<%BANNER%>

A multi-disciplinary evaluation of the invasion and management of Melaleuca quinquenervia

Permanent Link: http://ufdc.ufl.edu/UFE0024299/00001

Material Information

Title: A multi-disciplinary evaluation of the invasion and management of Melaleuca quinquenervia
Physical Description: 1 online resource (166 p.)
Language: english
Creator: Martin, Melissa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: bio, biological, control, ecosystem, exotic, herbicide, herbivory, invasion, litterfall, melaleuca, nutrient, soil
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Ecosystem invasion by exotic plant species poses a significant threat to community biodiversity, function, and stability in the Florida Everglades. One of the most problematic invasive species in this ecosystem is Melaleuca quinquenervia (Cav.) Blake. The goal of my dissertation was to use a multi-scale, interdisciplinary approach to evaluate the ecosystem-level consequences of the invasion and management of M. quinquenervia. The results presented include above and belowground storages of nutrients in invaded and non-invaded systems, alterations of soil microbial biomass and function, evaluations of biological and chemical control methods, and plant community response to a native disturbance regime in the context of management. Overall, both the invasion and management of M. quinquenervia altered basic ecosystem functions of nutrient storage and cycling, plant community diversity, and community response to native disturbance. We found that in the absence of top-down control from herbivory, M. quinquenervia trees were able to create a positive feedback loop to growth whereby increased quality and quantity of above- and below-ground biomass drove higher levels of storage and availability of below-ground resources. In addition, analysis of large-scale research plots in native, invaded, and managed sub-tropical wetland sites revealed that the use of herbicides to control M. quinquenervia invasion may damage long-term ecosystem structure and function. We found that after a seasonal fire plots that had been chemically treated to control M. quinquenervia populations had the lowest storages of critical plant nutrients and the fewest number of mature native trees. The results of this study will benefit local efforts to manage invasive species. Currently there is a need to develop a better understanding of the ecological consequences of exotic species invasion and methods for countering them. This issue is vital to the restoration of the Everglades because efforts are underway to restore hydrologic systems in Florida?s natural areas without a good understanding of ecosystem-level effects of exotic plants, which may ultimately hinder or even prevent restoration. Elucidation of the extent, duration, and impact of the changes caused by exotics will help in developing more effective restoration and management techniques.
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 Melissa Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Reddy, Konda R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024299:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024299/00001

Material Information

Title: A multi-disciplinary evaluation of the invasion and management of Melaleuca quinquenervia
Physical Description: 1 online resource (166 p.)
Language: english
Creator: Martin, Melissa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: bio, biological, control, ecosystem, exotic, herbicide, herbivory, invasion, litterfall, melaleuca, nutrient, soil
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Ecosystem invasion by exotic plant species poses a significant threat to community biodiversity, function, and stability in the Florida Everglades. One of the most problematic invasive species in this ecosystem is Melaleuca quinquenervia (Cav.) Blake. The goal of my dissertation was to use a multi-scale, interdisciplinary approach to evaluate the ecosystem-level consequences of the invasion and management of M. quinquenervia. The results presented include above and belowground storages of nutrients in invaded and non-invaded systems, alterations of soil microbial biomass and function, evaluations of biological and chemical control methods, and plant community response to a native disturbance regime in the context of management. Overall, both the invasion and management of M. quinquenervia altered basic ecosystem functions of nutrient storage and cycling, plant community diversity, and community response to native disturbance. We found that in the absence of top-down control from herbivory, M. quinquenervia trees were able to create a positive feedback loop to growth whereby increased quality and quantity of above- and below-ground biomass drove higher levels of storage and availability of below-ground resources. In addition, analysis of large-scale research plots in native, invaded, and managed sub-tropical wetland sites revealed that the use of herbicides to control M. quinquenervia invasion may damage long-term ecosystem structure and function. We found that after a seasonal fire plots that had been chemically treated to control M. quinquenervia populations had the lowest storages of critical plant nutrients and the fewest number of mature native trees. The results of this study will benefit local efforts to manage invasive species. Currently there is a need to develop a better understanding of the ecological consequences of exotic species invasion and methods for countering them. This issue is vital to the restoration of the Everglades because efforts are underway to restore hydrologic systems in Florida?s natural areas without a good understanding of ecosystem-level effects of exotic plants, which may ultimately hinder or even prevent restoration. Elucidation of the extent, duration, and impact of the changes caused by exotics will help in developing more effective restoration and management techniques.
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 Melissa Martin.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Reddy, Konda R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024299:00001


This item has the following downloads:


Full Text

PAGE 1

A MULTI-DISCIPLINARY EVALUATION OF THE INVASION AND MANAGEMENT OF MELALEUCA QUINQUENERVIA By MELISSA ROSEMARY MARTIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Melissa Rosemary Martin 2

PAGE 3

To the Boys 3

PAGE 4

ACKNOWLEDGMENTS Deepest thanks go to my major advisor, Dr K. Ramesh Reddy, without your academic and personal support this work would not have been po ssible. I would like to express my gratitude to my committee members, Dr. Samira M. Daroub, Dr Mark W. Clark, Dr. Ni cholas B. Comerford, and Dr. Michelle C. Mack, whose understanding a nd patience helped me walk through fire (well evaluated it at least). Special thanks go to the faculty, staff, and students of the Soil and Water Science Department in Gainesville, FL and the UF Research and Education Stations in Belle Glade and Fort Lauderdale, FL, especially Vivana Nadal, Brandi Schoefield, and Kelly Lewis. I thank Angelique Bochnak for being there to talk and listen every step of the way. This work would not have been possible without the support of the IGERT, AMW3 program at the University of Florida. In addi tion, this work was part ially fund by the Florida Exotic Pest Plant Councils Julia Morton, Invasive Plant Rese arch Grant Program, the Sam Polston Memorial Fellowship program, and th e Everglades Foundations Graduate Student Fellowship Program. Initial plant identificati ons were made by the Un iversity of Florida Herbarium at the Florida Museum of Natural History. To Gavin Wilson: I thank you for your help, patience, and friendship over the past five years. It has been a pleasure learning from a nd working beside someone so kind, intelligent, and insightful. I thank Miss Yu fo r your tireless efforts from the ve ry beginning to final end. To Shannon and Bill White: I thank you for opening your hearts and home to me over the last few years. My time at UF would have been half the fun and twice as hard without your support and friendship. To Philip W. Tipping: I thank you for your s upport, friendship, and humor over the last seven years. You have changed my life and I will be eternally grateful. 4

PAGE 5

I am especially appreciative of the invaluable insights and long hours of field support of staff of the USDA-ARS Invasive Plant Research Laboratory, including E ileen Pokorny, Danyelle Fitzgerald, Paul Madiera, Kayla Nimmo, Eli zabeth Bolton, Emily White, and Susan Keusch. Last but certainly not least, I thank my fam ily. Words cannot express your contribution to my happiness and sanity over the last five year s. You should know that this is as much your accomplishment as it is mine. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................12 ABSTRACT...................................................................................................................................14 CHAPTER 1 INTRODUCTION................................................................................................................. .16 Ecosystem Consequences and the Management of Plant Invasions.......................................16 Melaleuca quinquenervia : A Case Study of Invasion and Management in the Florida Everglades...........................................................................................................................17 Objective 1 An Investigation of Ecosystem-Alteration after Management of Melaleuca quinquenervia .............................................................................................21 Objective 2 Assessing the Impact of a Seasonal Fire on Native, Invaded, and Managed Plots..............................................................................................................21 Objective 3 An Analysis of Native and N on-native Litter Quality................................22 Objective 4 Recovery of Plant Community St ructure after a Seasonal Fire..................23 Figure......................................................................................................................................24 2 HERBIVORY INTERRUPTS INVADER SUCCESS BY DECREASING PLANT GROWTH AND ALTERING NUTRIENT DYNAMICS.....................................................25 Introduction................................................................................................................... ..........25 Materials and Methods...........................................................................................................28 Site Description...............................................................................................................28 Plot Description...............................................................................................................28 Litter and Soil Sampling..................................................................................................29 Litter Quality Analysis....................................................................................................30 Soil Characteristics..........................................................................................................30 Nutrient Analyses............................................................................................................31 Microbial Biomass...........................................................................................................31 Nutrient Availability........................................................................................................32 Statistical Analyses..........................................................................................................3 3 Results.....................................................................................................................................34 Discussion...............................................................................................................................36 Herbivory and Nutrient Dynamics..................................................................................36 Herbivory and Fire Interactions.......................................................................................40 Synthesis and Conclusion.......................................................................................................41 Tables and Figures............................................................................................................. .....43 6

PAGE 7

3 ASSESSING THE IMPACT OF NATIVE DISTURBANCE REGIMES IN FORESTS MANAGED TO CONTROL THE INVA SION OF AN EXOTIC TREE.............................59 Introduction................................................................................................................... ..........59 Materials and Methods...........................................................................................................62 Experimental and Statis tical Justification.......................................................................62 Site Description...............................................................................................................63 Litter and Soil Sampling..................................................................................................64 Litter Quality Analysis....................................................................................................65 Soil Characteristics..........................................................................................................65 Nutrient Analyses............................................................................................................65 Microbial Biomass...........................................................................................................66 Nutrient Availability........................................................................................................66 Statistical Analyses..........................................................................................................6 7 Results.....................................................................................................................................68 Discussion...............................................................................................................................71 Alteration of Aboveground Components........................................................................71 Alteration of Belowground Components.........................................................................74 Conclusions.............................................................................................................................76 Tables and Figures............................................................................................................. .....77 4 COMPARING NATIVE AND EXOTIC PL ANT QUALITY: IMPLICATIONS FOR NUTRIENT TURNOVER......................................................................................................89 Introduction................................................................................................................... ..........89 Materials and Methods...........................................................................................................91 Site Description...............................................................................................................91 Experimental Approach...................................................................................................91 Litter Component Analyses.............................................................................................92 Nutrient Analyses............................................................................................................92 Statistical Analyses..........................................................................................................9 3 Results.....................................................................................................................................93 Discussion...............................................................................................................................95 Organic Matter Turnover.................................................................................................95 Nutrient Turnover............................................................................................................96 Conclusion..............................................................................................................................98 Tables and Figures............................................................................................................. ...100 5 RECOVERY OF PLANT COMMUNITY STRUCTURE AFTER A SEASONAL FIRE.109 Introduction................................................................................................................... ........109 Materials and Methods.........................................................................................................111 Experimental and Statis tical Justification.....................................................................111 Site Description.............................................................................................................112 Diversity Plots...............................................................................................................113 Statistical Analyses........................................................................................................114 Results...................................................................................................................................114 7

PAGE 8

Discussion.............................................................................................................................115 Plant Community Structure...........................................................................................115 Plant Community Re-invasion......................................................................................116 Conclusion............................................................................................................................117 Tables and Figures............................................................................................................. ...119 6 SYNTHESIS.................................................................................................................... .....124 Objective 1 An Investigation of Ecosystem-Alteration after Management of Melaleuca quinquenervia ....................................................................................................................124 Objective 2 Assessing the Impact of a Seasonal Fire on Native, Invaded, and Managed Plots...................................................................................................................................125 Objective 3 An Analysis of Native and N on-native Litter Quality.....................................127 Objective 4 Recovery of Plant Community St ructure after a Seasonal Fire.......................128 Overall Conclusions............................................................................................................ ..128 Figures..................................................................................................................................130 APPENDIX A FULL MODEL STATISTICAL RESULTS........................................................................134 Chapter 2 Model Results......................................................................................................13 4 Chapter 3 Model Results......................................................................................................14 3 Chapter 4 Model Results......................................................................................................15 1 Chapter 5 Model Results......................................................................................................15 5 LIST OF REFERENCES.............................................................................................................157 BIOGRAPHICAL SKETCH.......................................................................................................166 8

PAGE 9

LIST OF TABLES Table page 2-1 Mean (S.E.) of Melaleuca quinquenervia litterfall biomass a nd nutrient transfer measured in the herbivory and non-herbivory plots..........................................................43 2-2 Mean (S.E.) of Melaleuca quinquenervia litterfall nutrient concentration and quality measured in the herb ivory and non-herbivory plots..............................................44 2-3 Mean ( S.E.) of preand post-fire litte r moisture, total biomass, and % litter loss measured in the herbivory and non-herbivory plots..........................................................45 2-4 Mean ( S.E.) of preand post-fire soil moisture, bulk density, and organic matter measured in the herbivory and non-herbivory plots..........................................................46 2-5 Mean ( S.E.) of preand post-fire litte r and soil nutrient concentration measured in the herbivory and non-herbivory plots...............................................................................47 2-6 Mean ( S.E.) of preand post-fire of litter nutrient storage measured in the herbivory and non-herbivory plots.....................................................................................48 2-7 Mean ( S.E.) of preand post-fire specifically minera lizable nitrogen and specifically mineralizable phosphorus leve ls measured in the herbivory and nonherbivory plots...................................................................................................................49 2-8 Mean ( S.E.) of preand post-fire soil microbial biomass carbon measured in the herbivory and non-herbivory plots.....................................................................................50 3-1 Mean ( S.E.) of preand post-fire litter moisture, biomass, and fire intensity in the non-invaded, herbicide, and biological sites......................................................................77 3-2 Mean ( S.E.) of preand post-fire fire soil moisture, bulk density, and organic matter in the non-invaded, herbic ide, and biological sites.................................................78 3-3 Mean ( S.E.) of preand post-fire fi re nutrient concentrations in the non-invaded, herbicide, and biological sites............................................................................................79 3-4 Mean ( S.E.) of preand post-fire litter nutrient storages in the non-invaded, herbicide, and biological sites............................................................................................80 3-5 Mean ( S.E.) of preand post-fire specifically minera lizable nitrogen and specifically mineralizable phosphorus levels measured in the non-invaded, herbicide, and biological sites.............................................................................................................81 3-6 Mean ( S.E.) of preand post-fire micr obial biomass carbon levels measured in the non-invaded, herbicide, and biological sites......................................................................82 9

PAGE 10

4-1 Mean ( S.E.) of litter nutrient concen trations at every sample time for each plant species........................................................................................................................ ......100 4-2 Mean ( S.E.) of mass nutrient ratios at every sample time for each plant species........101 4-3 Mean ( S.E.) of litter decomposition constants and residence times for each plant species reported in lite rature and this study.....................................................................102 5-1 Species list of all plants identified in the non-invade d, herbicide, and biologically controlled plots.................................................................................................................119 5-2 Diversity indices for the non-invaded, herbicide, and biologically controlled plots.......121 5-3 Woody species abundance for the noninvaded, herbicide, and biologically controlled plots.................................................................................................................122 A-1 Full model results for main effects and interactions for Melaleuca quinquenervia litterfall biomass and carbon, nitrogen, a nd phosphorus transfer in Chapter 2...............134 A-2 Full model results for main effects and interactions for litte r % moisture, litter biomass, and % litter loss in Chapter 2............................................................................135 A-3 Full model results for main effects and interactions for litte r carbon, nitrogen, and phosphorus concentration in Chapter 2............................................................................136 A-4 Full model results for main effects and interactions for litte r carbon, nitrogen, and phosphorus storage in Chapter 2......................................................................................137 A-5 Full model results for main effects and interactions for soil % moisture, bulk density, and organic matter in Chapter 2.......................................................................................138 A-6 Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus concentration in Chapter 2............................................................................139 A-7 Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus storage in Chapter 2......................................................................................140 A-8 Full model results for main effects and interactions for soil specifically mineralizable nitrogen and specifically minera lizable phosphorus in Chapter 2...................................141 A-9 Full model results for main effects and interactions for soil microbial biomass carbon in Chapter 2................................................................................................................... ...142 A-10 Full model results for main effects and in teraction for root biomass in Chapter 2.........142 A-11 Full model results for main effects and interactions for litte r % moisture, litter biomass, and % litter loss in Chapter 3............................................................................143 10

PAGE 11

A-12 Full model results for main effects and interactions for litte r carbon, nitrogen, and phosphorus concentration in Chapter 3............................................................................144 A-13 Full model results for main effects and interactions for litte r carbon, nitrogen, and phosphorus storage in Chapter 3......................................................................................145 A-14 Full model results for main effects and interactions for soil % moisture, bulk density, and organic matter in Chapter 3.......................................................................................146 A-15 Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus concentration in Chapter 3............................................................................147 A-16 Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus storage in Chapter 3......................................................................................148 A-17 Full model results for main effects and interactions for soil specifically mineralizable nitrogen and specifically minera lizable phosphorus in Chapter 3...................................149 A-18 Full model results for main effects and in teractions for soil microbial biomass carbon in Chapter 3................................................................................................................... ...150 A-19 Full model results for main effects and interactions for % mass loss, K value, and turnover time in Chapter 4...............................................................................................151 A-20 Full model results for main effects and interactions for ca rbon, nitrogen, and phosphorus concentration in Chapter 4............................................................................152 A-21 Full model results for main effects and in teractions for change in carbon, nitrogen, and phosphorus storage in Chapter 4...............................................................................153 A-22 Full model results for main effects and in teractions for litter chemical composition in Chapter 4...................................................................................................................... ....154 A-23 Full model results for main effects for non-woody plant species richness and diversity indices in Chapter 5..........................................................................................155 A-24 Full model results for main effects for Melaleuca quinquenervia live seedling, Melaleuca quinquenervia dead seedling, Pinus elliottii live seedling, and Taxodium distichum live seedling densities in Chapter 5.................................................................156 11

PAGE 12

LIST OF FIGURES Figure page 1-1 Conceptual model of a forest ecosys tem indicating the potential processes and components that may be altered by Melaleuca quinquenervia invasion and management.......................................................................................................................24 2-1 Maps of the study site in southwest Florida.......................................................................51 2-2 Mean ( S.E.) of total litterfall m easured during each sampling period in the herbivory and non-herbivory plots.....................................................................................52 2-3 Relationship between turnover time of litter and lignin:P ration of Melaleuca quinquenervia litterfall.......................................................................................................53 2-4 Mean ( S.E.) of preand post-fire litterpool biomass in the herbivory and nonherbivory plots...................................................................................................................54 2-5 Mean ( S.E.) of preand post-fire to tal carbon storage measured in the herbivory and non-herbivory plots.....................................................................................................55 2-6 Mean ( S.E.) of preand post-fire tota l nitrogen storage meas ured in the herbivory and non-herbivory plots.....................................................................................................56 2-7 Mean ( S.E.) of preand post-fire total phosphorus storage measured in the herbivory and non-herbivory plots.....................................................................................57 2-8 Conceptual models of the feedback cycle of Melaleuca quinquenervia biomass production in the two treatment plots................................................................................58 3-1 Map of the study site in southwest Florida showing the non-invaded, chemically controlled, and biologically controlled sites......................................................................83 3-2 Mean ( S.E.) of preand post-fire litterpool biomass by co mponents measured in the non-invaded, herbicide, and biological sites................................................................84 3-3 Relationship between microbial biomass carbon levels and % soil moisture in the three treatment sites.......................................................................................................... .85 3-4 Mean ( S.E.) of preand post-fire total storages of car bon measured in the noninvaded, herbicide, and biological sites.............................................................................86 3-5 Mean ( S.E.) of preand post-fire storages of nitrogen measured in the noninvaded, herbicide, and biological sites.............................................................................87 3-6 Mean ( S.E.) of preand post-fire storages of phosphorus measured in the noninvaded, herbicide, and biological sites.............................................................................88 12

