2 2.4 0.6 A 3.3 1.0 A Growing season burn Richness (n o. of families) Total 1 13.4 0.6 A 5.9 1.1 B 0.017 2 11.2 0.6 A 10.1 0.8 A Araneida 1 4.6 0.5 A 2.0 0.3 B 0.018 2 3.3 0.3 A 3. 3 0.3 A Roller chop/burn Richness (n o. of families) Orthoptera 1 2.0 0.3 A 1.4 0.3 A 0.027 2 1.4 0.2 A 2.2 0.2 A Blattodea 1 0.9 0.1 A 0.1 0.1 B 0.001 2 0.5 0.1 A 0.3 0.1 A Abundance (n o. of i ndiv iduals) Blattodea 1 2.4 0.7 A 0.1 0.1 B 2 1.0 0.3 A 1.0 0.6 A a Only arthropod families with richness or abundance significantly affected by a treatment time interaction presented ( P b.Ti m e since treatment appl ication (years) c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 101

101 Table 4 4 Effects of treatment season interactions on arthropod familial richness and abundance in Florida flatwoods, 2 007 2008. Arthropod Richness and Abundance by Treatment a Season Treatment ( SE) b P Control Treated Dormant season roller chop Richness ( no. of families) Hemiptera Winter 2.1 0.2 A 1.3 0.2 B 0.058 Spring 2.5 0.3 A 2.3 0.3 A Summer 2.0 0.3 A 2.2 0.3 A Diptera Winter 1.5 0.3 A 0.5 0.2 A 0.026 Spring 0.7 0.2 A 0.8 0.2 B Summer 1.1 0.3 A 1.0 0.2 B Hymenoptera Winter 0.5 0.1 A 0.5 0.1 A 0.005 Spring 1.5 0.2 A 0.5 0.2 B Summer 1.1 0.2 A 1.1 0.1 A Abundance ( no. of individuals) Diptera Winter 2.5 0.6 A 0.6 0.3 A 0.009 Spring 1.1 0.5 A 1.0 0.2 B Summer 1.8 05 A 2.3 0.7 B Growing season burn Richness ( no. of families) Blattodea Winter 0.9 0.1 A 0.1 0.1 B 0.062 Spring 0.1 0.1 A 0.1 0.1 A Summer 0.6 0.1 A 0.0 0.0 A Abundance (n o. of individuals) Blattodea Winter 2.7 0.7 A 0.1 0.1 B 0.052 Spring 0.1 0.1 A 0.2 0.2 A Summer 2.1 0.7 A 0.0 0.0 A Roller chop/burn Richness (n o. of families) Coleoptera Winter 0.5 0.2 A 1.2 0.3 A 0.016 Spring 1.7 0.3 A 0.9 0.2 B Summer 0.9 0.2 A 1.2 0.2 A Abundance (n o. of individuals) Coleoptera Winter 0.7 0.4 A 3.1 1.1 A 0.003 Spring 2.8 0.5 A 1.6 0.6 B Summer 1.8 0.5 A 3.7 1.3 B a Only arthropod families with richness or abundance significantly affected by a treatment season interaction presented ( P b Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 102

102 Table 4 5 Effects of treatment grazing interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 2008. Arthropod Richness and Abundance by Treatment a Grazing Treatment ( SE) b P Control Treated Dormant season roller chop Richness (n o. of families ) Blattodea Nongrazed 0.7 0.1 A 0.5 0.1 A 0.034 Grazed 0.3 0.1 A 0.5 0.1 A Abundance (n o. of individuals) Diptera Nongrazed 1.2 0.3 A 1.4 0.4 B 0.069 Grazed 2.4 0.5 A 1.3 0.4 B Blattodea Nongrazed 2.1 0.5 A 1.5 0.4 A 0.033 Grazed 0.4 0.1 A 1.4 0.4 A Growing season roller chop Richness (n o. o f families) Hemiptera Nongrazed 2.2 0.4 A 2.3 0.4 A 0.017 Grazed 2.8 0.4 A 1.9 0.6 A Dormant season burn Richness (n o. of families) Total Nongrazed 9.7 1.4 A 10.0 1.8 A Grazed 13.7 0.8 A 10.5 0.9 A Hymenoptera Nongrazed 0.8 0.1 A 1.0 0.2 A 0.094 Grazed 1.0 0.1 A 0.7 0.1 A Abundance (n o. of individuals) Coleoptera Nongrazed 1.6 1.1 A 2.7 0.7 B 0.060 Grazed 3.8 0.9 A 2 .6 0.4 A Blattodea Nongrazed 0.4 0.2 A 0.9 0.3 B 0.040 Grazed 2.1 0.5 A 0.7 0.3 A Growing season burn Abundance (n o. of individuals) Orthoptera Nongrazed 3.6 0.9 A 6.2 1.2 A 0.080 Grazed 3.9 1.1 A 4.9 1. 6 A a Only arthropod families with richness or abundance significantly affected by a treatment grazing interaction presented ( P b Means in a row followed by the same uppercase letter not significantly different ( P > 0.1)

PAGE 103

103 Table 4 6 Comparison of the effects of treatment time interactions on arthropod familial richness and abundance in Florida flatwoods, 2007 Ar thropod Richness and Abundance a Time b Treatment ( SE) c P Control Dormant Burn Growing burn Dormant Roller Chop Growing Roller chop Roller Chop/Burn Richness (no. of individuals) Total 1 13.8 0.5 A 7.1 0.9 BD 6.1 1.2 B 9.4 0.9 AC 9.3 1.1 BCD 9.8 1.9 AD 2 11.7 0.4 A 14.3 0.9 A 10.3 0.8 A 11.4 0.9 A 11.1 1.0 A 13.0 1.1 A Orthoptera 1 1.9 0.1A 0.7 0.1B 0.6 0.3B 1.9 0.2A 1.7 0.2B 1.4 0.3A 2 1.6 0.1 A 1.9 0.2 A 1.7 0.2 A 2.1 0.2 A 2.0 0.3 A 2.2 0.2 A Dip tera 1 1.2 0.1 A 0.5 0.1 A 0.3 0.2 A 0.4 0.1 A 0.8 0.5 A 0.9 0.3 A 0.001 2 1.1 0.1A 1.5 0.2A 1.2 0.2A 1.1 0.2A 0.7 0.2A 1.0 0.2A Araneida 1 4.1 0.2 A 2.5 0.3 A 2.0 0.3 A 2.6 0.3 A 2.9 0.4 A 2.6 0.6 A 2 3.6 0.1 A 4 .2 0.3 A 3.3 0.3 A 3.3 0.2 A 3.1 0.2 A 3.9 0.4 A Hymenoptera 1 1.0 0.1 A 0.6 0.1 B 0.8 0.2 A 0.6 0.2 A 0.7 0.1 AB 0.7 0.1 A 0.006 2 0.8 0.1A 1.0 0.2B 0.1 0.1A 0.8 0.1A 0.7 0.2A 1.0 0.2AB Blattodea 1 0.6 0.1 A 0.2 0. 1 B 0.0 0.0 B 0.6 0.1 A 0.3 0.2 AB 0.1 0.1 B 0.040 2 0.5 0.0 A 0.5 0.1 A 0.1 0.1 A 0.4 0.1 A 0.3 0.1 A 0.3 0.1 A Abundance (no. of individuals) Total 1 48.3 4.0 A 18.3 3.8 B 23.4 5.7 B 26.3 3.7 AB 40.6 16.6 AB 31.6 9.4 AB 2 34.5 2.1 A 43.8 5.2 A 24.2 3.3 A 32.1 4.8 A 24.8 2.8 A 42.3 7.3 A Orthoptera 1 6.5 0.5 A 1.3 0.4 B 1.4 0.9 B 7.0 1.2 A 3.4 0.7 BC 4.0 1.1 AC 2 5.3 0.4A 6.0 1.0A 4.8 0.9A 7.9 1.4A 5.1 1.0A 11.3 2.0A Hemiptera 1 6.7 0.8 A 2.1 0.6 BC 1.7 0.7 B 3.3 1.0 AB 2.4 1.0 AB 6.1 2.7 AC 0.001 2 6.4 0.5 A 7.0 1.2 A 4.6 0.9 A 6.3 1.1 A 7.0 1.4 A 7.8 1.6 A Diptera 1 2.5 0.4 A 1.1 0.4 A 0.4 0.2 A 0.7 0.3 A 2.2 1.9 A 3.3 1.7 A 0.010 2 2.9 0.4A 4.3 1.1A 2.6 0.8A 1.9 0.4A 1.0 0.3A 2.1 0.6A Araneida 1 17.6 1.6 A 7.1 1.7 AB 5.3 1.2 B 6.4 1.0 B 7.4 1.9 AB 10.5 3.6 A 2 12.9 1.0 A 17.8 2.5 A 8.4 1.2 A 9.3 1.7 A 6.4 0.8 A 13.6 2.7 A Hymenoptera 1 5.8 1.2 A 2.2 0.7 B 12.4 4.0 AC 3.5 1.3 AC 4.6 1.7 AC 3.1 0.8 C 0.041 2 2.9 0.6A 3.3 1.0A 1.9 0.7A 2.4 0.8A 1.4 0.4A 2.2 0.7A Bl attodea 1 2.4 0.4 A 0.2 0.1 B 0.0 0.0 B 1.9 0.5 A 1.2 0.8 AB 0.1 0.1 B 0.003 2 1.3 0.2A 1.3 0.5A 0.1 0.1A 1.0 0.4A 1.3 0.5B 1.0 0.6A a Only arthropod families with ri chness or abundance significantly affected by a treatment time int eraction ( P b Time since treatment application (years). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 104

104 Table 4 7 Comparison of the effects treatment season interactions on arthropod familial richness and abunda nce in Florida flatwoods, 2007 Arthropod Richness and Abundance a Season Treatment ( SE) b P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/Burn Richness (n o. of families) Orth optera Winter 1.5 0.1 A 1.1 0.2 A 1.8 0.4 A 1.7 0.2 A 2.0 0.4 A 2.7 0.3 B 0.078 Spring 1.5 0.1 A 1.3 0.3 AB 1.8 0.3 A 1.9 0.2 A 1.3 0.4 B 1.4 0.3 AB Summer 2.1 0.1 A 1.5 0.2 A 1.0 0.3 A 2.3 0.2 A 2.2 0.3 A 2.0 0.3 A Hymenopter a Winter 0.7 0.1 ABC 0.6 0.1 A 0.7 0.2 BC 0.5 0.1 A 1.2 0.3 C 1.2 0.3 C 0.001 Spring 1.1 0.1 A 1.0 0.2 B 0.7 0.3 A 0.5 0.2 B 0.4 0.2 A 0.6 0.2 B Summer 0.9 0.1 AB 0.7 0.1 AB 0.6 0.1 A 1.1 0.1 B 0.8 0.1 AB 0.9 0.1 AB Abundance (n o. of individuals) Coleoptera Winter 1.0 0.2 A 1.5 0.3 A 0.6 0.2 A 0.7 0.3 A 0.5 0.3 A 3.1 1.1 B 0.086 Spring 1.8 0.2 A 2.4 0.5 A 0.7 0.3 A 1.4 0.6 A 1.4 0.8 A 1.6 0.6 A Summer 3.3 0.6 AB 4.0 0.8 AB 1.3 0.3 A 4.3 1.4 AB 2.6 0.7 B 3.7 1.3 AB Araneida Winter 14.8 1.6 A 11.9 3.0 A 8.1 2.0 A 7.9 1.8 A 7.8 1.2 A 17.9 6.7 A 0.070 Spring 12.0 1.1 A 7.7 1.9 AB 1.6 1.3 B 7.9 2.1 AB 4.2 0.8 B 5.2 1.7 B Summer 17.0 1.6 A 16.9 3.0 A 5.8 0.9 A 8.2 1.6 A 8.1 1.3 A 15.6 3.0 A a Only arthropod families with richness or abundance significantly affected by a treatment season interaction presented ( P b Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 105

105 CHAPTER 5 EFFECTS OF PRESCRIBED BU R N ING AND ROLLER CHOPP ING ON AVIAN COMMUNITIES AND THEI R HABITAT ASSOCIATIO NS IN FLATWOODS Introduction Peninsular Floridas rangeland habitats, including its pine flatwoods provide habitat for a variety of resident and migrator y bird species (United States Fish and Wildlife Service [USFWS] 1999, Alsop 2002, Engstrom et al. 2005, Florida Fish and Wildlife Conservation Commission [FWC] 2005). Many of these avian species, some of which are federally and/or state listed as endanger ed or threatened, have decreasing populations and are considered by FWC to be of conservat ion concern ( Engstom 2005, FWC 2005). Declines in bird populations, particularly those of obligate and facultative species associated with open savanna and grassland habitats, have been reported across the eastern United States and attributed to a loss of suitable habitat (Askins 1993, Murphy 2003, Brennan and Kuvlesky 2005). Certainly, t he pine flatwoods which provide habitat for many of Floridas resident and migratory savanna and grassland affiliated species are in poor condition and declining in quality and quantity (Davis 1967, Abrahamson and Hartnett 1990, Kautz 1993, Cox et al. 1997, USFWS 1999, FWC 2005). As a result, they are listed by FWC (2005) as terrest rial habitats with a high relative threat status. C hanges in land management techniques (e.g., alteration of fire regimes and increases in grazing pressure) are thought to be contributing to the degradation of these areas (Abrahamson and Hartnett 1990, FW C 2005, Gordon et al. 2005). H istorically, pine flatwoods were subject to high freq uency, low intensity, lightning ignited fires during th e thunderstorm season (May July; Komarek 1968, Christensen 1981, Abrahamson and Hartnett 1990, Pyne et al. 1996). F requent fires are considered essential to maintain

PAGE 106

106 the structure and diverse herbaceous groundcover of many southeastern rangeland systems, such as pine flatwoods (Christensen 1981, Abrahamson and Hartnett 1990, Platt 1998). Fire exclusion, reductions in fire frequency, and/or a shift in fire season, commonly a result of human intervention, can result in excessive shrub growth and declines in the species rich herbaceous ground layer of these areas (Wade et al. 1980, Platt et al. 1988, Fitzgerald and Tanner 1992, Robbins and Myers 1992, Watts and Tanner 2003, Watts and Tanner 2006). On many private pine flatwoods in Florida the majority of which are used for livestock production, changes in land management have resulted in alterations to the natural fire r egime. This has frequently permitted the proliferation of shrubby vegetation such as saw palmetto ( Serenoa repens [Bartram] Small), gallberry ( Ilex glabra [L.] A. Gray ) and wax myrtle ( Morella cerifera [L.] Small) and caused reductions in grass and forb species Such changes threaten the integrity of remaining pine flatwoods habitats, potentially reducing their suitability for a variety of bird species, many of which are considered of conservation priority ( Engstrom 2005, FWC 2005). In Florida, FWC and t he United States Department of Agriculture (USDA) are utilizing assistance based programs such as the Farm Bills Environmental Quality Incentives Program and Wildlife Habitat Incentives Program, to encourage private landowners to better manage remaining areas of pine flatwoods. Management activities currently being promoted under these programs include the us e of prescribed fire and roller chopping during dormant ( November March) and growing (April October ) seasons. These practices, depending on season o f application, have been shown to improve the quality of southeastern rangelands by reducing shrub cover and promoting

PAGE 107

107 the growth and flowering of herbaceous vegetation ( Chapter 2; Kalmbacher and Martin 1984, Platt et al. 1988, Tanner et al. 1988, Robbins and Myers 1992, Watts and Tanner 2003, Watts and Tanner 2006). The impacts prescr ibed burning and roller chopping have on Floridas pine flatwoodsassociated avian communities are largely unknown. However, these practices are likely to lead to alterations in avian community composition through changes in vegetation composition and structure and available habitat (Johnston and Odum 1956, Anderson 1980, Norris et al. 2003, Venier and Pearce 2005). Certainly, in other North American grassland systems, distur bances such as prescribed burning have influenced bird abundance and species richness through changes to the plant community (Huber 1984, Madden et al. 1999). In Florida, a study examining the effect of management activities on bird communities found the application o f prescribed burning and roller chopping to a shrubdominated, former dry prairie site in south Florida had an acute effect on bird abundance and species composition with notable declines on roller chopped sites (Fitzgerald 1990, Fitzgerald a nd Tanner 1995). In light of current efforts to promote the use o f prescribed burning and roller chopping on rangelands through the use of assistance based programs, a clearer understanding of the effects these activities have on avian communities is need ed (Berkland et al. 2005, FWC 2005, Gray et al. 2005). Such research will establish whether these management practices are beneficial to the numerous bird species of greatest conservation need commonly associated with pine flatwoods habitats (FWC 2005), a nd determine whether their use to maintain and restore rangeland vegetation is also an effective avian conservation strategy.

PAGE 108

108 The objectives of my study were to 1) compare avian s pecies richness and abundance in treated ( management activities implemented) and untreated ( no management activi ties implemented) pine flatwoods during winter, migratory, and breeding seasons and 2) examine the relative importance of local pine flatwoods habitat characteristics (e.g., shrub cover, forb, cover, graminoid height art hropod abundance ) in determining avian species richness and abundance Methods Study Sites I conducted research o n 50 privately and publically owned pine flatwoods sites that were being prescribed burned and roller chopped by local landowners and land man agers using a variety of individual protocols. These sites had varying management and grazing histories and were located across 6 counties (Desoto, Highlands, Lee, Manatee, Osceola, and Sarasota) in central and south Florida. When grazed, both the treatm ent and paired control study sites were subject to similar grazing pressures at similar times. Floridas pine flatwoods are characterized as having a pure or combined overstory stand of scattered longleaf ( P inus palustris Mill.) and slash ( P. elliotti Enge lm.) pine and, when the shrub layer is relatively open, an often diverse herbaceous layer. This herbaceous layer contains a wide variety of grasses (e.g., Agrostis, Andropogon, Aristida, Eragrostis, Panicum, and Paspalum spp.). Common forbs include legum es (e.g., Cassia, Crotalaria, Galactia, Tephrosia spp.), milkweeds ( Asclepias spp.), milkworts ( Polygala spp.), and a wide variety of composites (e.g., Aster, Chrysopsis, Eupatorium, Liatris, and Solidago spp. ). U nderstory and shrub species include saw palmetto, wax myrtle ( Morella cerifera [L.] Small), gallberry ( Ilex glabra [Pursh] Chapm.),

PAGE 109

109 fetterbush ( Lyonia lucida [Lam.] K. Koch), staggerbush ( Lyonia fruticosa [Michx.], G. S. Torr), dwarf huckleberry ( Gaylussacia dumosa [Andrews] Torr. & A. Gray), dwar f live oak ( Quercus mimima [Sarg.] Small), and tarflower ( Bejaria racemosa Vent ; Abrahamson and Hartnett 1990, U SFWS 1999.). Treatment Types One of 5 treatment types was applied to treated sites: dormant season (November March) burn, growing season (April October) burn, dormant season roller chop, growing season roller chop, or a roller chop/burn combination treatment. The roller chop/ burn combination treatment (hereafter referred to as roller chop/burn) involved roller chopping in the dormant seaso n foll owed by burning within 6 months. I established a total of 11 dormant season burn, 9 growing season burn, 9 dormant season roller chop, 12 growing season roller chop, and 9 roller chop/burn and control pairs. Bird Surveys I assessed the effects of treatment (prescribed burning, roller chop ping, and roller chopping/ burning) on avian richness and abundance using a paired sampling approach. Richness and abundance were compared between sampling point s randomly located in paired treated (e.g., dormant season burned) and untreated (control) flatwoods sites. Paired treatment and control sampling point s were of similar current and past management (e.g., grazing), surrounding land use, plant community (e.g. overstory cover), and soil conditions, being located in t he same pasture or management unit I established 1 randomly selected sampling point within each treatment and control site. To minimize edge effects, I rejected and randomly relocated sampling point s that fell

PAGE 110

110 within 50 m of the edge of a treatment or c ontrol site. Sites within which treatment and control sampling points were located ranged from 2 20 ha. I surveyed the avian communit y at each sampling point 3 tim es each year, during each of 2 years (2007 and 2008) following treatment Sampling periods corresponded to presumed seasonal differences in avian habitat utilization: wintering (January March) and breeding (June September ) seasons and spring migration (April May ). I conducted bird surveys at each sampling point using unlimited radius point count methods (Gibbons et al. 1996, Bibby et al. 2000) After arriving at each point count location I waited 3 minutes before commencing sampling This provided time for birds to settle following my arrival. Point counts were conducted for 5 minutes, during which all birds detected to an unlimited distance were recorded (Gibbons et al. 1996, Bibby et al. 2000). Study sites were surveyed within 4 hours of sunrise, on mornings with little rain, no wind, and no fog, and with paired sampling point s always being counted on the same day (Gibbons et al. 1996, Bibby et al. 2000) During each sampling period, the paired point sampled first was alternated. I used careful observation, including recording the approximate position of detected birds and flyovers to reduce the likelihood of double c ounting (Gibbons et al. 1996, Gregory et al. 2004). I divided counts of avian abundance and richness into 4 categories and 7 guilds based on residency and breeding status. Depending on residency status, counts were assigned t o either a permanent resident or migrant category. The migrant category was divided into 2 guilds, short distance migrant, and neotropical migrant. Based on their breeding status in Florida, birds were also assigned to 1 of 2 categories, nonbreeding (o verwintering) or breeding. The breeding category was divided into 5 breeding habitat

PAGE 111

111 guilds: woodland, urban, successional scrub, wetland, and grassland (Table 51 ; (Peterjohn and Sauer 1993) ). Habitat Sampling I conducted habitat sampling once in winter (January March), spring (April May) and summer (June September) during each of 2 years (2 007 2008) following treatment at the same point s used for av i a n surveys At each sampling point plant community composition and structure, li t ter and soil variables, and vertical obstruction were examined in several strata (i.e., ground, herbaceous, shrub, understory, and overstory levels), using a 0.03 ha nested circular plot design similar to that described by Dueser and Shugart (1978) and Higgins et al. (2005). G round layer I assessed litter c over (%; ocular estimate) within 4 1 m2 sub sample plots, 1 randomly located in each quadrant of the 0.03 ha circular plot s, along with soil density (g/cm3), moisture (%), and pH. Litter cover was recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 99%, and 7 = 100% (Donhaue et al. 1971, Hays et al. 1981, Higgins et al. 2005). I measured s oil density as the dry weight density (g/cm3) of a 5 cm diameter, 10cm deep soil core sample a fter oven drying at 45C for 48 hours. Soil pH and moisture were measured using a Kelway soil tester (Rodewald and Yahner 2001). Herbaceous layer I determined s pecies richness, cover (%; ocular estimate), and maximum height (cm) of forbs and graminoids with in the 1m2 subsample plots. Forb and graminoid cover were recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 9 9%, and 7 = 100% (Hays et al.1981, Krebs 1999, Higgins et al. 2005).

PAGE 112

112 Vertical Obstruction Vertic al obstruction (%) from 0 2 m above ground was measured using a cover pole (Griffin and Yo utie 1988) centered on the 0.03ha circular plot at a distance of 5 and 10 m. Shrub layer I counted and measured the h e ight of all shrubs (woody vegetation <2.0 m i n height) in 2 perpendicular 20m2 quadrats centered on the 0.03ha plot to estimate species richness (no. of species), density (no./m2), and maximum height (cm) for individual species and all combined (Hays et al. 1981, Krebs 1999, Higgins et al. 2005). Shrub cover (%) was assessed along 2 perpendicular 20 m transects centered on the 0.03ha circular plot using the line intercept method (Hays et al. 1981, Higgins et al. 2005). Understory I counted, identified the species, and measured the diameter at br east height (dbh ) of all understory (woody vegetation <7.5 cm dbh, height) plants within the 0.03ha circular plots to estimate species richness, density (no./ha), and basal area (cm2/ha) for individual species and all combined (Krebs 1999). Und erstory canopy cover (%) was estimated from 41 evenly spaced, vertical ocular tube sightings taken at a height of 0.75 m along 2 perpendicular 20m transects centered on the 0.03ha circular plot (James and Shugart 1971). Overstory All overst ory (woody v egetation ) plants were also counted, species identified, and dbh measured within the 0.03ha circular plot to estimate species richness, density (no./ha), and basal area (cm2/ha) for individual species and all combined (Krebs 1999). I estimated overstory canopy cover (%) from 41 evenly spaced, vertical ocular tube sightings taken at a height of 1.5 m along 2

PAGE 113

113 perpendicular 20m transects centered on the 0.03ha circular plot (James and Shugart 1971). Arthropods. I collected arthropods occupying vegetation less than 30 cm above the ground from within 4 1m2 sub sample plots, randomly located in each quadrant of a 0.03ha nested circular plot centered on the sampling point ( Dueser and Shugart 1978, Higgins et al. 2005) Arthropods were sampled us ing a suction sampler (Wright and Stewart 1993, Ausden 1996). Within each 1m2 plot, the suction sampler was turned on and systematically moved around the subsample area, no more than 30 cm above the ground, for a 3minute period collecting arthropods. Suction sampling was not conducted if vegetation was damp or had been flattened by wind, rain, or trampling (Ausden 1996). I separated arthropods collected in each suction sub sample from coarse vegetation and combined them in a vial containing a preservation agent of 75% ethanol and 25% distilled water (Schauff 1986). I collected sub samples of mobile arthropods and arthropods occupying vegetation more tha n 30 cm above the ground along 2 perpendicular 2 0 m transects centered on the sampling point (Dueser and Shugart 1978 Higgins et al. 2005). Arthropods were sampled using a sweep net (Ausden 1996). I made 50 sweeps (1 sweep comprising a forward and backward stroke of the sweep net) along each of the 20 m transects, ensuring the sweep net did not pass wi thin 30 cm of the ground (Schauff 1986). I combined arthropods collected in each sweep net subsample and preserved as described for those collected using suction sampling. In the laboratory, I identified arthropods contained in each suction and sweep net sample to the family level using a

PAGE 114

114 microscope and appropriate identification keys (Triplehorn and Johnson 2005, Ubick et al. 2005). Suction and sweep net samples were combined prior to analyses. I recorded the presence of livestock on study sites for incorporation into analyses. Analyses I performed repeated measures mixed model regressions to examine differences in total avian abundance and species richness, and avian abundance and species richness by category and guild, between treated and control sites Differences were examined both within (e.g., dormant season roller chop) and among (i.e., dormant season roller chop, growing season roller chop, dormant season burn, growing season burn, and roller chop/burn) treatment types. Repeated measures were season and time since treatment (time). Study site pair was included as a blocking factor and presence of grazing as an additional influential independent variable. I used Fishers Protected LSD tests to make post hoc comparisons. In my results and discussion, I focus on treatment rather than repeated measures or grazing effects. I present ed results for 2, 3 and 4way treatment interactions when they occurred. However a s 3 and 4way treatment interaction effects are difficult to reliably interpret, they were not discussed further (Zar 1999) Due to the timing of data collection, it was not possible to test for 3way interactions for growing season burning and roller chopping treatments. If, when examining 2way treatment interactions, differences in linear combinations of groups or biologically meaningless compari sons (e.g., av i a n abundance in dormant season burn sites in year 1 versus avian abundance in control sites in year 2) arose, I stated that post hoc comparisons revealed no differences based on treatment and the interacting factor.

PAGE 115

115 To determine the c ombination of habitat characteristics that best described changes in avian abundance and species richness, both overall and by category and guild I used backward stepwise multiple regression. M ulticollinearity problems were reduced by subjecting all predictor variables involved in pairwise correlations with r 0.7, to a univariate, oneway analysis of variance (ANOVA) with each dependent variable. For each pair of highly correlated predictor variables, I retained the one with the greatest F value (Noon 1981, McGarigal et al. 2000). All regression models were fit using a Tolerance = 0.001, F to enter = 0.15, and F to remove = 0.15. These values are considered appropriate for predictor variables that are relatively independent (SYSTAT 2007). I considered r egression models statistically and biologically significant at P and R2 Only models considered significant are presented. The relative importance of each variable in the best model was assessed by examining standardized regression coefficients (SC; i.e., variables with higher coefficients made greater ind ividual contributions to the explanatory power of t he model). All data sets were rank transformed prior to analyses due to violations of normality and homogeneity of var iance assumptions (Conover 1998, Zar 1999 SYSTAT 2007). I concluded statistical signi ficance at P rather than the more common P to minimize the probabi lity of making a Type II error ( Mapstone 1995, Zar 1999). All statistical tests were performed using SYSTA T (2007) statistical software. Results Dormant Season Burn I recorded 52 bird species on dormant season burn sites over the course of the study (Table 51) Non breeding category richness was affected by dorma nt season burning alone (Table 5 2 ). Within the breeding category, a dormant season burn

PAGE 116

116 season grazing interaction affec ted woodland guild abundance ( P = 0.071). Successional scrub and wetland guild abundance were affected by a dormant season burning time season interaction ( P = 0.099 and P = 0.063, respectively). Total avian richness, permanent resident, migrant and breeding category richness, and short distance migrant, neo tropical migrant, urban, and grassland guild richness were unaffected by dormant season burning alone and in all combinations with time, season, and grazing ( P 0.118). Permanent resident categor y abundance was affected by a dormant season burning season grazing interaction ( P = 0.082). Migrant and non breeding category abundance were affected by a dormant season burning season interaction (Table 5 3 ). Within the migrant category, a dorman t season burning season interaction also affected neotropical migrant guild richness (Table 5 3 ). Within the breeding category, a dormant season burning grazing interaction affected successional scrub abundance (Table 5 4 ). However, no differences i n abundance for this guild based on burning and grazing were observed from post hoc comparisons. A dormant season burning time season interaction affected wetland guild abundance ( P = 0.072). Dormant season burning alone and in all combinations with time, season, and grazing had no affect on total avian abundance, breeding category abundance, and short distance migrant and urban guild abundance ( P Growing Season Burn I identified 46 bird species on growing season burn sites throughout the 2 years of the study (Table 51) Migrant category richness was affected by growi ng season burning alone (Table 5 2 ). Within the migrant category, gro wing season burning alone also affected neotropical migrant guild richness (Table 52 ). Non breeding category

PAGE 117

117 richness was affected by a growing season burning season interaction (T able 53 ). Within the breeding category, urban guild richness was affected by growing season burning alone (Table 5 2 ). W etland guild richness was affected by a growing season burning season interaction ( Table 53 ). However, examination of post hoc comparisons revealed no differences in richness for this guild based on burning an d season Total a vian richness, permanent resident and breeding category richness, and short distance migrant, woodland, successional scrub, and grassland guild richness were unaffected by growing season burning alone and in all combinations with time, season, and grazing ( P 0.170) Non breeding category abundance was affected by a growing season burni ng season interaction (Table 53 ). Within the breeding category, wetland guild abundance was affected by growi ng season burning alone (Table 52 ). Growing season burning alo ne and in all possible combinations with time, season, and grazing had no affect on total a vian abundance, permanent resident, migrant, and breeding category abundance, and short distance migrant, neo tropical migrant, woodland, successional scrub, and gra ssland guild abundance ( P 0.119). Dormant Season Roller Chop I detected 48 bird species on dormant season roller chop sites over the study period (Table 51) Permanent resident and breeding category richness were affected by a dormant season roller chopping grazing interaction (Table 54 ). However, examination of post hoc comparisons revealed no d ifferences in richness for these categories based on roller chopping and season. Within the breeding category, grassland guild richness was affected by a dormant season roller chop p ing time interaction (Table 55 ) Total avian richness, migrant and breeding category richness,

PAGE 118

118 and short distance migrant, neo tropical migrant, woodland, urban, successional scrub, and wetland guild richness were unaffected by dormant season roller ch opping alone and in all combinations with time, season, and grazing ( P 0.113). Breeding category abundance was affected by a dormant season roller chop ping time interaction (Table 55 ). However, no differences in abundance for this guild based on roll er chopping and time were observed from post hoc comparisons. Dormant season roller chopping alone and in all possible combinations with time, season, and grazing had no affect on total avian abundance, permanent resident, migrant, and nonbreeding category abundance, and short distance migrant, neotropical migrant, woodland, urban, successional scrub, wetland, and grassland guild abundance ( P 0.101). Growing Season Roller Chop I observed 39 bird species on growing season roller chop sites over the course of the study (Table 51) Within the migrant category, short distance migrant guild richness was affected by a growing season roller choppi ng season grazing interaction ( P = 0.090). Within the breeding category grassland guild richness was affected by a growing season roller chopping time interaction (Table 55 ). However, examination of post hoc comparisons revealed no d ifferences in richness for this guild based on roller chopping and time. A growing season roller chopping time season grazing interaction affected urban guild richness ( P = 0.087). Total avian richness, permanent resident, migrant, nonbreeding and breeding cat egory richness, and neo tropical migrant, woodland, successional scrub, wetland, and grassland guild richness were unaffected by growing season roller chopping alone and in all combinations with time, season, and grazing ( P ).

PAGE 119

119 Permanent resident cat egory abundance was affected by growing seaso n roller chopping alone (Table 52 ). Within the migrant guild, short distance migrant guild abundance was affected by a growing season roller chopping season grazing interaction ( P = 0.090). Growing season roller chopping alone affected breeding category abundance (Table 52 ). With the breeding category, successional scrub, and grassland guild abundance were also affected by growing seaso n roller chopping alone (Table 52 ). Growing season roller chopping alone and in all combinations with time, season, and grazing had no affect on total avian abundance, migrant and nonbreeding category abundance, and neot r opical migrant, woodland, urban, and wetland guild abundance ( P Roller Chop/Burn I recorde d 48 bird species were recorded on roller chop/ burn sites during the 2 years of the study (Table 51) Non breeding category richness was affected by a roller chopping/burning grazing interaction (Table 54 ). However, post hoc comparisons revealed no differences in richness for this category based on roller chopping/burning and grazing. Within the breeding category, grassland guild richness was affected by roller chopping/burning alone (Table 52 ). Total avian richness, permanent resident, migrant, an d breeding category richness, and short distance migrant, neo tropical migrant, woodland, urban, successional scrub, and wetland guild abundance were unaffected by roller chopping/burning alone and in all combinations with time, season, and grazing ( P 109). Total avian abundance was affected by roller chopping/burning alone (Table 52 ). Roller chopping/burning alone also affected permanent resident category abundance (Table 52 ). Migrant category abundance was affected by a roller chopping/burning

PAGE 120

120 t ime season grazing interaction ( P = 0.063). A roller chopping/burning grazing interaction affected nonbreeding category abundance. However, no differences in abundance for this category based on roller chopping and grazing were observed from post hoc comparisons. Breeding category abundance was affected by roller chopping/burning alone (Table 52 ). Within the breeding category, grassland guild abundance was also affected by roller chopping/burning alone (Table 52 ). Urban guild abundance was aff ected by a roller chopping/burning season grazing interaction ( P = 0.026). Roller chopping/burning alone and in all combinations with time, season, and grazing had no affect on short distance migrant, woodland, successional scrub, and wetland guild ab undance ( P Treatment Type Comparisons Total avian richness was affected by treatment type alone (Table 56 ). Treatment type alone also affected permanent res ident category richness (Table 56 ). Non breeding category richness was affected by a t reatment type season interaction (Table 5 7 ). Treatment type alone affected breeding category richness (Table 5 6 ). Within the breeding category, woodland guild richness was affected by a treatment type time (Table 5 8 ) and a tr eatment type grazing (Table 59 ) interaction. Urban guild richness was also affected by a treatment ty pe grazing interaction (Tale 59 ). A treatment type time season interaction affected successional scrub, wetland and grassland guild richness ( P 0.087). Migrant category ( P = 0.045) and neotropical migrant guild ( P = 0.022) richness were affected by a treatment type time season grazing interaction. Treatment type alone and in all combinations with time, season, and grazing had no affect on short distance migrant guild richness ( P 0.366).

PAGE 121

121 Total avian abundance was affected by a treatment ty pe* grazing interaction (Table 59 ). A treatment type season grazing interaction affected permanent resident category abundance ( P = 0.030) M ig rant category ( P = 0.013) and neotropical migrant guild ( P = 0.011) abundance were affected by a treatment type time season grazing interaction. A treatment type season interaction had an effect on nonbreeding category abundance (Table 5 7 ). Br eeding category abundance was affected by a treatment type season grazing interaction ( P = 0.033). Within the breeding category, woodland guild abundance was affected by a treatment type* time (Table 58 ) and a treatment type grazing interaction (Ta ble 59 ). Successional scrub and grassland guild abundance were also affected by a treatment type grazing interaction (Table 59 ). However, post hoc comparisons revealed no differences in grassland guild abundance based on treatment type and grazing. Grassland abundance was affected by a treatment type grazing interaction (Table 57 ). A treatment type season grazing interaction affected urban guild abundance ( P = 0.018). S hort distance migrant and wetland guild abundance were unaffected by trea tment type alone and in all combinations with time, season, and grazing ( P 0.108). Avian Habitat Relationships Relationships among avian species richness and abundance and habitat characteristics were generally weak and best models contained large number s of variables. The combination of vegetation characteristics having the greatest affect on avian abundance and richness differed by category and guild (Table s 5 10). Vegetation characteristics that positively affected richness and abundance with many av ain categories and guilds were related to shrub richness, cover, and density, forb cover and height, and arthropod familial richness Vegetation characteristics that negatively

PAGE 122

122 affected species richness and abundance within many avain categories and guilds were shrub richness, cover, and density, forb richness,cover, and height, and graminoid cover. Discussion There are negative effects of dormant season burning on the nonbreeding, overwintering avian community of pine flatwoods habit at s, including declin es in species richness and abundance. F ew studies have shown strong effects of burn season on avian communities (Engstom et al. 1996, King et al.1998) However, i t has been propo sed that declines in the survival of overwintering species can occur following dormant season burning, due in part to decrease s in herbaceous ground cover important as a food and cover source ( Robbins and Myers 1992, Engstrom et al. 2005, Thatcher et al. 2006) In pine flatwoods, dormant season burning has little impact on unders tory and shrubby vegetation density and cover, while causing reductions in graminoid and forb species richness, cover and height and li tter cover and depth for at least the first year post treatment (Chapter 2). It is suspected that dormant season burning may exert a negative impact on nonbreeding, overwintering species by removing herbaceous cover and foraging substrate of species that are active close to the ground ( Engstrom et al. 2005, Thatcher et al. 2006). T he effects of growing season burning on the non breeding, overwintering avian community of pine flatwoods were in contrast largely positive, with increases observed in species richness and abundance. G rowing season bur n ing tends to ca use a greater reduction in understory and shrubby ve getation than dormant season bur n ing (Chapter 2 and 3) especially if repeated fires are implem ented over multiple years (Waldrop et al. 1987) In addition, although declines in forb and graminoid species richness, cover and height may be observed for

PAGE 123

123 2 years fol lowing treatment application (Chapter 2), growing season burning can stimulate the growth, flowering and seeding of certain herbaceous species common in pine flatwoods. These include wiregrass ( Aristida stricta Michx.), c utthroat grass ( Panicum abscissum Swallen ) toothache grass ( Ctenium aromaticum (Walter) Alph. Wood ), l ittle bluestem ( Schizachyrium rhizomatum [Swallen] Gould) and other bluestem grasses ( Andropogon spp .; Myers and Boettcher 1987, Platt et al. 1988, Outcault 1990, Streng et al. 1993) Resulting increases in seed and insect production following growing season burning may be of potential benefit to ground foraging species during the winter months (Engstrom et al. 2006, Thatch er et al. 2006), helping to maintain and even increase species richness and abundance of nonbreeding, overwintering species. G rowing season burning had no effect on species richness and abundance within the breeding category. However, 2 guilds within th is category urban and wetland, exhibited increases in richness or abundance following growing season burning In addition, richness and abundance of all ot her breeding category guilds were maintained. G rowing season burning also resulted in increases in migrant category and neotropical migrant guild species richness. These increases may reflect changes in the occurrence of some migratory breeding birds Although research suggests growing season burning may be beneficial to nonbreeding, overwintering birds, concerns have been raised that the practice may be detrimental to nesting species This concern arises from the potential threat fires set during the growing season may have on the nests of breeding birds, with damage to the nests of game species o ften of particular issue (Stoddard 1931, Sisson et al. 1990, Hermann et al. 1998, Engstrom et al. 2005). However, the indirect benefits of habitat alteration are likely beneficial to breeding species and are

PAGE 124

124 thought to outweigh potential nest losses (Engstrom et al. 2005, Robbins and Myers 1992, Cox and Widner 2008). G rowing season burning may lead to improved late summer and fall brood habitat (Cox and Widener 2008) through potential increases in herbaceous growth (Myers and Boettcher 1987, Platt et al. 1988, Outcault 1990, Streng et al. 1993, Provencher et al. 1998) and insect food sources (Hanula and Fanzeb 1998, Provencher et al. 1998, Collins et al 2003, Hardy 2003). However, i n pine flatwoods growing season burni ng can result in reduced herbaceo u s grow th through the second year post treatment (Chapter 2). In addition, insect res p ons e to growing season burning can be highly variable and decline s in f amilial richness and abundance, particu l a rly within the Hemiptera, Araneida, and Blattodea, have been observed (Chapter 4) potentially causing reductions in food. Therefore, there is a need to be cautious in assuming improved habitat following growing season burning as declines in av i a n abundanc e have been observed soon after treatment application (Fitz gerald and Tanner 1992) However, in general r ecent research has concluded growing season burning may not be as problematic to breeding birds as once thought (Cox and Jones 2007, Tucker et al. 2004). The results of this study suggest avian communities w ere largely unaf fected by dormant season roller chopping. However, growing season roller chopping resulted in increases in permanent resident and breeding category abundance. Within the breeding category, increases in abundance were observed for successi onal scrub and grassland guilds. The effects of ro ller chopping on avian communities have been examined in dry prairie habitats, with reductions in species richness and abundance observed on dormant and growing season roller chopped sites (Fitzgerald and Tanner

PAGE 125

125 1992). Prop o sed reas ons for these declines include reduced perch availability and vertical vegetation diversity Reductions in shrub cover following treatment application significantly reduced the abundance of shrublevel inhabitants such as whiteeyed vireo ( Vireo griseus Boddaert ) northern cardinal ( Cardinalis cardinalis L. ) and grey catbird ( Dumetella caroliniensis L ; Fitzgerald and Tanner 1992) In pine flatwoods, g rowing season roller chopping cause s prolonged ( 2 year) reductions in shrub cover, height and density and herbaceous plant species richness, cover and height ( Chapter 2 and 3). However, these changes appear to maintain or enhance abundance and species richness within cert ain av i a n categories and gui lds. This includes increases in successional scrub guild abundance even t hough this guild contains species observed to decline following roller chopping of dry prairie (Fitzgerald and Tanner 1992). Roller chopping/ burning had a positive effect on a numb er of avian categories and guild s. Grassland guild richness and total av ian, permanent resident and breeding category, and grassland guild abundance increased following the use of this practice. No other study exam ining the effect of this combination tre atment on avian communities has been conducted. Prescribed burns were applied following roller chopping, typically in the early growing season. This as for growing season burning and roller chopping resulted in declines in herbaceous species richness, cover, and abundance. However, due to the timing of burning there is potential for increased seed ( Myers and Boettcher 1987, Platt et al. 1988, Outcault 1990, Streng et al. 1993, Hardy 2003) and insect production ( Streng et al. 1993 Hanula and Fanzeb 199 8 Provencher et al. 1998, Collins et al. 2003, Hardy 2003) This may improve food and cover conditions for certain species and explain the increases in species richness and abundance observed.

PAGE 126

126 Grazing did not interact with the effects of burning and roll er chopping to impact av i a n communities. However, grazing should be applied cautiously in areas where avian conservation and management are a priority. Grazing can result in decreased spatial heterogeneity through reductions in plant biomass and cover and changes in structural conditions (e.g., plant density and height, and litter cover and depth; Chapter 2 ; Vallentine 1990, Milchunas and Lauenroth 1993, Fuhlendorf and Engle 2001, Derner et al. 2009). In addition, species such as creeping bluestem ( Schiz achyrium scoparium [Michx.] Nash var. stoloniferum [Nash] Wipff ), chalky bluestem ( Andropogon capillipes Nash) and wiregrass decline when grazed immediately following burning (White and Terry 1979, Sievers 1985). These changes in the plant community can r educe the suitability and availability of food and cover resources for a variety of avian species (Saab et al. 1995, Brennan and Kuvalesky 2005, Coppedge et al. 2008, Derner et. al. 2009). However, rather than managing l ivestock for uniform use of vegetat io n or management to the middle, with extremes in vegetations structure (e.g., low sparse and high dense ) absent (Derner et al. 2009), there is the potential for them to be used as ecosystem engineers. Herbivores naturally exhibit preference for the consumption of certain plants over others (Van Soest 1996). If stocking rates are appropriate and pastures of a sufficient size, this results in differential patterns of use of individual plant species across a pasture (Launchbaugh and Howery 2005), maintaining or increasing the heterogeneity of vegetation and potential benefiting av i a n species (Derner et al. 2009). Habitat characteristics that were most often identified as positively related to av i a n category and guild species richness were shrub richness, c over, and density, for b cover

PAGE 127

127 and height, and arthropod familial richness. All of the treatments examined cause d reductions in some, if not all of these habitats characteristics (Chapter 2, 3, and 5). However, some avian categories and guilds were neg atively affected by these habitat characteristics as well as graminoid cover and forb richness. Treatments therefore, stand to benefit some avian communities through reductions in these habitat characteristics. Generally, treatment applications need to be carefully considered to most appropriately benefit the avian community of primary concern. To maintain biodiversity, it seems most appropriate to adopt a strategy of fire and roller chopping application that is diverse in season, frequency, and space. Management Implications Dormant season burning should be used cautiously in situations where conservation and management of non breeding, overwintering species is a priority. The use of this practice can result in decreases in species richness and abundance within this category, potentially due to reductions in ground level food and cover resources. In situations where the maintenance or enh ancement of the non breeding av i a n community is desired, growing season burning appears more beneficial and should be used in preference to dormant season bur n ing. This practice also has the potential to benefit the migrant bird community and promote species richness and abundance within certain breeding bird guilds. The use of growing season roller chopping appears t o provide a potential treatment alternative in situations where positive effects on permanent resident and breeding species are desired. However, studies examining the effects of this treatment on avian communities have been contradictory and further rese arch should be conducted before it is widely applied as an avian management and co nservation

PAGE 128

128 strategy. The same applies to roller chopping/prescribed burning, the effects of which have been little studied. Maintenance of avian biodiversity may be best ac hieved through the diverse application of fire and roller chopping based on season, frequency, and space.