PAGE 13

4-1 Mean ( S.E.) of litter mass remaining at every sample time for each plant species......103 4-2 Mean ( S.E.) of the chemical composition for each plant species.................................104 4-3 Relationship between litter residence tim e and final nutrient concentration for each plant species.....................................................................................................................105 4-4 Relationship between litter residence tim e and chemical and nutrient ratios at the final sample period for each plant species.......................................................................106 4-5 Mean ( S.E.) percent cha nge in the pools of nutrients from the initial storage at every sample time for each plant species.........................................................................107 4-6 Mean ( S.E.) pools of nutrients at every sample time for each plant species................108 5-1 Frequency data for each non-woody plant sp ecies in the non-invaded, herbicide, and biological plots.................................................................................................................123 6-1 Conceptual model of a forest ecosystem dominated by mature Taxodium distichum trees.......................................................................................................................... ........130 6-2 Conceptual model of a forest ecosystem dominated by early stage Melaleuca quinquenervia trees..........................................................................................................131 6-3 Conceptual model of a forest ecosystem dominated by early stage Melaleuca quinquenervia trees controlled with biological agents....................................................132 6-4 Conceptual model of a forest ecosystem dominated by early stage Melaleuca quinquenervia trees controlled with herbicides...............................................................133 13

PAGE 14

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A MULTI-DISCIPLINARY EVALUATION OF THE INVASION AND MANAGEMENT OF MELALEUCA QUINQUENERVIA By Melissa Rosemary Martin May 2009 Chair: K. Ramesh Reddy Major: Soil and Water Science Ecosystem invasion by exotic plant species poses a significant th reat to community biodiversity, function, and stabil ity in the Florida Everglades. One of the most problematic invasive species in this ecosystem is Melaleuca quinquenervia (Cav.) Blake. The goal of my dissertation was to use a multi-scale, interdiscip linary approach to evaluate the ecosystem-level consequences of the invasion and management of M. quinquenervia The results presented include above and belowground storages of nut rients in invaded and non-invaded systems, alterations of soil microbial biomass and function, evaluations of biologica l and chemical control methods, and plant community response to a nati ve disturbance regime in the context of management. Overall, both the invasion and management of M. quinquenervia altered basic ecosystem functions of nutrient storage a nd cycling, plant community divers ity, and community response to native disturbance. We found that in the ab sence of top-down control from herbivory, M. quinquenervia trees were able to create a positive feedback loop to growth whereby increased quality and quantity of aboveand below-ground biomass drove higher levels of storage and availability of below-ground resources. In additi on, analysis of large-scale research plots in native, invaded, and managed sub-tropical wetland s ites revealed that the use of herbicides to 14

PAGE 15

control M. quinquenervia invasion may damage long-term ecosystem structure and function. We found that after a seasonal fire plots that had been chemically treated to control M. quinquenervia populations had the lowest storages of critical plant nutrients a nd the fewest number of mature native trees. The results of this study will benefit local effo rts to manage invasive species. Currently there is a need to develop a bett er understanding of th e ecological consequences of exotic species invasion and methods for countering th em. This issue is vital to th e restoration of the Everglades because efforts are underway to restore hydrologic systems in Floridas natural areas without a good understanding of ecosystem-level effects of e xotic plants, which may ultimately hinder or even prevent restoration. Eluc idation of the extent, duration, and impact of the changes caused by exotics will help in developing more effec tive restoration and management techniques. 15

PAGE 16

CHAPTER 1 INTRODUCTION Ecosystem Consequences and the Management of Plant Invasions Ecosystem invasion by exotic plant species poses a significant th reat to community diversity, function, and stability (Kohli et al. 200 4, Yurkonis et al. 2005). Exotic plants often thrive in new habitats where they are free of top-down regulation from herbivory which allows them to out-compete native plant species a nd dominate ecosystems (D'Antonio and Meyerson 2002). Although it is clear that exotic species in vasion can alter the basi c structure of native plant communities, the resulting consequences for ecosystem function are less predictable. Declines in native species diversity and abunda nce may reduce or eliminate an ecosystems ability to provide ecological goods and services, such as waste processing and carbon sequestration (Fenn et al. 2003). However, studies have found both positive and negative changes in the rates of nutrient storage and cy cling of invaded ecosystems (Ehrenfeld 2003). These changes may encompass functional changes in plant structure and growth rate which can alter aboveand belowground nutri ent pool sizes (Ehrenfeld 2003). For example, grassland invasion by woody plants has been shown to increas e the storage of carbon in standing biomass (Jackson et al. 2002). In addition, exotic plants may differ from native species in litter nutrient concentration and relative decomposability or li tter quality (Ehrenfeld 2003). A sample of 30 invasive species from Hawaii was found to have higher foliar nutrient levels as compared to native plants, potentially altering the rate of ecosystem nutrien t fluxes (Baruch and Goldstein 1999). Federal, state, and local governments have cr eated comprehensive weed control programs to reduce the deleterious effects of these exotic plant inva sions. The most common method of controlling exotic plants is by us ing herbicides. Applications mu st be made in perpetuity to 16

PAGE 17

maintain satisfactory control and several programs have touted the maintenance control concept whereby annual costs are reduced because populations are treated regularly and thereby kept at relatively low levels of density and abundance (Ramey and Hassell 2005). However, herbicides can injure adjacent na tive vegetation and, despite the m aintenance control concept, require large investments of resources. Mechanic al control, or the physical removal of exotic populations, is costly and time consuming thereby limiting its effectiveness for large scale efforts. This approach can al so have significant negative collat eral effects on local flora and fauna. Classical biological control attempts to re unite weeds with their coevolved natural enemies, most of which are insects (Center et al. 2002). However alt hough successful programs are self-sustaining, they can ta ke a decade to implement and th e relatively slow action of the biological control agents is incompatible with current management practices using herbicides (Center et al. 1999). While the goa l of these all these approaches is to reduce exotic populations and restore ecosystem integrity, little work has been done to monitor and evaluate their impact on ecosystem function. Melaleuca quinquenervia: A Case Study of Invasion and Management in the Florida Everglades Melaleuca quinquenervia (Cav.) Blake, otherwise known as the paper-bark tree, cajeput, punk tree, or white bottlebrush tree is a member of the Myrtaceae family, sub family Leptospermoidae. This tall evergreen tree hist orically occupies tropic al wetland sites throughout its native range along the eastern coast of Australia (Kaufman and Smouse 2001). It was introduced into South Florida in 1886 (Dray et al. 2006), or iginally for sale as an ornamental, but later was used for erosion cont rol, as a forestry crop, and as an agricultural windrow plant (Meskimen 1962, Stocker and Sander s Sr. 1981, Bodel et al. 1994). 17

PAGE 18

The exotic tree colonized and thrived in most natural areas of South Florida, including bayhead tree islands, sawgrass prairies, pine flatwoods, pastures, and cypress forests (Bodel et al. 1994). M. quinquenervia has several morphological adaptations that make it well suited to the dynamic environmental conditions found in South Florida including its to lerance to variable water pH, salinity, and depth (Kaufman and Smouse 2001). Dense stands of M. quinquenervia are found in well drained, seasonally saturate d, and permanently flooded sites (Bodel et al. 1994). Plant survival is enhanced by large numbers of arenchymenous surface roots produced shortly after flooding (SerbesoffKing 2003). Trees can produce sinker roots that extend down to the water table during dry periods (Serbesoff-King 2003). M. quinquenervia is also a competitive rooter and it aggressively colonizes areas dominated by the roots of native plants (Lopez-Zamora et al. 2004). Once established, M. quinquenervia trees have a high growth rate and reach reproductive maturity at three years after which they may flower two to five times a year (Bodel et al. 1994). A single M. quinquenervia tree can hold an estimated 5.6 million viable seeds in its canopy which, once released, may remain viable in the soil for at least three years (Van et al. 1998, Rayamajhi et al. 2002). Seeds are held in capsules that open in re sponse to desiccation, frost, or fire (Serbesoff-King 2003). The high concentr ations of essential oils found in mature M. quinquenervia trees fuel canopy fires that can kill na tive vegetation. These perturbations are followed by massive releases of M. quinquenervia seed and may lead to the creation of M. quinquenervia monocultures (Serbesoff-King 2003). By the early 1980s it became clear that M. quinquenervia was significantly altering the plant communities of the Florida Everglades. In 1980, a survey completed by the U.S. Forestry Service revealed that 161, 874 hect ares of land was covered by pure M. quinquenervia stands 18

PAGE 19

(Bodle et al. 1994). LaRoche and Fe rriter (1992) estimated that once M. quinquenervia populations reached a critical ecosy stem concentration of two to fi ve percent, ninety-five percent infestation occurred within 25 years. Native plant communities are altered significantly following M. quinquenervia invasion (Mazzotti et al. 1997). Di Stefano and Fisher (1983) found that the relative density, fr equency, and dominance of several native plant species were diminished significantly in sites invaded by M. quinquenervia as compared to neighboring noninvaded sites. This difference was explained in part by potential allelopathic properties of M. quinquenervia leaf extracts that reduced seed germin ation and seedling growth of native plant species. In addition, Mazzotti et al. (1981) found decreased di versity and abundance of small mammals in M. quinquenervia stands as compared to native plant communities. Myers (1984) identified the Pinus elliottii Taxodium distichum ecotone as one of South Floridas most susceptible habitats for M. quinquenervia invasion. This is the transition zone between upland P. elliottii -dominated sites and depressional T. distichum -dominated swamps where neither species can grow to its full potenti al (Myers 1983). Greenhouse and field research in this ecotone found high surv ival and growth rates for M. quinquenervia seedlings, indicating the suitability of this area for M. quinquenervia invasion (Myers 1984). Ewel (1986) agreed with Myers assessment and stated that ecosystem a lteration after such introductions was inevitable and that understanding the interactions betw een native and introduced species was key in developing effective management strategies. Perhaps the most successful integrated pest ma nagement project in South Florida to date has been the effort to control M. quinquenervia in the Everglades. Inter-agency cooperation has led to the integration of chemical, mechanical and biological control methods to reduce the impact of M. quinquenervia (Ferriter et al. 2005). The South Florida Water Management District 19

PAGE 20

(SFWMD) has sustained a multi-year campaign to chemically and mechanically control M. quinquenervia on public lands. In 2003 alone th e SFWMD treated 1,795 hectares of M. quinquenervia with ground application and 4,118 hectares using aerial applicati ons of herbicides (Ferriter et al. 2005). In a ddition, the SFWMD partially funded a biological control project headed by the United States Department of Ag riculture, Agricultural Research Service (ARS) (Ferriter et al. 2005). This pr oject, begun in 1986 at the ARS Invasive Plant Research Lab in Fort Lauderdale, is responsible for the release of four biological contro l agents to reduce or eliminate the capacity of M. quinquenervia to invade (Ferriter et al. 2005). Two of the biological agents, Oxyops vitiosa Pascoe (Coleoptera: Curculionidae) and Boreioglycaspis melaleucae Moore (Hemiptera: Psyllidae), have successfully established and are suppressing M. quinquenervia reproduction, growth, and recruitment on a landscape scale (Tipping et al. 2009). Although there have been significant reductions in the rates of growth, reproduction, and spread of M. quinquenervia no work has been done investigating the ecosystem impacts of different management strategies (Tipping et al. 2008). Invasive plant management programs are often evaluated based on the quantity of plant biomass removed or reduction in rates of exotic population spread. However, other factors need to be considered in program evaluation including the impact on non-target vegetation, recovery of native pl ant communities, and alteration of ecosystem function (Denslow a nd DAntonio 2005). My Ph.D. dissertation investigated the implications of M. quinquenervia invasion on the decomposition of organic matter, above and belowground storages of nutri ents, soil microbial community biomass and function, and plant community response to a native disturbance. In addition, this work explored the consequences and benefits provided by curr ent management strategies. The key processes measured and presented in the following chapters are identif ied in Figure 1-1. 20

PAGE 21

Objective 1. An Investigation of Ecos ystem-Alteration after Management of Melaleuca quinquenervia. The first objective of my P h.D. dissertation was to investigate the impact of the M. quinquenervia biological control agents on ecosystem processes. My previous research examined the potential impact of M. quinquenervia insect biological control agents on ecosystem processes (Martin et al. 2009). In 2005, litterf all was collected from an area dominated by mature M. quinquenervia trees and saplings that had been under attack by O. vitiosa since 1998 and B. melaleucae since 2002. In these plot s litterfall was approxima tely 60 g dry weight m-2 year-1. Although not quantified, the M. quinquenervia leaves collected du ring litterfall sampling had noticeable insect feeding damage. Post biologi cal control litterfall values were significantly lower than rates reported in othe r studies. Van et al. (2002) repor ted pre-biological control litter production of 750 to 930 g dry weight m-2 year-1. The nutrient poor, sandy surface soils of many of Floridas forests do not have a large capac ity to store nutrients vital for ecosystem maintenance. Reductions in litterfall after i nvasion and subsequent biological management of M. quinquenervia may result in changes to soil nutrient stor age and availability which in turn could have significant consequences for co mmunity and ecosystem restoration. The work presented in Chapter 2 tested two main hypotheses: 1) herbivory from the biological control agents will lower M. quinquenervia litter quality and ra tes of litter production and 2) herbivore-induced changes in litter quality will lower soil nutrient storage and availability before and after a seasonal fire. Objective 2. Assessing the Impact of a Se asonal Fire on Native, Invaded, and Managed Plots. Many ecosystems depend on regular disturbances such as seasonal fires, to maintain community structure and function. This is espe cially true in Florida where many native plant communities depend on seasonal fire s (Wade et al. 1980). Fires influence ecosystem function by 21

PAGE 22

opening canopies and triggering the seed release and germination of plant species (Neary et al. 1999). Fire-adapted species also rely on temporar y pulses of soluble nutrie nts and reductions in plant competition (Neary et al. 1999). Howeve r, many of Florida natural areas have been invaded by exotic plants, including the fire-adapted M. quinquenervia which have altered native fire patterns (Wade et al. 1980). While knowledge of the consequences of fire on aboveground processes is vital, belowground alterations may be of equal importance to understanding ecosystem function. Little work has been done to evaluate the effects of fire on nutrient storage and availability in Floridas invaded and manage d ecosystems. Therefore, the second objective of my Ph.D. dissertation was to elucidate the ch anges in nutrient storag es and availability in native, invaded, and managed sites after a seasonal fire. The work presented in Chapter 3 tested two main hypotheses were tested: 1) M. quinquenervia invasion and treatment with an herbicide will reduce the quantity and availability of nutrients before and after a seasonal fire compared to an non-invaded site and 2) M. quinquenervia invasion and treatment with biological control agents wi ll not alter the quantity and availability of nutrients befo re and after a seasonal fire compared to an non-invaded site. Objective 3. An Analysis of Nati ve and Non-native Litter Quality. The third objective of my Ph.D. dissertation was to investigate differences in rates of organic matter turnover and nutrient release from M. quinquenervia litter and the l itter two native tree species. Organic matter decomposition and the subsequent release of plant available nutrients is a vital ecosystem process that is controlled by biotic and ab iotic factors such as environmental conditions (temperature, pH, and available moisture) and the chemical composition of plant litter. For example, th e ratio of soluble fibers (e.g. sugars and carbohydrates) to resistant materi als (e.g. lignin) in organic ma tter can affect the rate of decomposition and subsequent release of plant available nutrients. The alteration of organic 22

PAGE 23

inputs after a disturbance, such as exotic plant invasions, can a lter basic ecosystem structure and function (Ehrenfeld 2003, Mack and D'Antonio 2003). Baruch and Goldstein (1999) and Ehrenfeld (2003) found that the litter of exotic plants may differ from native species in chemical composition, which could alter the rate of ecosystem nutrient fluxes. The work presented in Chapter 4 tested two main hypotheses were tested: 1) M. quinquenervia will have the slowest rate of decomposition and 2) M. quinquenervia litter will release least amount of carbon, nitr ogen, and phosphorus compared to T. distichum and P. elliottii litter. Objective 4. Recovery of Plant Communi ty Structure after a Seasonal Fire The fourth objective of my Ph.D. dissertati on was to investigate the impact of the M. quinquenervia invasion and management on plant communi ty structure in invaded and managed forest plots. It is often assumed that cont rol methods reduce the competitive advantage of invasive plants and allow native plant comm unities to restore pr e-invasion conditions. Approaches to restoring communities after the ma nagement of exotics can run the gamet from passive to active. The most passive approach relies on native plant communities recovering on their own, while more active methods involve dir ected efforts like replan ting natives. Selecting the best approach depends on seve ral factors including cost and the impact of the management itself on the native plant community. Mechanical chemical, and biological control programs have contained the spread and eliminat ed the invasive potential of existing M. quinquenervia populations (Ferriter et al. 2005, Ti pping et al. 2008, Tipping et al. 2009). However, live noninvasive M. quinquenervia trees remain part of vegetative la ndscape and are targets for future management. Treatment of remnant M. quinquenervia populations with chemical or mechanical methods may cause significant collateral damage to native plant communities and may 23

PAGE 24

negatively influence ecosystem function. Further work is needed to determine if communities would actually benefit from the removal of this exotic, but now less invasive plant. The work presented in Chapter 5 will test two main hypotheses: after a seasonal fire 1) plant community structure will not be different in an invaded and biologica lly controlled site but will be different in an invaded and chemically controlled site compared to the non-invaded site and 2) the re-invasion of M. quinquenervia will be most severe in the chemically controlled site compared to the biologically controlled and non-invaded sites. Figure Aboveground Biomass Belowground Biomass Litterfall Biomass Litter Biomass Soil Organic Matter Soil Microbial Biomass Soil Soil Bio-available Nutrients Figure 1-1. Conceptual model of a forest ec osystem indicating the potential processes and components that may be altered by Melaleuca quinquenervia invasion and management. 24