PAGE 129

129 Table 5 1 Avian migratory and breeding cat egory composition and abundance in prescribed burned and roller chopped Florida flatwoods, 2007 2008 G uild a Common Name Scientific Name Abundance (n o. of individuals) Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn C b T c C T C T C T C T RE, BD(WD) American crow Corvus brachyrhynchos 66 84 32 27 21 23 13 10 34 14 MI(SD), NBD American robin Turdus migratorius 21 50 0 6 0 3 2 1 1 4 RE, BD(WD) Bachmans sparrow Aimophila botterii 42 86 41 39 24 29 28 33 33 32 MG(NM), NBD Barn swallow Hirundo rustica 6 0 0 0 3 1 0 0 1 5 RE, BD(WD) Barred owl Strix varia 0 0 1 0 1 1 0 0 0 0 MG(NM), NBD Black and white warbler Mniotilta varia 0 1 0 0 0 0 0 0 0 0 RE, BD(WT) Black bellied whistling duck Dendrocygna autumnalis 0 0 0 6 0 0 0 0 0 0 RE, BD(WD) Blue gray gnatcatcher Polioptila caerulea 18 5 4 1 1 5 7 2 8 7 MG(SM), BD(SS ) Brown headed cowbird Molothrus ater 17 1 0 0 0 0 0 0 0 0 RE, BD(WD) Brown headed nuthatch Sitta pusilla 18 13 6 7 14 8 14 33 8 14 RE, BD(SS) Black vulture Coragyps atratus 3 0 0 0 0 0 0 0 1 0 RE, BD(UB) Blue jay Cyanocitta cristata 19 20 16 15 11 14 1 4 5 7 16 RE, BD(WT) Boat tailed grackle Quiscalus major 1 38 0 0 5 3 0 0 0 1 RE, BD (SS) Brown thrasher Toxostoma rufum 1 1 0 1 2 0 1 0 3 0 RE, BD(WD) Carolina chickadee Poecile carolinensis 1 2 0 0 0 0 0 1 0 0 RE, BD(SS) Carolina wren Thryothorus ludo vicianus 26 26 21 13 12 10 8 6 18 11 RE, BD(WT) Cattle egret Bubulcus ibis 4 5 0 1 1 1 0 0 0 3 MG(NM), NBD Chipping sparrow Spizella passerina 7 9 3 21 2 7 1 2 2 2 RE, BD(UR) Common grackle Quiscalus quiscula 72 3 6 1 6 5 3 4 4 16 RE, BD(SS) Common gro und dove Columbina passerine 11 20 7 8 15 23 7 7 12 14 MG(NM), BR(SS) Common nighthawk Chordeiles minor 14 15 12 3 9 9 4 5 11 10 MG(SM), NBD Common snipe Gallinago gallinago 0 1 0 0 0 0 0 0 2 3 RE, BD(SS) Common yellowthroat Geothlypis trichas 83 76 46 46 27 25 23 20 22 29 RE, BD(WD) Downy woodpecker Picoides pubescens 6 5 2 4 3 2 5 3 0 1 RE, BD(GR) Eastern bluebird Sialia sialis 11 22 6 6 25 8 22 30 8 21 MG(NM), BD(GR) Eastern kingbird Tyrannus tyrannus 4 0 0 0 0 0 0 0 2 0 RE, BD(GR) Eastern meadowl ark Sturnella magna 67 87 53 69 160 178 121 132 96 142

PAGE 130

130 Table 5 1. Continued Guild a Common Name Scientific Name Abundance (n o. of individuals) Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn C b T c C T C T C T C T MG(SM), NBD Eastern phoebe Sayornis phoebe 1 0 0 0 0 0 0 0 0 1 RE, BD(SS) Eastern towhee Pipilo erythrophthalmus 174 213 77 116 105 106 91 132 119 154 RE, BD(UB) Eurasian collard dove Streptopelia decaocto 0 0 0 3 0 0 0 0 0 0 RE, BD(UB) European star ling Sturnus vulgaris 0 0 0 0 0 0 2 0 0 0 RE, BD(WD) Fish crow Corvus ossfragus 4 3 2 1 2 2 1 0 1 1 MG(NM), NBD Gray catbird Dumetella carolinensis 31 19 2 5 5 4 2 1 3 1 MG(NM), BD(WD) Great crested flycatcher Myiarchus crinitus 19 16 5 2 12 4 15 13 10 20 RE, BD(WD) Great horned owl Bubo virginianus 0 0 0 0 0 0 1 0 0 0 MG(NM), NBD House wren Troglodytes aedon 0 0 0 0 0 1 0 0 0 0 RE, BD(GR) Killdeer Charadrius vociferous 5 6 0 0 2 1 1 0 1 4 RE, BD(GR) Loggerhead shrike Lanius ludovicianus 5 5 4 6 9 10 9 10 5 4 RE, BD(WT) Mottled duck Anas fulvigula 4 5 2 4 3 4 1 0 3 3 RE, BD(UB) Mourning dove Zenaida macroura 22 38 23 23 42 37 19 27 23 27 RE, BD(SS) Northern bobwhite Colinus virginianus 38 44 38 34 36 36 24 18 22 24 RE, BD(SS) Northern cardinal Car dinalis cardinalis 69 68 45 57 30 28 10 8 15 19 RE, BD(UB) Northern mockingbird Mimus polyglottos 9 20 4 6 27 25 10 1 14 21 MG(NM), BD(WD) Northern parula Parula americana 5 5 6 8 4 3 3 2 4 2 RE, BD(WD) Palm warbler Dendroica palmarum 13 15 0 2 6 3 2 1 0 0 RE, BD(WD) Pileated woodpecker Dryocopus pileatus 7 11 10 10 7 7 7 10 2 6 RE, BD(WD) Pine Warbler Dendroica pinus 4 9 0 1 2 1 0 1 4 3 RE, BD(WD) Red bellied woodpecker Melanerpes carolinus 51 70 34 35 37 34 34 33 23 36 RE, BD(WD) Red cockaded woodp ecker Picoides borealis 4 7 6 6 6 3 4 4 1 2 RE, BD(WD) Red headed woodpecker Melanerpes erythrocephalus 6 4 1 2 0 1 2 1 1 1 RE, BD(WD) Red shouldered hawk Buteo lineatus 8 4 3 4 1 3 4 3 3 2 RE, BD(WD) Red tailed hawk Buteo jamaicensis 8 2 2 6 7 10 17 12 10 16 RE, BD(WT) Red winged blackbird Agelaius phoeniceus 45 42 7 12 64 71 44 62 65 74 RE, BD(WT) Sandhill crane Grus canandensis 28 32 18 19 23 22 32 20 36 28 MG(SM), NBD Savannah sparrow Passerculus sandwichensis 17 13 2 2 8 6 0 0 1 1

PAGE 131

131 Table 5 1 Con tinued Guild a Common Name Scientific Name Abundance (n o. of individuals) Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn C b T c C T C T C T C T MG(NM), BD(WD) Swallow tailed kite Elanoides forficatus 0 0 0 0 0 0 1 0 0 0 MG(SM), NBD Tree swallow Tachycineta bicolor 6 0 0 0 0 0 0 0 0 0 RE, BD(WD) Tufted titmouse Baeolophus bicolor 6 8 6 9 4 5 2 1 2 8 RE, BD(WD) Turkey vulture Cathartes aura 0 2 2 3 0 0 1 0 1 1 RE, BD(SS) White eyed vireo Vireo giseus 24 16 15 1 1 4 5 2 12 12 6 RE, BD(WD) Wild turkey Meleagris gallopavo 7 6 0 1 0 1 2 2 1 1 RE, BD(WT) Wood stork Mycteria americana 0 0 1 0 0 0 1 0 0 0 MG(SM), NBD Yellow bellied sapsucker Sphyrapicus varius 0 1 0 0 0 0 0 0 0 0 MG(NM), NBD Yellow rumped warbler De ndroica coronate 3 5 0 0 1 0 0 1 0 1 RE, BD(WD) Yellow shafted flicker Colaptes auratus 3 4 6 4 3 5 3 0 3 2 MG(NM), NBD Yellow throated warbler Dendroica dominica 1 0 0 0 0 0 0 0 0 0 a Migrant status categories and guilds: RE = Resident category MG = M igrant category (SM = Short distance migrant guild, NM = Neotropical migrant guild. Breeding status categories and guilds: NBD = Nonbreeding category, BD = Breeding category (WD = Woodlan d guild, UB = Urban guild, SS Successional scrub guild, WT = Wetland guild, GR = Grassland guild) b C = Control c T = Treated.

PAGE 132

132 Table 5 2 Effects of treatment on avian richness and abundance in Florida flatwoods, 2007 2008 Avian Abundance and Richness by Treatment Type a, b Treatment ( SE) P Control Treated Dormant season burn Richness (n o. of species) Non breeding 1.2 0.2 0.9 0.2 0.004 Growing season burn Richness (n o. of species) Migrant 0.6 0.1 1.1 0.2 0.057 Neo tropical migrant 0.6 0.1 0.9 0.1 0.085 Urban 1.1 0.1 0. 8 0.1 0.092 Abundance (n o. of individuals) Wetland 0.8 0.3 1.2 0.4 0.078 Growing season roller chop Abundance (n o. of individuals) Permanent resident 14.2 1.0 15.6 1.1 0.058 Breeding 14.7 1.0 16.0 1.0 0.060 Succe ssional scrub 4.2 0.5 4.8 0.4 0.054 Grassland 3.2 0.5 3.5 0.4 0.086 Roller chop/burn Richness (n o. of species) Grassland 0.9 0.1 1.1 1.1 0.021 Abundance (n o. of individuals) Total 15.4 1.4 19.1 1.1 Permanent resident 14.6 1.4 17.9 1.1 0.017 Breeding 15.2 1.4 18.7 1.1 0.017 Grassland 2.5 0.4 3.5 0.4 0.058 a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categor ies and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds) b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by treatment presented ( P 0.1).

PAGE 133

133 Table 5 3 Effects of treatment season interactions on avian richness and abundance in Florida flatwoods, 2007 2008 Avian Abundance and Richness by Treatment Type a, b Season Treatment ( SE) c P Control Treated Dorman season burn Abundance (n o. of individuals) Migrant Winter 4.6 0.9 A 5.3 1.7 A 0.038 Spring 2.1 0.5 A 0.8 0.2 B Summer 1.6 0.3 A 1.3 0.2 A Neo tropical migrant Winter 1.9 0.3 A 2.2 0.4 A 0.024 Spring 2.1 0.5 A 0.8 0.2 B Summer 1.3 0.3 A 1.1 0.1 A Non breeding Winter 3.8 0.8 A 5.3 1.7 A 0.007 Spring 1.1 0.4 A 0.1 0.1 B Summer 0.4 0.1 A 0.2 0.1 A Growing season burn Richness (n o. of species) Non breeding Winter 0.3 0.2 A 1.3 0.4 B 0.039 S pring 0.0 0.0 A 0.2 0.1 A Summer 0.1 0.1 A 0.1 0.1 A Wetland Winter 0.3 0.2 A 0.6 0.3 A 0.075 Spring 0.3 0.2 A 0.6 0.3 A Summer 0.5 0.2 A 0.5 0.1 A Abundance (n o. of individuals) Non breeding Winter 0.6 0.3 A 3.0 1.0 B 0.058 Spring 0.0 0.0 A 0.8 0.7 A Summer 0.1 0.1 A 0.1 0.1 A a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment season interaction presented ( P 0.1). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 134

134 Table 5 4. Effects of treatment grazing interactions on avian richness and abundance in Florida flatwoods, 2007 2008. Avian Abundance and Rich ness by Treatment Type a, b Grazing Treatment ( SE) c P Control Treated Dorman season burn Abundance (n o. of individuals) Successional scrub Nongrazed 7.1 1.1 A 9.6 1.2 A 0.080 Grazed 8.0 0.7 A 7.3 0.5 A Dormant season roller chop Richness (n o. of species) Permanent resident Nongrazed 8.1 0.8 A 9.1 0.9 A 0.072 Grazed 8.0 0.5 A 7.8 0.6 A Breeding Nongrazed 8.7 0.9 A 9.6 0.9 A 0.053 Grazed 8.4 0.5 A 8.2 0.7 A Roller chop/burn Richness (n o. of species) Non breeding Nongrazed 0.3 0.1 A 0.2 0.1 A 0.025 Grazed 0.1 0.1 A 0.3 0.1 A Abundance (n o. of individuals) Non breeding Nongrazed 0.3 0.1 A 0.3 0.2 A 0.031 Grazed 0.2 0.1 A 0.5 0.3 A a Migrant status categories and guilds: r esident c ategory m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment grazing interaction presented ( P 0.1). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1). Table 5 5. Effects of treatment time interactions on avian richness and abund ance in Florida flatwoods, 2007 2008. Avian Abundance and Richness by T reatment Type a, b Time c Treatment ( SE) d P Control Treated Dorman season roller chop Abundance (n o. of individuals) Non breeding 1 1.1 0.4 A 0.7 0.3 A 0.087 2 0.1 0.1 A 0.4 0.2 A Growing season roller chop Richness (n o. of species) Grassland 1 1.0 0.1 A 1.2 0.2 A 0.059 2 1.2 0.1 A 1.1 0.1 A a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment time interac tion presented ( P 0.1). c Time s ince treatment application (years) d Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 135

135 Table 5 6. Comparison of the effects of treatment on avian richness and abundance in Florida flatwoods, 200 7 Avian Richness and Abundance a ,b Treatment Type ( SE) c P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn Richness (no. of species) Total 9.7 0.3 ABC 11.0 0.5 AB 10.2 0.6 AB 9.3 0.6 AB 8 .6 0.5 B 9.7 0.5 C 0.001 Permanent resident 8.5 0.2 ABC 9.2 0.4 A 9.1 0.6 BC 8.4 0.5 A 8.0 0.5 C 8.7 0.5 AC 0.059 Breeding 9.1 0.2 AB 9.8 0.5 AC 9.7 0.7 B 8.8 0.6 AC 8.5 0.5 B 9.4 0.5 C 0.039 a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian reside ncy and breeding categories and guilds with richness or abundance significantly affected by treatment presented ( P 0.1). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1). Table 5 7. Comparison of the effects of treatment season interactions on avian richness and abundance in Florida flatwoods, 2007 Avian Richness and Abundance a,b Season Treatment Type ( SE) c, P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop / Burn Richness (n o. of species) Non breeding Winter 1.2 0.2 A 2.3 0.3 AB 1.3 0.4 B 1.3 0.4 AB 0.4 0.3 A 0.8 0.3 AB 0.095 Spring 0.3 0.1 A 0.1 0.1 B 0.2 0.1 AB 0.1 0.1 AB 0.0 0.0 AB 0.1 0.1 AB Summer 0.2 0.0 A 0.3 0.1 A 0.1 0.1 A 0.1 0.1 A 0.0 0.0 A 0.1 0.1 A Abundance (n o. of individuals) Non breeding Winter 1.7 0.3 A 5.3 1.7 AB 3.0 1.0 B 1.8 0.6 AB 0.5 0.3 A 1.1 0.5 AB 0.058 Spring 0.5 0.1A 0.1 0.1B 0.8 0.7AB 0.1 0.1AB 0.0 0.0AB 0.4 0.3AB Summer 0.2 0.1 A 0.2 0.1 A 0.1 0.1 A 0.1 0.1 A 0.0 0.0 A 0.1 0.0 A Grassland Winter 2.0 0.4 AB 1.4 0.4 B 1.6 0.6 B 2.4 0.6 AB 4.3 0.8 A 4.9 1.0 A 0.016 Spring 2.8 0.4A 1.3 0.4B 2.9 1.1AB 4.8 0.8A 3.3 1.0AB 2.0 0.4AB Summer 2 .0 0.4 A 2.0 0.4 A 1.9 0.5 A 4.0 0.1 A 3.2 0.5 A 2.5 0.4 A a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment season interaction presented ( P 0. 1). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 136

136 Table 5 8 Comparison of the effects of treatment time interactions on avian richness and abundance in Florida flatwoods, 2007 Avian Richness and A bundance a,b Time c Treatment Type ( SE) d P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn Richness (n o. of species) Woodland 1 3.5 0.2A 4.0 0.4A 5.1 0.9A 2.6 0.5A 2.8 0.7B 2.9 0.6A 0. 085 2 3.0 0.2 A 3.4 0.3 A 3.2 0.3 A 3.0 0.4 A 3.1 0.3 A 3.0 0.4 A Abundance (n o. of individuals) Woodland 1 4.0 0.3 A 6.1 0.5 AB 5.4 0.9 B 2.9 0.3 AB 2.4 0.6 A 3.5 0.7 AB 0.022 2 4.5 0.3AC 5.6 0.6A 4.5 0.5A 4.4 0.6B 5.3 0.7C 4.9 0.8ABC a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment time interaction presented ( P 0.1). c Time since treatment application (years) d Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 137

137 Table 5 9 Comparison of the effects of treatment grazing interactions on avian richness and abundance in Florida flatwoods, 2007 Avian Richness and Abundance a,b Grazing Treatment Type ( SE) c P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn Richness (n o. of species) Woodland Nongrazed 3.6 0.2AB 3.4 0.7A 3.9 0.4AB 3.7 0.5B 3.9 0.3C 3.0 0.4ABC 0.010 Grazed 2.8 0.2 A 3.8 0.3 A 3.0 0.7 A 2.1 0.3 A 1.7 0.5 A 3.0 0.5 A Urban Nongrazed 1.0 0.1 AB 1.5 0.3 A 0.9 0.2 AB 1.0 0.2 AB 1.3 0.3 AB 0.7 0.2 B 0.004 Grazed 0.9 0.1 A 0.9 0.1 A 0.5 0.3 A 1.0 0.2 A 0.9 0.1 A 1.5 0.2 A Abundance (no. of individuals) Total Nongrazed 16.4 0.8 A 23.0 1.1 A 19.4 1.8 A 15.9 1.3 A 16.1 1.4 A 18.9 1.7 A 0.053 Grazed 17.4 1.0 A 21.7 2.1 B 15.7 1.9 AB 17.9 1.4 A 16.2 1.7 AB 19.4 1.5 A Woodland Nongrazed 4.8 0.3A 4.7 0.9B 4.9 0.5A 4.4 0.6C 5.6 0.7D 4.3 0.7A 0.034 Grazed 4.0 0.3 A 6.1 0.4 A 4.1 1.2 A 3.2 0.6 B 3.0 0.9 AB 6.1 0.4 A Successional scrub Nongrazed 5.5 0.4 A 9.6 1.2 B 8.1 0.9 AB 4.8 0.6 AB 4.7 0.6 A 6.2 0.7 AB 0.003 Grazed 6.5 0.4AB 7.3 0.5A 8.9 1.0B 5.2 0.6AB 5.0 0.7A 7.3 0.5AB Grassland Nongrazed 2.4 0.2 A 1.3 0.6 A 2.3 0.5 A 2.8 0.3 A 2.7 0.4 A 3.9 0.6 A 0.054 Grazed 2.4 0.2 A 1.6 0.3 A 1.4 0.6 A 4.8 0.6 A 4.8 0.7 A 1.6 0.3 A a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding categor y, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Only avian residency and breeding categories and guilds with richness or abundance significantly affected by a treatment grazing interaction presented ( P 0.1). c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 138

138 Table 5 10. Habitat characte ristics that best predict total, category and guild specific avian abundance and richness in Florida flatwoods 2007 2008 Dependant Variable a Independent Model Variables b Standardized Regression Coefficient R2 P Richness (n o. of species) Total Mean soil moisture (%) Shrub density (no./m2) Mean forb richness (no. of species) Mean forb height (cm) Shrub cover (%) Mean soil density (g/cm3) Mean graminoid cover (%) Mean litter depth (cm) Arthropod family richness (no. of families ) Shrub richness (no. of species) Variance of soil density (g/cm3) Variance of forb cover (%) Variance of soil pH Variance of graminoid richness (n o. of species) 0.233 0.187 0.175 0.158 0.158 0.133 0.125 0.118 0.103 0.096 0.084 0.077 0.062 0.052 0.242 0.001 Permanent resident Mean soil moisture (%) Shrub cover (%) Mean forb height (cm) Mean for b richness (no. of species) Shrub density (no./ m2) Mean soil density (g/cm3) Mean graminoid cover (%) M ean litter depth (cm) Shrub richness (no. of species) Arthropod family richness (no. of families) Variance of forb cover (%) Mean forb cover (%) Variance of soil pH Variance of soil density g/cm 3 ) 0.2 12 0.175 0.166 0.160 0.155 0.139 0.136 0.109 0.106 0.104 0.101 0.097 0.077 0.076 0.246 0.001 Breeding Mean soil moisture (%) Shrub density (no./m2) Mean forb height (cm) Shrub cover (%) Mean graminoid cover (%) Mean forb richness (n o. of species) 0.219 0.172 0.168 0.163 0.158 0.153 0.264 0.001

PAGE 139

139 Table 5 10 Cont inued Dependant Variable a Independent Model Variables b Standardized Regression Coefficient R 2 P Mean soil density (g/cm 3 ) Arthropod family ric hness (no. of families) Shrub richness (no. of species) Mean litter depth (cm) Variance of soil density (g/m3) Variance of forb cover ( n o of species) Variance of pH Mean forb cover (%) 0.147 0.112 0.107 0.106 0.097 0.090 0.090 0.084 Woodland Mean soil moisture (%) Shrub richness (no. of species) Mean litter depth (cm) Mean forb richness (no. of species) Shrub cover (%) Mean graminoid cover (%) Mean forb cover (%) Overstory basal area (cm2/ha) Mean soil density (g/cm3) Variance of soil density (g/cm3) Variance of litter depth (cm) Variance of litter cover (%) Shrub density (no./m2) Variance of pH 0.311 0.198 0.197 0.195 0.182 0.157 0.150 0.136 0.126 0.101 0.093 0.079 0.078 0.069 0.288 0.001 Successional scrub Mean forb height (cm) Shrub density (no./m2) Mean forb richness (no. of species) Shrub cover (%) Variance of graminoid height (cm) Mean litter cover (%) Visual obstruction from (% from 5 m distance) Mean graminoid height (cm) Var iance of graminoid cover (%) Mean graminoid cover (%) Variance of forb height (cm) Variance of pH Arthropod family richness (n o. of families) 0.196 0.190 0.181 0.150 0.124 0.108 0.101 0.086 0.085 0.059 0.050 0.037 0.026 0.241 0.001

PAGE 140

140 Table 5 10 Con tinued Dependant Variable a Independent Model Variables b Standardized Regression Coefficient R 2 P Wetland Variance of forb cover (%) Shrub richness (no. of species) Mean forb richness (no. of species) Understory density (no./ha) Shrub density (no./m2) Ar thropod abundance (no. of individuals) Shrub height (cm) Mean graminoid height (cm) Mean litter cover (%) Shrub cover (%) Variance of graminoid height (cm) Mean forb height (cm) Mean soil moisture (%) Variance of soil density Mean soil density (g/cm 3 ) 0.234 0.207 0.132 0.125 0.121 0.118 0.104 0.098 0.088 0.083 0.076 0.069 0.064 0.058 0.053 0.208 0.001 Grassland Mea n graminoid richness (n o. of species) Shrub cover (%) Shrub height (cm) Overstory basal area (cm2/ha) Variance of graminoid richness (n o. of species) Mean soil density (g/cm3) Understory density (no./ha) Mean soil moisture (%) Mean graminoid cover (%) Arth ropod family richness (no. of families) Mean forb cover (%) Variance of forb cover (%) Variance of graminoid height (cm) Variance of soil density (g/cm3) Mean pH 0.208 0.165 0.147 0.147 0.132 0.125 0.125 0.105 0.105 0.097 0.095 0.078 0.072 0.065 0.061 0.221 0.001 Abundance (No. of Individuals) Woodland Mean litter depth (cm) Mean forb richness (No. of species) Overstory density (No./ha) Shrub cover (%) Mean soil moisture (%) Mean forb cover (%) 0.240 0.207 0.198 0.190 0.176 0.165 0.239 0.001

PAGE 141

141 Table 5 10 Continued Dependant Variable a Independent Model Variables b Standardized Regression Coefficient R 2 P Shrub richness (n o. of species) Arthropod abundance (no. of individuals) Mean graminoid richness (no. of species) V ariance of graminoid ric hness (no. of species) Variance of litter height (cm) Variance of litter cover (%) Shrub density (no./m2) Mean graminoid cover (%) Variance of forb height (cm) 0.158 0.151 0.130 0.110 0.096 0.092 0.089 0.074 0.057 Successional scrub Mean forb heigh t (cm) Shrub cover (%) Shrub richness (n o. of species) Mean litter cover (%) Mean forb richness (no. of species) Overstory basal area (cm3/ha) Variance of graminoid height (cm) Mean pH Variance of soil moisture (%) Understory cover (%) Variance of forb hei ght (cm) Mean soil density (g/cm3) Variance of litter depth (cm) V ariance of graminoid richness (no. of species) Variance of graminoid cover (%) 0.220 0.214 0.149 0.145 0.138 0.106 0.105 0.101 0.090 0.074 0.057 0.048 0.047 0.045 0.044 0.252 0.001 Wetland Mean forb richness (no. of species) Variance of forb cover (%) Shrub richness (no. of species) Mean litter cover (%) Understory density (no./ha) Visual obstruction (% from 10 m distance) Mean soil density (g/cm3) Shrub height (cm) Variance of soil density (g/cm3) Mean graminoid height (cm) Mean forb height (cm) Variance of pH Variance of graminoid height (cm) 0.190 0.186 0.142 0.130 0.105 0.104 0.098 0.097 0.096 0.077 0.076 0.072 0.068 0.224 0.001

PAGE 142

142 Table 5 10 Cont inued Dependant Variable a Independent Model Variable s b Standardized Regression Coefficient R 2 P Overstory cover (%) Arthropod abundance (n o. of individuals ) 0.060 0.052 Grassland Mean forb richness (no. of species) Understory density (no./ha) Overstory basal area ( cm3/ha) Shrub height (cm) Mean soil density (g/cm3) Variance of graminoid richness (n o. of species) Shrub cover (%) Variance of forb cover (%) Variance of forb height (cm) Mean pH Mean graminoid richness (no. of species) Variance of graminoid height (cm) V isual obstruction (% from 10 m distance) Mean litter depth (cm) Variance of soil density (g/cm 3 ) 0.208 0.146 0.138 0.123 0.117 0.099 0.086 0.083 0.064 0.060 0.057 0.057 0.055 0.028 0.023 0.214 0.001 a Migrant status categories and guilds: r esident category m igrant category (short distance migrant and n eotropical migrant guilds). Breeding status categories and guilds: nonbreeding category, breeding category (woodland, urban, successional scrub, wetland and grassland guilds). b Vegetation characteristics selected using backward stepwise multiple linear regression.

PAGE 143

143 CHAPTER 6 DIURNAL LEPIDOPTERAN RESPONSE TO PRESCRIBED BURNING A ND ROLLER CHOPPING IN FLORIDA FLATWOODS Introduction Lepidoptera (butterflies and moths) play an essential role in natural systems as herbivores and pollinators (Scott 1986, Hendrix and Kyhl 2000), and have an ability to foster public sympathy, something that is unusual among insects. They are often considered a flagship or umbrella taxa for the conservation of other wildlif e and have the potential to act as indicators of habitat type, quality, and/or condition (Erhardt 1985, Kremen 1992, Launer and Murphy 1994, Nelson and Anderson 1994, New 1997) and the presence of certain bird species (Swengel and Swengel 1999). As a result, they are frequently the target of invertebrate management, research, and conservation efforts (New 1997). Many species of Lepidoptera inhabit Floridas pine flatwoods (Gerberg and Arnett 1989, Covel l 1984, Opler 1998), including 2 species of butterfly, the arogos skipper ( Atrytone arogos arogos ) and southern dusted skipper ( Atrytonopsis hianna loammi Whitney ), listed by the Florida Fish and Wildlife Conservation Commission (FWC) as species having declining populations and of conservation need (FWC 2005) U nfortunately, the pine flatwoods that provide habitat for these Lepidoptera are also deteriorating and in recent years have exhibited considerable declines in quantity and quality (Abrahamson and Hartnett 1990, United States Fish and Wildlife Service [ US FWS] 1999, FWC 2005). Seventy five percent of Floridas pine flatwoods (Cox et al. 1997) are privately owned (FWC 2005) and used primarily for livestock production. Changes to management practices in these flatwoods and pinelands across the southeastern U S partic u la rly the modification of historic fire regimes (i.e., deviations in

PAGE 144

144 fire intensity, retur n frequency and seasonality) have resulted in excessive shrub growth and the loss of herbaceous vegetation (Wade et al. 1980, Robbins and Myers 1992, Glitzenstein et al. 1995, Abrahamson and Abrahamson 1996, Platt 1998 ). These vegetative changes have alter ed the structure and composition of flatwoods, contributing to their degradation and reducing their suitability for associated wildlife, including g rasslandand open country associated Lepidopteran species (FWC 2005) In an attempt to maintain and restore remaining areas of privately owned pine flatw oods, FWC and the United States Department of Agriculture are utilizing assistance based programs su ch as the Farm Bills Environmental Quality I ncentives Program and Wildlife Habitat Incentives Program, to encourage landowners to implement appropriate management activities on their lands. Management activities promoted under these programs inc lude the use of prescribed fire and roller chopping duri ng dormant (November March) and growing (April October) seasons. Prescribed fire and roller chopping are management techniques that have been shown to reduce shrub cover and encourage the growth and flowering of grasses and forbs ( Chapter 2; Platt et al. 1988, Tanner et al. 1988, Fitzgerald and Tanner 1992, Glitzenstein et al. 1995, Watts and Tanner 2003, Watts et al. 2006) helping to improve the quality of degraded pineland and prairie habitats where changes in fire use have permitted the proliferation of shrubby vegetation to the detriment of herbaceous groundcover species. The effects management activities have on Lepidopteran communities occupying Floridas pine flatwoods have not been studied. Many rare b utterflies require vegetation management in some form to maintain their populations (New 1991, Oates 1995, Robertson et al. 1995). In Floridas flatwoods it might be suspected that prescribed

PAGE 145

145 burning and roller chopping may benefit grasslandassociated butterfly species through a reduction in shrub cover and an increase in herbaceous plant species ( Chapter 2; Platt 1988, Tanner et al. 1988, Fitzgerald and Tanner 1992, Glitzenstein et al. 1995, Watts and Tanner 2003, Watts et al. 2006), potentially providing supplementary food sources in the form of nectar producing plants and additional sites for egg laying and caterpillar development. However, in other areas of North A merica, declines in Lepidopteran species richness and abundance have been observed following fire, at least in the short term, suggesting that this management practice may threaten populations of locally endangered species (Dunwiddie 1991, Siemann 1997). Swengel (1996, 1998) also observ ed declines in many Lepidopteran species following fire, and proposed that nonfire vegetation management strategies (e.g., mowing) might be more favorable for maintenance of specialist butterflies than prescribed burning. With assistance based management programs currently encouraging the use of fire and roll er chopping on Florida rangelands, we need to investigate further the impacts these practices have on pine flatwoodsassociated Lepidoptera. Such research will determine whether the use of these practices to manage pine flatwoods vegetation is also appropriate for the conservation of this insect group. The objectives of my study were to 1) compare diurnal Lepidopter a n species richness and abundance on treated (management activities implemented) and untreated (no management activi ties implemented) pine fla twoods sites, 2) contrast the species richness and abundance of nectar producing plants on treated and untreated sites, and 3) i nvestigate the relative importance of local pine flatwoods habitat characteristics (e.g., flowering forb and shrub abundance, sh rub

PAGE 146

146 density, forb cover, graminoid height) may play in determining Lepidoptera n species richness and abundance. Methods Study Sites I conducted research on 50 privately and publicly owned, paired treatment and control sites across 6 counties (Desoto, Hig hlands, Lee, Manatee, Osceola, and Sarasota) in central and south Florida. Study sites consisted of pine flatwoods habitat s with varying management histories and grazing regimes that were being prescribed burned and roller chopped by local landowners and land managers using varying, individual protocols. Floridas pine flatwoods are characterized as having an overstory of scattered slash ( Pinus elliotti Engelm.) and longleaf ( P. palustris Mill. ) pine, either in pure stands or in combination. The understory and shrub layer includes saw palmetto ( Serenoa repens [ Bartram] Small) wax myrtle ( Morella cerifera [L.] Small), gallberry ( Ilex glabra [Pursh] Chapm.), fetterbush ( Lyonia lucida [Lam.] K. Koch), staggerbush ( Lyonia fruticosa [Michx.], G. S. Torr), dwa rf huckleberry ( Gaylussacia dumosa [Andrews] Torr. & A. Gray), dwarf live oak ( Quercus mimima [Sarg.] Small), and tarflower ( Bejaria racemosa Vent.). An appreciable herbaceous layer exists when the shrub layer is relativel y open. This layer contains a wi de variety of grasses ( e.g., Agrostis Andropogon, Aristida Eragrostis Panicum, and Paspalum spp. ). Common forbs include legumes ( e.g., Cassia Crotalaria Galactia Tephrosia spp.), milkweeds ( Asclepias spp.), milkworts ( Polygala spp.), and a wide vari ety of composites ( e.g., Aster, Chrysopsis, Eupatorium, Liatris, and Solidago sp p. ; Abrahamson and Hartnett 1990, U S Fish and Wildlife Service 1999).

PAGE 147

147 Treatment Types Treatment types included dormant season ( November March) bur n, growing season (April Oc tober ) burn, dormant season roller chop, growing season roller chop, and a roller chop/burn combination treatment. The roller chop/burn combination treatment (hereafter referred to as roller chop/burn) involved roller chopping in the dormant seaso n follow ed by burning within 6 months. I established a total of 11 dormant season burn, 9 growing season burn, 9 dormant season roller chop, 12 growing season roller chop, and 9 roller chop/bur n site pairs Lepidoptera n Surveys I used a paired sampling approach t o examine the effects of treatment type (i.e., prescribed burning, roller chopping, and roller chopping/burning) on Lepidopter a n species richness and abundance. R ichness and abundance were compared between sampling point s randomly located in paired treate d (e.g., dormant season roller chopped) and untreated (control) flatwoods sites. Paired treatment and control sampling point s were adjacent being located in the same pasture or management unit. In addition, they were of similar current and past manageme nt (e.g., grazing intensity), surrounding landuse, plant community (e.g., overstory cover), and soil conditions. I established 1 randomly selected sampling point within each treatment and control site. To minimize edge effects, I rejected and randomly r elocated sampling point s that fell within 50 m of the edge of a treatment or control site. Sites within which treat ment and control sampling point s were located ranged from 2 20 ha in size I conducted Lepidopteran surveys in s pring (March April) of 2008 following the application of dormant and growing season burning and roller chopping treatments. Lepidoptera were surveyed using line transect techniques along 2 100 m perpendicul a r

PAGE 148

148 transects c entered on the sampling point (Swengel 1998, Giuliano et al. 2004, Shepherd and Dubinsky 2005). I walked each transect at a steady pace of 10 m/min for a total survey time of 20 min (Shepherd and Debinski 2005). During this period, I recorded a ll L epidoptera observed through binoculars or captured in a sweep net (Op ler and M aluku 1998, Glassberg 1999) The 20 min sampling period did not include capture, processing, or recording of individuals. I only conducted surveys on calm (winds <17km/h), sunny (cloud cover < 60%) and warm (temperature > 18C) days between 100 0 and 1500 hours (Shepherd and Debinski 2005). Nectar Producing Plant Sampling I also used a paired sampling approach to examine the effects of treatment type on nectar producing plant species richness and abundance, which were compared at the same sampling points used for Lepidopteran sur veys. Nectar producing plant sampling was conducted on the same day as Lepidopteran surveys and involved count ing the number of nectar producing forbs and shrubs exhibiting inflorescence within a 0.03ha nest ed plot centered on the sampling point ( Dueser and Shugart 1978, Higgins et al. 2005) Habitat Sampling I conducted habitat sampling in spring (April May) 2008 at the same point s used for Lepidoptera n surveys. At each sampling point plant community composition and st ructure, litter and soil variables, and vertical obstruction were examined in several strata (i.e., ground, herbaceous, and shrub) within the 0.03 ha nested circular plot (Dueser and Shugart 1978, Higgins et al. 2005) Ground layer I assessed litter cover (%; ocular estimate) within 4 1 m2 sub sample plots, 1 randomly locat ed in each quadrant of the 0.03ha circular plots, along

PAGE 149

149 with soil density (g/cm3), moisture (%), and pH. Litter cover was recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 99%, and 7 = 100% (Donhaue et al. 1971, Hays et al. 1981, Higgins et al. 2005). I measured soil density as the dry weight density (g/cm3) of a 5 cm diameter, 10cm deep soil core sample after oven drying at 45C for 48 hour s. Soil pH and moisture were measured using a Kelway soil tester (Rodewald and Yahner 2001). Herbaceous layer I determined species richness, cover (%; ocular estimate), and maximum height (cm) of forbs and graminoids within the 1m2 subsample plots. Fo rb and graminoid cover were recorded on a scale: 0 = 0%, 1 = 1 5%, 2 = 5 25%, 3 = 25 50%, 4 = 50 75%, 5 = 75 95%, 6 = 95 9 9%, and 7 = 100% (Hays et al.1981, Krebs 1999, Higgins et al. 2005). Shrub layer I counted and measured the height of all shrubs (woody vegetation <2.0 m in height) in 2 perpendicular 20m2 quadrats centered on the 0.03ha plot to estimate species richness (no. of species), density (no./m2), and maximum height (cm) for individual species and all combined (Hays et al. 1981, Krebs 1999, Higgins et al. 2005). Shrub cover (%) was assessed along 2 perpendicular 20m transects centered on the 0.03ha circular plot using the line intercept method (Hays et al. 1981, Higgins et al. 2005). Analyses I used mixed model regressions to examine differences in Lepidopteran and nectar producing plant species richness and abundance between untreated (control) and treated sites, both within (e.g., dormant season burn) and among (i.e., dormant season burn, growing season burn, dormant season roller chop, g rowing season roller chop, and roller chop/burn) treatment types Study site pair was included in analyses as

PAGE 150

150 a blocking factor Significant com p a risons among treatment type were followed by Fishers Protected LSD tests Multiple linear regressions were used to examine which combination of habitat characteristics best described changes in Lepidopteran species richness and abundance I subjected all predictor variables involved in pair wise correlations with r 0.7 to a univariate, one way analysis of variance (ANOVA) with each dependent variable to reduce issues with m ulticollinearity For each pair of highly correlated predictor variables, I retained the one with the greatest F value ( Noon 1981, McGarigal et al. 2000). All regression models were fit using a backward stepwise procedure with Tolerance = 0.001, F to enter = 0.15, and F to remove = 0.15. These values are considered appropriate for predictor variables that are relatively independent (SYSTAT 2007). I considered r egression models statistically and biologically significant at P 0.1 and R2 Only models considered significant are presented. I assessed t he relative importance of each variable in the best model by examining standardized regression coefficients (SC; i.e., variables with higher coefficients made greater individual contributions to the explanatory power of t he model). I rank transformed all data sets prior to analyses due to violations of normality and homogeneity of variance assumptions (Conover 1998, Zar 1999, SYSTAT 2007). Statistical significance was concluded at P I used this value, rather than the more common P ( Mapstone 1995, Zar 1999). All statistical tests were performed using SYSTAT (2007) statistical software.

PAGE 151

151 Results Lepidoptera n Species Richness and Abundance I identified 20 Lepidopteran species during the study (Table 61) Lepidoptera n species richness was unaffected by dormant and growing season burning and roller chopping and roller chopping/burning ( P ). Lep idoptera n abundance was affected by dormant season burning, decreasing by 64% from 3.3 0.6 ( SE) individuals on burn to 1.2 0.5 individuals on control sites ( P = 0.026). No other treatment had an effect on Lepidoptera n abundance ( P 424). Flowering Forb and Shrub Species Richness and Abundance Flowering forb species richness was affected by g rowing season roller chopping, i ncreasing by 67% from 4.2 0.8 species on control to 7.0 0.8 species on roller chop sites ( P = 0.011). Dorman t and growing season burning, dormant season roller chopping and roller chopping/ burning had no effect on flowering forb species richness ( P 0.146) Dormant season roller chopping affected flowering forb abundance, which increased by 682% from 38.7 18.1 individuals on control to 302.7 159.1 individuals on roller chop sites ( P = 0.017). Flowering forb species richness was unaffected by al l other treatments ( P 0.106). Growing season roller chopping affected flowering shrub species richness, which decreased by 39% from 2.8 0.5 species on control to 1.7 0.6 species on roller chop sites ( P = 0.084) Flowering shrub species richness was also affected by roller chopping/burning, decreasing by 33% from 2.4 0.3 species on control to 1.6 0.3 species on roller chop/burn sites ( P = 0.099). Comparisons among treatments revealed the effects of growing season roller chopping and roller chopping/burning on flowering

PAGE 152

152 shrub species richness were similar (Table 62). Dormant and growing season burning and dormant season roller chopping had no effect on flowering shrub species richness ( P 0.249). Growing season roller chopping affected flowering shrub abundance which decreased by 22% from 72.3 16.3 individuals on control to 55.8 51.3 individuals on roller chop sites ( P = 0.086). Roller chopping/burning also affected flowering shrub abundance, which decreased by 53% from 21.3 8.0 individuals on control to 10.0 5.0 individuals on burn/roller chop sites ( P = 0.037). Again, c omparisons among treatments revealed the effects of growing season roller chopping and roller chopping/burni ng on flowering shrub abundance were similar (Table 62). Dormant and growing season burning and dormant season roller chopping had no effect on flowering shrub abundance ( P 0.366). Lepidoptera Habitat Relationships Relationships among Lepidoptera n species richness and abundance and habitat characteristics were generally weak and best models contained large numbers of variables. Habitat characteristics that best predicted Lepidoptera n species richness were variance of litter depth (SC = 0.393), variance of graminoid cover (SC = 0.320), variance of litter cover (SC = 0.293), shrub height (SC = 0.254), mean graminoid height (SC = 0.251), flowering shrub abundance (SC = 0. 249), variance of graminoid height (SC = 0.235) mean graminoid richness (SC = 0.215), variance of soil density (SC = 0.202), mean pH (SC = 0.196), mean forb height (SC = 0.193), and mean litter cover (SC = 0.184; P R2 = 0.358). M ean forb height (SC = 0.334), variance of graminoid cover (SC = 0.299), shrub height (SC = 0.226), mean litter cover (SC = 0.214), variance of litter cover (SC = 0.210), flowering shrub abundance (SC = 0.207),

PAGE 153

153 variance of soi l density (SC = 0.203), variance of litter height (SC = 0.199), shrub cover (SC = 0.192), variance of forb cover (SC = 0.192), mean soil moisture (SC = 0.184), and mean forb cover (SC = 0.171; P = 0.026, R2 = 0.273) were the combination of variables that best predicted Lepidopteran abundance Discussion Lepidopteran species richness and abundance were largely unaffected by treatment type, the exception being dormant season burn sites where declines in abundance were observed. However, results should be interpreted cautiously as the high mobility of adult butterflies may mask management affects on this insect group (Swengel 1998). In prairie habitats, while leaving areas entirely unmanaged rarely benefits Lepidoptera, regular prescribed burning may result in low numbers particularly for more specialized species (Swengel 1998) W ildfires or less frequent burns that resemble wildfires appear more appropriate in maintaining and increasing Lepidoptera n abundance in prairie habitats (Swengel 1998) Infrequent burns that create new habitat patches to be occupied by Lepidoptera during firefree intervals, rather than repeated fires that maintain existing habitat already occupied, may be most favorable for maintenance of species richness and abundance (New 1991, 1993). In contrast, Lepidopteran abundance is higher in regularly burned compared to control Texas pine forests, with regularly burned sites having a more open midand undserstory (Rudolph and Ely 2000) The effect of season of burn on Lepidoptera has not been widely examined. However, differences in the response s of members of this insect group may depend on whether individuals are in a stage of activity or diapause (Swengel 2001). Burns that occur when Lepidoptera are in diapause may result i n greate r mortality of immature s and reduced abundance of adults later

PAGE 154

154 Prescribed burning alone had no effect on flowering f orb or shrub species richness and abundance. In Texas forests, incr e ases in flowering plants were observed following regular prescribed burning, providing additional valuable nectar sources and correspon ding to greater Lepidoptera abundance (Rudolph and Ely 2000) Platt et al. (1988) and Streng et al (1993) demonstrated that flowering in the herbaceous plant community of southeastern pine forest s increased following burning. The pattern of flowering depended on season of burn, with many forb species including Liatris Pityopsi s, and Solidago spp. flowering more profusely f ollowing growing season burning. Similar increases in flowering for b abundance might have been expected on pine flatwoods sites subject to growing season burning. If this had been the case, concurrent increases in Lepidoptera n species richness or abundance may have been observed. Growing season roller chopping resulted i n increase s in flowering forb species richness and dormant season roller chopping to increases in flowering forb abundance. These practices result in shrub cover, height, and occasionally density reductions (Chapter 2). The opening up of the ground layer and reduced shading of herbaceous vegetation that result may allow for increased flowering. How ever, there were no corresponding increases in Lepidopteran species richness and abundance on these sites. Flowering shrub species richness and abundance decr eased on growing season roller chop and roller chop/burn sites. Lepidopteran sp ecies richness and abundance were negatively affected by these habitat characteristics. Therefore, despite there being no increase in Lepidoptera n species richness and abundance on roller chopped sites, there is the potential for Lepidoptera to be positively affected by these treatments.

PAGE 155

155 Lepidopteran abundance and species richness were also negatively affected by graminoid and shrub height and cover. These habitat characteri stics are often reduced on prescribed burn and roller chopped sites (Chapter 2 and 3), potentially improving habitat for Lepidoptera. However, regression models also suggest that species richness and abundance of these insects is often positively related to forb cover and height and litter depth. These habitat characteristics also decline on prescribed burn and roller chop sites (Chapter 2 and 3), potentially hindering Lepidopteran management efforts Management Implications D ormant season burning may neg atively affect Lepidoptera n abundance in pine flatwoods. However, until further research examining the response of both immature and adult Lepidoptera to prescribed burning is conducted, the application of all treatments over large areas in situations wh ere the management of these insects is a priority should be carefully considered. Examination of only mobile adults may result in treatment effects being masked. Until further research is conducted, application of prescribed burning and roller chopping practices in pine flatwoods where active Lepidopteran management is occurring, should b e done on smaller areas in a mosaic arrangement. This will ensure a variety of pine flatwoods habitats are available for occupation by a range of Lepidopteran species.

PAGE 156

156 Table 6 1. Lepidoptera n abundance (no. of individuals) in prescribed burned and roller chopped Florida flatwo ods, 2008 Common Name Scientific Name Abundance (n o. of individuals) Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Rolle r Chop/ Burn C a T b C T C T C T C T Black swallowtail Papilio polyxenes 3 0 7 4 2 0 3 4 1 3 Cabbage white Pieris rapae 2 2 1 1 0 3 2 2 3 1 Checkered white Pontia protodice 0 1 4 4 4 4 0 3 9 6 Cloudless sulphur Phoebis sennae 2 0 1 0 0 0 1 0 1 1 Commo n buckeye Junonia coenia 0 2 0 2 2 1 4 1 4 4 Eastern tiger swallowtail Papilio glaucus 1 0 0 0 0 0 0 0 0 0 Giant swallowtail Papilio cresphontes 0 0 0 0 0 0 0 0 3 0 Gray hairstreak Strymon melinus 0 0 2 1 0 0 0 0 0 0 Gulf fritillary Agraulis vanillae 0 0 0 1 0 0 0 1 0 0 Least skipper Ancyloxypha numitor 0 0 2 0 0 1 2 0 0 1 Little metalmark Cale phelis virginiensis 3 0 0 1 0 1 2 4 0 2 Monarch Danaus plexippus 0 0 0 0 0 0 1 0 0 0 Obscure sphinx Errinyis obscura 0 0 0 5 0 0 0 0 0 0 Orange sulphur Colia s eurytheme 0 0 0 0 0 0 0 0 1 0 Palmetto skipper Euphyes arpa 0 0 0 0 0 1 0 0 0 0 Red admiral Vannesa atalanta 0 0 2 1 0 0 0 0 1 0 Short lined chocolate Agnomonia anilis 0 0 0 0 0 0 0 0 1 0 Southern broken dash Wallengrenia otho 0 0 0 0 0 0 1 0 0 0 So uthern skipperling Copaeodes minimus 2 0 2 1 0 2 0 2 0 1 Zebra swallowtail Eurytides marcellus 13 7 10 19 5 2 3 3 12 7 a C = Control b T = Treated.

PAGE 157

157 Table 6 2. Comparison of th e effects of treatment on lepidoptera n flowering forb and flowering shub species richness and abundance in Florida flatwoods, 2007 Lepidoptera n F l owering Forb and F low ering Shrub A bundance and Species R ichness a Treatment Type ( SE) b P Control Dormant Burn Growing Burn Dormant Roller Chop Growing Roller Chop Roller Chop/ Burn Abundance (n o. of individuals) Lepidoptera 2.7 0.4 A 1.2 0.5 B 4.3 1.7 A 1.7 0.3 A 2.0 0.7A 3.0 0.9A 0.065 Flowering forb 61.3 14.7A 61.1 29.7A 15.2 6.2A 302.7 159.1AB 238.4 115.6B 322.4 163.9B 0.015 Flowering shrub 72.3 16.3 A 118.3 30.5 A 71.8 16.2 A 30.0 13.2 A 55.8 51.3 B 10.0 5.0 B 0.002 Richness (n o. of species) Flowering forb 3.9 0.4 A 3.2 0.4 AB 2.7 0.4 A 6.0 1.2 AB 7.0 0.8 B 5.8 1.1 B 0.035 Flowering shrub 3.0 0.2A 4.4 0.5A 2.9 0.4A 3.1 0.5A 1.7 0.6B 1.6 0.3B 0.002 a Only Lepidopteran flowering forb, and flowering shrub species richness or abundance significantly affected by treatment presented ( P 0.1). b Means in a row followed by the same uppercase letter not significantly different ( P > 0.1).