PAGE 25

CHAPTER 2 HERBIVORY INTERRUPTS INVADER SUCC ESS BY DECREASING PLANT GROWTH AND ALTERING NUTRIENT DYNAMICS Introduction Invasive plants can thrive in new habitats where they out-compete native plants and dominate communities and whole ecosystems (Mack and D'Antonio 1998, Mack et al. 2000, Mitchell and Power 2003). The reasons for their su ccess are debated and have led to the creation of hypotheses that roughly break do wn into either top-down or bo ttom-up centric explanations of invasion or some combinati on of both (Shea and Chesson 2002, Blumenthal 2006). The enemy release hypothesis attributes their success to reductions in top-down forces where natural enemies are left behind when plants inva de into a new area (Keane and Crawley 2002). Conversely, the resource hypothesis suggests that invasion is f acilitated by more bottom-up forces such as resource avai lability (Davis et al. 2000). The enemy release hypothesis forms the cornerstone for the practice of classical biological control of weeds whereby coevol ved, host specific insect herb ivores from the plants native range are reunited with the plant in its extant range, thereby restoring some level of top-down regulation via herbivory (Mack et al. 2000). This practice has lead to th e successful control of many exotic species including the floating aquatic plant Salvinia molesta Mitchell (Room et al. 1981). However, bottom-up forces may affect topdown controls when the presence of excessive resources like eutrophication allows plants to compensate for herbivory (Heard and Winterton 2000). Blumenthal (2006) suggested that both resource availability and enemy release act in concert to facilitate invasion. He also pred icted that fast-growing, presumably high resource plant species like Melaleuca quinquenervia (Cav.) Blake would be nefit more from enemy release. 25

PAGE 26

Melaleuca quinquenervia otherwise known as the paper-ba rk tree, cajeput, punk tree, or white bottlebrush tree is a member of the Myrtac eae family, sub-family Leptospermoidae. This evergreen tree historically occ upies tropical wetland sites th roughout its native range along the eastern coast of Australia (Kaufman and Smouse 2001). It was introduced into South Florida in 1886 (Dray et al. 2006), originally for sale as an ornamental, but later was used for erosion control, as a forestry crop, and as an agri cultural windrow plant (Meskimen 1962, Stocker and Sanders Sr. 1981, Bodel et al. 1994). Once established, M. quinquenervia trees have a high growth rate and reach reproductive maturity in three years after which they may flower two to five times a year (Bodel et al. 1994). A single M. quinquenervia tree can hold an estimated 5.6 million viable seeds in its canopy which, once release d, may remain viable in the soil for at least three years (Van et al. 1998, Raya majhi et al. 2002). The high concentrations of essential oils found in mature M. quinquenervia trees fuel canopy fires that can kill native vegetation. These perturbations are followed by massive releases of M. quinquenervia seed and may lead to the creation of M. quinquenervia monocultures (Serbesoff-King 2003). Although there are over 450 known herbivores that provide top-down regulation of M. quinquenervia in its native range, there are no known native specialist he rbivores that feed on the plant in its introduced range (Burrows and Balciunas 1999, Costello et al. 2003). However, currently the growth a nd reproductive capacity of M. quinquenervia are being suppressed by two intentionally introduced, specialized insect herbivores, Oxyops vitiosa Pascoe (Coleoptera: Curculionidae) and Boreioglycaspis melaleucae Moore (Hemiptera: Psyllidae) (Tipping et al. 2008). Oxyops vitiosa larvae and adults feed on buds and ne wly flushing leaves (Balciunas et al. 1994). The feeding activity of the larvae is th e most damaging to the plant and causes long, window-like scars on leaves (Purcell and Balciuna s 1994). Larvae are covered in a thick viscous 26

PAGE 27

coating consisting of essen tial oils sequestered from M. quinquenervia which provides a potent anti-predator defense (W heeler et al. 2002). Boreioglycaspis melaleucae feed on plant phloem and third through fifth instar nymphs secrete a wa xy flocculence (Purcell et al. 1997, Wineriter et al. 2003). Both species are established thr oughout the State of Florida and are suppressing M. quinquenervia on a landscape level (Center et al. 2006, Tipping et al. 2008). Many of Floridas natural areas that have been invaded by M. quinquenervia depend on regular disturbances like fires to maintain community structure and function. Native plant communities in these areas depend on seasonal fires to open canopies and trigger the seed release and germination of plant species, provide temporar y pulses of soluble nutr ients, and reduce plant competition (Wade et al. 1980, Neary et al. 1999). The invasion of M. quinquenervia and other exotic plants have altered native fire pattern s which may have long-term consequences for ecosystem function (Wade et al. 1980). While understanding the consequences of fire on aboveground processes is vital, understanding belowground alterations to ecosystem functions is no less important particularly after the inva sion and management of exotic plants. This work was part of a larger study that ex amined how the absence of top-down controls affects M. quinquenervia growth, survival, and inter-specific competition in a resource-limited ecosystem (Tipping et al. 2009). To date, the st udy has found increases in plant density, height, and survival in herbivore-free pl ots (Tipping et al. 2009). The ma jor objective of this work was to elucidate the mechanisms by which M. quinquenervia can dominate a resource limited ecosystems when freed from the top-down control of herbivory. In particular, we were interested in the consequences for aboveand below-ground nutrient transfer, storage, and availability before and after a seasonal fire. Two main hypotheses were tested: 1) herbivory from the biological control agents will lower M. quinquenervia litter quality and ra tes of litter production 27

PAGE 28

and 2) herbivore-induced changes in litter quality will lower soil nutrient storage and availability both before and after a seasonal fire. Materials and Methods Site Description The study site was located in the Belle Meade Tract of the Picayune Strand State Forest in Collier County, Florida (Figure 2-1). This area c onsists of nearly level, poorly drained, low fertility soils which are loamy, siliceous, hypertherm ic Arenic Glassoqualfs. The soil series is Pineda-Boca-Hallandale which is characterized by moderately to poorly drained sands over-lying limestone bedrock at a depth of approximately 1.4 m (USDA 1998). The water table fluctuates annually between greater than 15 cm below the soil surface to approximately 25 cm above. The area has a distinct wet season from approximate ly July to December and a dry season from January to June. Average annual rainfall in th is region is approxima tely 1.36 m (SERC 2007). Vegetation in this area was a mixed Pinus elliottii Englem-Taxodium distichum (L.) L.C. Rich var. nutans (Ait.) Sweet forest with a hard wood under-story and a few mature M. quinquenervia trees. Over the past several decades, many areas in this landscape have been invaded more extensively by M. quinquenervia and are now characterized by sp arse populations of mature trees with dense understories of seedlings and saplings that can exceed 100 plants per square meter. Plot Description This work was part of a larger study that ex amined how the absence of top-down controls affects M. quinquenervia growth, survival, and inter-specific competition. The experimental design is presented here briefl y and in full in Tipping et al. 2009. In order to assess the establishment and spread of B. melaleucae four permanent transects were established in March 2002 radiating out in cardinal directions from a central location. Pairs of 9 m2 plots, separated 28

PAGE 29

by at least 10 m, were demarcated at the cen tral location and at 30, 60, and 100 m intervals in each direction (n = 26). However, because the insect quickly spread beyond the limits of the transects, the plots were used to assess the impact of the introduc ed agents (Center et al. 2006). Two treatments were assigned randomly to each pair of 9 m2 plots #1) sapling or prereproductive M quinquenervia with herbivory (hereafter referred to as the herbivory plots) and area #2) sapling M. quinquenervia with no herbivory (hereafter refe rred to as the non-herbivory plots). In both areas saplings average about 2 m in height and ha d not yet flowered and produced seeds. Plants in the herbiv ory plots have been under attack by O. vitiosa since 1998 and B. melaleucae since 2002, while those in the non-herbivor y plots have been protected from herbivory by the biological control agents sin ce 2002 by monthly foliar app lications of acephate (OSdimethyl acetylphosphor-amidothiote). A concentration of 0.367% (v/v) acephate was applied to all plant foliage until runoff ev ery 4 weeks during 2003 through 2007 using a hand pressurized backpack sprayer. In early May of 2007 the Great Basal fire burned approximately 8, 000 hectares in southwest Florida and all of the established experimental plots. As the fire was not planned no direct measurements of the fire intensity were taken. However, all of the plots were equally affected as well as the entire surrounding lands cape. The aboveand belowground samples were taken 24 to 48 hours after the fire. Litter and Soil Sampling On September 3, 2006, two 0.25 m2 litter-traps were deployed in the twenty-six 9 m2 plots. Litterfall was collected every month until the fire in May 2007, air dried to a constant weight, separated into component parts, and reported on a dry weight basis. On March 27, 2007 and May 15, 2007, the litterpool wa s sampled in every 9 m2 plot (n=13 in each treatment) by placing two 0.1 m2 frames on the surface of the soil and collecti ng all of the organic material therein. 29

PAGE 30

Litterpool samples were separated into undecompos ed Oi, moderately decomposed Oe, humified Oa, and woody biomass layers. Litterpool sample s were air dried to a constant weight and reported on a dry weight basis. Fire intensity was estimated by calculating the percentage of the total litterpool that was lo st after the May 2007 fire. On March 27, 2007 and May 15, 2007, a belowground biomass was estimated by taking a 5 cm diameter soil core from a randomly selected location within each plot. Each belowground biomass core was separated at two depths, 0 cm and 5 cm, and sieved to separate fine roots which were then rinsed, air drie d to a constant weight, and weighed. A second 5 cm diameter soil core was taken as above and also separated into 0-5 cm and 5-15 cm depths. These soil samples were returned to the laboratory, siev ed to remove roots and large plant debris, homogenized, and kept at 4C for a maximu m of 10 days before microbial analysis. Litter Quality Analysis Litter quality was measured with a sequen tial extraction using an Ankom A200 Fiber Analyzer (Rowland and Roberts 1994). 0.5 g of co arsely ground litter material was weighed and sealed into Ankom filter bags. The bags were extracted with a neutral detergent to remove soluble cellular contents (sugars, carbohydrates, lipids, etc.) followed by an acid detergent to removal hemi-cellulose. The residual (hemi-cellulose, lignin, and ash) was combusted at 550C for 4 hours to determine ash content. The resist ant pool is composed of the hemi-cellulose and lignin components. Litter quality was calculated on a dry mass basis. Soil Characteristics Percent moisture and bulk density of soils were determined by drying 20 to 30 g subsamples of field-moist, sieved, and homogenized soil at 70C for three days. Bulk density and percent moisture were determined on a wet soil weight basis and pH was measured on a 2:1 water: soil slurry with an Accumet Research, AR50 dual ch annel pH/ion/conduc tivity meter. 30

PAGE 31

Soil organic matter was measured by loss on igni tion from 0.2 to 0.5 g samples of dried and ground soils, which were first measured into 50 mL beakers (Luczak et al. 1997). The beakers were placed in a muffle furnace and brought to 250C for 30 minutes. The furnace temperature was then increased to 550C for 4 hours. Orga nic matter content was calculated as the mass loss on ignition on a dry weight basis. Nutrient Analyses Dried and ground soil and plant material were analyzed for percent carbon and nitrogen on a Thermo-Electron, 1112 Series elemental analyz er. Total phosphorus was determined by a twophase acid extraction after lo ss on ignition (Andersen 1976). Th e ash remaining in the 50 mL beaker was moistened with 2 to 3 mL distilled de-ionized water and then extracted with 20 mL of 6 N hydrochloric acid (HCl). All of the water was removed and the hot plate was placed on high for 30 minutes to completely dry the samp les. After cooling, 2.25 mL of 6N HCl was added to each beaker and the beakers placed on a hot plate until almost boiling. Extracts were then filtered through a #41 Whatma n filter into 50 mL volumetric flasks that were then brought to volume with distilled de-ionized water. Total phosphorus was measured with an automated ascorbic acid method on a Bran and Luebbe Auto Analyzer 3, Digital Colorimeter (Method 365.4; USEPA 1993). Microbial Biomass Microbial biomass carbon (MBC) was measured by a modified chloroform-fumigation extraction method (Vance et al. 1987). Two repl icates of 1 g field-moist soil samples were weighed into 50 mL centrifuge tubes. One of the duplicates was immediately extracted with 25 mL of 0.5 M potassium sulfate (K2SO4), shaken for 1 hour, and then filtered through a Whatman #41 filter. The second replicate of soil underw ent chloroform fumigation where 0.5 mL of chloroform was added to each tube and the tubes were placed in a dessicator with a beaker 31

PAGE 32

containing 30 mL of chloroform and boiling ch ips. The dessicator was vacuum-sealed for 24 hours then alternatively f illed with air and evacuated ten times to remove all residual chloroform. The tubes were then extracted as described above and all extracts were stored at 4C until analysis for MBC. This was calculated as the difference between the total carbon in the fumigated and un-fumigated soil extracts as measured on a Shimadzu TOC 5050C, total organic carbon analyzer. Nutrient Availability Potentially mineralizable nitrogen (PMN) wa s measured with a modified incubationextraction method that calculates a minerali zation potential based on the net production of ammonium during a 10 day anaerobic incuba tion (White and Reddy 2000). In order to determine the initial amount of ammonium in the soil, a set of 1 g field-moist soil samples were weighed into 50 mL centrifuge tubes. The tubes received 25 mL of 0.5M K2SO4, shaken on a longitudinal shaker for one hour, and extracts were filtered through a #41 Whatman filter. An additional set of 1 g field-moist soil samples were weighed into 50 mL serum bottles with 5 mL of deionized water. These bottles were sealed with butyl rubber stoppers and aluminum crimps and their head-space was purged with oxygen (O2)-free nitrogen (N2) gas for 2-5 minutes. Bottles were incubated in the dark at 35C for 10 days then extracted as described above with 25 mL of 0.5 M K2SO4 to determine final ammonium c oncentration. Tota l ammonium was determined with an automated colorimetric method on a Bran and Luebbe Auto Analyzer 3, Digital Colorimeter (Method 353.2; USEPA 1993). Specifically mineralizable nitrogen (SMN) was determined by dividing PMN by the concentration on nitrogen in the soil (SMN = PMN TN-1). Potentially mineralizable phosphorus (PMP) wa s measured with a modified incubationextraction method under aerobic conditions (Grierson et al. 1999). Soils were incubated at a 32

PAGE 33

temperature of 35C for a period of 10 days in capped 125 mL polyethylene bottles that contained 5 g of soil held under optimal percent mo isture conditions (4-7%). Soils were aerated every three days and moisture content was adju sted if necessary. On day 10, soil sub-samples were collected from each incubation bottle and analyzed for available phosphorus. Soil samples were extracted with 0.1 M potassium chloride (K Cl) at a ratio of 10:1, shaken for one hour, and then filtered with 0.45 m me mbrane filters. Available phosphorus was measured with an ascorbic acid method on a Shimadzu Spect rophotometer UV-160 (Method 365.4; USEPA 1993). Specifically mineralizable phosphorus (SMP ) was determined by dividing PMP by the concentration on phosphorus in the soil (SMP = PMP TP-1). Statistical Analyses Values of measured soil and ve getation characteristics, as we ll as microbial population size and activity, were calculated as a mean for each plot. ANOVA tests were used to detect any differences between the measured parameters. Di fferences are reported as significant for tests with p values 0.05. Data that varied from normal distributions were transformed with square root(x), arcsin(x), or log(x+1) functions. The datasets that were tran sformed with the square root(x) function were: preand pos t-fire litterpool biomass and mois ture, preand post-fire litter carbon, nitrogen, and phosphorus storage, preand post fire soil organic matter (0-5 cm), preand post fire soil nitrogen concentration (0-5 and 5-15 cm), pre and post fire nitrogen availability (0-5 cm), and preand post-fire microbial biom ass carbon (0-5 and 5-15 cm). The fire intensity dataset was transformed with the square arcsin(x) function. The da tasets that were transformed with the square log(x+1) function were: pre-and po st fire carbon concentration (0-5 cm), preand post-fire phosphorus concentration (0-5 and 5-15 cm), preand pos t-fire soil carbon storage (0-5 and 5-15 cm), and preand post-fi re phosphorus storage (5-15 cm). All statistical analyses were preformed using JMP 7.0.1 software (SAS Institute, North Carolina, USA). 33

PAGE 34

Results Full model results for the main effects and in teractions are reported in Tables A-1 through A-10. Based on the magnitude of the results the ma in effects of treatment, site, and depth were the most important determinants of the measured response variables. There were no consistent transect effects for any of the measured variables. Both the total litterfall biomass and M. quinquenervia litterfall biomass were greater in the non-herbivory plot compared to the herbivor y plot (Table 2-1, Figure 2-1). Higher concentrations of carbon, nitroge n, and phosphorus were found in the M. quinquenervia leaf litterfall in the herbivory plots (Table 2-2). There were no differences in the moisture, total biomass, and % biomass loss of the litter between the treatment areas (Table 2-3). Both treatment areas had lower levels of litter mo isture after the fire (96.5 and 94.7% less for the herbivory and non-herbivory plots, respectively) (Table 2-3). Preand post-fire values for litter carbon, nitrogen, and phosphorus storage were not different between treatment areas (Figure 2-5, 2-6, and 2-7). Preand post-fire values for soil moisture, bulk density, and organic matter are reported in Table 2-4. Unlike litterfall, there were no differences on th e preor post-fire c oncentrations of the carbon or nitrogen in the litter be tween treatment areas (Table 2-5). Th e litter in the nonherbivory plots had the highest concentration of phosphorus both before and after the fire (Table 2-5). The soil in the non-herbi vory plots had higher concentrations of carbon, nitrogen, and phosphorus at the 0-5 and 5-15 cm depths before and after the fire with one exception, namely post-fire phosphorus at 515 cm. (Table 2-5). The storage of carbon in the 0-5 cm soil depth was higher after the fire in both treatment areas (1.4 and 2.3 times in the he rbivory and non-herbivory plots, respectively) (Figure 2-5). The storage of carbon in the 5-15 cm soil depth was also higher after the fire in both treatment 34

PAGE 35

areas (2.2 and 2.1 times in the he rbivory and non-herbivory plots, respectively) (Figure 2-5). Nitrogen storage in the 0-5 cm soil depth was high er after the fire in both treatment areas (1.3 and 1.9 times in the herbivory and non-herbivory plots, respectively) (Figure 2-6). The same was true for nitrogen storage in the 5-15 cm soil depth which was also higher after the fire in both treatment areas (1.3 and 1.4 times in the he rbivory and non-herbivory plots, respectively) (Figure 2-6). Phosphorus storag e in the 0-5 cm soil depth was higher after the fire in both treatment areas (1.4 and 1.8 times in the herbivor y and non-herbivory plots, respectively) (Figure 2-7). Similarly, phosphorus storage in the 5-15 cm soil depth was higher after the fire in both treatment areas (1.8 and 1.4 times in the herbivor y and non-herbivory plots, respectively) (Figure 2-7). Although nitrogen availability in creased within both treatment areas after the fire, there were no differences between treatments (Table 27). Phosphorus availa bility did not change within treatment plots post-fire and there were no differences between treatments (Table 2-7). Microbial biomass carbon (MBC) was not different in the 0-5 cm soil depth between treatment areas, both before and after the fire, despite the fact that MBC was lower in the 0-5 cm soil depth in the herbivory plots before the fire (Table 2-8). Overall, post-fire MBC was lower within individual treatment areas at bot h soil depths (Table 2-8). The data collected on the storages of carbon, nitrogen, and phosphorus were combined to give a broader ecosystem pictur e (Figures 2-5, 2-6, and 2-7). Before the fire, there was no difference in the total carbon storage between the two treatm ents (p = 0.48, Figure 2-5). However, after the fire, the non-he rbivory plots stored more carbon than the herbivory plots (p = 0.0343, Figure 2-5). The total storage of nitrogen did not differ between treatments before or after the fire (p = 0.41 and 0.30, Figure 2-6). Ho wever, nitrogen storage increased in both 35

PAGE 36

treatment plots after the fire (p = 0.22 and 0.0008, Figure 2-6). The total storage of phosphorus did not differ between the herb ivory and non-herbivory plots. Although post-fire phosphorus storage increased at both depths within tr eatments (p = 0.02 and 0.003 for herbivory and nonherbivory, respectively), there were no differences between treatment s either before or after the fire (Figure 2-7). Mean (+ SE) root biomass was greater in the 0-5 cm soil depth in the nonherbivory plots (254.43 89.5 g m-2) compared to the herbiv ory plots (70.55 15.07 g m-2). The same was true in the 5-15 cm soil depth where mean root biomass was 240.37 45.00 g m-2 in the non-herbivory plots and 47.01 14.06 g m-2 in the herbivory plots. Discussion Herbivory and Nutrient Dynamics Fast-growing, high resource adapted plants ma y disproportionately benefit from the lack of the top-down regulation of herbivory when introduced into new habitats (Fine et al. 2004, Blumenthal 2006). One such species, the Australian tree M. quinquenervia, has successfully colonized and invaded natural areas within Flor ida regardless of resour ce conditions. However, there is a dearth of studies on how these species maintain higher growth and reproduction rates in low resource environments. The results of this study elucidated how M. quinquenervia altered aboveand belowground nutrient dynamics in a low resource environment in a way which may act to promote and maintain site dom inance in the absence of herbivory. As hypothesized, both the total litterfall and M. quinquenervia leaf biomass in the litterfall was greater in the non-herbivory plots (Table 2-1, Figure 2-1). The feeding activity of O. vitiosa larvae caused a significant amount of damage to th e leaf material in the herbivory plots. A 2003 study showed that a single large O. vitiosa larvae can remove over 200 mg of fresh leaf biomass before it pupates (Wheeler 2003). Tipping et al. (2008) found that the inse ct exclusion treatment removed 98% of the larvae in the non-herbivory plots, resulting in 42 times fewer large larvae on 36

PAGE 37

the M. quinquenervia trees. Plant density in the herbi vore plots declined steadily over the previous five years as a di rect result of herbivory by O. vitiosa (Tipping et al. 2009). These lower plant densities, coupled with continuing defoliation by he rbivores, likely reduced the standing biomass that could contribute to litterfall in this study. The M. quinquenervia leaf litterfall in the non-herbivory plots had lower tissue carbon and nitrogen concentrations compared to the herbivor y plots (Table 2-2). Several studies have shown that herbivory can increase the quality of litterfall in forest ecosystems (Chapman et al. 2003, Chapman et al. 2006). Franks et al (2006) found that high densities of B. melaleucae caused M. quinquenervia seedlings to drop leaves prematurely which could interrupt the natural cycle of nutrient resorption during senescen ce and result in higher quality l eaf litter. Overall, there was no difference in the concentration of phosphorus in the leaf litterfall which resu lted in relatively phosphorus-rich tissues in the non-herbivory plots (T able 2-2). However, the greater volume of litterfall in the non-herbivory plots offset its lower overall nutr ient concentration, resulting in the transfer of 9 times more carbon, 10 times more nitrogen, and 17 times more phosphorus to the soil than in the herbivory plots during the six month sampling period (Table 2-1) Despite the 5 fold difference in litterfall betw een treatments, the total litter biomass on the soil surface was not different (Table 2-3). The carbon and nitrogen concentrations of the litter were the same between treatments, while the co ncentration of phosphorus was greater in the nonherbivory plots (Table 2-5). This is most lik ely due to the relative composition of the litter between the plots where 78% of th e litter in the herbivory plot s consisted of relatively lownutrient wood and humified organic material compar ed to only 49% in th e non-herbivory plots (p = 0.002) (Figure 2-4). Although the variability of the litter composition complicated individual 37

PAGE 38

treatment comparisons, there was clear trend of higher storages of carbon, nitrogen, and phosphorus in the herbivory plots (Figures 2-5, 2-6, and 2-7). The high rate of litterfall and low litter biom ass indicates that the organic matter in the non-herbivory plots is decomposing at a relatively rapid rate. In order to estimate turnover or residence time of the litter, it was assumed that th e ecosystem was in a steady state (Chapin et al. 2002). Projecting the measured li tterfall rates over a year-long period, the turnover time of the litter in the herbivory and non-herb ivory plots was estimated at 4.4 and 1.6 years, respectively (p = 0.002). In addition, the turnover time of the litter in both treatment areas was positively correlated with resistant:P ratios of the M. quinquenervia leaf litterfall (Figure 2-3). Several studies have found that the ratio of soluble fibe rs (e.g. sugars and car bohydrates) to resistant materials (e.g. lignin) in organic matter can a ffect the rate of decomposition and subsequent release of plant available nutri ents (Baruch and Goldstein 1999, Ehrenfeld 2003). The leaves in the non-herbivory plots contai ned less resistant material s and more phosphorus thereby increasing the rate of litter turnove r (Table 2-2). The faster tur nover of the litter in the absence of herbivory should result in significantly mo re nitrogen and phosphorus being mineralized and made available for plant growth. Although the absence of herbivory significantly increased the organic matter and nutrient additions to the soil, there were no differences in nitrogen storage or availability between treatments in the 0-5 or 5-15 cm soil depths (Table 2-7). In contrast while the storage of phosphorus was higher at 0-5 cm in the non-herbivor y plots, phosphorus availability was greatest in the herbivory plots at both so il depths (Table 2-7). This re sult was contrary to our hypothesis that storage and availability of nutrients would be higher in non-herbivory plots. Several studies have shown that herbivory can alter the soil nutrient availability th rough the addition of high 38

PAGE 39

quality materials (Franzluebbers 1999, Chapman et al. 2003, Ehrenfeld 2003). Chapman et al. (2006) found that herbivory by scale insects increased litter quality and nutrient cycling rates in a forested ecosystem. In contrast, we found that herbivory decreased the qu antity and quality of nutrient inputs to the soil. The observed decrea se in phosphorus availabi lity in this study may indicate higher levels of mi crobial and plant uptake in the non-herbivory plots. Alterations in soil microbial communities can have significant impacts on ecosystem functions like nutrient storage a nd cycling (Kourtev et al. 2003). Wardle et al. (2004) detailed how plant community structure drives changes in soil microbial community size and composition via a series of aboveground-belowground feedback loops. Several studies have shown that through the addition of high quality materials (e.g. frass and presenescent leaf litterfall), herbivory can increase levels of soil microbial biomass (Bardgett and Wardle 2003, Chapman et al. 2003). In contrast, we found that the absence of herbivory increased the addition of high quality leaf litter to the soil surface which supported higher leve ls of soil microbial biomass (Table 2-8). This may lead to a more rapid turnover of soil resources thereby increasing nutrient availability and potential for plant growth. A potential mechanism whereby M. quinquenervia out-competes native species is through a highly competitive root system. Lopez-Zamora et al. (2004) found that M. quinquenervia plants developed higher root densities than native grass competitors and maintained these densities throughout the soil profile, independent of moisture conten t. Similarly, Martin et al. (2009) found that the ratio of aboveground to belowground biomass was higher in a M. quinquenervia invaded forest compared to a native forest. In this study there were 4.2 times more root biomass in the non-herbivory plots (p = 0.001). This may be a function of plant density with 29.2 saplings growi ng in the herbivory plots and 70.8 saplings growing in the non39

PAGE 40

herbivory plots (Tipping, unpublished data). Ho wever, trees in the non -herbivory plots still produced 1.73 times more root biomass per plant after adjusting for dens ity (p = 0.004). This relatively higher root biomass per tree may result in greater capture of nutrient resources. Herbivory and Fire Interactions Many of Floridas low resource forests depend on disturbances such as fire to maintain ecosystem structure and function (Wade et al. 1980, Neary et al. 1999). In the past M. quinquenervia has been able to benefit from nativ e fire regimes and dominate systems (Serbesoff-King 2003). To date, no work has be en done evaluating the combined effects of herbivory and fire on ecosystem function in M. quinquenervia dominated forests. The results of this study show that the herbivore exclusion treatment altere d both the above and belowground ecosystem response to a fire. Not surprisingly, both the herbivory and non-herbivory plots lost a significant amount of litter biomass after the fire, specifically 73 a nd 79%, respectively (Table 2-3). While both treatment areas experienced re ductions in carbon and nitrogen storage in the litter, the nonherbivory plots lost the largest percentage of the litter carbon (74%, p = 0.42) and nitrogen (50%, p = 0.4). Other studies have reported simila r losses of aboveground nitrogen after fires (Kauffman et al. 1993, Hughes et al. 2000, Wan et al. 2001, Wanthongchai et al 2008). Nutrient losses from aboveground litter after fire can be caused by volatiliza tion and particulate transport, transformation of organic to inorganic forms, a nd transport by wind and water (Certini 2005). In contrast, litter storage of phosphorus increased after the fire, a fi nding contrary to other reported studies (Kauffman et al. 1993, Hughe s et al. 2000, Wanthongchai et al. 2008). It is possible that the intensity of the fire in our experimental plots volatilized car bon and nitrogen, while depositing phosphorus rich ash onto the soil surface. Unfortunately, as this fire was not planned, there were no intensity measuremen ts taken during the burn. 40

PAGE 41

Carbon, nitrogen, and phosphorus storage increased after the fire in both treatments at both soil depths (Figures 2-5, 2-6, and 2-7). However, soils in the non-herbivory plots stored more nutrients with the greatest increases in car bon (64%, p = 0.44), nitrogen (55%, p = 0.35), and phosphorus (83%, p = 0.11) storages occurring at 0-5 cm. The observed differences are likely the result of the incorporation of nutrient rich ash produced fro m the combustion of the larger litter pool into the surface soils of the non-he rbivory plots. In addi tion, the non-herbivory plots had a larger per plant root biomass. The trees killed by the fire may have released nutrients from the root biomass into the soil (Certini 2005). As expected, both plots experienced an increase in nitrogen availability post-fire, a finding supported by many studies (Giardina and R hoades 2001, Carter and Foster 2004, Wanthongchai et al. 2008) (Table 2-7). However, while th ere was no difference in the availability of phosphorus in the herbivory plots, the availability of phosphorus increased in the non-herbivory plots after the fire in both soil depths (Table 2-7). As discusse d above, this is most likely the result of the combustion of th e larger, relatively phosphorus ri ch litter pool. Changes in the storage and availability of nutri ents may have unpredictable futu re effects on ecosystem structure and function. Ross et al. (1997) found that the higher nitrogen and phosphorus availability after a fire increased the foliar nutrient content of th e re-colonizing plant communities. The increases found in nutrient availability post-fire may accelerate M. quinquenervia growth and reproduction thereby increasing the severity of the problem. Synthesis and Conclusion This study identified an indirect mechanism whereby M. quinquenervia out-completes native plants and maintains a dominant position in low-resource ecosystems. When freed from the top-down regulation of herbivory, M. quinquenervia creates a positive feedback loop to growth and reproduction (Figur e 2-8). Initially, high aboveand below-ground biomass 41

PAGE 42

production allows M. quinquenervia to out-compete native plants fo r nutrient and light resources. Higher rates of nutrient uptake pr oduce greater amounts of high qua lity standing biomass, which eventually falls to the soil at a much higher volu me. This high quality litter maintains a larger population of soil microbial biomass that processes litter and soil nutrients at a faster rate. The resultant mineralized nutrients are quickly taken up by the extensive M. quinquenervia root biomass, further increasing the production of vegetative and re productive biomass. The no herbivory advantages formerly experienced by M. quinquenervia have now been removed with the introduction and establishment of two specialized herbivores. As a result, all ages of M. quinquenervia are now under continuous attack by insect herbivores, from the most recently recruited seedlings to the tallest and fully mature trees. This relentless herbivory slows the rate of above-ground biomass production and reduces the size of the root zone (Figure 2-8). In addition, herbivores remove a significant am ount of standing biomass, thereby preventing it from falling to the forest floor. The remaining litterfall has higher concentrations of resistant materials such as lignin which, in turn, increase s its turnover time. C oncomitantly smaller pools of soil microbial biomass are supported, which furthe r reduces the rate of nut rient turnover. This is clear evidence of how herbivory not only controls populations of M. quinquenervia directly by reducing plant growth and reproduction, but also i ndirectly by interruptin g its positive feedback growth cycle which would otherwise mainta in its dominance in the ecosystem. We predict that M. quinquenervia populations exposed to herb ivory will be less invasive after native disturbances such as fire. A post-fire census of the experimental plots revealed that 73% of the trees were killed in the herbivory pl ots compared to only 41% of the trees in the nonherbivory plots (Tipping, unpublished data). Surviving trees in the herbivory plots may have been weakened by herbivory perhaps resulting in less root biomass which should reduce their 42

PAGE 43

ability to efficiently scavenge the pulse of av ailable nutrients produced immediately after the fire. Native plants should benefit from the reduced competition for light, space, and nutrients. In contrast, without herbivory, populations of M. quinquenervia will likely benefit disproportionally from the increases in storage a nd availability of nutrients, compared to native plants. Although some plant mortality would be expected from fires, over the longer term populations may actually benefit from reduced in tra-specific competition. Thus M. quinquenervia would continue to invade and dominate hi gh and low resource plant communities. Tables and Figures Table 2-1. Mean (S.E.) of Melaleuca quinquenervia litterfall biomass a nd nutrient transfer measured in the herbivory and non-herbivory plots. Variable Herbivory Non-herbivory P -----------------------g m-2 6 months-1----------------------M. quinquenervia biomass 11.0 3.79 104 5.35 <0.0001 Carbon 5.11 1.57 47.6 2.31 <0.0001 Nitrogen 0.09 0.03 0.91 0.05 <0.0001 ----------------------mg m-2 6 months-1---------------------Phosphorus 1.33 0.39 22.9 1.85 <0.0001 43

PAGE 44

Table 2-2. Mean (S.E.) of Melaleuca quinquenervia litterfall nutrient con centration and quality measured in the herbivory and non-herbivory plots. Variable Herbivory Non-herbivory P ----------------------------------mg g-1--------------------------------Carbon 511 17.0 458 9.99 0.016 Nitrogen 11.7 0.86 8.75 0.22 0.007 ----------------------------------mg kg-1--------------------------------Phosphorus 225 23.5 219 12.0 0.82 --------------------------------Mass ratio-----------------------------C:N 45 2.68 52 1.17 0.028 N:P 50 3.92 41 2.86 0.11 C:P 2342 246 2170 138 0.57 ------------------------------------%------------------------------------Soluble fiber 56.8 1.31 63.1 2.27 0.03 Hemi-cellulose 6.10 0.98 5.91 0.33 0.86 Resistant 30.7 0.52 26.7 2.01 0.01 Ash 6.41 0.13 6.31 0.29 0.75 Resistant : N 27.3 2.38 28.4 2.38 0.75 Resistant : P 1403 142 1159 97.6 0.21 44

PAGE 45

Table 2-3. Mean ( S.E.) of preand post-fire litter moisture, total biomass, and % litter loss measured in the herbivory and non-herbivory plots. Variable Fire Herbivory Non-herbivory P -----------------------------------%----------------------------------Moisture Pre-Fire 49.1 2.30 47.5 2.68 0.62 Post-Fire 1.74 0.43 2.51 0.53 0.27 ---------------------------------g m-2--------------------------------Biomass Pre-Fire 649 93.2 683 193 0.84 Post-Fire 182 53.8 186 50.3 0.86 ----------------------------------%---------------------------------% Litter loss 72.6 9.17 79.7 7.87 0.80 45

PAGE 46

Table 2-4. Mean ( S.E.) of preand post-fi re soil moisture, bulk density (BD), and organic matter (OM) measured in the he rbivory and non-herbivory plots. Variable Fire Depth Herb ivory Non-herbivory P ---------------------------------% --------------------------------Moisture Pre-Fire 0-5 cm 8.36 0.56 9.44 0.32 0.11 5-15 cm 5.61 0.09 5.89 0.18 0.17 Post-Fire 0-5 cm 1.36 0.24 1.94 0.48 0.30 5-15 cm 3.07 0.17 2.80 0.22` 0.34 ------------------------------g cm-3-----------------------------BD Pre-Fire 0-5 cm 1.20 0.05 1.02 0.06 0.035 5-15 cm 1.42 0.01 1.33 0.03 0.018 Post-Fire 0-5 cm 1.08 0.04 1.00 0.05 0.22 5-15 cm 1.50 0.03 1.45 0.03 0.31 ----------------------------------%-------------------------------OM Pre-Fire 0-5 cm 1.39 0.14 1.96 0.31 0.06 5-15 cm 0.51 0.05 0.55 0.07 0.32 Post-Fire 0-5 cm 2.23 0.30 3.35 0.51 0.06 5-15 cm 0.94 0.06 1.04 0.06 0.26 46