PAGE 158

158 CHAPTER 7 AVIAN COMMUNITY RESP ONSE TO GRAZING INTE NSITY ON MONOCULTURE AND MIXED FLORIDA PA STURES Introduction There are approximately 1.2 million ha of monoculture and mixed pasture in Florida, 78% of which occurs on private lands and is used primarily for cattle grazing (Florida Fish and Wildlife Conservation Commission [FWC] 2005) Monoculture pastures, also known as nonnative or improved pastures, are dominated by nonnative forage species and usually comprise d of former native pastures that have been cleared, tilled, and reseeded with improved forage types, so that few native vegetative species remain. Mixed pastures, also known as semi native or semi improved pastures, are comprised of a mixture of nonnativ e improved forage species interspersed with substantial quantities of native grasses and forbs. Mixed pasture conversion is less intense than for monoculture pastures and management inputs (e.g. fertilizer, weed control) are lower (FWC 2005) Monocultur e and mixed pastures are not a native part of Floridas landscape but, if grazed appropriately, have the potential to provide significant habitat for and be utilized by a diversity of wildlife species including resident and migratory birds (Alsop 2002, Eng strom et al. 2005, FWC 2005). Many of these species, some of which are federally and/or state listed as endangered or threatened, have decreasing populations and are of conservation concern (FWC 2005). Grazing livestock most typically affect avian commu nities indirectly through changes in vegetation composition, structure, and biomass, and can cause decreases in spatial heterogeneity of the vegetative community (Brennan and Kuvlesky 2005, Coppedge et al. 2008, Derner et al. 2009). Reductions in spatial heterogeneity caused

PAGE 159

159 by grazing imply the loss of habitat diversity (Adler et al. 2001) and changes in food abundance, foraging conditions, and breeding, thermal, and escape cover (Bock and Webb 1984, Vallentine 1990, Milchunas and Lauenroth 1993, Saab et al. 1995). Such changes have the potential to cause declines in avian species richness of grazed pastures, although avian abundance may be little affected, as species adapted to the grazed conditions often become highly abundant (Kantrud and Kologiski 1982). Therefore, avian community structure has the potential to be strongly influenced by the degree of structural heterogeneity in associated plant communities (Wiens 1974), with certain birds being attracted to habitats with specific vegetative attributes (Cody 1985). The majority of research on the effects of grazing on plant communities and bird abundance and species richness has been conducted in the central and western United States (Bock and Webb 1984, Vallentine 1990, Adler et al. 2001, Brennan and K uvlesky 2005, Coppedge et al. 2008, Derner et al. 2009) However, the transfer of results among environmentally divergent lands should be done cautiously. If management activities of benefit to birds associated with monoculture and mixed pastures are to be promoted in Florida, the impact grazing of these lands has on avian communities needs further investigation. The objectives of this study were: 1) to compare avian species richness and abundance on monoculture and mixed pastures subject to 4 grazing in tensities (nongrazed, low, medium, and high) and 2) to explore the role structural habitat attributes play in determining avian species richness and abundance on monoculture and mixed grazed pastures.

PAGE 160

160 M ethods Study Sites Research was conducted at the MacArthur Agro Ecology Research Station (MAERC), a 4 170 ha cattle ranch operated by Archbold Biological Station, located in Highlands County, south Florida (27o 09N, 81o 12W) One monoculture pasture and 1 mixed pasture study area were selected at MAER C. During 1996 1998, the 162ha monoculture pasture study area was subdivided using fences into eight, approximately 20ha experimental past ure units used for summer grazing (May monoculture pasture units were comprised almost entirely of bahiagrass ( Paspalum notatum Flgg), but included scattered wetlands, the majority of which were ditched and consisted of grasses, sedges, and miscellaneous wetland species (Werner et al. 1998). Wetland dominants included carpetgrass ( Axonopus furcatus Flgg Hitchc.), maidencane ( Panicum hemitomon Schult.), soft rush ( Juncus effuses L.), yellow eyed grass ( Xyris sp.), pickerelweed ( Ponteder ia cordata L.), and sawgrass ( Cladium jamaicense Crantz). A number of the monoculture pastures also contained small cabbage palm ( Sabal palmetto [Walt.] Lodd. ex J.A. and J.H. Schultes ) hammocks (Werner et al. 1998) During 1996 1998, the 260 ha mixed pasture study area was subdivided using fences into eight, approximately 32ha experimental pasture units used for winter grazing (November of bahiagrass and a variety of native species such as broomsedge ( An dropogon virginicus L. ) and bushy bluestem ( A glomeratus [Walt] B.S.P). The mixed pasture units were interspersed with seasonal wetlands, nearly all within 30 m of existing ditches, and comprised of grasses, sedges, and miscellaneous wetland species. Do minants in these wetlands included carpetgrass, maidencane, red top panicum ( P.

PAGE 161

161 rigidulum Bosc ex Nees), hat pins ( Eriocaulum sp.), yelloweyed grass, and some pickerelweed and soft rush. Cabbage palm hammocks occurred in the western third of this mixed p asture unit array. Mixed pastures were not as intensively drained as monoculture pastures and were frequently flooded or saturated during the June to October rainy season (Werner et al. 1998). Swain et al. (2007) provide additional details on pastures Forage production was higher on summer grazed monoculture pastures than on winter grazed mixed pastures. Therefore, to provide similar amounts of forage in both seasons and accommodate consistent grazing intensities, i t was necessary for mixed pasture uni ts to be larger than monoculture pasture units (Capece et al. 2007, McSorley and Tanner 2007). In 1998, a grazing study was initiated on the monoculture and mixed pasture units. Pasture units were subject to 4 cattle grazing intensities: 0 = nongrazed ( control), 15 = low, 20 = medium, or 35 = high animal units (AU) per pasture unit (no cattle, 1.3, 1.0, and 0.6 ha AU1 on monoculture pastures and no cattle, 2.1, 1.6, and 0.9 ha AU1 on mixed pastures). Stocking densities were selected based on input from the Florida Cattlemans Association and the University of Florida Institute of Food and Agricultural Sciences to reflect typical regional stocking densities, which average 1.42 ha AU1 (Gornak and Zhang 1999). Each grazing intensity was replicated twice for each pasture type and randomly assigned to the pasture units. The monoculture pasture units were summer grazed from May October and mixed pasture units were winter grazed from November April each year. Cattle were introduced and grazing intens ities initiated on mixed pasture units in October 1998 and on monoculture pasture units in May 1999. Monoculture pasture units were not grazed from October 1998 through April 1999. Prior

PAGE 162

162 to the initiation of study grazing treatments, cattle grazed the 2 pasture types during these same seasons at an average stocking density of approximately 1 ha AU1 and 1.6 ha AU1 on monoculture and mixed pastures, respectively. Prescribed burning was conducted in all mixed pasture units during November December 1998 and in all monoculture pastures units during February 1999 with similar affects observed across the study areas. All mixed pasture units were prescribed burned again in February 2002 and monoculture pastures in April 2002. All monoculture pasture units were mowed at a height of 35 cm for general weed control between September and November in 1998, 2000, and 2002. Dog fennel ( Eupat o r ium capillifolium Lam.) in the monoculture pasture units was treated with a combination of dimethylamine salts of dicamba and 2,4D (WEEDMASTER) at 4.6 L ha -1 plus 7.5 mL L1 of nonionic surfactant from May pasture management are provided in Swain et al. (2007). Vegetation Sampling Vegetation sampling was conducted quarter ly during spring (March 15 April 15), summer (June 1 July 1), fall (October 1 November 1), and winter (January 1 February 1) for 4 years (1999 2003). Vegetation sampling methods were similar to those described by Wiens (1969) and utilized a transect system. One 800m line transect was established in each pasture and divided into 4 sampling units of equal length. Within each sampling unit, I randomly located 4 vegetation sampling subpoints on either side of the transect. Sampling was repeated in each of the transect sampling units for a total of 32 sub points per transec t. At each subpoint, I visually estimated percent canopy coverage of grasses, forbs, litter, and bare ground to the nearest 5% within a 2.4m2 circular plot (Wiens 1969,

PAGE 163

163 Higgins et al. 20 05). I measured vertical stem density in the center of each plot by recording the number of vegetation contacts with a pole at 10 cm height intervals ( Wiens 1969). This data was used to calculate vertical stem density 0 30, 30 60, 60 90, and 90 120 cm above ground (stems/30 cm). Litter depth to the nearest 10 cm was measured on the same pole. Bird Surveys The avian communities within pastures were sampled using strip transect methods. Two parallel, 50 x 800m strips were positioned centrally in each of the 16 pasture units by marking the start and end points, as well as each 200m interval, with 3 m tall PVC pipe. Strip transects were separated by a 50m buffer and adjacent pasture (Wiens and Rotenberry 1981, Eberhardt 1978 Gibbons et al. 1996, Bibby et al. 2000). Vegetation sampling and avian surveys were conducted concurrently. I sampled avian transects quarterly between 1999 and 2003, corresponding to presumed seasonal differences in avian habitat utilization: spring (March 15 April 15) and fall (October 1 November 1) migrations, and breeding (June 1 July 1) and wintering (January 1 February 1) seasons. The June 1 July 1 breeding season selected for this study is the same as that used by the North American Breeding Bird Survey (Peterjohn and Sauer 1993). Each transect was sampled twice per season beginning at sunrise and ending no later than 3.5 h after sunrise. Sampling time per transect was 25 30 min. Sampling was not conducted during excessive wind, rain, fog, or periods of other unusual weather conditions (Gibbons et al. 1996, Bibby et al. 2000). Transects within the same pasture unit were not sampled on the same day to reduce the chance of counting the same bird twice (Wiens and Rotenberry 1981). I alternated sampling order each sampling period to reduce bias of counting the same transect at

PAGE 164

164 the same time of the morning. O nly birds recorded within the 50m strip boundaries were counted (Bibby et al. 2000). I counted flushed birds at the point they were first observed. I used careful observation, including recording of the location of flushed birds, to reduce the likelihood of double counting (Gregory et al. 2004). Average height of the herbaceous vegetation was approximately 75 cm and 150 cm in the summer and winter pasture arrays respectively, allowing observers consistent sighting ability within the pasture types. I di vided counts of avian abundance and species richness into guilds prior to analyses. Eight guilds were utilized, each falling into 1 of 2 major categories based on breeding habitat or migrant status (Peterjohn and Sauer 1993). Grassland, wetland and open water, successional scrub, woodland, and urban species guilds comprised the breeding habitat category and short distance migrant, neotropical migrant, and permanent resident species guilds the migrant status category. Not all species fell within a breedi ng habitat and migrant status guild. Within each category, guilds were independent and did not overlap in species composition (Table 7 1). Analyses I performed repeated measures analyses using mixed model regressions, with season and time since introducti on of grazing as repeated measures, followed by Fishers Protected LSD tests, to examine differences in vegetation attributes (mean, variance, and maximum percent coverage of grasses, forbs, bare ground, and litter, litter depth, and vertical stem density) total avian abundance and species richness, and avian abundance and species richness by guild among grazing intensities for monoculture and mixed pastures. I focused on grazing intensity effects rather than repeated measures effects. Twoand threeway grazing intensity interactions were

PAGE 165

165 noted in the results, if they occurred. Because threeway interactions are difficult to reliably interpret, they were not discussed further (Zar 1999 SYSTAT 2007). On both monoculture and mixed pastures, multiple linear regression was used to examine which combination of vegetation attributes best described changes in avian abundance and species richness, both overall and by guild. To reduce m ulticollinearity problems, all predictor variables involved in pairwise corr elations with r subjected to a univariate, oneway analysis of variance (ANOVA) with each dependent variable. For each pair of highly correlated predictor variables, the variable retained was the one with the greatest F value (Noon 1981, McGarigal et al. 2000). All regression models were fit using a forward stepwise procedure with Tolerance = 0.001, F to enter = 0.15, and F to remove = 0.15. These values are considered appropriate for predictor variables that are relatively independent (SYSTA T 2007). Regression models were considered statistically and biologically significant at P R2 Only models considered significant are presented. The relative importance of each variable in the best model was assessed by examining standardized regression coefficients (SC; i.e., variables with higher coefficients made greater indi vidual contributions to the explanatory power of t he model). All data sets were rank transformed prior to analyses due to violations of normality and homogeneity o f variance assumptions (Zar 1999, Conover 1998 SYSTAT 2007). Statistical significance was c oncluded at P rather than the more conservative P Type II error (1998). All statistical tests were performed using SYST AT (2007) statistical software.

PAGE 166

166 R esults Vegetation Monocult ure pasture. Variance of grass cover and litter depth were affected by grazing, decreasing as grazing intensity increased. Grazing also affected mean, variance, and maximum forb cover, mean and variance of litter cover, maximum vegetation height, mean li tter depth, and mean, variance, and maximum stem density at 60 90 cm and 90 120 cm. All generally decreased in the presence of grazing compared to control pasture units, but effects at low, medium, and high grazing intensities were similar. Mean grass co ver was also affected by grazing, increasing as grazing intensity increased (Table 7 2). A grazing intensity season interaction affected mean vegetation height, maximum litter depth, maximum litter cover, mean bare ground cover, variance in stem density at 0 30 cm, and mean, variance, and maximum stem density at 30 60 cm (Table 7 3). During the spring, maximum stem density at 30 60 cm and maximum litter depth decreased in the presence of grazing compared to control pasture units. Mean vegetation height and mean and variance of stem density at 30 60 cm typically decreased in the presence of grazing in winter and spring. Maximum litter cover decreased in the presence of grazing in all seasons. For all of these vegetation attributes, effects at low, mediu m, and high grazing intensities were similar. In spring, variance of stem density at 0 30 cm decreased in the presence of grazing compared to control pasture units, with reductions similar at low and high and greatest at medium grazing intensities. Mean bare ground cover decreased at high grazing intensity compared to control pasture units in fall, winter, and summer. Depending on season, reductions in this variable were also observed at medium grazing intensity (Table 7 3).

PAGE 167

167 Mean, variance, and maximum stem density at 0 30 cm were affected by a grazing intensity time interaction (Table 7 4). Decreases in mean stem density at 0 30 cm were typically observed at medium and high grazing intensities compared to control pasture units 1 4 years after introduction of grazing. Similar decreases were seen in variance and maximum stem density at 0 30 cm, but generally only immediately following and up to 1 year after introduction of grazing (Table 7 4). Variance in vegetation height and bare ground cover were a ffected by a grazing intensity season time interaction ( P = 0.019 and P = 0.098, respectively). Grazing alone and grazing intensity time, grazing intensity season, and grazing intensity season time interactions had no impact on maximum grass or bare ground cover ( P Mixed p asture. Grazing affected mean and variance of forb cover, which increased at low grazing intensity compared to controls. In contrast, mean grass cover decreased at low grazing intensity. No clear trend in mean and variance of forb cover or mean grass cover was observed with further increases in grazing intensity. Mean litter depth and vegetation height were affected by grazing. Mean litter depth decreased at high grazing intensity and mean vegetation height decreased at medium and high grazing intensities when compared to control pasture units. Grazing affected maximum forb cover, but no clear trend was observed with increasing grazing intensity (Table 7 2). Mean stem density at 0 30 cm was affected by a grazing intensity season interaction, decreasing at high grazing intensity in fall and summer (Table 7 3). This vegetation attribute was also effected by a grazing intensity time interaction, with decreases observed 2, 3, and 5 years after introduction of gr azing at high, and

PAGE 168

168 sometimes medium and low, grazing intensities compared to control pasture units (Table 7 4). Grazing intensity alone and grazing intensity time, grazing intensity season, and grazing intensity season time interactions had no im pact on variance and maximum grass cover, variance and maximum vegetation height, variance and maximum litter depth, mean, variance, and maximum litter cover, mean, variance, and maximum bare ground cover, variance and maximum stem density at 0 30 cm, or m ean, variance, and maximum stem density at 30 60 cm, 60 90 cm, and 90 120 cm ( P 0. 127). Avian Abundance and Species Richness Monoculture pasture. Sixty nine bird species were observed on monoculture pasture units (Table 7 1). Grazing affected wetland guild abundance ( P = 0.075), which decreased at low and medium grazing intensiti es and increased at high grazing intensity compared to control pasture units. Short distance migrant ( P = 0.028) and permanent resident ( P abundance at low and high grazing intensities However, at medium grazing intensity, abundance within these guilds was similar to that observed on control pasture units. Grazing affected neotropical migrant guild abundance ( P low, medium, and high grazing intensities. Declines were similar at low and medium grazing intensities and greatest at high grazing intensity (Figure 7 1). Total species richness was affected by grazing ( P in the presence of grazing compared to control pasture units R eductio ns were simila r at low and high grazing intensities and g reatest at medium grazing intensity Grazing also affected short distance migrant ( P = 0.001) and permanent resident ( P guild species richness, which decre ased in the presence of grazing. D eclines were similar at

PAGE 169

169 low and medium grazing intensities and greatest at high grazing intensity Species richness within successional scrub ( P and neotropical migrant ( P guilds was affected by grazing. Both decreased in the presence of grazing compared to control pasture units, but reductions at low medium, and high grazing intensities were similar (Figure 7 2). A grazing intensity season interaction affected grassland ( P = 0.002) and woodland ( P = 0.055) guild abundance and woodl and species richness ( P = 0.086). In the fall, g rassland g uild abundance increased in the presence of grazing compared to control pasture units. However, increases at low, medium, and high grazing intensities were similar. Woodland guild abundance and s pecies richness decreased in the fall at low grazing intensity, but no differences were observed between control pasture units and those subject to medium and high grazing intensities (Figure 7 3). Abundance within successional scrub ( P = 0.053) and woodla nd ( P = 0.074) guilds was affected by a grazing intensity time interaction. Successional scrub guild abundance decreased in the presence of grazing compared to control pasture units 34 years after the introduction of grazing. However, reductions at low medium, and high grazing intensitie s were similar. Woodland guild abundance decreased at low grazing intensity 2 years after the introduction of grazing However, abundance at medium and high g razing intensities was similar to that observed on control pasture units ( Figure 7 4). Grazing intensity alone and grazing intensity time, grazing intensity season, and grazing intensity season time interactions had no effect on total avian abundance, urban guild abundance, and wetland, grassland, and urban guild species richness ( P

PAGE 170

170 Mixed p asture. Seventy eight bird species were observed on mixed pasture units (Table 7 1). Grazing affected total avian abundance ( P = 0.017), which decreased at medium and high grazing intensities. At low grazi ng intensity, total avian abundance was similar to that observed on control pasture units. Within the grassland guild, abundance was also affected by grazing ( P = 0.045), increasing at low and medium grazing intensities ( P = 0.045). However, at high graz ing intensity abundance was similar to that observed on control pastures units. Grazing affected urban ( P = 0.046) and neotropical migrant ( P = 0.002) guild abundance. For both guilds, abundance decreased at high grazing intensity, but no differences were observed between control pastures and those subject to low and medium grazing intensities. Successional scrub abundance was affected by grazing ( P = 0.013). However, no clear trend in abundance was observed with increasing grazing intensity. Short di stance migrant guild abundance was affected by grazing ( P = 0.071), but was similar at low, medium, and high grazing intensities to that observed on control pasture units (Figure 7 5) Grazing affected successional scrub guild richness ( P = 0.009), which decreased at high grazing intensity. However, no differences in richness were observed between control pastures and those subject to low and medium grazing intensities. Neotropical migrant guild abundance was affected by grazing ( P = 0.0 70), decreasing at medium and high grazing intensities. However, at low grazing intensity abundance was similar to that observed on control pasture units. Grazing affected urban species richness ( P = 0.027), which increased at high and low grazing intensities. No differ ences in urban species richness were observed between control pasture units and those subject to medium grazing intensity. Species richness within the short distance migrant guild was

PAGE 171

171 affected by grazing intensity ( P = 0.015). However, no clear trend was observed as grazing increased (Figure 76 ). A grazing intensity season interaction affected woodland ( P = 0.048) and permanent resident ( P = 0.055) guild abundance. In fall, within the woodland guild, abundance increased at medium grazing intensity, but was similar to control pasture units at low and high grazing intensities. In all seasons, abundance within the permanent resident guild was similar on control pasture units and those subject to low, medium, and high grazing intensities. Total avian species richness ( P = 0.075) and species richness within the woodland guild ( P = 0.028) were also affected by a grazing intensity season interaction. However, in all seasons, total species richness was similar on control pasture units and those subject to low, medium, and high grazing intensities In fall, s pecies richness within the woodland guild increased at medium grazing intensity, but was similar to control pastures at low and high grazing intensities. Woodland guild richness also increased in spring at high grazing intensity, but was similar on control pasture units and those subject to low and medium grazing intensities (Figure 7 7 ). Total species richness was affected by a grazing intensity time interaction ( P = 0.092). However, richness was si milar at low medium, and high grazing intensities compared to controls at all times following introduction of grazing (Figure 7 8) Grassland guild species richness was affected by a g razing intensity season time interaction ( P = 0.087). Grazing int ensity alone and grazing intensity time, grazing intensity season, and grazing intensity season time interactions had no effect on

PAGE 172

172 wetland guild abundance and wetland and permanent resident guild abundance and species richness ( P Avian Habitat Relationships Monoculture pasture. Vegetation attributes that best predicted wetland guild abundance were mean vegetation height (SC = 0.425), minimum bare ground cover (SC = 0.262), variance of litter depth (SC = 0.182), and var iance of stem density at 0 30 cm (SC = 0.161; P R2 0.387 ). Mean vegetation height (SC = 0.358), variance of stem density at 0 30 cm (SC = 0.231), variance of litter depth (SC = 0.153), and mean forb cover (SC = 0.143; P R2 = 0.368) were the vegetation attributes that best explained neotro pical migrant guild abundance. Vegetation attributes that best predicted species richness within the wetland guild were mean stem density 0 30 cm (SC = 0.631), maximum stem density at 60 90 cm (SC = 0. 330), mean bare ground cover (SC = 0.282), mean litter cover (SC = 0.201), maximum stem density at 0 30 cm (SC = 0.197), and mean stem density at 90 120 cm (SC = 0.166, P R2 = 0.358). The combined effects of variance of stem density at 0 30 cm (SC = 0.641), maximum stem density at 0 30 cm (SC = 0.558), mean vegetation height (SC = 0.558), mean litter depth (SC = 0.476), mean stem density at 0 30 cm (SC = 0.329) mean litter cover (SC = 0.327), and variance of grass cover (SC = 0.166; P R2 = 0 .359 ) best predicted grassland guild species richness. Variance of litter depth (SC = 0.354), variance of vegetation height (SC = 0.333), maximum litter cover (S C = 0.218), and mean forb cover (SC = 0.175; P R2 = 0.308) were the vegetation attributes that best explained succesional scrub guild species richness. Vegetation attributes that best predicted short distance migrant guild species richness were mean litter depth (SC = 0.422), variance of stem density at

PAGE 173

173 0 30 cm (SC = 0.404), mean vegetation height (SC = 0.366), mean litter cover (SC = 0.334), maximum forb cover (SC = 0.278), and maximum vegetation height ( SC = 0.268; P R2 = 0.326) Mean vegetation height (SC = 0.458), maximum litter depth (SC = 0.185), and mean forb cover (SC = 0.138; P R2 = 0.324) were the vegetation attributes that best explained neotropical m igrant guild species richness. Mixed p asture. Vegetation a t tributes that best explained successional scrub guild abundance were m aximum stem density at 90 120 cm (SC = 0.413), mean grass cover (SC = 0.403), variance of litter depth (SC = 0.373), mean litter cover (SC = 0.365), variance of grass cover (SC = 0.223), and maximum grass cover (SC = 0.118; P R2 = 0.452). Vegetation attributes that best predicted neo tropical migrant guild abundance were maximum litter depth (SC = 0.385), variance of vegetation density at 60 90 cm, mean litter cover (SC = 0.275), variance of forb cover (SC = 0.102), and maximum grass cover (SC = 0.097, P R2 = 0.475). Neo tropical migrant guild species richness was best explained by maximum litter depth (SC = 0.320), maximum stem density at 90 120 cm (SC = 0.301), variance of grass cover (SC = 0.196), and m aximum litter cover (SC = 0.149; P R2 = 0.306 ). Discussion Vegetation Mon o culture pasture. The only vegetation attribute that increased as grazing intensity increased on monoculture pasture was mean grass cover. The dominant grass in monocultur e pastures was the improved species bahiagrass. Persistence of monoculture pastures and improved grasses, when subject to grazing, is a crucial factor in their sustainability Bahiagrass is capable of forming a highly persistent sward which tolerates sev ere defoliation (Beaty et al 1977, Stanley et al 1977, Hirata 1993, 2000;

PAGE 174

174 Hirata and Pakiding 2001) and, when grown in regions with warm summers and cool winters, often shows large seasonal variations in herbage mass under grazing (Pakiding and Hirata 2002) Eighteen of the other vegetation attributes examined exhibited some degree of decline in the presence of grazing. A decrease in the variance of a considerable number of vegetation attributes in the presence of grazing suggests a loss of spatial heter ogeneity. Grazing can result in decreased spatial heterogeneity through reductions in plant biomass and cover and changes in structural conditions (e.g., plant density and height, and litter cover and depth; Vallentine 1990, Milchunas and Lauenroth 1993, Fuhlendorf and Engle 2001, Derner et al. 2009). Current grazing practices often neglect to recognize the importance of maintaining spatial heterogeneity in plant structure and composition to biodiversity conservation. Livestock have typically been managed for uniform use of vegetation or management to the middle with extremes in vegetations structure (e.g., low sparse and high dense) absent (Derner et al. 2009). However, if used appropriately, grazing offers a potentially important tool for conservation management because of its influence on habitat structure and composition (Collins et al 1998, Adler et al. 2001). L ivestock have the potential to be used as ecosystem engineers, altering the heterogeneity of vegetation (Derner et al. 2009). Herbivores naturally exhibit preference for the consumption of certain plants over others (Van Soest 1996). If stocking rates are appropriate and pastures of a sufficient size, this results in differential patterns of use of individual plant species across a pastur e (Launchbaugh and Howery 2005). Typically, declines in heterogeneity are only observed at very low or very high intensities of grazing as

PAGE 175

175 respectively, livestock remove almost none or all of the vegetation. At medium grazing intensity, heterogeneity is maintained or increased as livestock selectively alter and remove a greater proportion of the vegetation in certain areas compared to others (Ausden 2007, Derner et al. 2009). I did not observe such a trend on monoculture pasture during this study and fur ther investigation is needed to understand grazing intensity, pasture sizes, and other livestock management activities that might permit maintenance of spatial heterogeneity in vegetation structure and composition on monoculture pastures in Florida. Possi ble methods proposed for enhancing spatial heterogeneity at the pasture scale include the strategic placement of supplemental feed, implementation of patch burns, and manipulation of water sources to alter vegetation structure in certain locations across t he pasture area (Derner et al. 2009). Mixed p asture. Far fewer vegetation attributes were affected by grazing on mixed than monoculture pastures, and then often only at low or high grazing intensities. Only 3 vegetation attribute s (mean grass cover, mean litter depth, and mean vegetation height) exhibited decreases based on grazing intensity alone on this pasture type. Mean and variance of forb cover increased at low grazing intensities. Other studies have shown moderate livestock grazing can result in increased forb cover, abundance, and species richness. These changes, as in this study, are often concomitant with decreases in vegetation height and litter depth (Talbot et al. 1939, Fensham et al. 1991, McNaughton 1993, Hayes and Holl 2003). No vegetati on attributes exhibited a decline in variance on mixed pasture because of grazing, suggesting that spatial heterogeneity of plant structure and composition may have been better maintained than on monoculture pastures. Grazing of native pasture

PAGE 176

176 systems ten ds to reduce their heterogeneity by favoring the most productive and palatable forage species for domestic cattle (Fuhlendorf and Engle 2001). On mixed pasture, such changes may have been observed at higher grazing intensities and over a longer time period. However, they may not have occurred during the relatively short duration of this study as a result of interannual and seasonal fluctuations in vegetation composition and quality and spatial and temporal patterns in diet selection observed in more compl ex vegetation (Ash and Smith 1996). Certainly, many of the native bunch grasses present on mixed pasture, such as broomsedge bluestem, chalky bluestem ( Andropogon cappilipes Nash ) and little bluestem ( Schizachyrium scoparium [Michx.] Nash var. stoloniferum [Nash] Wipff), can grow to considerable heights compared to bahiagrass. During the winter, when this pasture type was grazed, these grasses become largely dormant, leaving dry, rank vegetation above ground. This vegetation is largely unpalatable to and not grazed by livestock. These taller grasses in combination with shrubs and lower growing and newly sprouting grasses and forbs may help maintain structural variability in this habitat (Wi llcox personal observation). Avian Abundance and Species Richness Monoculture p asture. Although g razing had no impact on total avian abundance on monoculture pastures, total species richness decreased as grazing intensity increased. Heavy grazing can reduce overall species richness in grassland ecosystems (Kantrud 1981, Kantrud and Kologiski 1982) as spatial heterogeneity in the plant community is reduced (Derner et al. 2009). Reductions in spatial heterogeneity caused by grazing imply the loss of habit at diversity (Adler et al. 2001), and can influence the suitability and availability of food and cover resources for a variety of avian species (Saab et al. 1995, Brennan and Kuvalesky 2005, Coppedge et al. 2008, Derner et. al.