PAGE 47

Table 2-5. Mean ( S.E.) of preand post-fire li tter and soil nutrient con centration measured in the herbivory and non-herbivory plots. Variable Fire Depth Herbivory Non-herbivory P ----------------------------------mg g-1--------------------------------Carbon Pre-Fire Litter 461 25.1 410 7.08 0.07 0-5 cm 6.23 0.83 9.01 1.61 0.07 5-15 cm 1.36 0.17 1.36 0.23 0.99 Post-Fire Litter 240 25.1 290 38.3 0.40 0-5 cm 9.42 1.53 11.0 0.92 0.16 5-15 cm 2.63 0.27 3.42 0.31 0.07 Nitrogen Pre-Fire Litter 9.50 0.53 8.81 0.52 0.36 0-5 cm 0.37 0.04 0.49 0.09 0.27 5-15 cm 0.14 0.01 0.17 0.02 0.39 Post-Fire Litter 10.2 1.75 12.0 1.28 0.36 0-5 cm 0.56 0.10 0.85 0.13 0.06 5-15 cm 0.17 0.03 0.19 0.02 0.48 ----------------------------------mg kg-1---------------------------------Phosphorus Pre-Fire Litter 152 9.85 208 16.3 0.008 0-5 cm 10.9 0.80 14.1 0.73 0.007 5-15 cm 6.21 0.17 7.84 0.56 0.02 Post-Fire Litter 318 53.1 526 35.3 0.005 0-5 cm 12.5 1.10 31.0 4.24 0.0001 5-15 cm 10.3 1.35 9.59 0.81 0.89 47

PAGE 48

Table 2-6. Mean ( S.E.) of preand post-fire of litter nutrient storage measured in the herbivory and non-herbivory plots. Variable Fire Herbivory Non-herbivory P ----------------------------------g m-2---------------------------------Carbon Pre-Fire 319 50.8 295 83.0 0.81 Post-Fire 59.8 20.8 71.7 19.1 0.68 Nitrogen Pre-Fire 6.21 0.92 6.16 1.71 0.98 Post-Fire 2.65 0.95 2.93 0.72 0.82 ---------------------------------mg m-2--------------------------------Phosphorus Pre-Fire 96.3 14.6 149 45.4 0.28 Post-Fire 85.2 31.7 123 30.8 0.41 48

PAGE 49

Table 2-7. Mean ( S.E.) of preand post-fire specifically mineralizab le nitrogen (SMN) and specifically mineralizable phosphorus (SMP) levels measured in the herbivory and non-herbivory plots. Variable Fire Soil Depth Herbivory Non-herbivory P -----------------------------g PMN mg-1 N soil ----------------------------SMN Pre-Fire 0-5 cm 15.3 1.88 18.6 2.07 0.26 5-15 cm 6.99 2.85 4.07 1.49 0.38 Post-Fire 0-5 cm 39.7 4.26 36.1 6.99 0.46 5-15 cm 57.1 10.6 31.3 5.56 0.05 ------------------------------g PMP mg-1 P soil ----------------------------SMP Pre-Fire 0-5 cm 15.8 7.51 3.53 0.46 0.2 5-15 cm 12.5 3.38 4.16 2.06 0.09 Post-Fire 0-5 cm 12.5 6.32 8.17 1.63 0.55 5-15 cm 11.2 5.27 6.33 0.65 0.42 49

PAGE 50

Table 2-8. Mean ( S.E.) of preand post-fire soil microbial biomass carbon measured in the herbivory and non-herbivory plots. Variable Fire Soil Depth Herbi vory Non-herbivory P -----------------------------mg MBC kg-1-----------------------------MBC Pre-Fire 0-5 cm 720 121 1016 189 0.21 5-15 cm 215 46.7 392 67.1 0.03 Post-Fire 0-5 cm 110 15.6 132 18.5 0.39 5-15 cm 124 31.7 106 19.1 0.86 50

PAGE 51

U.S. Interstate 75 City of Naples Figure 2-1. Maps of the study site in southwest Florida (large photograph credit South Florida Water Managem e nt District, inset photograph credit Google Earth). 51

PAGE 52

0 20 40 60 80 100Litterfall Biomass g m-2 Herbivory Non-herbivory O N DJF M Total Litterfall Biomass (g m-2) Month Figure 2-2. Mean ( S.E.) of total litterfall measured during each sampling period in the herbivory and non-herbivory plots. 52

PAGE 53

R2 = 0.6169 R2 = 0.5651 0 1 2 3 4 5 6 7 8 600800100012001400160018002000 Resistant : PTurnover Time (years) Herbivory Non-herbivory Figure 2-3. Relationship between turnover time of litter and resistant:P ratio of Melaleuca quinquenervia litterfall in the non-herbi vory and herbivory plots. 53

PAGE 54

0 100 200 300 400 500 600 700 800 PrePostPrePost Herbivory Non-herbivory Litterpool Biomass g m-2 Litterpool Biomass (g m-2) Oi Oe Oa WoodA a A b Figure 2-4. Mean ( S.E.) of preand post-fi re litterpool biomass (woody, undecomposed Oi, moderately decomposed Oe, and humified Oa) in the herbivory and non-herbivory plots (lower and capital letters indicate significant differences for each analysis). 54

PAGE 55

0 250 500 750 1000 1250 1500 1750 2000 Pre Post Pre Post Litter g 0-5 soil g 5-15 soil gA a A b Carbon Storage (g m-2) Herbivory Non-herbivory Figure 2-5. Mean ( S.E.) of preand post-fire total carbon storage measured in the herbivory and non-herbivory plots (lower and capital letters indicate significant differences for each analysis). 55

PAGE 56

0 20 40 60 80 100 Pre Post Pre Post Litter g 0-5 soil g 5-15 soil gA a A b Nitrogen Storage (g m-2) Herbivory Non-herbivory Figure 2-6. Mean ( S.E.) of preand post-fire total nitrogen st orage measured in the herbivory and non-herbivory plots (lower and capital letters indicate significant differences for each analysis). 56

PAGE 57

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Pre Post Pre Post Litter g 0-5 soil g 5-15 soil gA a A b Phosphorus Storage (g m-2) Herbivory Non-herbivory Figure 2-7. Mean ( S.E.) of preand post-fi re total phosphorus storage measured in the herbivory and non-herbivory plots (lower and capital letters indicate significant differences for each analysis). 57

PAGE 58

Microbial population size and activity Litterfall quantity and quality Nutrient turnover and u p take Biomass p roduction + + + + A Herbivory Microbial population size and activity Litterfall quantity and quality Nutrient turnover and u p take Biomass p roduction +B Figure 2-8. Conceptual models of the feedback cycle of Melaleuca quinquenervia biomass production in the two treatment plots. A) A positive feedback cycle without herbivory (solid line). B) A negative fee dback cycle with herbivory (dashed line). 58

PAGE 59

CHAPTER 3 ASSESSING THE IMPACT OF NATIVE DISTURBANCE REGIMES IN FORESTS MANAGED TO CONTROL THE INVA SION OF AN EXOTIC TREE Introduction The invasion of exotic species alters the ba sic structure of native plant communities with unpredictable consequences for ecosystem func tions (Mack and D'Antonio 1998, Mack et al. 2000). Studies have found both positive and negativ e changes in the rates of nutrient storage and cycling of invaded ecosystems, two crucial ecosystem services (Ehrenfeld 2003). For example, grassland invasion by woody plants can increas e the storage of carbon in standing biomass (Jackson et al. 2002). Exotic plan ts may also differ in the rela tive nutrient concentration and decomposability of thei r litter, so called litte r quality (Ehrenfeld 2003). A sample of 30 invasive species from Hawaii had higher foliar nutrient levels compared to native plants, potentially altering the rate of ecosystem nutrient fluxes (Baruch and Goldstein 1999). Federal, state, and local governments have cr eated comprehensive weed control programs to reduce the populations of exotic plants. The most common method of controlling exotic plants is by using herbicides. Often, applications must be made in perpetuity to maintain satisfactory control and severa l programs have touted the maintenance control concept whereby annual costs are reduced by maintaining popul ations at relative ly low densities by regular and consistent herbicide applications (Ramey and Hassell 2005). However, herbicides can injure adjacent native vegetation and, despite the maintenance control concept, require large, repeated investments of re sources. Mechanical control, or the physical removal of exotic populations, is costly and time consuming thereby limiting its effectiveness for large scale efforts. This approach can al so have significant negative collat eral effects on local flora and fauna. 59

PAGE 60

Classical biological control attempts to re unite weeds with their coevolved natural enemies, most of which are insects (Center and Hill 2002). Although there have been some notable successes with this self -sustaining method, progress can be slow, taking up to a decade to implement. In addition, this method can be in compatible with current herbicide management practices by preventing the establishment and bu ilding up of bio-control agents because of dramatic and unpredictable declines in their food supply (Center et al. 1999). While the goal of all of these approaches is to reduce exotic populations and restore ecosystem integrity, little work has been done to monitor and evaluate their impact on ecosystem function. Perhaps one of the most successful integrated pest management projects to date has been the effort to control Melaleuca quinquenervia (Cav.) Blake in the Florida Everglades. This Australian tree was introduced into South Florida in 1886 and has since co lonized and thrived in most natural areas of South Fl orida, including bayh ead tree islands, sawgrass prairies, pine flatwoods, pastures, and cypress forests (Bodel et al. 1994, Dray et al. 2006). A high growth rate, early reproductive maturity, multiple annu al flowering periods, and serotiny increase the population potential of this plan t following perturbations like fire s (Bodel et al. 1994). The high concentrations of essential oils found in mature M. quinquenervia trees can fuel more destructive canopy fires that can kill native vegetation whic h are adapted to cooler ground fires. Simultaneous and massive releases of M. quinquenervia seed from the canopy seed bank occur after fires, invariably leading to the creation of M. quinquenervia monocultures in many areas (Serbesoff-King 2003). Inter-agency cooperation promoted the integrat ion of chemical, mechanical, and biological control methods to reduce the impact of M. quinquenervia (Ferriter et al. 2 005). The biological control component began in 1986 with the United States Department of Agriculture, Agricultural 60

PAGE 61

Research Service (ARS) taking th e lead. Research conducted at the USDA-ARS Invasive Plant Research Lab in Fort Lauderdale le d to the release of four biologi cal control agents (Ferriter et al. 2005). Two of the biological agents, Oxyops vitiosa Pascoe (Coleoptera: Curculionidae) and Boreioglycaspis melaleucae Moore (Hemiptera: Psyllidae), have successfully established and are suppressing M. quinquenervia reproduction, growth, and recr uitment on a landscape scale (Tipping et al. 2009). The other two species have only recently been released so information on their impact is unavailable. A lthough these programs have been evaluated based on the quantity of plant biomass removed or reduction in rate s of exotic population spread, other evaluation factors need to be considered including the impact on non-target vegetation, recovery of native plant communities, alteration of ecosystem func tion, and interactions with native disturbance regimes like seasonal fires (Denslow and D'Antonio 2005). Many of Floridas natural areas that have been invaded by M. quinquenervia depend on regular disturbances like fires to maintain community structure and function. Native plant communities in these areas depend on seasonal fires to open canopies and trigger the seed release and germination of plant species, provide temporar y pulses of soluble nutr ients, and reduce plant competition (Wade et al. 1980, Neary et al. 1999). The invasion of M. quinquenervia and other exotic plants have altered native fire pattern s which may have long-term consequences for ecosystem function (Wade et al. 1980). While understanding the consequences of fire on aboveground processes is vital, understanding belowground alterations to ecosystem functions is no less important particularly after the inva sion and management of exotic plants. The major objective of this work was to el ucidate aboveand belowground changes to a Taxodium distichum (L.) L.C. Rich var. nutans (Ait.) Sweet dominated ec o-tone forest after the invasion by and subsequent management of M. quinquenervia In particular, we were interested 61

PAGE 62

in the management consequences for nutrient storage and availability before and after a seasonal fire. Two main hypotheses were tested: 1) M. quinquenervia invasion and treatment with an herbicide will reduce the quantity and availability of nutrients before and after a seasonal fire compared to an non-invaded site; 2) M. quinquenervia invasion and treatment with biological control agents will not alter the quantity and availa bility of nutrients before and after a seasonal fire compared to an non-invaded site. Materials and Methods Experimental and Statistical Justification In order to assure the proper assignment of treatment causality in experiments several fundamental assumptions must be met includi ng: random assignment of treatments across experimental units and treatment replication (Beyers 1998). Ra ndom assignment of treatments reduces the impact of non-treatment factors that could confound results a nd replication reduces the likelihood that random variati on is the cause of measured differences between treatments. Although the most effective field studies have randomly assigned, repl icated experimental treatments many times limitations of money, labor, and time make these conditions impossible. Without a proper experimental design, the use of in ferential statistics may only reveal differences between un-replicated plots and not the desired tr eatment effect (Hurlbert 1984). In other words, the null hypothesis becomes that there is no diffe rence between plots NOT that the treatment has no effect (Hurlbert 1984). In ecosystems all over the world land manage rs are manipulating natural areas in an attempt to restore function, provide habitat, or mitigate anthropogenic disturbance. Often times these treatments are done on one large tract of land or single e xperimental unit. Scientific analysis of these areas can be complex because both of the assumptions mentioned above are violated. Still large-scale field studies investigating these treatm ents must be done in order to 62

PAGE 63

assess the impact of natural resource management d ecisions. There is nothing to be gained from limiting scientific investigation when manageme nt must continue. Insight gained from determining large-scale treatment effects can be used to guide further research and prevent undesirable consequences of management decisions. Site Description The study site was located in the Belle Meade Tract of the Picayune Strand State Forest in Collier County, Florida (Figure 2-1). This area consists of nearly level, poorly drained, low fertility soils which are loamy, siliceous, hypertherm ic Arenic Glassoqualfs. The soil series is Pineda-Boca-Hallandale which is characterized by moderately to poorly drained sands which overly limestone bedrock to a depth of approximately 1.4 m (USDA 1998). The water table fluctuates annually between greater than 15 cm below the soil surface to approximately 25 cm above. The area has a distinct wet season from about July to December and a dry season from January to June. Average annual rainfall in th is region is approxima tely 1.36 m (SERC 2007). Historically, the vegetation in this area was a mixed T. distichum-Pinus elliottii Englem forest with a hardwood under-story. Over the past seve ral decades, many areas in this landscape have been invaded with M. quinquenervia and are now comprised of sp arse populations of mature trees with dense understories of seedlings and saplings that can exceed densities of 100 plants per square meter. On March 27, 2007, twenty-five, 1 m2 plots were established along five transects in each of three contiguous study areas (Figure 3-1). The areas sample d were: area #1) reproductive M. quinquenervia treated with herbicide in the summer of 2003 (hereafter referred to as herbicide site), area #2) reproductive M. quinquenervia treated with biological control (hereafter referred to as biological site), and area #3) native forest with no M. quinquenervia (hereafter referred to as non-invaded site). The two biologi cal control agents mentioned above, O. vitiosa and B. 63

PAGE 64

melaleucae, were common in all areas. The herb icide site consisted of reproductive M. quinquenervia that was treated aerially with Velpar (Hexazinone, 3-cyclohexyl-6(dimethylamino)-1-methyl-1,3,5-tr iazine-2,4(1H,3H)-dione) in 2003. The biological site consisted of reproductive M. quinquenervia which was not treated with herbicides and has been under attack by O. vitiosa since 1998 and B. melaleucae since 2002. The non-invaded site was a P. elliottii T. distichum forest with no history of M. quinquenervia invasion. Plots were revisited and all parameters re-sampled after a seasonal fire on May 15, 2007. In early May of 2007 the Great Basal fire burned approximately 8, 000 hectares in southwest Florida and all of the established experimental plots. As the fire was not planned no direct measurements of the intensity of the fire were taken. However, all of the plots were equally affected as well as the entire surrounding landscape. The aboveand belowground samples were taken 24 to 48 hours after the fire. Litter and Soil Sampling The litterpool was sampled in every m2 plot (n=25 in each study site) by placing two 0.1 m2 frames on the surface of the soil and collecting all of the organic material therein. Pre-fire litterpool samples were separated into undecompos ed Oi, moderately decomposed Oe, humified Oa, and woody biomass layers. Post-fire litter pool samples were also separated into an additional after layer which sampled leaves that fell from the tree after the fire. Litterpool samples were air dried to a cons tant weight and reporte d on a dry weight basi s. Fire intensity was estimated by calculating the percentage of th e total litterpool that was lost after the May 2007 fire. One 5 cm diameter soil core was taken in each plot and separated into 0-5 cm and 515 cm depths. Soil samples were returned to th e laboratory, sieved to remove roots and large plant debris, homogenized, and kept at 4C for a maximum of 10 days before microbial analysis. 64

PAGE 65

Litter Quality Analysis Litter quality was measured with a sequen tial extraction using an Ankom A200 Fiber Analyzer (Rowland and Roberts 1994). 0.5 g of co arsely ground litter material was weighed and sealed into Ankom filter bags. The bags were extracted with a neutral detergent to remove soluble cellular contents (sugars, carbohydrates, lipids, etc.) followed by an acid detergent to removal hemi-cellulose. Cellulose was remove d by soaking the bags in 72% sulfuric acid (H2SO4) for 3 hours. The residual (lignin and ash) was combusted at 550C for 4 hours to determine ash content. Litter quality was calculated on a dry mass basis. Soil Characteristics Percent moisture and bulk density of soils were determined by drying 20-30 g sub-samples of field-moist, sieved, and homogenized soil at 70C for three days. Bulk density and percent moisture were determined on a wet soil weight basis and pH was measur ed on a 2:1 water:soil slurry with an Accumet Research, AR50 dual ch annel pH/ion/conductivity meter. Soil organic matter was measured by loss on ignition from 0.2 to 0.5 g samples of dried and ground soils, which were first measured into 50 mL beakers (L uczak et al. 1997). The beakers were placed in a muffle furnace and brought to 250C for 30 minutes. The furnace temperature was then increased to 550C for 4 hours. Organic matte r content was calculated as the mass loss on ignition on a dry weight basis. Nutrient Analyses Dried and ground soil and plant material were analyzed for percent carbon and nitrogen on a Thermo-Electron, 1112 Series elemental analyz er. Total phosphorus was determined by a twophase acid extraction after lo ss on ignition (Andersen 1976). Th e ash remaining in the 50 mL beaker was moistened with 2 to 3 mL distilled de-ionized water and then extracted with 20 mL of 6 N hydrochloric acid (HCl). All of the water was removed and the hot plate was placed on 65

PAGE 66

high for 30 minutes to completely dry the samp les. After cooling, 2.25 mL of 6 N HCl was added to each beaker and the beakers placed on a hot plate until almost boiling. Extracts were then filtered through a #41 Whatma n filter into 50 mL volumetric flasks that were then brought to volume with distilled de-ionized water. Total phosphorus was measured with an automated ascorbic acid method on a Bran and Luebbe Auto Analyzer 3, Digital Colorimeter (Method 365.4; USEPA 1993). Microbial Biomass Microbial biomass carbon (MBC) was measured by a modified chloroform-fumigation extraction method (Vance et al. 1987). Two repl icates of 1 g field-moist soil samples were weighed into 50 mL centrifuge tubes. One of the duplicates was immediately extracted with 25 mL of 0.5 M potassium sulfate (K2SO4), shaken for 1 hour, and then filtered through a Whatman #41 filter. The second replicate of soil underw ent chloroform fumigation where 0.5 mL of chloroform was added to each tube and the tubes were placed in a dessicator with a beaker containing 30 mL of chloroform and boiling chip s. The dessicator was then vacuum-sealed for 24 hours then alternatively filled with air a nd evacuated ten times to remove all residual chloroform. The tubes were extracted as descri bed above and all extracts were stored at 4C until analysis for MBC. This was calculated as the difference between the total carbon in the fumigated and un-fumigated soil extracts as measured on a Shimadzu TOC 5050C, total organic carbon analyzer. Nutrient Availability Potentially mineralizable nitrogen (PMN) wa s measured with a modified incubationextraction method (White and Reddy 2000) that ca lculates a mineralizat ion potential based on the net production of ammonium during a 10 day an aerobic incubation. In order to determine the initial amount of ammonium in the soil, a set of 1 g field-moist soil samples were weighed into 66

PAGE 67

50 mL centrifuge tubes. The tubes received 25 mL of 0.5 M K2SO4, shaken on a longitudinal shaker for one hour, and extracts were filtered through a #41 Whatman filter. An additional set of 1 g field-moist soil samples were weighed in to 50 mL serum bottles with 5 mL of deionized water. These bottles were sealed with butyl rubber stoppers and alum inum crimps and their head-space was purged with oxygen (O2)-free nitrogen (N2) gas for 2-5 minutes. Bottles were incubated in the dark at 35C for 10 days then extracted as described a bove with 25 mL of 0.5 M K2SO4 to determine final ammonium concentration. Total ammonium was determined with an automated colorimetric method on a Bran and Lu ebbe Auto Analyzer 3, Digital Colorimeter (Method 353.2; USEPA 1993). Specifically mine ralizable nitrogen (SMN) was determined by dividing PMN by the concentration on nitr ogen in the soil (SMN = PMN TN-1). Potentially mineralizable phosphorus (PMP) wa s measured with a modified incubationextraction method under aerobic conditions (Grierson et al. 1999). Soils were incubated at a temperature of 35C for a period of 10 days in capped 125 mL polyethylene bottles that contained 5 g of soil held under optimal percent mo isture conditions (4-7%). Soils were aerated every three days and moisture content was adju sted if necessary. On day 10, soil sub-samples were collected from each incubation bottle and analyzed for available phosphorus. Soil samples were extracted with 0.1 M potassium chloride (K Cl) at a ratio of 10:1, shaken for one hour, and then filtered with 0.45 m me mbrane filters. Available phosphorus was measured with an ascorbic acid method on a Shimadzu Spect rophotometer UV-160 (Method 365.4; USEPA 1993). Specifically mineralizable phosphorus (SMP ) was determined by dividing PMP by the concentration on phosphorus in the soil (SMP = PMP TP-1). Statistical Analyses Values of measured soil and ve getation characteristics, as we ll as microbial population size and activity, were calculated as a mean for each plot. ANOVA and Tukey means separation tests 67

PAGE 68

were used to detect any differences among the measured parameters. Diff erences are reported as significant for tests with p values 0.05. Data that varied fr om normal distributions were transformed with square root(x), log(x), or log(x+1) functions. The datasets that were transformed with the square root(x) function were preand post-fire moistu re (0-5 and 5-15 cm), preand post-fire litterp ool biomass and moisture, preand po st-fire nitrogen concentration (0-5 and 5-15 cm), preand post-fire phosphorus concentration (0-5 a nd 5-15 cm), preand post-fire SMN (5-15 cm), preand post-fire nitrogen st orage (0-5 and 5-15 cm), preand post-fire phosphorus storage (5-15 cm), and preand post-fire SMP (5-15 cm). The datasets that were transformed with the log(x+1) f unction were preand post-fire or ganic matter (0-5 and 5-15 cm), preand post-fire carbon concentr ation (0-5 and 5-15 cm), and preand post-fire carbon storage (0-5 and 5-15 cm). The datasets that were tr ansformed with the log(x) function were preand post-fire microbial biomass carbon (0 -5 and 5-15 cm). All statis tical analyses were preformed using JMP 7.0.1 software (SAS In stitute, North Carolina, USA). Results Full model results for the main effects and in teractions are reported in Tables A-11 through A-18. Based on the magnitude of the results the ma in effects of treatment, site, and depth were the most important determinants of the measured response variables. There were no consistent transect or plot effects for any of the measured variables. The litter in the non-invaded site had the highest percent moisture followed by the herbicide and then the biological sites. All site s had lower levels of li tter moisture and nutrient storage after the fire (-91, -96, and -92% for the non-invaded, herbicide, and biological sites, respectively) (Tables 3-1 and 3-4). Litter biom ass was greatest in the non-invaded site, both before and after the fire (Figur e 3-1). Percent biomass loss was greater in the herbicide site 68

PAGE 69

compared to the non-invaded and biological sites (T able 3-1). Preand post-fire values for soil moisture, bulk density, and organic matter are reported in Table 3-2. The litter in the non-invaded si te had the highest concentra tion of carbon before and after the fire (Table 3-3). Among site s the soil in the herbicide site had a lower concentration of carbon, nitrogen, and phosphorus at both depths before and after the fire compared to the noninvaded and biological sites (Tab les 3-3). Conversely, the soil in the biological site had the highest concentration of nitrogen and phosphorus both before and af ter the fire compared to the other two sites (Table 3-3). The concentration of carbon in the 0-5 cm soil depth was lower after the fire within all sites (-20, -55, and -16% in the non-invaded, herbicide, and bi ological sites, respectively) whereas in the 5-15 cm soil depth carbon con centration was lower in the non-invaded and herbicide sites but higher in th e biological site after the fire (-1, -2, and +35%, respectively) (Table 3-3). Within sites, nitrogen concentrati on in the 0-5 cm soil depth was lower after the fire (-16, -30, and -3% in the non-invaded, herbicide, and biolog ical sites after the fire, respectively) but was only lower in the herbicide site while be ing higher in the non-inva ded and biological site in the 5-15 cm soil depth (-13, +15, and +5%, respec tively) (Table 3-3). With sites, phosphorus concentration in the 0-5 cm soil depth soil was higher after the fire (+ 27, +31, and +2% in the non-invaded, herbicide, and biolog ical sites, respectivel y) while in the 5-15 cm soil depth it was lower in the herbicide site but higher in the non-invaded and bi ological sites (-1, +16, and +35% respectively) (Table 3-3). There were no differences in SMN among sites in the 0-5 cm or 5-15 cm soil depths before or after the fire (Table 3-5). However, SMN in the 0-5 cm soil depth was higher in all sites after the fire (+20, +102, and +52% in the non-invaded, herbicide, and bi ological sites after the fire, 69

PAGE 70

respectively) (Table 3-5). SMN was lower in the 5-15 cm soil depth in the herbicide site but higher in the non-invaded and herbicide site after the fire (+23, 6, and +5%, respectively). Pre-fire SMP was highest in the non-invaded site in the 0-5 cm soil depth and in the biological site in the 5-15 cm soil depth (Tab le 3-5). However, post-fire there were no differences among sites in the 0-5 cm soil depth and the herbicide site had th e highest availability of phosphorus at both depths. Within sites, SMP in the 0-5 cm soil depth was higher post-fire in the herbicide and biological sites but lower in the non-inva ded site (+107, +46, and -1%, respectively). SMP in the 5-15 cm soil depth was lo wer in the biological si te but higher in the non-invaded and herbicide sites af ter the fire (-78, +41, and +76% respectively) (Table 3-5). Microbial biomass carbon (MBC) was lower in the 0-5 cm soil depth in the herbicide site compared to the non-invaded and biological sites, both before and after th e fire (Table 3-6). Before the fire, MBC was lower in the 5-15 cm soil depth in the herbicide site compared to the non-invaded and biological sites (Table 3-6). However, MBC was lower in the non-invaded site after the fire. Within sites, MBC at in th e 0-5 cm soil depth was lower in the non-invaded, herbicide, and biological sites af ter the fire (-76, -68, and -78%, respectively). The same was true in the 5-15 cm soil depth after the fire (76, -63, and -20%, respecti vely) (Table 3-6). Before the fire, the non-invaded site contained a larger pool of total carbon compared to the herbicide and biological sites (Figure 3-4). However, after the fire, the biologic al site had the larger pool of total carbon compared to the he rbicide and non-invaded s ites (Figure 3-4). The pool of total nitrogen was largest in the noninvaded site compared to the herbicide and biological sites, both before and af ter the fire (Figure 3-5). Befo re the fire, the non-invaded site contained the largest pool of tota l phosphorus, while after the fire the largest pool occurred in the biological site (Figure 3-6). 70

PAGE 71

Discussion Alteration of Aboveground Components Many natural systems are managed to maximize the delivery of ecosystem services such as space for recreational activitie s, wildlife habitat, and maintenance of biological diversity. However, few studies have evaluated the effects of management programs on the delivery of less obvious ecosystem services like nut rient storage. In this study, the invasion and management of M. quinquenervia caused significant and persistent cha nges in the aboveground litter storage of a sub-tropical wetland forest. Li tter accumulates on the forest floor when organic matter inputs (e.g. litterfall) exceed outputs (e.g. decomposition of organic matter). The main source of the organic matter input to the litter in these forests was from litterfall. Before the release of the two biological control agents, litterfall in a mature M. quinquenervia forest measured 750 to 930 g dry weight m-2 year-1 (Van et al. 2002). In Australia, where M. quinquenervia is attacked by many insect herbivores, litterfall rang ed from 675 to 809 g dry weight m-2 year-1 in seasonally inundated sites (Greenway 1994). Litterfall by M. quinquenervia was not measured in this study although rates in the biological site were presumed to be lower than reported by the previous studies because of the smaller si zed trees. Martin et al. (2009) reported litterfa ll rates of 257.5 g dry weight m-2 year-1 in a T. distichum forest adjacent to the non-invaded study area. The herbicide site had no new litterfall inputs, as all the trees were killed with the treatment and had not yet re-grown. The main source of the organic matter output from the litter in these forests is from litter decomposition. Previous studies have shown that M. quinquenervia has a lower rate of litter decay compared to T. distichum because leaves of M. quinquenervia contain antibacterial phenolics which hinder microbial decompos ition (Greenway 1994). Although no direct measurements of M. quinquenervia decomposition have been publis hed, we re-analyzed litterfall 71

PAGE 72

and litter accumulation data from Greenway (1994) to obtain an estimate of 4.3 years for the turnover time of M. quinquenervia litter. Nessel and Bayley (1984) estimated 3 years for the turnover time of T. distichum litter, the dominant litterfall co mponent in the non-invaded study area. The proposed higher litterfall and lower decomposition rates would suggest that greater amounts of litter would accumulate in the bi ological and non-invaded areas compared to the herbicide area. As hypothesized, there was a reduction in both litter biomass a nd nutrient concentration in the herbicide treated site despite the fact that it was treated four years ago (Figure 3-1). These differences resulted in the herbicide site st oring 10 times less carbon, 10 times less nitrogen, and 9 times less phosphorus in the litter compared to th e non-invaded site. This is most likely due to the relative composition of the litter between the sites where 62% of the litter in the herbicide site consisted of relatively low-nutrient wood material compared to only 13% in the non-invaded and 26% in the biological sites (p = 0.02) (Figure 3-2). However, in contrast to our hypotheses, the biological site also contained less litter bi omass and stored less nutrients compared to the non-invaded site (2 times less carbon, nitrog en, and phosphorus). We suggest two possible explanations for the observed differences: 1) the biological control agents may be altering quality of nutrient inputs to the biological site accelerating the decomposition of the M. quinquenervia litter and/or 2) another low quality litterfall component in the n on-invaded area is slowing litter decomposition rates. Franks et al. (2006) found th at higher densities of the biological control agent B. melaleucae caused M. quinquenervia seedlings to drop leaves prematurely, which could interrupt the natural cycle of nutrient reso rption during senescence and result in higher quality leaf litter inputs. In addition, the frass produced by both biological agents may be rich in nitrogen and 72

PAGE 73

phosphorus thereby potentially short-circuiting the nutrient limitations and accelerating litter decomposition (Frost and Hunter 2007). The second possibility is th at the accumulation of P elliottii needles on the forest floor in the non-inva ded area is slowing the decomposition of the litter. Martin et al. (2009) f ound that 7% of the litterfall in the adjacent non-invaded area was P elliottii needles. P elliottii needles decompose at a much slower rate that T. distichum and may accumulate on the forest floor (Gholz et al. 1985). This could increase total biomass inputs while slowing the deco mposition rate of the T. distichum litter, resulting in greater litter accumulation than expected. All of the sites lost a significant amount of litter biomass and aboveground nutrient storage after the fire (Tables 3-1 and 3-4). However, th e herbicide site lost more of the total biomass compared to the other two sites (p = 0.001). This may indicate that the intensity of the fire was greatest in the herbicide site. Unfortunately, as this fire was not planned, there were no intensity measurements taken during the burn. The herbicide site also lost the largest percentage of the litter carbon (95%, p = 0.27), nitrogen (88%, p = 0.21), and phos phorus (86%, p = 0.03) pools (Table 3-1). Other studies have reported similar losses of aboveground nitrogen after fires (Kauffman et al. 1993, Hughes et al. 2000, Wan et al. 2001, Wanthongchai et al. 2008). However, phosphorus losses in the herbicide si te were higher than previously reported (Kauffman et al. 1993, Hughes et al. 2000, Wanthongc hai et al. 2008). Nutrient losses from aboveground storages after fire can be cause d by volatilization and pa rticulate transport, transformation of organic to inorganic forms, and transport by wind and water (Neary et al. 1999, Certini 2005). Before the fire the landscape in the herbicide site consisted of sparsely distributed herbaceous plants and large amounts of dry woody debris on the soil surface. The fire reduced woody debris by 94% and killed mo st of the standing vegetation, leaving bare 73

PAGE 74

mineral soil that was extremely vulnerable to wind transport. While some of the litter phosphorus was incorporated into the mineral soil after the fire, it is possible that some was blown off site by wind (Neary et al 1999, Certini 2005). Alteration of Belowground Components The invasion and treatment of M. quinquenervia significantly altered th e biotic and abiotic components of the surface soils both before and after the fire. Befo re the fire, the herbicide site had significantly less moisture at 0-5 cm compared to the biologi cal and non-invaded sites. This may be explained by the litter composition in the herbicide site which, unlike the mostly undecomposed and moderately decomposed leav es in the non-invaded and biological sites, consisted primarily of woody materi al that left large portions of the mineral soil exposed to wind and solar radiation (Figure 3-2). Post-fire the herbicide site had lower levels of soil moisture at 0-5 cm compared to the non-native and biological sites. This ma y be the result of lower initial soil moisture and increased fire intensity in the herbicide site. The herbicide site also had the lowest st orages of carbon, nitrogen, and phosphorus preand post-fire at both soil depths (Figure 3-4, 3-5, and 3-6). The largest differences were recorded in the storages of nitrogen and phosphorus at 05 cm before the fire where the biological and non-invaded sites stored approxi mately 1.6 and 2 times more nitrogen, respectively, and 1.5 times more phosphorus than the herbicide site. Fi ndlay et al. (2003) noted a similar pattern with the invasive reed Phragmites australis whose removal from a marsh by herbicides and cutting disrupted the systems ability to process and st ore nitrogen. Similarly, the chemical treatment and cutting of the invasive plant Pteridium aquilinum led to a loss of ecosystem carbon storage in an European moorland (Marrs et al. 2007). In this study, despite a trend of increased phosphorus storage after the fire in every site and soil depth, th e only significant increase was found in the herbicide site at 0-5 cm. As discu ssed above the high intensity fire in this site 74

PAGE 75

probably caused the significantly higher loss of th e litter phosphorus. It is likely that the phosphorus rich ash produced from the combustion of the litter was incorporated into the 0-5 cm soil depth by wind events and microbial communiti es (Neary et al. 1999, Carter and Foster 2004, Certini 2005). Nitrogen availability increased in all sites after the fire in every soil depth (Table 3-5). Several studies have found similar results with in creasing nitrogen availability after fire (Carter and Foster 2004, Wanthongchai et al. 2008). Giardina and Rhoade s (2001) found that nitrogen availability increased significantl y after a prescribed fire in a pi ne forest. In contrast, although the availability of phosphorus was c onsistently lower at 0-5 cm in the biological site, the fire did not alter the availability in any site or soil dept h. These changes in the storage and availability of nutrients may have unpredictable effects on ecosys tem structure and function in the future. For example, minimum levels of nitrogen and phosphorus availability are needed for maintenance of soil microbial communities and re-vegetation of nativ e plants after disturbances such as fire. However, any nitrogen or phosphorus produced in excess of these needs may be vulnerable to loss by leaching and erosion. Further study is required to determine the cu rrent requirements of the plant and microbial communities in these sites to ensure that invasive species management strategies do not lead to leaching of nutrie nts with negative consequences for downstream ecosystems. In addition to alterations to nutrient cy cling and storage caused by the invasion and management of exotic species, microbial communities also exhibit changes that may ultimately influence future ecosystem services. Wardle et al. (2004) detailed how plant community structure drives changes in soil microbial co mmunity size and composition via a series of aboveground-belowground feedback loops. In this study, MBC was lower at 0-5 cm in the 75