PAGE 177

177 2009). It has been proposed that declines in grassland birds may, in part, be associated with grazing driven reductions in vegetation heterogeneity that minimize the heavily disturbed and undisturbed plant communities different species require (Brennan and Kuvlesky 2005) Despite declines in species richness, total avian abundance is often little affected. Although some species are negatively affected by grazing, others respond positively (Saab et al. 1995). Species adapted to the grazed conditions become highly abundant resulting in little change in the total number of birds present (Kantrud and Kologiski 1982). On monoculture pastures, grazing proved detrimental to many of the avian guilds examined, negatively affecting species richness and abundance. Most notably affected were short distance migrant, neotropical migrant, and permanent resident guilds, all of which exhibited a decrease in species richness with increasing grazing intensity. In addition, these guilds also exhibited decreases in abundance as grazing intensity incr eased. Management to the middle places emphasis on the homogenous use of vegetation by grazers. The results of this study indicated grazing of monoculture pasture led to a trend of decreasing heterogeneity for a variety of habitat attributes. Loss of heterogeneity typically results in a lack of suitable habitat for birds that occupy the extremes of the vegetation structure gradient, e.g., low sparse and highdense vegetation, many of which are in these guilds (Kantrud and Kologiski 1982, Bollinger and Gavin 1992, Wilkins and Swank 1992, Saab et al. 1995, Guzy and Ritchison 1999, Derner et al. 2009). This results in loss of species richness and, if remaining guild members do not increase in number, decreases in abundance. The use of livestock as ecosystem engineers at the pasture scale has the potential to alter and maintain

PAGE 178

178 vegetation structure, particularly at the extremes of the structure gradient. Grazing management of this type would permit the creation and maintenance of a variety of habitat types and bird species (Derner et al. 2009). Wetland guild abundance increased with increasing grazing intensity. At high grazing intensity, abundance within this guild was higher than in control pastures. Studies have shown that waterfowl are tolerant of li ght to medium grazing, although optimal habitat conditions probably occur in the absence of grazing (Kirsc h 1969, Kruse and Bowen 1996). On monoculture pasture, the successional scrub guild did not exhibit declines in abundance until 3 4 years and the woodland guild until 2 years after the introduction of grazing. This suggests that the birds within these guilds were only sensitive to the vegetation effects and decreased heterogeneity of monoculture pasture habitats after prolonged, high intensity grazing. High adult breeding site fidelity is typical for many migratory birds. These species will often return yearly to the same areas to nest despite declining habitat conditions. After failing to reproduce successfully for a number of years, they may make t he decision to move to new breeding habitats (Hass 1998, Hoover 2003) or die, resulting in declines in avian abundance ( Greenwood 1980 Greenwood and Harvey 1982, Beheler et al. 2003, Ortega et al. 2006). However, it should be noted that pasture management strives to reduce woody plant dominance, so declines in shrubassociated species are to be expected. Attributes that were most often identified as positively related to avian abundance a nd richness within guilds on monoculture pasture were mean vegetation height and cover of forbs and variance of litter depth and stem density 0 30 cm above ground. All

PAGE 179

179 of these attributes exhibited declines of some degree in the presence of grazing, with some declines being seasonal and others affected by time since introduction of grazing. All of these attributes may be important to a variety of bird species due to their role in providing food and cover resources. Cover of forbs, litter, and bare ground are likely to affect seed and invertebrate food availability Vegetation height and stem density may influence the type and availability of cov er present (Saab et al. 1995). Mixed p asture. Total avian abundance decreased at medium and high grazing intensities but total species richness was not affected until 3 years after introduction of grazing at which point declines were observed at high grazing intensity. Grazing of mixed pastures had a detrimental effect on species richness and abundance within som e avian guilds. However, in this pasture type, the number of guilds negatively affected by grazing was fewer than in monoculture pastures. In addition, in mixed pastures, I only observed negative impacts on species richness and abundance at high grazing intensity, compared to monoculture pastures where detrimental effects were frequently observed at low and medium grazing intensities. Species richness and abundance within successional scrub and neotropical migrant guilds only decreased at high grazing i ntensity. Few vegetation attributes on mixed pasture were affected by grazing and typically only at high grazing intensity. Short distance migrant and urban guild species richness increased over that of control pastures at medium and high grazing intensi ties respectively. Grassland guild abundance increased at low and medium grazing intensities. Members of the grassland guild are of particular concern due to recent population declines (Brennan and Kuvalesky 2005). T heir declines may be associated with grazingdriven reductions in vegetation heterogeneity and the

PAGE 180

180 suitability and availability of food and cover resources (Saab et al. 1995, Brennan and Kuvalesky 2005, Coppedge et al. 2008, Derner et. al. 2009) However, spatial heterogeneity in plant struc ture and composition was largely maintained on mixed pastures throughout the study. This likely resulted in a diversity of food and cover resources which helped maintain and increase avian abundance within the grassland guild and species richness within the short distance migrant and urban guilds (Saab et al. 1995, Coppedge et al. 2008, Derner et al. 2009). This study suggests that, on mixed pastures, management and conservation of species within the grassland guild may be compatible with low to medium g razing intensities, and that livestock have the potential to serve as ecosystem engineers for members of this and other guilds (Derner et al. 2009). Attributes that were most often identified as positively related to avian abundance and richness on mixed pasture were mean and maximum grass cover and maximum litter depth and vegetation density 90 120 cm above ground. Mean grass cover was the only one of these attributes to decrease in the presence of grazing and may be important to birds as a food and cover resource. Within many guilds, abundance declined as maximum litter depth increased. Therefore, methods that reduce litter present on the ground may benefit many species. Management Implications On monoculture and mixed pasture, increasing grazing intens ity resulted in changes in a variety of vegetati on attributes. There was a trend toward increasing homogeneity of plant structure and composition as grazing intensity increased, particularly on monoculture pasture, and, depending on guild, this resulted i n increases or decreases in abundance and richness within particular avian guilds If the

PAGE 181

181 management and conservation of certain avian guilds is a priority, grazing intensity should be tailored to fit their needs. On monoculture and mixed pasture, the mi nimization of grazing intensity would be advantageous and likely result in increased abundance and species richness of many guilds. Based on the results of this study, a grazing intensity of 1.3 and 2.1 ha AU1 on monoculture and mixed pasture, respecti vely is recommended. However, some decline in species richness may still be expected. Ultimately, if habitat diversity is to be maximized and a range of avian species supported on monoculture and mixed pastures, the goal should be to maintain spatial het erogeneity in plant structure and composition, potentially using liv estock as ecosystem engineers.

PAGE 182

182 Table 7 1. Avian guild composition and seasonal abundance on monoculture and mixed pastures at MacArthur AgroEcology Research Station, Highlands County, F lorida, 1999 2003. Guild a, b Common Name Scientific Name Abundance (n o. of individuals) Fall Winter Spring Summer Mono c Mixed d Mono Mixed Mono Mixed Mono Mixed WT American bittern Botaurus lentiginosus 6 9 1 1 0 0 0 0 WD, SD American crow Corvus b rachyrhynchos 50 32 63 53 23 32 22 13 SS, SD American goldfinch Carduelis tristis 0 0 31 15 57 28 0 0 SD American kestrel Falco sparverius 16 11 14 20 21 12 0 0 UB, SD American robin Turdus migratorius 0 0 96 122 28 26 0 0 WT Anhinga Anhinga anhinga 0 1 2 0 0 0 0 0 WT, SD Bald eagle Haliaeetus leucocphalus 0 0 0 1 0 0 0 0 GR, SD Barn owl Tyto alba 0 1 0 0 0 0 0 0 NM Barn swallow Hirundo rustica 83 52 0 0 4 0 0 0 WD, RE Barred owl Strix varia 0 2 0 2 0 0 0 0 WT Belted kingfisher Ceryle alcyon 1 0 5 0 0 0 0 0 RE Black vulture Coragyps atratus 4 7 26 19 8 22 0 0 WT Black crowned nt. heron Nycticorax nycticorax 5 1 0 0 0 0 0 0 WT Black winged teal Anas discors 1 0 0 0 0 0 0 0 UB, SD Blue jay Cyanocitta cristata 0 3 0 0 0 1 0 0 WD, NM Blue gray gnat catcher Polioptila caerulea 12 20 6 8 0 0 0 0 WT, RE Boat tailed grackle Quiscalus major 130 191 28 30 36 12 30 36 GR, NM Bobolink Dolichonyx garrulus 64 40 0 0 0 0 0 0 SD Brown headed cowbird Molothrus ater 0 0 2 0 0 0 0 1 GR, NM Burrowing owl Athene cunicularia 0 0 0 0 1 0 1 0 SS, RE Carolina wren Thryothorus ludovicianus 4 3 0 0 3 0 0 0 WT Cattle egret Bubulcus ibis 1063 249 23 6 25 3 237 25 UB, SD Common grackle Quiscalus quiscula 5 287 0 4 14 7 34 19 SS, RE Common ground dove Columbina passerin e 1 2 1 0 0 2 7 11 WT Common moorhen Gallinula chloropus 0 1 0 0 0 0 1 0 WT Common snipe Gallinago gallinago 146 27 147 88 26 20 0 26 SS, NM Common yellowthroat Geothlypis trichas 226 499 46 54 27 71 18 72

PAGE 183

183 Table 7 1 Continued Guild a,b Common Name S cientific Name Abundance (n o. of individuals) Fall Winter Spring Summer Mono c Mixed d Mono Mixed Mono Mixed Mono Mixed RE Crested caracara Caracara cheriway 7 5 5 5 5 15 1 4 WT Double crest. cormorant Phalacrocorax auritus 0 1 1 0 2 0 0 2 NM East ern kingbird Tyrannus tyrannus 0 0 0 0 0 0 0 7 GR, SD Eastern meadowlark Sturnella magna 7141 568 394 265 723 465 648 606 SD Eastern phoebe Sayornis phoebe 17 15 27 11 1 2 0 0 SS Eastern towhee Pipilo erythrophthalmus 0 1 0 0 0 0 0 0 UB, SD European st arling Sturnus vulgaris 0 4 8 0 0 8 0 0 WT, Glossy ibis Plegadis falcinellus 3 1 37 8 0 0 0 0 GR, NM Grasshopper sparrow Ammodramus savannarum 0 1 0 5 2 19 0 0 SS, NM Gray catbird Dumetella carolinensis 1 4 0 0 0 0 0 0 WT Great blue heron Ardea herodias 3 2 3 7 3 1 0 3 WT Great egret Ardea alba 43 22 28 7 11 0 33 11 WT, NM Greater yellowlegs Tringa melanoleuca 3 0 3 3 19 18 0 19 WT Green heron Butorides virescens 4 0 0 0 1 0 3 1 SS, NM House wren Troglodytes aedon 3 28 5 37 5 18 0 0 SS, NM Indigo b unting Passerina cyanea 0 1 0 0 0 0 0 0 SD Killdeer Charadrius vociferous 6 0 16 18 12 9 0 4 WT King rail Rallus elegans 10 5 0 0 0 0 0 0 GR, SD Le Contes sparrow Ammodramus leconteii 0 0 0 11 1 0 0 0 WT, NM Least sandpiper Calidris minutilla 3 0 0 0 0 3 0 0 WT, NM Lesser yellowlegs Tringa flavipes 2 0 0 0 0 0 0 0 WT Little blue heron Egretta garzetta 38 10 5 2 2 1 9 2 SD Loggerhead shrike Lanius ludovicianus 9 19 3 7 0 4 3 6 WT, SD Marsh wren Cistothorus palustris 9 25 1 1 0 0 0 0 WD, NM Merlin F alco columbarius 2 0 0 0 0 0 0 0 WT Mottled duck Anas fulvigula 58 63 28 18 26 70 19 26 UB, SD Mourning dove Zenaida macroura 7 12 2 14 18 25 27 32 SS, RE Northern bobwhite Colinus virginianus 158 17 60 6 142 29 145 35 SS, RE Northern cardinal Cardinal is cardinalis 5 7 2 5 1 5 3 5

PAGE 184

184 Table 7 1 Continued Guild a,b Common Name Scientific Name Abundance (n o. of individuals) Fall Winter Spring Summer Mono c Mixed d Mono Mixed Mono Mixed Mono Mixed GR, SD Northern harrier Circus cyaneus 0 3 7 20 4 11 0 0 UB, RE Northern mockingbird Mimus polyglottos 4 18 10 3 2 16 6 8 WT, SD Osprey Pandion haliaetus 0 0 0 0 1 0 0 1 SS Palm warbler Dendroica palmarum 264 310 341 152 94 72 0 0 WD, RE Pileated woodpecker Dryocopus pileatus 0 1 0 1 0 1 0 0 WD, RE Red bellied woodpecker Melanerpes carolinus 8 10 2 18 0 15 12 7 WD, SD Red shouldered hawk Buteo lineatus 24 12 9 9 11 16 19 10 WT, SD Red winged blackbird Agelaius phoeniceus 2156 1672 671 213 490 523 711 490 NM Rough winged swallow Stelgidopteryx serripen nis 0 0 0 0 0 0 2 0 WT Sandhill crane Grus canandensis 1 6 44 19 15 13 2 15 GR, SD Savannah sparrow Passerculus sandwichensis 126 142 359 654 423 393 1 0 GR, SD Sedge wren Cistothorus plantensis 32 232 152 301 93 192 0 0 WT, Snowy egret Egretta thula 3 1 10 1 1 0 0 0 NM Solitary sandpiper Tringa solitaria 0 0 2 0 0 0 0 0 SS, SD Song sparrow Melospiza melodia 0 0 0 0 2 0 0 0 WT Sora Porzana carolina 9 2 3 0 0 0 0 0 SD Swamp sparrow Melospiza georgiana 1 12 70 727 14 124 0 0 WD Swallow tailed kite El anoides forficatus 0 0 0 0 0 1 0 0 SD Tree swallow Tachycineta bicolor 68 144 631 379 349 198 0 0 WT Tricolored heron Egretta tricolor 14 1 4 0 0 0 0 0 SD Turkey vulture Cathartes aura 2 40 12 29 2 9 0 5 WT Virginia rail Rallus limicola 3 2 0 0 0 0 0 0 WT White ibis Endocimus albus 958 119 61 2 0 0 110 0 RE White tailed kite Elanus leucurus 0 0 0 4 0 0 0 0 WT Wood stork Mycteria americana 71 23 19 11 3 2 0 3 WD Yellow rumped warbler Dendroica coronate 0 0 2 15 0 2 0 0 WD Yellow throated warbler Den droica dominica 0 0 0 1 0 0 0 0 a Breeding habitat guilds: WT = wetland, GR = grassland, SS = successional scrub, WD = woodland, and UB = urban. b Migrant Status guilds: RE = resident, SM = short distance migrant, and NM = neotropical migrant. c Monocul ture pasture. d Mixed pasture.

PAGE 185

185 Table 7 2. Effects of grazing intensity on vegetation attributes of monoculture and mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Pasture Type a Vegetation Attributes b Cattle Grazing Intensity ( SE) c P Control d Low e Medium f High g Mono Mean grass cover (%) 78.9 1.7A 90 0.8B 86.9 1.1B 90.0 0.9B Variance of grass cover (%) 734.7 62.9A 400.0 42.1B 533.8 1.0AC 482.0 57.2BC 0.003 Mean forb cover (%) 38.0 20.7 A 6.5 0.6 B 9.0 1.0 C 6.6 0.9 B Variance of forb cover (%) 552.3 63.3 A 226.2 28.0 B 3 12.3 36.2 B 285.7 46.1 B 0.001 Maximum forb cover (%) 85.9 3.4A 65.9 4.4B 74.0 3.8B 70.3 4.0B 0.003 Maximu m vegetation height (cm) 53.4 6.4 A 30.4 3.0 B 34.6 5.6 B 30.8 4.8 B Mean litter depth (cm) 3.3 0.6 A 2.3 0.5 BC 2.1 0. 4 B 2.3 0.6 C Variance of litter depth (cm) 13.2 4.3 A 4.4 1.5 B 8.2 5.8 C 3.2 1.2 C Mean litter cover (%) 34.5 5.2 A 31.2 5.4 B 29.6 5.3 B 24.8 5.4 B 0.001 Variance of litter cover (%) 343.0 60.9 A 255.2 48.6 BC 300.0 69.4 B 177.0 48.1 C 0.001 Mean stem density 60 90 cm (stems/30 cm) 0.2 0.1 A 0.03 0.01 B 0.04 0.2 B 0.03 0.01 AB 0.006 Variance of stem density 60 90 cm (stems/30 cm) 0.8 0.3 A 0.1 0.03 B 0.2 0.1 B 0.1 0.1 AB 0.007 Maximum stem density 60 90 cm (stems/30 cm) 3.4 0.7 A 0.9 0.3 B 1.0 0.4 B 2.3 1.2 AB 0.009 Mean stem density 90 120 cm (stems/30 cm) 0.1 0.0 A 0 .0 0.0 B 0.0 0.0 B 0.0 0.0 B 0.001 Variance of stem density 90 120 cm (stems/30 cm) 0.3 0.1A 0.0 0.0B 0.1 0.0B 0.1 0.1B 0.001 Maximum stem density 90 120 cm (stems/ 30 cm) 1.8 0.5 A 0.4 0.2 B 0.4 0.2 B 0.7 0.4 B 0.001 Mixed Mean grass c over (%) 90.8 10.4 A 85.7 1.9 B 88.2 2.4 AB 90.7 2.2 AB 0.033 Mean forb cover (%) 4.16 0.5A 7.8 1.0B 5.5 0.7ABC 3.8 0.6AC 0.011 Variance of forb cover (%) 111.5 16.2 A 248.8 38.1 B 157.9 24.3 ABC 99.2 0.023 AC 0.023 Maximum forb cover (%) 50.3 4.7 AB 62.3 4.9 A 57.9 4.9 A 41.1 5.2 B 0.043 Mean vegetation height (cm) 23.5 2.9 A 21.6 2.8 AB 20.7 2.5 B 19.3 2 .1 B 0.029 Mean litter depth (cm) 4.0 0.1 A 3.2 0.7 AB 2.7 0.5 AB 2.3 0.4 B 0.020 a Pasture type: Mono = m onocultur e pasture, Mixed = mixed pasture. b Only vegetation a ttributes significantly affected by grazing presented ( P 0.1) c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1). d Nongrazed. e 1.3 ha AU-1 on monoculture pasture and 2.1 ha AU-1 on mixed pasture. f 1.0 ha AU-1 on monoculture pasture and 1.6 ha AU-1 on mixed p asture. g 0.6 ha AU-1 on monoculture pasture and 0.9 ha AU-1 on mixed pasture.

PAGE 186

186 Table 7 3. Effects of a grazing intensity season interaction on vegetation attributes of monoculture and mixed pastures at MacArthur Agro Ecology Research Station, Highlands County, Florida, 1999 2003. Pasture Type a Vegetation Attributes b Season Cattle Grazing Intensity ( SE) c P Control d Low e Medium f High g Mono Mean vegetation height (cm) Fall 26.3 4.4A 19.6 3.2A 19.1 3.9A 17.6 3.2A 0.052 Winter 14.0 0.8 A 11.6 0.3 A 10.4 0.3 B 10.2 0.2 B Spring 15.0 1.4 A 10.8 0.6 B 10.9 0.6 B 10.5 0.4 B Summer 24.4 5.4 A 16.4 3.3 A 14.7 3.3 A 11.5 2.2 A Maximum litter depth (%) Fall 12.9 3.4A 59.5 10.7A 18.6 9.8A 8.4 3.4A 0.084 Winter 11.2 4.0 A 11.3 5.0 A 5.7 1.4 A 5.7 1.3 A Spring 13.3 3.7 A 4.8 1.3 B 6.0 1.4 B 4.3 1.3 A Summer 15.0 8.1 A 5.4 1.9 A 4.5 1.8 A 6.3 3.8 A Maximum litter cover (%) Fall 80.0 6.7 A 59.5 10.7 B 56.5 11.5 A 46. 0 10.5 B 0.067 Winter 84.2 11.0A 75.0 13.1B 71.7 13.5A 80.0 11.3B Spring 73.6 11.0 A 56.9 12.3 B 61.9 12.3 A 53.1 9.5 B Summer 72.0 9.8 A 70.5 8.0 B 69.0 9.1 A 54.0 8.7 B Mean bare ground cover (%) Fall 5.6 1.4 A 3.12 1.0 A 6.4 1.3 A 5.3 1.2 B 0.064 Winter 2.7 0.8 A 2.9 1.0 A 3.1 0.8 AB 3.4 1.0 B Spring 2.6 0.9A 2.6 0.7A 1.5 0.3A 2.0 0.8A Summer 3.4 2.1 A 4.4 1.7 A 3.8 1.0 A 2.5 0.8 B Variance of stem density 0 30 cm (stems/30 cm) Fall 102.7 34.3 A 46.6 6.9 A 51.7 7.0 B 62.5 22.6 AB 0.005 Winter 68.1 22.3 A 43.6 9.0 A 29.0 6.2 A 25.6 7.3 A Spring 71.6 23.6 A 19.5 3.1 B 18.6 3.8 C 20.0 2.7 BC Summer 40.1 11.5A 23.1 8.4A 25.6 8.4A 19.5 10.1 A Mean stem density 30 60 cm (stems/30 cm) Fall 2.2 0.6 A 1.2 0.5 A 1.0 0.4 A 1.0 0.4 A 0.015 Winter 0.2 0.1 A 0.0 0.0 AB 0.0 0.0 B 0.1 0.1 AB Spring 0.5 0.2 A 0.0 0.0 B 0.0 0.0 B 0.1 0.0 B Summer 1.4 0.1 A 1.0 0.7 A 0.7 0.4 A 0.4 0.3 A Variance of stem density 30 60 cm (stems/30 cm) Fall 15.2 6.0A 4.3 1.8A 4.4 2.1A 4.1 1.8A 0.031 Winter 0.7 0.2 A 0.1 0.1 AB 0.0 0.0 B 0.7 0.6 AB Spring 2.5 1.3 A 0.0 0.0 B 0.2 0.1 B 0.2 0.1 B Summer 4.5 2.3 A 3.9 3.1 A 3.0 2.0 A 1.5 1.1 A

PAGE 187

187 Table 7 3. Continued Pasture Type a Vegetation Attributes b Season Cattle Grazing Intensity ( SE)c P Control d Low e Medium f High g Mono Maximum stem density 30 60 cm (stems/30 cm) Fall 15.6 4.2 A 8.2 2.0 A 8.6 2.4 A 8.1 1.6 A 0.052 Winter 4.5 1.0A 2.0 1.0A 0.3 0.2A 3.7 2.2A Spring 7.6 2.6 A 0.4 0.3 AB 1.6 1.0 B 2.4 0.9 AB Summer 26.6 19.4 A 4.0 2.2 A 4.4 2.4 A 2.8 1.5 B Mixed Mean vegetation density 0 30 cm ( stems/30 cm) Fall 18.9 3.4 A 15.7 2.1 A 18.7 2.8 A 18.9 2.7 B 0.012 Winter 15.0 2 .5 A 13.5 2.4 A 10.1 2.1 A 9.9 2.2 A Spring 13.5 2.1 A 8.0 1.5 A 9.9 2.0 A 8.1 1.5 A Summer 9.6 1.3 A 8.0 1.0 A 8.8 1.0 A 8.4 1.5 B a Pasture type: Mono = m onoculture pasture, Mixed = mixed pasture. b Only vegetation a ttributes significantly affected by a grazing season interaction presented ( P 0.1) c Means in a row followed by the same uppercase letter not significantly different ( P > 0.1). d Nongrazed. e 1.3 ha AU-1 on monoculture pasture and 2.1 ha AU-1 on mixed pasture. f 1.0 ha AU-1 on monoculture pasture and 1.6 ha AU-1 on mixed pasture. g 0.6 ha AU-1 on monoculture pasture and 0.9 ha AU-1 on mixed pasture.

PAGE 188

188 Table 7 4. Effects of a grazing in tensity time interaction on vegetation attributes of monoculture and mixed pastures at MacArthur Agro Ecology Research Station, Highlands County, Florida, 1999 2003. Pasture Type a Vegetation Attributes b Time c Cattle Grazing Intensity ( SE) d P Control e Low f Medium g High h Mono Mean stem density 0 30 cm (stems/30 cm) 0 22.3 5.4 A 10.8 3.4 A 10.3 4.0 A 9.9 4.6 A 1 13.5 1.6 A 8.9 1.0 ABC 7.2 0.7 B 40.3 7.2 C 2 7.4 1.8 A 5.9 1.6 A 5.0 1.4 A 4.4 1.3 B 3 11.0 1.9 A 8.9 1.4 AB 8.0 1.2 B 7.0 1.3 B 4 11.6 0.4 A 12.0 1.0 A 14.5 1.5 B 11.9 1.0 A Variance of stem density 0 30 cm (stems/30 cm) 0 179.5 74.2 A 47.1 22.7 B 43.1 23.9 B 84.0 61.4 B 0.005 1 97.7 20.7 A 49.1 8.5 AB 40.3 7. 2 A 40.3 7.2 B 2 28.4 7.7 A 20.6 6.3 A 15.5 5.1 A 13.9 4.6 A 3 54.6 15.0 A 27.1 3.2 A 28.5 4.6 A 20.7 4.0 A 4 40.8 9.6 A 25.3 5.1 A 41.4 8.8 A 33.5 5.9 A Maximum stem density 0 30 cm (stems/30 cm) 0 45.5 5.1 A 26.3 7.7 B 23.5 9.5 B 26.5 12.5 B 0.013 1 38.5 3.7 AB 27.4 2.5 B 40.3 7.2 A 26.0 3.0 B 2 19.3 4.4 A 16.9 4.2 A 28.1 14.0 A 12.8 3.3 A 3 35.4 8.4 A 24.0 1.6 A 24.3 2.5 A 20.8 2.4 A 4 32.8 6.0 A 23.5 2.4 A 33.5 3.1 B 27.7 2.5 A Mixed Mean s tem density 0 30 cm (stems/30 cm) 0 45.0 0.0 A 4.9 0.0 B 38.8 0.0 A 42.4 0.0 A 1 15.5 3.7 A 15.2 3.3 A 14.7 4.0 A 13.1 3.3 A 2 15.3 2.8 A 12.8 3.1 AB 14.6 2.3 B 14.4 1.9 B 3 13.6 0.9 A 9.3 2.4 B 9.4 1.0 B 8.2 1.0 B 4 11.9 2.4 A 9.3 2.4 A 9.4 2.2 A 8.1 2.5 A 5 11.2 2.8 A 10.3 1.2 A 10.1 1. 0 A 11.4 1.3 B a Pasture type: Mono = m onoculture pasture, Mixed = mixed pasture. b Only vegetation a ttributes significantly affected by a grazing time interaction presented ( P 0.1) c Time since introduction of grazing (years) d Means in a row foll owed by the same uppercase letter not significantly different ( P > 0.1). e Nongrazed. f 1.3 ha AU-1 on monoculture pasture and 2.1 ha AU-1 on mixed pasture. g 1.0 ha AU-1 on monoculture pasture and 1.6 ha AU-1 on mixed pasture. h 0.6 ha AU-1 on m onoculture pasture and 0.9 ha AU-1 on mixed pasture. .

PAGE 189

189 Figure 7 1. Effects of grazing intensity on avian abundance by guild i n monoculture pastures at MacArthur Agro Ecology Research Station, Highlands County, Florida 1999 2003. Bars present + SE. Bars topped by different letters are significantly different ( P s: control = nongrazed, low = 1.3 ha AU1, medium = 1.0 ha AU1, and high = 0.6 ha AU1.

PAGE 190

190 Figure 7 2. E ffects of grazing intensit y on total avian species richness and avian species richness by guild i n monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Bars present + SE. Bars topped by different letters are signific antly different ( P Grazing intensities: control = nongrazed, low = 1.3 ha AU1, medium = 1.0 ha AU1, and high = 0.6 ha AU1.

PAGE 191

191 Figure 7 3. Effects of a grazing intensity season interaction on avian abundance and species richness by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Bars present + SE. Bars topped by different letters are significantly different ( P Grazing intensities: control = nongrazed, low = 1.3 ha AU1, medium = 1.0 ha AU1, and high = 0.6 ha AU1.

PAGE 192

192 Figure 7 3 Continued

PAGE 193

193 Figure 7 4. Effec ts of a grazing intensity time interaction on avian abundance by guild in monoculture pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Bars present + SE. Bars topped by different letters are signi ficantly different ( P s: control = nongrazed, low = 1.3 ha AU1, medium = 1.0 ha AU1, and high = 0.6 ha AU1.

PAGE 194

194 Figure 7 5. Effects of grazing intensity on total avian abundance and avian abundance by guild i n m ixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Bars present + SE. Bars topped by different letters are significantly different ( P Grazing intensitie s: control = nongrazed, low = 2.1 ha AU1, medium = 1.6 ha AU1, and high = 0.9 ha AU1.

PAGE 195

195 Figure 7 6 Effects of grazing intensity on avian species richness by guild i n m ixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida 1999 2003. Ba rs present + SE. Bars topped by different letters are significantly different ( P s: control = nongrazed, low = 2.1 ha AU1, medium = 1.6 ha AU1, and high = 0.9 ha AU1.

PAGE 196

196 Figure 7 7. Effects of a grazing intensity season interaction o n total avian species richness and avian abundance and species richness by guild i n mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida 1999 2003. Bars present + SE Bars topped by different letters are significantly different ( P : control = nongrazed, low = 2.1 ha AU1, medium = 1.6 ha AU1, and high = 0.9 ha AU1.

PAGE 197

197 Figure 7 7 Continued

PAGE 198

198 Figure 7 8. Effects of a grazing intensity time interaction on avian species richness by guild on mixed pastures at MacArthur AgroEcology Research Station, Highlands County, Florida, 1999 2003. Bars present + SE. Bars topped by different letters are significantly different ( P s: control = nongrazed, low = 2.1 ha AU1, medium = 1.6 ha AU1, and high = 0.9 ha AU1.

PAGE 199

199 CHAPTER 8 CO NCLUSIONS The r esults of this study suggest prescribed burning and roller chopping, if appropriately applied, have an important role to play in maintaining and enhancing pine flatwoods habitat and associated wildlife communities. In areas of pine flatwoods, where active management has been lacking and shrub proliferation has become a concern, roller chopping provides a means of quickly reducing shrub cover, height, and density to achieve conditions that are more favorable for herbaceous plant growth. In a ddition, prescribed burning and roller chopping practices, depending on season of application, can shift the structure and composition of pine flatwoods vegetative communities towards conditions preferred by overwintering, permanent resident, migrant, and breeding avian guilds. The study also highlights the need to consider the effect of prescribed burning and roller chopping on arthropod communiti es, which are frequently negatively affected by these practices Many arthropods are important pollinators or pests and the practices examined can alter their abundance, potentially hindering or benefiting conservation and control efforts. Given the lack of knowledge, many questions surrounding vegetation, avian, and arthropod response to prescribed burning and r oller chopping still need to be addressed to advance wildlife management. There is a need to investigate the effect frequency of roller chopping h as on the structure and compo siton of the pine flatwoods plant community and, in turn, wildlife. A greater understanding of the interaction of prescribed burning and roller chopping would also be useful in determining the likely impacts of this combination treatment This study focused on the examination of changes in species richness and abundance of various t axa following prescribed burning and roller

PAGE 200

200 chopping and examined the ir habitat preferences to enable better tailoring of management activities to priority species or groups. As such, the study was primarily exploratory in nature, providing a strong found ation on which to build future research efforts. To allow practice effects to be more fully understood, future research should consider not only measures of species richness and abundance but demographic variables such as survival and fecundity Inclusio n of these variables would allow a more definitive assessme nt of habitat quality and help deepen the understanding of habitat preferences Grazing had considerable impacts on avian communities occupying monoculture and mixed pastures. Monoculture pastures in particular, exhibited decreasing spatial homogeneity of the vegetative community as grazing level increased. Loss of spatial heterogeneity typically results in a lack of suitable habitat for birds that occupy the extremes of the vegetation structure gradient. This can lead to a loss of species richness and abundance. F or the majority of avian guilds, a low grazing intensity of 1.3 and 2.1 ha AU1 on monoculture and mixed pasture, respectively is recommended to maintain abundance. However, these grazing intensities may result in declines in species richness. Ultimately, if a range of avian species are to be supported on monoculture and mixed pastures, spatial heterogeneity of plant structure and composition must be maintained. These findings sug gest that further investigation of the role of livestock as ecosystem engineers could benefit certain avian groups.

PAGE 201

201 LIST OF REFERENCES Abrahamson, W. G. 1984a. Post fire recovery of Florida Lake Wales Ridge vegetation. Am erican Journal of Bot any 71: 9 21. Abrahamson, W. G. 1984b. Species responses to fi re on the Lake Wal es Ridge. American Journal of Bot any 71: 35 43. Abrahamson W. G., and D. C. Hartnett. 1990. Pine flatwoods and dry prairies. Pages 103 149 in R. L. Myers and J. J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando, USA. Adler, P. B., D. A. Raff, and W. K. Lauenroth. 2001. The effect of grazing on the spatial heterogeneity of vegetation. Oecologia 128:465 479 Ahlgren, C. E. 1960. Effects of fire on reproduction and growth of vegetation in northeastern Minnesota. Ecology 41: 431 445. Alsop F. J. III. 2002. Birds of Florida. S mithsonian Handbooks. Dorling Kindersley, New York, N ew Y ork USA Anderson R. C., T. Leahy, and S. S. Dhillion. 1989. Numbers and biomass of selected insect groups on burned and unburned sand prairie. American Midland Naturalist 122:151 Anderson, S. H. 1980. Habitat selection, succession, and bird community organization. Pages 13 26 in R. M. D egraff, editor. Management of western forests and grasslands for nongame b irds. U S Forest Service General Technical Report INT 86 Intermountain Forest and Range Experiment Station, Ogden, Utah, USA. Ash, A. J. and D. M. S. Smith. 1996. Evaluating stocking rate impacts on rangelands: animals dont practice what we preach. Journal of Rangelands 18:216 243. Askins, R. A. 1993. Population trends in grassland, shrubland and forest birds in eastern North America. Current Ornithology 11:1 34. Ausden M. 1996. Invertebrates. Pages 139 177 in W. J. Sutherland, editor. Ecological census techniques: a handbook. Cambridge University Press, United Kingdom. Ausden, M. 2007. Habitat management for conservation: a handbook of techniques. Oxford University Press, Oxford, U nited K ingdom Beaty, E. R., J. L. Engel, and J. D. Powell. 1977. Yield, le af growth, and tillering in bahiagrass by N rate and season Agronomy Journal 69: 308 311.

PAGE 202

202 Beheler, A. S., O. E. Rhodes Jr., H. P. Weeks Jr., and F. R. Moore. 2003. Breeding site and mate fidelity in Eastern Phoe bes ( Sayornis phoebe) in Indiana. Auk 120:990 999. Berkland, M. W., and C. A. Rewa. 2005. Environmental quality incentives program contributions to fish and wildlife conservation. Pages 171192 in J. B. Haufler, editor. Fish and wildlife benefits of Farm B ill conservation programs: 20002005 u pdate. The Wildlife Society, Bethesda, Maryland, USA. Bibby, C. J., N. D. Burges s, D. A. Hill, and S. H. Mustoe. 2000. Bird Census Techniques. Academic Press, London U nited K ingdom Bock, C. E., and B. Webb. 1984. Birds as grazing indicator species in southeastern Arizona. Journal of Wildlife Management 48:1045 1049. Bollinger, E. K., and T. A. Gavin. 1992. Eastern bobolink populations: ecology and conservation in an agricultural landscape. Pages 497506 i n: J. M. Haga n III, and D. W. Johnston, editors Ecology and conservation of neotropical migrant landbi rds. Smithsonian Institute Pres s, Washington, D C USA Bowman R. 200 2. Common GroundDove ( Columbina passerina) Account 645 i n A. Poole, and F. Gill, editor s. Birds of North America The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists Union, Washington, D.C., USA. Bozzo, J. A. S. L. Beasom, and T. E. Fulbright. 1992. Vegetation responses to two brush management practices in South Texas. Journal of Range Management 45:170 175. Brennan L. A. 1999. Northern Bobwhite ( Colinus virginianus ) Account 397 i n A. Poole, and F. Gill, editor s. Birds of North America The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists Union, Washington, D.C., USA. Brennan, L. A., and W. P. Kuvlesky. 2005. North American grassland birds: an unfolding conservation crisis? Journal of Wildlife Management 69: 1 13. Bulan C A and G. W. Barrett. 1971. The effects of two acute stresses on the arthropod component of an experimental grassland ecosystem. Ecology 52:597 605 Bullock, J. 1996. Plants. Pages 111 138 in W. J. Sutherland, W.J., editor. Ecological census techniques: a handbook. Cambridge University Press, U nited K ingdom Cancelado, R an d T. R. Yonke 1970. Effect of fire on prairie insect populations. Journal of the Kansas Entomol ogical Soc iety 43:274 281.

PAGE 203

203 Capece, J. C., K L. Campbell, P. J. Bohlen, D. A. Graetz, and K. M. Portier. 2007. Soil phosphorus, cattle stocking rates and water quality in subtropical pastures in Florida, USA. Rangela nd Ecology and Management 60:19 30. Chambers B Q and M. J. Samways. 1998. Grasshopper response to a 40year experimental burning and mowing regime, with recommendations for invertebrate conservation management. Biodivers ity Conserv ation 7:985 1012 Christensen, N. L. 1981. Fire regimes in southeastern ecosystems. Pages 112 136 i n H. A. Mooney T. M. Bonnicksen, N. L. Christensen Jr., J. E. Lotan and W. A. Reiners editors. Proceeding s of the Conference of Fire Regimes and Ecosystem Properties. U.S. Forest Service Ge neral Technical Report WO 26, Washington, D C ., USA. Christensen, N. L. 2004. Fire and soil plant nutrient relations in a pine wiregrass savanna on the coastal plain of North Carolina. Oecologia 31:27 44. Cody, M. L. 1985. Habitat selection in grassland and opencountry birds. Pages 191 226 i n M. L. Cody, editor Habitat selection in birds. Florida Academic Press, Orlando, USA. Collins, C. S., R. N. Conner, and D. Saenz. 2002. Influence of hardw ood midstory and pine species on pine bole arthropods. Forest Ecology and Maangement 164:211 220. Collins, S. L., and D. J. Gibson. 1990. Effects of fire on community structure in tallgrass and mixedgrass prairie. Pages 81 98 in S. L. Collins, and L L. Wallace editors. Fire in North American Tallgrass Prairie. University of Oklahoma Press, Norman, USA Collins, S. L., A. K. Knapp, J. M. Briggs, J. M. Blair, and E. M. Steinauer. 1998. Modulation of diversity by grazing and mowing in native tall grass prairie. Science 280:745 747. Conover, W J. 1998 Practical nonparametric s tatistics John Wiley and Sons, New York, N ew Y ork USA Coppedge, B. R., S. D. Fuhlendorf, W. C. Harrell, and D. M. Engle. 2008. Avian community response to vegetation and structural features in gr asslands managed with fire and grazing. Biological Conservation 141:1196 1203 Covell, C. V. 1984. A field guide to the moths of e astern North America. Houghton Mifflin, New York New York, USA.

PAGE 204

204 Cox, J., R. Kautz, M. MacLaughlin, and T. Gilbert. 19 94. Closing the gaps in Floridas wildlife habitat conservation s ystem: r ecommendations to meet minimum conservation goals for declining wildlife species and rare plant and animal communities. Florida Game and Fresh Water Fish Commission, Tallahassee, US A. Cox, J., and B. Widener. 2008. Lightning season burning, friend or foe of breeding birds? Tall Timbers Research Station Miscellaneous Publication 17, Tallahassee, Florida, USA. Davis, J. H. Jr. 1967. General map of the natural vegetation of Flori da. University of Florida Institute of Food and Agricultural Sciences Cir cular S 178, Agricultural Experiment Station, Gainesville, Florida, USA. Derner, J. D., W. K. Lauenroth, P. Stapp, and D. J. Augustine. 2009. Livestock as ecosystem engineers for grassland bird habitat in the western Great Plains of North America. Rangeland Ecology and Management 62:111 118. Donahue, R. L., J. C. Shickluma, and L. S. Robertson. 1971. Soils. Prentice Hall, Upper Saddle River New Jersey, U SA. Drewa, P. B., W. J. Platt, and E. B. Moser. 2002. Fire effects on resprouting of shrubs in headwaters of southeastern longleaf pine savannas. Ecology 83: 75 767. Dueser, R. D., and H. H. Shugart. 1978. Microhabitats in a forest floor sm all mammal fauna. Ecology 59:89 98. Dunning J. B. 2006. Bachmans Sparrow ( Aimophila aestivalis ) Account 38 i n A. Poole, editor. Birds of North America The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists Union, Washington, D.C., USA Dunwiddie, P. W. 1991. Comparisons of aboveground arthropods in burned, mowed and untreated sites in sandplain grasslands of Nantucket Island. American Midland Naturalist. 125:206 212 Eberhardt, L. L. 1978. Transect methods for population studies Journal of Wildlife Management 42:131. Engstrom, R. T., P. D. Vickery, D. W. Perkins, and W. G. Shriver. 2005. Effects of fire regime on birds in southeastern pine savannas and native prairies. Studies in Avian Biology 30:147 160. Erhardt, A. 1985. Diurnal Lepidoptera: sensitive indicators of cultivated and abandoned grasslands. Journal of Applied Ecol ogy 22:849 861.

PAGE 205

205 Evans E W 1984. Fire as a disturbance to grasshopper assemblages of tallgrass prairie. Oikos 43:9 16 Ewell J. J. Introdu ction. Pages 8 10 in R. L. Myers and J. J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando, USA Facelli, J. M., and T. A. Pickett. 1991. Plant litter: its dynamics and effects on plant community structure. Bot anic al Rev iew 57:1 32. Fensham, R. J., J. E. Holman, and M. J. Cox. 1991. Plant species responses along a grazing disturbance gradient in Australian grassland. Journal of Vegetation Science 10:77 86. Fitzgerald, S. M. 1990. Small mammal, bird and plant responses to t he rehabilitation of a dry prairie grassland association using fire and roller chopping applied in two seasons. Thesis, University of Florida, Gainesville, USA. Fitzgerald, S. M., and G. W. Tanner. 1992. Avian community response to fire and mechanical shrub control in south Florida. Journal of Range Management 45:396 400. Florida Biodiversity Task Force. 1993. Conserving Floridas biological diversity. The Florida Biodiversity Ta sk Force, Tallahassee, USA. Florida Fish and Wildlife Conservation Commission. 2005. Floridas comprehensive wildlife strategy. Flori da Fish and Wildlife Conservation Commission, Tallahassee, USA. Freeman, J. E., and S. Jose 2009. The role of herbicide in savanna restoration: effects of shrub reduction treatments on the understory and overstory of a longleaf pine flatwoods. Forest Ecology and Manag ement 257: 978 986. Fuhlendorf, S. D., and D. M. Engle. 2001. Restoring heterogeneity on rangelands: ecosystem restoration based on evolutionary grazing patterns. BioScience 51:625 632. Gerberg, E. J. and R. H. Arnett. 1989. Florida Butterflies. Natural Sci ence Publications Baltimore, Maryland, USA. Gibbons, D. W., D. A. Hill, and W. J. Sutherland. 1996. Birds. Pages 227259 in W. J. Sutherland, editor Ecological census techniques: a handbook. Cambridge University Press, U nited K ingdom Giuliano, W. M., A. K. Accamando, and E. J. McAdams. 2004. Lepidopterahabitat relationships in urban parks. Urban Ecosystems 7:361 370.

PAGE 206

206 Glass berg, J. 1999. Butterflies through binoculars: a fiel d guide to the birds of eastern North America. Oxford University Press, New York, USA. Glitzenstein, J., D. R. Streng, and W. J. Platt. 1995. Evaluating the effects of season of burn on vegetation in longleaf pine savannas. Florida Game and Freshwater Fish Commission, Tallahassee, USA. Gordon, D., D. Shaw, and E. Contreras. 2005. Problems adversely affecting Floridas wildlife habitats and species of greatest conservation need. The Nature Conservancy, Gainesville, Florida, USA Gornak, S., and J. Zhang. 1999. A summary of landowners surveys and water quality data from improved pasture sites in the northern Lake Okeechobee watershed. Applied Engineering in Agriculture 15:121 127 Gray, R. L., S. L. Benjamin, and C. A. Rewa. 2005. Fish and wildlife benefits of the wildlife habitat incentives program. Pages 155 170 in J. B. Haufler, editor. Fish and wildlife benefits of farm bill c o nservation programs: 2000 2005 u pdate. T he Wildlife Society, Bethesda, Maryland, USA. Greenwood, P. J 1980 Mating systems, philopatry and dispersal in birds and mammals. Animal Behaviour 28: 1140 1162 Greenwood, P. J. and P. H. Harvey 1982 The natal and breeding dispersal of birds. Annual Review of Ecology and Systematics 13: 1 21 Gregory, R. D., D. W. Gibbons, and P. F. Donald. 2004. Bird survey and census techniques. Pages 17 55 i n W. J. Sutherland, I Newton, and R. E. Green, editors Bird ecology and conservation: a handbook of techniques. Oxford University Press, U nited K ingdom Grelan, H. E. 1975. Vegetative response to twelve years of seasonal burning on a Louisiana longleaf site. U S Forest Service Research Note SO 192 Southern Forest Experimental Station, Ashvil le, North Carolina, USA. Griffin, B., and B. A. Youtie. 1988. Two devices for estimating f oliage density and deer hiding cover. Wildlife Society Bulletin 16:206 210. Guzy, M. J. and G. Ritchison. 1999. Common yellow throat ( Geothlypis trichas ). Ac count 488 i n A. Poole, and F. Gill, editors Birds of North America The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists Union, Washington, D.C., USA. Haas, C. A. 1998. Effects of prior nesting success on sit e fidelity and breeding dispersal: an experimental approach. Auk 115: 929 936.