PAGE 76

herbicide site both before and after the fire (T able 3-6). Nutrient availability did not differ significantly between sites suggesting that envi ronmental factors such as soil moisture and temperature controlled levels of MBC. Although pre-fire levels of MBC were highly correlated with soil moisture, no relationship existed post-fire (Figure 3-3). This lack of a relationship postfire may reflect higher soil temperatures normally found right after a fire that would reduce MBC (Neary et al. 1999, Certini 2005). Prieto-Fernandez et al. (1998) found a significant decrease in the size of the soil microbial biomass after a fire in a pine forest. In this study all sites lost approximately 70% of the microbial biomass in the 0-5 cm soil depth and 65% in the 5-15 cm soil depth after the fire. Change s in microbial community struct ure may significantly affect the ability of these systems to r ecover after the fire. Conclusions Determining the best management practices for exotic species requires consideration of a broad array of factors and their potential interactions, including fu ture interactions with natural events like fires. Most efforts to date ha ve emphasized above ground factors like plant and animal diversity and richness, w ith little to no consideration of below ground factors like nutrient storage, nutrient cycling, and microbial comm unity diversity. This study clearly shows how these foundational ecosystem compon ents were affected by the management of exotics in the backdrop of a natural fire event. Biological control of M. quinquenervia using insect herbivores has proven to be effective at controlling plant growth a nd reproduction (Tipping et al. 2 009). The results of this study suggest that this method had less of an impact on nutrient storage and cycling than herbicides. Additional questions remain including how both methods affect re-vegetation over the longer term. Although herbicides remain a valuable tool in the management of invasive species, more attention needs to be paid to the resulting consequences for ecosystem structure and function. 76

PAGE 77

Practices such as active re-vegetation with native plants may mitigate the deleterious impacts of the treatment and help to prevent future invasion. If evaluations of the below-ground side effects of exotic plant management remain rare, then advancing our understand ing of basic ecosystem structure and restoration will suffer and any alterations to foundational communities like microbes may permanently and negatively alter ecosystem function. Tables and Figures Table 3-1. Mean ( S.E.) of preand post-fire li tter moisture, biomass, and fire intensity in the non-invaded, herbicide, and biological site s (lower case letters indicate significant differences for each analysis). Variable Fire Non-invaded Herbicide Biological P ---------------------------------% --------------------------------Moisture Pre-fire 38.7 1.40 a 25.8 11.4 b 17.1 6.69 c <0.0001 Post-fire 3.33 0.40 a 1.15 0.53 b 1.42 0.15 b <0.0001 --------------------------------kg m-2 ------------------------------Biomass Pre-fire 2.24 0.29 a 0.71 0.08 c 1.36 0.13 b <0.0001 Post-fire 0.35 0.04 a 0.12 0.04 c 0.18 0.03 b <0.0001 ---------------------------------% --------------------------------Litter Loss 81.5 2.27 b 93.7 2.15 a 85.9 2.24 b 0.001 77

PAGE 78

Table 3-2. Mean ( S.E.) of preand post-fire fire soil moisture, bulk density (BD), and organic matter (OM) in the non-invaded, herbicide, and biological sites (lower case letters indicate significant differences for each analysis). Variable Fire Depth Non-invaded Herbicide Biological P ---------------------------------% --------------------------------Moisture Pre-fire 0-5 11.2 1.17 a 7.05 0.65 b 11.3 1.64 a 0.048 5-15 7.86 0.46 a 6.07 0.17 b 5.48 0.37 b <0.0001 Post-fire 0-5 1.37 0.25 a 0.29 0.06 b 1.17 0.20 a <0.0001 5-15 3.96 0.25 a 2.54 0.13 b 3.17 0.21 b <0.0001 ---------------------------------g cm-3--------------------------------BD Pre-fire 0-5 1.11 0.06 a 1.20 0.05 a 0.89 0.05 b 0.0004 5-15 1.30 0.05 b 1.43 0.16 a 1.35 0.03 ab 0.017 Post-fire 0-5 1.13 0.60 a 1.19 0.03 a 0.88 0.03 b <0.0001 5-15 1.49 0.03 a 1.60 0.02 b 1.49 0.03 a 0.004 ---------------------------------% --------------------------------OM Pre-fire 0-5 1.87 0.22 b 1.70 0.17 b 3.71 0.49 a 0.0001 5-15 0.95 0.09 a 0.52 0.04 b 0.82 0.07 a 0.0001 Post-fire 0-5 2.46 0.32 b 1.07 0.04 c 3.49 0.37 a <0.0001 5-15 1.38 0.14 a 0.68 0.05 b 1.05 0.10 a <0.0001 78

PAGE 79

Table 3-3. Mean ( S.E.) of preand post-fire fire nutrient concentrations in the non-invaded, herbicide, and biological site s (lower case letters indicate significant differences for each analysis). Variable Fire De pth Non-invaded Herbicide Biological P ---------------------------------mg g-1--------------------------------Carbon Pre-fire Litter 470 9.79 a 460 8.82 a 421 8.82 b 0.008 0-5 12.9 1.80 8.45 1.06 16.8 2.82 0.08 5-15 3.77 0.61 a 1.74 0.21 b 1.92 0.2 b 0.003 Post-fire Litter 273 29.6 203 26.7 300 26.8 0.07 0-5 10.3 1.56 a 3.78 0.33 b 14.1 1.79 a <0.001 5-15 3.73 0.55 a 1.70 0.16 b 2.59 0.34 a 0.008 Nitrogen Pre-fire Litter 9.45 0.29 b 7.98 0.40 c 11.1 0.28 a <0.001 0-5 0.93 0.14 a 0.43 0.07 b 0.96 0.19 a 0.01 5-15 0.33 0.05 a 0.15 0.02 b 0.19 0.03 b 0.002 Post-fire Litter 12.3 0.75 a 8.94 1.19 b 13.2 0.96 ab 0.01 0-5 0.78 0.13 a 0.30 0.04 b 0.93 0.13 a <0.001 5-15 0.38 0.06 a 0.13 0.01 b 0.20 0.02 b <0.001 ---------------------------------mg kg-1--------------------------------Phosphorus Pre-fire Litter 135 6.33 b 141 8.83 b 197 6.66 a <0.001 0-5 15.8 1.98 ab 10.3 0.98 b 23.5 3.09 a 0.004 5-15 9.03 0.83 7.26 0.52 8.64 0.65 0.21 Post-fire Litter 296 5.54 b 280 37.1 b 387 24.3 a 0.006 0-5 20.1 2.64 ab 13.4 1.55 b 23.9 2.49 a 0.006 5-15 10.5 1.03 ab 7.18 0.50 b 11.7 1.36 a 0.01 79

PAGE 80

Table 3-4. Mean ( S.E.) of preand post-fi re litter nutrient storag es in the non-invaded, herbicide, and biological site s (lower case letters indicate significant differences for each analysis). Variable Fire Non-invaded Herbicide Biological P --------------------------------g m-2 -------------------------------Carbon Pre-fire 1076 125 a 367 52.4 c 592 60.4 b <0.0001 Post-fire 65.0 10.5 33.6 65.9 40.4 7.89 0.12 Nitrogen Pre-fire 22.5 2.96 a 6.15 1.03 c 15.0 1.40 b <0.0001 Post-fire 3.35 0.46 a 1.41 0.58 b 1.74 0.31 b 0.01 ---------------------------------mg m-2 -----------------------------Phosphorus Pre-fire 322 26.4 a 96.9 13.4 b 260 26.4 a <0.0001 Post-fire 77.0 9.15 46.6 18.7 53.2 9.36 0.26 80

PAGE 81

Table 3-5. Mean ( S.E.) of preand post-fire specifically mineralizab le nitrogen (SMN) and specifically mineralizable phosphorus (SMP) levels measured in the non-invaded, herbicide, and biological site s (lower case letters indicate significant differences for each analysis). Variable Fire Depth Native Herbicide Biological P ------------------------g PMN mg-1 N soil ----------------------SMN Pre-fire 0-5 18.7 3.35 14.8 1.67 18.1 4.01 0.66 5-15 12.4 2.45 12.6 2.76 16.0 2.41 0.55 Post-fire 0-5 22.5 3.53 29.8 4.01 27.6 3.04 0.33 5-15 15.2 3.11 11.9 2.24 16.9 3.07 0.45 -------------------------g PMP mg-1 P soil -----------------------SMP Pre-fire 0-5 16.3 2.14 a 8.38 4.31 ab 4.49 1.01 b 0.03 5-15 3.41 0.99 8.69 2.89 12.6 5.88 0.31 Post-fire 0-5 16.1 4.81 17.3 3.50 6.54 0.82 0.09 5-15 4.82 2.14 a 15.3 4.56 b 2.73 0.69 a 0.04 81

PAGE 82

Table 3-6. Mean ( S.E.) of preand post-fire microbial biomass carbon levels measured in the non-invaded, herbicide, and biological site s (lower case letters indicate significant differences for each analysis). Variable Fire Depth Native Herbicide Biological P ----------------------------------mg C kg-1----------------------------------MBC Pre-fire 0-5 785 116 330 46.4 791 150 0.08 5-15 408 45.4 a 295 29.5 a 181 34.3 b 0.001 Post-fire 0-5 186 20.4 a 104 9.75 b 174 19.5 a 0.0018 5-15 99.3 10.7 108 18.5 144 24.92 0.495 82

PAGE 83

U.S. Inter s ta te 7 5 City of Naples N C B Figure 3-1. Map of the study site in southwes t Florida showing the (N) non-invaded, (C) chem ically controlled, an d (B ) biologically cont rolled sites. Diagram not to scale (photograph credit Google Earth). 83

PAGE 84

0 500 1000 1500 2000 2500 3000 PrePostPrePostPrePost Non-invadedHerbicideBiological Oi Oe Litterpool Biomass g m-2 Oa After WoodA B B a b c Litterpool Biomass (g m-2) Figure 3-2. Mean ( S.E.) of preand post-fi re litterpool biomass by components (after, wood, undecomposed Oi, moderately decomposed Oe, and humified Oa) measured in the non-invaded, herbicide, and bi ological sites (different cap ital and lower case letters indicate significant differences for pr eand post-fire means, respectively). 84

PAGE 85

R2 = 0.8907 0 500 1000 1500 2000 2500 01020 Figure 3-3. Relationship between mi crobial biomass carbon (MBC) levels and % soil moisture in the three treatment sites. A) The rela tionship in the non-invaded site. B) The relationship in the herbicide site. C) The relationship in the biological site. 30 R2 = 0.8168 0 500 1000 1500 2000 2500 3000 01020304 0 R2 = 0.5189 0 200 400 600 800 1000 05Moisture (%) 1 0A B C MBC g g-1 Microbial Biomass Carbon (mg kg-1 soil) 85

PAGE 86

0 500 1000 1500 2000 2500 PrePostPrePostPrePost Non-invadedHerbicide Biological Litter 0-5 soil A Carbon Storage (g m-2) Carbon Storage (g m-2) 5-15 soil B b B a a Figure 3-4. Mean ( S.E.) of preand post-fire total storages of car bon measured in the noninvaded, herbicide, and biological sites (d ifferent capital and lower case letters indicate significant differences for pr eand post-fire means, respectively). 86

PAGE 87

0 20 40 60 80 100 120 PrePostPrePostPrePost Litter 0-5 soil Non-invadedHerbicide Biological Nitrogen Storage (g m-2) 5-15 soilA Nitrogen Storage (g m-2) B B a b a Figure 3-5. Mean ( S.E.) of preand post-fire storages of nitr ogen measured in the non-invaded, herbicide, and biological site s (different lower and capital letters indicate significant differences for each analysis). 87

PAGE 88

0 0.5 1 1.5 2 2.5 3 3.5 PrePostPrePostPrePost Litter 0-5 soil 5-15 soilA A A ab b a Phosphorus Stor age (g m-2) Phosphorus Storage (g m-2) Non-invaded Herbicide Biological Figure 3-6. Mean ( S.E.) of preand post-fi re storages of phosphorus measured in the noninvaded, herbicide, and biological sites (d ifferent lower case and capital letters indicate significant differences for each analysis). 88

PAGE 89

CHAPTER 4 COMPARING NATIVE AND EXOTIC PLANT QUALITY: IMPLICATIONS FOR NUTRIENT TURNOVER Introduction Organic matter decomposition and the subsequent release of plant avai lable nutrients is a vital ecosystem process. The decomposition of plant derived organic ma tter occurs in three phases: 1) initial losses due to the leaching of soluble chemical compounds, 2) microbial colonization and degradation, and 3) physical and biological fragme ntation (Chapin et al. 2002). All three phases of d ecomposition are controlled by both biotic and abiotic factors (Berg 2000). For example, in general, plant residues with higher concentrations of nitrogen and phosphorus and lower ratios of resistant materials (e.g. lignin) to soluble fibers (e.g. sugars and carbohydrates) may have faster decomposition ra tes (Berg 2000, Chapin et al. 2002, Mack and D'Antonio 2003). In addition, environmental factor s such as temperature, moisture, and pH can alter the structure and function of microbial communities that process organic matter (Couteaux et al. 1995, Katterer et al. 1998, Chapin et al. 20 02). Several studies have shown that the alteration of organic inputs af ter a disturbance, such as e xotic plant invasions, can alter decomposition rates resulting in changes in the structure and f unction ecosystems (Ehrenfeld 2003, Mack and D'Antonio 2003). Melaleuca quinquenervia (Cav.) Blake, otherwise known as the paper-bark tree, cajeput, punk tree, or white bottlebrush tree, is a me mber of the Myrtaceae family, sub-family Leptospermoidae. This tall evergreen tree hist orically occupies tropic al wetland sites throughout its native range along the eastern coast of Australia (Kaufman and Smouse 2001). It was introduced into South Florida in 1886 (Dray et al. 2006), or iginally for sale as an ornamental, but later was used for erosion cont rol, as a forestry crop, and as an agricultural windrow plant (Meskimen 1962, Stocker and Sanders Sr. 1981, Bode l et al. 1994). The exotic tree colonized 89

PAGE 90

and thrived in most natural areas of South Fl orida, including bayhead tree islands, sawgrass prairies, pine flatwoods, pa stures, and cypress forest s (Bodel et al. 1994). Myers (1984) described the Pinus elliottii EnglemTaxodium distichum (L.) L.C. Rich var. nutans (Ait.) Sweet ecotone in Sout h Florida. This is the tr ansition zone between upland P. elliottii -dominated sites and depressional T. distichum dominated swamps where both trees codominate but neither grows to its full potential (Myers 1983). This ecotone has also been extensively invaded by M. quinquenervia (Myers 1984). These infestations have been managed using a combination of chemical, mechanical, and biological methods. However, to date, no research has been done on the rate of M. quinquenervia litter decomposition and its effect on nutrient turnover in South Florida ecosystems like this one. A 1999 study suggested that ecosystem invasion by M. quinquenervia alters rates of decomposition and increases the storage of organic material (Anonymous 1999). Melaleuca alternafolia a close relative of M. quinquenervia is known to produce volatile essential oils with antimicrobial properties (Carson et al. 2006). In addition, DiStefano and Fisher (1983) found that extracts of M. quinquenervia leaves reduced fungal infections of plant embryos. The antimicrobial properties of li tter are often cited as a mechanism through which M. quinquenervia dominates ecosystems. The objective of this work, theref ore, was to elucidate temporal changes in the litter quality of M. quinquenervia T. distichum and P. elliottii and quantify their effects on nutrient turnover. Two main hypotheses were tested: 1) M. quinquenervia will have the slowest rate of decomposition; and 2) M. quinquenervia litter will release least am ount of carbon, nitrogen, and phosphorus compared to T. distichum and P. elliottii litter. 90

PAGE 91

Materials and Methods Site Description The study site was located in the Belle Meade Tract of the Picayune Strand State Forest in Collier County, Florida (Figure 2-1). This area consists of nearly level, poorly drained, low fertility soils which are loamy, siliceous, hypertherm ic Arenic Glassoqualfs. The soil series is Pineda-Boca-Hallandale which is characterized by moderately to poorly drained sands which overly limestone bedrock to a depth of approximately 1.4 m (USDA 1998). The water table fluctuates annually between greater than 15 cm below the soil surface to approximately 25 cm above. The area has a distinct wet season from approximately July to December and a dry season from January to June. Average annual ra infall in this region is approximately 1.36 m (SERC 2007). Historically, the ve getation in this area was a mixed T. distichum-Pinus elliottii forest with a hardwood under-story. Over the past several decades, many areas in this landscape have been invaded with M. quinquenervia and are now comprised of sparse populations of mature trees with dense understories of seedlings and saplings that can exceed densities of 100 plants per square meter. Experimental Approach Six by six inch litter bags were constructed us ing 1 mm mesh fibergla ss screen. Leaves or needles that were senescent but st ill attached were collected from M. quinquenervia T. distichum and P. elliottii and air dried to a constant weight Leaves from M. quinquenervia were collected in an area that is dominated by mature M. quinquenervia trees while T. distichum and P. elliottii needles were collected from an adjacent non-inva ded area. Two gram samples of leaves or needles from each species were placed into sepa rate litter bags which were then deployed at three locations within a noninvaded, non-burned, mixed T. distichum-P. elliottii forest. Each species was replicated 3 times w ithin a location and held in place using staples. Samples were 91

PAGE 92

collected from each location at 6 week, 3 m onth, 6 month, and 12 month intervals. The decomposition constants (k) were determined for each plant species by assuming an exponential rate of mass loss (Mf = Mi e-kt) where Mi is the initial mass of the litter, Mf is the final mass of the litter, and t is the time at Mf. The mean residence time, or time required for the litter to decompose under steady state, was calc ulated as 1/k (Chapin et al. 2002). Litter Component Analyses Percent organic matter was measured by loss on ignition from 0.2 to 0.5 g samples of dried and ground plant material, which were first measur ed into 50 mL beakers (Luczak et al. 1997). The beakers were placed in a muffle furnace and brought to 250C for 30 minutes. The furnace temperature was then increased to 550C for 4 ho urs. Organic matter content was calculated as the mass loss on ignition on a dry weight basis. Litter quality was measured with a sequen tial extraction using an Ankom A200 Fiber Analyzer (Rowland and Roberts 1994). Half of a gram of coar sely ground litter material was weighed and sealed into Ankom filter bags. The ba gs were extracted with a neutral detergent to remove soluble cellular contents (sugars, ca rbohydrates, lipids, etc.) followed by an acid detergent to removal hemi-cellulose. Cellulo se was removed by soaking the bags in 72% sulfuric acid (H2SO4) for 3 hours. The residual (lignin and ash) was combusted at 550C for 4 hours to determine ash content. Litter quality was calculated on a dry mass basis. Nutrient Analyses Dried and ground plant material was analy zed for percent carbon and nitrogen on a Thermo-Electron, 1112 Series elemental analyzer Total phosphorus was determined by a twophase acid extraction after lo ss on ignition (Andersen 1976). Th e ash remaining in the 50 mL beaker was moistened with 2 to 3 mL distilled de-ionized water and then extracted with 20 mL of 6 N hydrochloric acid (HCl). All of the water was removed and the hot plate was placed on 92