PAGE 207

207 Hall D. F, and D. F. Schweitzer, D. F. 1992. A survey of the moths, butterflies, and grasshoppers of four Nature Conservancy preserves in southeastern North Carolina. The Nature Conservancy, Durham, North Carolina, USA. Hanula, J. L., and K. Franzreb. 1998. Source, distribution, and abundance of macroarthropods on the bark of longleaf pine: potential prey of the redcockaded woodpecker. Fores t Ecology and Management 102:89 102. Hanula J L, and D. D. Wade. 2003. Influence of long term dormant season burning and fire exclusion on grounddwelling arthropod populations in longleaf pine flatwoods ecosystems. Forest Ecology and Management 175:163 184. Hardwood R F and M. T. James. 1979. Entomology in Human and Animal Health. MacMillan, New York New York, USA. Hardy, C. 2003. Flora and fauna community response to season of prescribed fire treatments in longl eaf pine forest of the North Carolina sandhills. Dissertation, Mississippi State University, Starkville, USA. Hayes, G F., and K. D. Holl. 2003. Cattle grazing impacts on annual forbs and vegetation composition of mesic California grasslands Conserv ation Biology 17:1694 1702. Hays, R. L., C. Summers, and W. Seitz. 1981. Estimating wildlife habitat variables. U .S. Fish and Wildlife Service FWS/OBS 81/47, Washing ton, D.C., USA Hendricks, J. J., C. A. Wilson and L. R. Boring. 2002. Foliar litter de position and deco mposition in a fire maintained longleaf pinewiregrass ecosystem. Canadian Journal of Forest Res ources 32: 928 941. Hendrix, S. D., and J. F. Kyhl. 2000. Population size and reproduction in Phlox pilosa Conservation Biology 14: 304 313. Hermann, S., T V. Hook, R. W. Flowers, L. A. Brennan, J. S. Glizenstein, D. R. Streng, J. L. Walker, and R. L. Myers. 1998. Fire and biodiversity: studies of vegetation and arthropods. Transactions of the North American Wildlife and Natural Resources Conference 63:3 84 401 Higgins, K. F., K. J. Jenkins, G. K. Clambey, D. W. Uresk, D. E. Naugle, J. E. Norland, and W. T. Barker. 2005. Vegetation sampling and monitoring. Pages 524 553 i n C. E. Braun editor Techniques for wildlife investigations and management. T he Wildlife Society, Baltimore, Maryland, USA. Hilmon, J. B. 1968. Autecology of saw palmetto ( Serenoa repens ). Dissertation, Duke University, Durham, North Carolina, USA.

PAGE 208

208 Hilmon J. B., and R. H. Hughes. 1965. Fire and forage in the wiregrass type. Journal of Range Manag ement 18:251 254. Hilmon, J. B., and C. E. Lewis. 1962. Effect of burning on south Florida range. U.S. Forest Service Paper 146, Southeast Forest Experimental Station. Ashville, North Carolina, USA. Hirata, M. 1993. Response of bahia grass ( Paspalum notatum Fl gge) sward to cutting height. 3. Density of tillers, stolons and primary roots. Journal of the Japanese Society of Grassland Science 39: 196 205. Hirata, M. 2000. Effects of nitrogen fertilizer rates and cutting height on leaf appearance and extension in bahia grass ( Paspalum notatum ) swards. Tropical Grasslands 34: 7 13. Hirata, M. and W Pakiding. 2001. Tiller dynamics in a bahia grass ( Paspalum notatum ) pasture under cattle grazing. Tropical Grasslands 35:151 160. Hoover, J. P. 2003. Decision rules for site fidelity in a migratory bird the prothonotary warbler. Ecology 84:416 430. Hough, W. A. 1968. Carbohydrate reserves of saw palmetto: seasonal variation and effects of burning. Forest Science 4:399 405. Huber, G. E., and A. A. Steuter 1984. Vegetation profile and grassland bird response to spring burning. Prairie Naturalist 16:5 5 61. Huffman, J. M., and S. W. Blanchard. 1991. Changes in w oody vegetation in Florida dry prairie and wetlands during a period of fire exclusion, and a fter dry growing season fire. Pages 75 83 in S. C. Nodvin, and T. A. Waldrop, editors. Fire and th e environment: ecological and cultural perspectives. U.S. Forest Service General Technical Report SE 69, Southeast Forest Experiment al Station, A sheville, North Carolina, USA. Huffman, J. M., and P. A. Werner. 2000. Restoration of Florida pine savanna: flowering response of Lilium catesbaei to fire and roller chop ping. Natural Areas Journal 20:12 23. Hunter, W. C., D. A. Buehler, R. A. Canterbury, J. L. Confer, and P. B. Hamel. 2001. Conservation of disturbancedependent birds in eastern North America. W ildlife Society Bulletin 29:440 455. Hurst G. A. 1971. The effects of controlled burning on arthropod density and biomass in relation to bobwhite quail brood habitat on a right of way. Proceedings of the Tall Timbers Fire Ecology Conference 2:173 183.

PAGE 209

209 James, F. C., and H. H. Shugart. 1971. A quantitative method of habitat description. Audubon Field Notes 24:727 736. Johnston, D. W., and E. P. Odum. 1956. Breeding bird populations in relation to plant succession on the pied mont of Georgia. Ecology 37:50 62. Kalmbacher, R. S., K. R. Long, M. K. Johnson, and F. G. Martin. 1984. Botanical composition of diets of cattle grazing south Florida r angeland. Journal of Range Management 37:334 340. Kalmbacher, R. S. and F. G. Martin. 1984. Chopping and tebuthiuron effects on saw palmetto. Proceedings of the Soil Crop Sci ence Society of Florida 43: 86 69. Kantrud, H. A. 1981. Grazing intensit y affects the breeding avifauna of North Dakota native grasslands. C anadian Field Naturalist 95:404 417. Kantrud, H. A., and R. L. Kologiski. 1982. Effects of soils and grazing on breeding birds on uncultivated upland grasslands of the northern Great P lains. U.S. Fish and Wildlife Service Research Report 15, Washington, D C. USA. Kautz, R. S. 1993. Trends in Florida wildlife habitat 1936 1997. Florida Scientist 56:7 23. King, T. G. M. A. Howel l, B. R. Chapman, K. V. Miller, R. A. Schorr. 1998. Comparisons of wintering bird communities in mature pine stands managed by prescribed burning. Wilson Bulletin 110 :570 574. Kirsch L. M. 1969. Waterfowl production in relation to grazing. Journal of Wildlife Management 33:821 828. Komarek, E. V. 1968. Lightning and lightning fires as ecological forces. Proceedings of the Tall Timbers Fire Ecology Conference 8:169 Krebs, C. J. 1999. Ecological methodology. Benjamin Cummings, New York, New York, USA. Kremen, C. 1992. Assessing the indicat or properties of species assemblages for natural areas monitoring. Ecological Applications 2:203 217. Kruse, A. D., and B. S. Bowen. 1996. Effects of grazing and burning on densities and habitats of breeding ducks in North Dakota. Journal of Wildlife Management 60: 238 246. Langdon, O. G. 1981. Some effects of prescribed fire on understory vegetation in loblolly pine stands. Pages 143 153 in G. W Wood, editor. Prescribed fire and wildlife in southern f orests. T he Belle W. Baruch Forest Science Institute of Clemson University Symposium Proceedings Georgetown, South Carolina, USA.

PAGE 210

210 Launchbaugh, K. L. and L. D. Howery. 2005. Understanding landscape use patterns of livestock as a consequence of foraging behavior. Rangeland Ecology and Management 58:99 108. Launer, A. E. and D. D. Murphy. 1994. Umbrella species and the conservation of habitat fragments: a case of a threatened butterfly and a vanishing grassland ecosystem. Biological Conservation 69:145 153. Lett, M. S., and A. K. Knapp. 2003. Consequences of shrub expansion in mesic grassland: resource alterations and graminoid responses. Journ al of Vegetation Science 14:487 496. Lewis, C. E. 1970. Responses to chopping and rock phosph ate on sout h Florida ranges. Journal of Range Management 23: 276 282. Lewis, C. E., and T. J. Harshbarger. 1976. Shrub and herbaceous vegetation after 20 years of prescribed burning in the S outh Carolina Coastal Plain. Journal of Range Management 29: 13 18. Lloyd, J., R. W. Mannan S. Destefano, and C. Kirkpatrick. 1998. The effects of mesquite invasion on a southeastern grassland bird community. Wilson Bulletin 110: 403 408. Lotti, T. 1956. Eliminating understory hardwoods with summer prescribed fires in co astal plain loblolly pine stands. Journal of Forestry 54:191 192. Lussenhop J 1976. Soil arthropod response to prair ie burning. Ecology 57:88 98. Madden, E. M., A. J. Hansen, and R. K. Murphy. 1999. Influence of prescribed fire history on habitat and abundance of passerine birds in nor thern mixedgrass prairie. Canadian Field Nat uralist 113: 627 640. Mapstone, B. D. 1995. Scalable decision rules for environmental impact studies: effect size, type I, and type II errors. Eco logical Appl ications. 5: 401 410. McGarigal, K., S. Cushman, and S. Stafford. 2000. Multivariate statistics for wildlife and ecology research. Springer, New York, N ew Y ork USA. McNair, D. B. 1998. Response of Henslows sparrows and sedge wrens to a dormant season presc ribed fire. Florida Field Naturalist 26:46 47. McNaughton, S. J. 1993. Grasses and grazers, science and management. Ecological Applications 3:17 20. McSorley, R., and G. W. Tanner. 2007. Effects of cattle stocking rates on nematode communities in S outh Florida. Rangeland Ecology and Management 60:31 35

PAGE 211

211 Means, D. B. 1996. Longleaf pine forests, going, goingPages 210229 in M. E. Davis, editor. Eastern oldgrowth forests. Island Press, Washington, D.C. USA. Milchunas, D. G., and W. K. Lauenroth. 1993. Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63:327 366. Millsap, B. A., J. A. Gore, D. E. Runde, and S. I. Cerulean. 1990. Setting priorities for the conservation of fi sh and wildlife in Florida Wildlife Monographs 111:157. Moore, W. H. 1974. Some effects of chopping saw palmettopineland threeawn range in south Florida. Journal of Range Management 27:101 104. Morris M. G 1975. Preliminary observations on the effects of burning on the Hemiptera (Heteroptera and Auchenorryhcha) of limestone grassland. Bio logical Cons ervation 7:311 319 Murphy, M. T. 2003. Avian population trends within the evolving agricultural landscape of eastern and central United States. Auk 120: 20 34. Myers, R. L., and S. E. Boettcher. 1987. Flowering response of cutthroat grass ( Panicum abscissum ) following fire. Bulletin of the Ec ological Society of America 68: 375. Nagel H. G 1973. Effect of spring prairie burning on herbivorous and non herbivorous arthropod populations. Journal of the Kansas Entomol ogical Soc iety 46: 485 497 Nelson, S. M., and D. C. Anderson. 1994. An assessment of riparian environmental quality by using butterflies and disturbance susceptibility scores. Southwest Naturalist 39:137142. New, T. R. 1991. Butterfly Conservation. Oxford University Press, United Kingdom. New, T. R. 1993. Conservation biology of Lycaenidae (butterflies). IUCN, Gland, Switzerland. New, T. R. 1997. Are Lepidoptera an e ffective umbrella group for biodiversity Conservation? Jou rnal of Insect Conservation 1:5 12. Nodvin S. C, and T. A. Waldrop. Fire and the environment: ecological and cultural perspectives. U .S. Forest Service General Technical Report SE 69, Southeast Forest Experiment al Station, Asheville, North Carolina, USA Noon, B. R. 1981. The dist ribution of an avian guild along a temperate elevational gradient: the importance and expression of competition. Ecological Monographs 51:105 125.

PAGE 212

212 Norris, W. R., L. M. Hemesath, D. M. Debinski, and D. R. Farrar. 2003. Does bird community composition var y along a disturbance gradient in northeast Iowa. Natural Areas Journal 23:262 273. Noss, R. F. 1989. Longleaf pine and wiregrass: keystone components of an endangered ecosystem. Na tural Areas Journal 9:211 Oates, M. R. 1995. Butterfly conser vation within the management o f grassland habitats. Pages 98 112 in A. S. Pullin, editor. Ecology and conservation of b utterflies. Chapman and Hall, London, U nited K ingdom Olson, M. S., and W. J. Platt. 1995. Effects of habitat and growing season fir es on resprouting of shrubs in longleaf pi ne savannas. Vegetatio 119:101 118. Opler, P. A. 1998. A field guide to e astern butterflies. Houghton Mifflin, New York, New York, USA. Orr, D. M., and C. J. Paton. 1997. Using fire to manage species compos ition in Heteropogon contortus (black speargrass) pastures: enhancing the effects of fire with grazing management. Australian Journal of Agricultural Research 48:803 810. Ortega, Y. K., K. S. McKelvey, and D. L. Six. 2006. Invasion of an exotic forb im pacts reproductive success and site fidelity of a migratory songbird. Oecologia 149:340 351 Outcalt, K. W. 1994. Seed production in wiregrass in central Florida following growing season prescri bed burns. International Journal of Wildland Fire 4: 123 125. Pakiding, W., and M. Hirata. 2002. Canopy dynamics in a bahia grass ( Paspalum notatum Flgge) pasture under cattle grazing: structural components responsible for variations in herbage mass. Grassland Science 48:311 316. Panzer R 1988. Managin g prairie remnants for insect conservation. Nat ural Areas Journal 8:83 90. Pet erjohn, B. G., and J. R. Sauer. 1993. North American breeding bird survey annual summary 19901991. Bird Populations 1:52 67. Platt, W J., G. W. Evans, and M. M. Davis. 1 988. Effect of fire season on flowering of forbs and shrubs in longleaf pine forests. Oecologia 76: 353 363. Pyne, S. J., P. L. Andrews, and R. D. Laven. 1996. Introduction to Wildland Fire. Wiley, New York, New York, USA.

PAGE 213

213 Reed C C 1997. Response of prairie insects and other arthropods to prescription burns. Natural Areas Journal 17:380 385. Rice L. A. 1932. The effect of fire on the prairie animal communities. Ecology 13:392 401 Robbins, L. E., and R. L. Myers. 1992. Seasonal effects of prescribed burning in Florida: a r eview. Tall Timbers Research Station Miscellaneous Publication No. 8, Tallahassee, Florida, USA. Robertson, P. A., S. A. Clarke, and M. S. Warren. 1995. Woodland management and butterfly diversity. Pages 113 122 in A. S. Pullin, editor. Ecology and conservation of butterflies. Chapman and Hall, London, United Kingdom. Rodewald, A. D., and R. H. Yahner. 2001. Avian nesting successes in forested landscapes; influence of landsca pe composition, stand and nest patch m icrohabitat, and biotic interactions. Auk 118:1018 1028 Rudolph, D. C., and C. A. Ely 2000. The influence of fire on Lepidopteran abundance and community structure in forested habitats of eastern Texas. Texas Journal of Science 52:127 138. Saab, V. A., C. E. Bock, T. D. Rich, and D. S. Dobkin. 199 5. Livestock grazing effects in western North America. Pages 311 356 i n T. E Martin, and D. M. Finch, editors. Ecology and management of neotropical migratory birds. Oxford University Press U nited K ingdom Schauff G. C. 1986. Collecting and preser ving insects and mites: techniques and tools. U .S. Department of Agriculture, Systematic Entomology Laboratory, Washington D C. USA. Scott, J. A. 1986. The Butterflies of North America. Stanford UniversityPress, California, USA Seastedt T 1984. Belowground macroarthropods of annually burned and unburned tallgrass prairie. Am erican Midl and Nat uralist 111 :405 408 Shepherd, S., and D. M. Debinski. 2005. Evaluation of isolated and integrated prairie reconstructions as habitat for prairie butterflies. Biological Conservation. 126:51 61. Siemann, E., J. Haarstad, and D. Tilman. 1997. Short term and lo ngterm effects of burning on oak savanna arthropods. Amer ican Midland Naturalist 137:349 361. Sievers, E. 1985. Burning and grazing Flori da flatwoods. Rangelands 7:208 209

PAGE 214

214 Sisson, D. C., D. W. Speake, J. L. Landers, and J. L. Buckner. 1990. Effec ts of prescribed burning on wild turkey habitat preference and nest site selection in south Georgia. National Wild Turkey Symposium 6:4450. Sparks, J. C., E. C. Masters, D. M. Engle, M. W. Palmer, and G. A. Bukenhofer. 1998. Effects of late growing se ason and late dormant season prescribed fire on herbaceous vegetation in pinegrassland communities. Jour nal of Vegetation Science 9:133 142. Stanley, R. L., E. R. Beaty, and J. D. Powell. 1977. Forage yield and percent cell wall constituents of Pensacola bahiagrass as related to N fertilization and clipping height. Agronomy Journal 69:501 504. Stephenson, S. N., and M. F. Buell 196 5. The reproducibility of shr ub cover sampling. Ecology 46: 379 380. Stoddard, H. L. 1931. The bobwhite quail. Charles Scribner's Sons, New York, New York, USA. Streng, D. R., J. S. Glitzenstein, and W. J. Platt. 1993. Evaluating effects of season of burn in longleaf pine forests: a critical literature review and some results from an ongoing long term study. Proceedings of the Tall Timbers Fire Ecology Conference 18:227 264 Swain, H. M., P. J. Bohlen, K. L. Campbell, L. O. Lollis, and A. D. Stein man. 2007. Integrated ecological and economic analysis of ranch management systems: an example from south central Florida. Rangeland Ec ology and Management 60:1 11. Swengel, A. B. 1996. Effects of fire and hay management on abundance of prairie butterflies. Biological Conservation 76:73 85. Swengel, A. B. 1998. Effects of management on butterfly abundance in tallgrass prairie and pine barrens. Biological Conservation 83:77 89. Swengel, A. B. 2001. A literature review of insect responses to fir e compared to other conservation managements of open habitat. Biodiversity and Conservation 10:11411169. Swengel, S. R., and A. B. Swengel. 1999. Correlations in abundance of grassland songbirds and prairie butterflies Biological Conservation 90:1 1 1 SYSTAT. 2007. Statistics II Version 12. SYSTAT Soft ware, San Jose, California, USA.

PAGE 215

215 Talbot, M. W., H. H. Biswell, and A. C. Hormany. 1939. Fluctuations in annual vegetation of California. Ecology 20:394 402. Tanner, G., Mullahey J. J., Maehr, D. 1996. Saw palmetto: an ecologically and economically important native palm. Institute of Food and Agricultural Sciences Extension Publication WEC 109 University of Florida, Gainesville, USA. Tanner, G. W., J. M. Wood R. S. Kalmbacher, and F. G. Martin. 1988. Mechanical shrub control on flatw oods range in south Florida. Journal of Range Manag ement 41: 245 248. Tester J. R and W. H. Marshall. 1961. A study of certain plants and animal interrelations on a native prairie in northwestern Minnesota. Bell Museum of Natural History Paper No. 8, Minneapolis, Minnesota, USA Thatcher, B. S., D. G. Krementz, and M. S. Woodrey. 2006. Henslows sparrow winter survival estimates and response to prescribed burning. Journal of Wildlife Management 70:198 Towne, G., and C. Owensby. 1984. Long term effects of annual burning at different dates in ungrazed Kansas tallgrass prairie. Journal of Range Management 37:392 397. Triplehorn C A, and N. F. Johnson. 2005. Borror and DeLong's introduction to the study of insects Brooks Cole, Florence, Kentucky, USA. Tucker, J. W. Jr., W. D. Robinson, and J. B. Grand. 2004. Influence of fire on Bachmans Sparrow, and endemic North American songbird. Journal of Wildlife Manag ement 68:11141123. Ubick D., P. Paquin, P. E. Cushing, and V. Roth, editors. 2005. Spiders of North America: an identification manual. American Arachnological Society, Keene, New Hampshire, USA U.S. Census Bureau. 2000. 2000 Census of Population and Housing. U.S. Census Bureau, Washington, D.C., USA. U.S. Department of Agriculture. 1997. Five year national resources inventory. U.S. Department of Agriculture, Washington, D.C., USA. U.S. Fish and Wildlife Service. 1999. South Florida multi species recovery plan. U.S. Fish and Wildlife Service, Atlanta, Georgia, USA. Vallentine, J. F. 1990. Grazing management. Academic Press, San Diego, California USA.

PAGE 216

216 Van Amburg G L ., J. A. Swaby, and R. H. Pemble. 1981. Response of arthropods to a spring burn of a tallgrass prairie in northwestern Minnesota. Ohio Biol ogical Surv ey Biology Notes 15:240 243. Van Auken, O. W. 2000. Shrub invasions of North American semi arid grasslands. Annual Review of Ecology and Systematics 31: 197 215. Van Soest, P. J. 1996. Allometry and ecology of feeding behavior and digestive capacity in herbivores: a r eview. Zoo Biology 15: 255 479. Venier, L. A., and J. A. Pearce. Boreal bird community response to jack pine forest succession. Fores t Ecology and Management 217:19 36. Vickery, P. D. 1996. Grasshopper Sparrow ( Ammodramus savannarum ). Account 239 i n A. Poole and F. Gill, editors Birds of North America The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists Union, Washington, D.C., USA. Wade, D., J. Ewel, and R. Hofstetter. 1980. Fire in south Florida ec osystems. U S Forest Service General Technical Report SE 17, Southeastern Forest Experimental Station, Asheville, North Carolina, USA. Waldrop, T. A., D. H. Van Lear, F. T. Lloyd, and W. R. Harms. 1987. Long term studies of prescribed burning in lobl olly pine forests of the Southeastern Coastal Plain. U.S. Forest Service General T echnical Report SE 45 Southeastern Forest Experiment al Station, Asheville, North Carolina, USA Warren S D ., C. J. Scifres, and P. D. Teel. 1987 Response of grassland arthropods to burning: a review. Agr iculture, Ecosyst ems and the Environment 19:105 125 Watts, A. C., and G. W. Tanner. 2003. Fire and roller chopping have varying effects on dry prairie plant species (Florida). Ecological Restoration 22:229 230. W atts, A. C. G. W. Tanner, and R. Dye. 2006. Restoration of dry prairies using fire and roller chopping. Pages 225 230 in R. Noss, e d itor. Land of fire and water: t he dry prairie ecosystem. Proceedings of the Dry Prairi e Restoration Conference, Sebring, Florida, USA Werner, P. A., G. W. Tanner, J. J. Mullahey, and S. P. Chrisman. 1998. Biological monitoring for the Buck Island agroecology study: preexperimental treatment phase. South Florida Water Management District Report for Contract No. C 664 9, West Palm Beach US A. White, L. D., and W. S. Terry. 1979. Creeping bluestem response to prescribed burning and grazing in south Florida. Journal of Range Management 32:369 371

PAGE 217

217 Wiens, J. A. 1969. An approach to the study of ecological relations hips among grassland birds. Ornithological Monographs 8:1 93. Wiens, J. A. 1974. Habitat heterogeneity and avian community structure in North American grasslands. American Midland Naturalist 91:195 213. Wiens, J. A. and J. T. Rot enberry. 1981. Cen susing and the evaluation of avian habitat occupancy. Studies in Avian Biology 6:522 232. Wilkins, R. N., and W. G. Swank. 1992. Bobwhite habitat use under short duration and deferred rotation grazing. Journal of Range Management 45:549 553. Wright A F, and A. J. A. Stuart. 1992. A study of the efficiency of a new inexpensive type of suction apparatus in quantitative sampling of grassland arthropod populations. British Ecological Society Bulletin 23:116 120 Yarlett, L. L. 1965. Control of saw palmetto and recovery of native grasses. Journal of Range Management 18:344 345. Yarlett, L. L., and Roush R. D. 1970. Creeping bluestem ( Andropogon stolonifer (Nash] Hitchc.). Journal of Range Management 23:117 Zar, J. H. 1999 Biostatistical a nalysis. Prentice Hall, Upper Saddle River N ew Jersey USA

PAGE 218

218 BIOGRAPHICAL SKETCH Emma graduated with a B S in z oology from the University of Cardiff, Wales, United Kingdom. Her M S in conservation biology is from the University of Kent, Canterbury, United Kingdom. Emma is an applied ecologist with a strong interest in wildlife ecology, management, and conservation on private, primarily agricultural, lands. Her primary research focus is in the area of wildlife habitat management. In addition, she has a strong interest in undergraduate teaching and wildlife extension activites.