PAGE 93

high for 30 minutes to completely dry the samp les. After cooling, 2.25 mL of 6 N HCl was added to each beaker and the beakers placed on a hot plate until almost boiling. Extracts were then filtered through a #41 Whatma n filter into 50 mL volumetric flasks that were then brought to volume with distilled de-ionized water. Total phosphorus was measured with an automated ascorbic acid method on a Bran and Luebbe Auto Analyzer 3, Digital Colorimeter (Method 365.4; USEPA 1993). Nutrient ratios we re all calculated on a mass basis. Statistical Analyses Values of measured vegetation characteristics and decomposition rates were calculated as a mean for each sample date. ANOVA and Tukey means separation tests were used to detect any differences between the measured parameters. Di fferences are reported as significant for tests with p values 0.05. All statistical analyses were pr eformed using JMP 7.0.1 software (SAS Institute, North Carolina, USA). Results Full model results for the main effects and in teractions are reported in Tables A-19 through A-22. Based on the magnitude of the results the ma in effects of species and week were the most important determinants of the measured response va riables. There were co nsistent block effects for many of the measured variables however in no circumstances did this effect change the pattern between the three plants species over time. The tissue concentrations of carbon, nitroge n, and phosphorus for each plant species are presented in Table 4-1. Although there was no difference in the initial concentr ation of carbon among the three species, there wa s more initial nitrogen in the T. distichum and M. quinquenervia litter compared to the P. elliottii litter (Table 4-1). The phosphorus concentration was highest in the T. distichum litter. The concen tration of nitrogen and phosphorus increased over time for each plant species. Pr ior to placement in the field, the P. elliottii litter had the 93

PAGE 94

highest ratio of carbon to nitrogen followed by the T. distichum and then M. quinquenervia litter (Table 4-2). Pinus elliottii litter had the lowest ratio of nitr ogen to phosphorus compared to the other two species. There was no difference in the initial ratio of carbon to phosphorus among any of the plant species. The ratio of carbon to nitrogen, nitrogen to phosphorus, and carbon to phosphorus decreased in all plant spec ies over time. The T. distichum litter had the highest decomposition constant compared to both the M. quinquenervia and P. elliottii litter (Table 4-3, Figure 4-1). It also had the shortest residence time or fastest rate of decomposition followed by M. quinquenervia and then P. elliottii (Table 43, Figure 4-1). At the end of the year, the T. distichum litter lost 53%, the M. quinquenervia litter lost 37%, and the P. elliottii litter lost 3 1% (Figure 4-1). The M. quinquenervia litter initially had the highest pe rcentage of soluble fiber, followed by the T. distichum litter, and then P. elliottii litter (p = <0.0001, Figure 4-2). This relationship was maintained in the final sample where soluble fiber was also highest in the M. quinquenervia litter, followed by the T. distichum and then P. elliottii (p = <0.0001, Figure 4-2). The P. elliottii litter had the highest percentage of lignin both initially and in the last sampling date, followed by the T. distichum litter, and then M. quinquenervia litter (p = 0.0001, Figure 4-2). Each plant species lost carbon compared to the initial pool at each sa mple period (Figure 45a). At 6 weeks and 1 year, the M. quinquenervia litter lost nitrogen comp ared to the initial pool (Figure 4-5b). However, M. quinquenervia litter gained nitrogen at 3 and 6 months. In contrast, T. distichum litter lost nitrogen at each sample period. Pinus elliottii litter gained nitrogen at 6 weeks, 3 months, and 6 months compared to the in itial pool but lost nitrogen at 1 year. Each plant species gained phosphorus compared to the initial pool at each sample pe riod (Figure 4-5c). 94

PAGE 95

Discussion Organic Matter Turnover Baruch and Goldstein (1999) and Ehrenfeld (200 3) found that the litter of exotic plants may differ from native species in chemical comp osition, which could alter the rate of ecosystem nutrient fluxes. For example, the ratio of so luble fibers (e.g. sugars and carbohydrates) to resistant materials (e.g. lignin) in organic ma tter can affect the rate of decomposition and subsequent release of plant available nutrients. In this st udy, differences were found among the chemical compositions of the M. quinquenervia T. distichum and P. elliottii litters. However, contrary to our hypothesis that M. quinquenervia would have the highest concentrations of lignin, M. quinquenervia had the highest concentration of soluble materials and the lowest concentration of lignin, both ini tially and at the one year samp le period (Figure 4-2). Lignin increased in all three plant species by end of one year, probably as a re sult of the preferential degradation of soluble carbon compounds during organic matter decomposition (Chapin et al. 2002). As hypothesized, the residence time of the M. quinquenervia litter was significantly longer than the T. distichum litter (Table 4-3). However, contra ry to our hypotheses, the residence time of the M. quinquenervia litter was significantly shorter compared to P. elliottii litter (Table 4-3). Many studies have found that ecosystem invasion by exotic plants signific antly alters rates of organic matter decomposition (Pidgeon and Cair ns 1981, Baruch and Goldstein 1999, Ehrenfeld 2003, Rothstein et al. 2004, Standish et al. 2004). Standish et al. (2004) found that the invasion of the exotic herb Tradescantia fluminensis increased rates of organic matter decomposition and nutrient availability in a remnant forest However, these data i ndicate that invasion by M. quinquenervia into a P. elliottii T. distichum ecotone forest may not si gnificantly alter rates of 95

PAGE 96

organic matter turnover. Further study is need ed to investigate the effects of mixed litter decomposition in areas dominated by exotic plants. The residence time of all three species was some what faster than reported in the literature (Table 4-3). Reported estimat es of the turnover time of T. distichum litter range 1.2 to 3.1 years compared to the 1.34 years in this stu dy. Although no direct measurements of M. quinquenervia decomposition have been published, we reanalyzed litterfall and litter accumulation data from Greenway (1994) and calculated a residence time of 4.3 years, compared to the 2.19 years in this study. Wienand and Stock (1995) found that the decomposition of P. elliottii needles can be relatively slow compared to co-existing native species. For example, Gholz et al. (1985) reported that P. elliottii litter lost only 13-17% of the dry litter in one year. However, decomposition of P. elliottii litter was faster in this study where 31% of the litter had decomposed after 12 months. This faster ra te may have been caused by the seasonal hydroperiod of the study area. The largest losses of orga nic material occurred during or directly after inundation (Figure 4-1). These fluc tuations in the water table may have increased the leaching of soluble compounds and provided moisture to the microbial communities. Nutrient Turnover Nutrient storage and cycling in the decom posing organic matter was different among the three plant species. Nitrogen was immob ilized at every sample period in the T. distichum and M. quinquenervia litter which led to an increase in the ni trogen concentration of the litter at the 1 year sample period (1.72 and 1.53 times greater for T. distichum and M. quinquenervia respectively, Table 4-1). In contrast, nitrogen was immobilized in the P. elliottii litter for the first 6 months, then mineralized at the 1 year sample, for an overall increase in nitrogen concentration (1.31 times greater, Table 4-1). Despite the increase in nitrogen concentration with T. distichum at every sample period, the significan t mass loss of litter led to an overall 96

PAGE 97

decline in nitrogen storage (Figure 4-5b). In co ntrast, the storage of nitrogen increased in M. quinquenervia and P. elliottii litter at the 3 month and 6 month sample periods and then decreased at 1 year (Figure 4-5b). Overall, diffe rences in nutrient concentration offset mass loss and led to the greater total nitr ogen storage of nitrogen in the T. distichum and M. quinquenervia litter compared to the P. elliottii litter (p = <0.0001, Figure 4-6b). The rates of nitrogen storage and cycling can differ between native and e xotic plant dominated ecosystems (Mack and D'Antonio 2003, Rothstein et al. 200 4, Standish et al. 2004). Roth stein et al. (2004) found that the invasion of the exotic tree Fraxinus uhdei into the Hawaiian rainfore st doubled the release of nitrogen of the litter compar ed to a native forest. Howe ver, in this study invasion by M. quinquenervia should not significantly alter the st orage and cycling of nitrogen in a P. elliottiiT. distichum ecotone forest. All three plant species immobilized phosphorus at every sample period and increased overall phosphorus storage indicating a phosphorus limitation in the system (Table 4-1 and Figure 4-5c). At the one year sample period, th e total storage of phosphor us was highest in the T. distichum followed by M. quinquenervia and then P. elliottii litter (p = 0. 012, Figure 4-6c). The measured differences in nutrient dynamics expl ain the differences in the residency times of the litter, namely the final nitrogen concentration, phosphorus concentration, and the final nutrient ratios of the litter mate rial (Figure 4-3 and 4-4). These relationships suggest that both nitrogen and phosphorus limitations, as well as, or ganic matter quality may be controlling the decomposition of the litter material of all three species. Another mechanism whereby M. quinquenervia litter could alter nutrient turnover rates is through changes in microbial communities. Micr obes are responsible for the decomposition of organic matter and cycling of nutrients critical to plant growth (Chapin et al. 2002). However, 97

PAGE 98

before a nutrient is released into the ecosystem, internal needs of microbial communities must be met. For example, it is estimated that microbial communities will immobilize nitrogen in order to meet internal needs if the C:N molar ratio of the substrate is greater than 25:1 to 30:1. This figure is based on two assumptions: 1) a microbial substrate use efficiency of 40% and 2) a C:N microbial biomass ratio of 10:1 (Chapin et al. 2002). Martin et al. (2009) found that M. quinquenervia altered the soil microbial community composition, which in turn may change rates of ecosystem nutrient cycling. For exampl e, if fungal biomass were to become more prevalent in the invaded soils, then substrate us e efficiency could be greater thereby increasing the availability of carbon substrates (Dav et 2004). In addition, microbial communities with higher C:N biomass ratios have lower requir ements for nitrogen and will therefore mineralize nitrogen at lower concentrations (Eviner and Chapin 2003). Furthe r research is n eeded to assess the structure of microbial communities growing on the various litter types. Conclusion The invasion of M. quinquenervia has been shown to alter ecosystem structure and function (Myers 1983, 1984, Bodel et al. 1994, Martin et al. 2009). However, this work has shown that M. quinquenervia may not significantly alter the basic ecosystem processes of organic matter decomposition and nutrient turnover in invaded P. elliottii T. distichum ecotone forests. This indicates the need for ecosystem-s pecific studies to evaluate the impact of plant invasions. Although M. quinquenervia has colonized and th rived in most natural areas of South Florida, the consequences for ecosystem func tion may not be the same for each community (Bodel et al. 1994). Currently there is an inte grated plant management program to control M. quinquenervia in South Florida ecosystems. Mechanical, chemical, and biological control programs have contained the spread and eliminated the invasive potential of existing M. quinquenervia populations (Ferr iter et al. 2005, Tipping et al 2008, Tipping et al. 2009). 98

PAGE 99

However, live non-invasive M. quinquenervia trees remain part of ve getative landscape and are targets for future management. Treatment of remnant M. quinquenervia populations with chemical or mechanical methods may cause sign ificant collateral dama ge to native plant communities and may negatively influence ecosys tem function. Further work is needed to determine if communities would actually benefit fr om the removal of this exotic, but now less invasive plant. 99

PAGE 100

Tables and Figures Table 4-1. Mean ( S.E.) of litter nutrient concentrations at every sample time for each plant species (lower case letters indicate si gnificant differences for each analysis). Sample period M. quinquenervia T. distichum P. elliottii P Carbon -------------------------------mg g-1----------------------------initial 479 12.3 523 28.4 475 10.80 0.22 6 weeks 484 2.92 a 464 1.81 b 470 5.39 b 0.0035 3 months 483 3.50 a 459 2.07 b 470 4.00 b <0.0001 6 months 472 2.41 a 457 2.03 b 469 1.63 a <0.0001 12 months 494 3.10 a 461 3.26 b 463 4.89 b <0.0001 Nitrogen initial 7.44 0.18 a 9.06 0.68 a 5.38 0.11 b 0.002 6 weeks 8.44 0.20 b 12.1 0.21 a 6.62 0.03 c <0.0001 3 months 9.20 0.40 b 13.8 0.34 a 6.75 0.22 c <0.0001 6 months 9.70 0.31 b 14.5 0.42 a 8.44 0.41 b <0.0001 12 months 11.4 0.71 b 15.6 0.34 a 7.03 0.26 c <0.0001 Phosphorus -------------------------------mg kg-1----------------------------initial 103 1.56 b 127 3.10 a 106 1.73 b 0.0006 6 weeks 188 11.0 c 526 19.8 a 266 18.3 b <0.0001 3 months 284 14.4 b 650 50.7 a 321 35.7 b <0.0001 6 months 289 15.2 b 623 38.7 a 322 23.7 b <0.0001 12 months 410 48.8 b 660 42.0 a 338 16.8 b <0.0001 100

PAGE 101

Table 4-2. Mean ( S.E.) of mass nutrient ratio s at every sample time for each plant species (lower case letters indicate significant differences for each analysis). Variable M. quinquenervia T. distichum P. elliottii P C : N 0 weeks 64.4 0.24 b 57.9 1.15 c 88.3 1.42 a <0.0001 6 weeks 57.5 1.15 b 38.6 0.69 c 72.0 2.80 a <0.0001 3 months 53.3 2.17 b 33.0 0.93 c 67.2 3.19 a <0.0001 6 months 49.1 1.55 b 31.9 1.00 c 56.7 2.83 a <0.0001 12 months 44.4 2.51 b 29.7 0.60 c 66.4 1.90 a <0.0001 N : P 0 weeks 72.4 2.38 a 71.9 6.86 a 50.8 1.67 b 0.019 6 weeks 46.0 2.22 a 23.1 0.66 b 25.8 1.88 b <0.0001 3 months 32.7 1.52 a 22.0 1.43 b 21.8 1.26 b <0.0001 6 months 34.2 1.67 a 23.6 0.96 b 26.8 1.23 b <0.0001 12 months 29.4 2.14 a 24.1 1.05 b 21.0 0.58 b 0.0009 C : P 0 weeks 4662 151 4143 306 4486 171 0.31 6 weeks 2646 149 a 892 32.2 c 1862 174 b <0.0001 3 months 1649 62.5 a 733 60.4 c 1435 119 b <0.0001 6 months 1676 93.5 a 757 47.9 b 1534 132 a <0.0001 12 months 1325 146 a 718 41.4 b 1397 68.9 a <0.0001 101

PAGE 102

Table 4-3. Mean ( S.E.) of litter decompositi on constants and residence times for each plant species reported in literatu re and this study (lower case letters indicate significant differences for each analysis). Citation Species Decomposition Residence time constant (k) years (1/k) Greenway 1994 Melaleuca quinquenervia 0.23 4.3 Gholtz et al. 1984 Pinus elliottii 0.13 7.7 Hendricks et al. 2002 Pinus elliottii 0.10 10.3 Polglase et al. 1992 Pinus elliottii 0.27 3.7 Wienand and Stock 1995 Pinus elliottii 0.21 4.8 Wienand and Stock 1995 Pinus elliottii 0.35 2.9 Battle and Golladay 2001 Taxodium distichum 0.62 1.6 Day 1982 Taxodium distichum 0.33 3.1 Deghi et al. 1980 Taxodium distichum 0.83 1.2 Middleton 1994 Taxodium distichum 0.32 3.1 Nessel and Bayley 1984 Taxodium distichum 0.33 3.0 This study Melaleuca quinquenervia 0.47 0.03 b 2.19 0.13 b Taxodium distichum 0.76 0.04 a 1.34 0.07 c Pinus elliottii 0.37 0.02 b 2.75 0.13 a P <0.0001 <0.0001 102

PAGE 103

0 20 40 60 80 100 01020304050 Weeks % Remaining Mass Remaining (%) Wet Wet Dry 60 T. distichum M. quinquenervia P. elliottii Figure 4-1. Mean ( S.E.) of litter mass remainin g at every sample time for each plant species. 103

PAGE 104

Chemical Composition (%) a 0 20 40 60 80 100 Ash Lignin Cellulose Hemicellulose Fiber b 0 20 40 60 80 100 M. quinquenerviaT. distichumP. elliottii A B Figure 4-2. Mean ( S.E.) of the chemical compos ition for each plant species. A) The chemical composition at the initial sample period. B) The chemical composition at the final sample period. 104

PAGE 105

0 1 2 3 4 0.50.70.91.11.31.51.71.9 Litter Nitrogen (mg g-1) M. quinquenervia T. distichum P. elliottii R2 = 0.9Residence time (years) 0 1 2 3 4 02004006008001000 Litter Phosphorus (mg kg-1) R2 = 0.8 R2 = 0.9 R2 = 0.76 A B Figure 4-3. Relationship between litter residenc e time and final nutrient concentration for each plant species. A) Relationship between residence time and litter nitrogen concentration at the final sample period. B) Relationship between residence time and litter phosphorus concentration at the final sample period (circles indicate individual plant species). 105

PAGE 106

Figure 4-4. Relationship between l itter residence time and chemical and nutrient ratios at the final sample period for each plant species A) Relationship between residence time and lignin to nitrogen ratios. B) Relati onship between residence time and lignin to phosphorus ratios. C) Relati onship between residence ti me and carbon to nitrogen ratios (circles indicate individual plant species). 0 1 2 3 4 02 04 06 08 Lignin : Nitrogen 5 0 M. quinquenervia A T. distichum P. elliottii R2 = 0.7 0 1 2 3 4 5 20040060080010001200140016001800 Lignin : Phosphorus R2 = 0.7 0 1 2 3 4 5 0 20406080100 Carbon : Nitrogen R2 = 0.9Residence time (years) R2 = 0.74 R2 = 0.71 R2 = 0.88 C B 106

PAGE 107

-100 -80 -60 -40 -20 0 6 weeks 3 months 6 months 1 year -30 -20 -10 0 10 20 30 40 0 50 100 150 200 250 300M. quinquenerviaT. distichumP. elliottiiChange Carbon Pool (%) Change Nitrogen Pool (%) Change Phosphorus Pool (%) A B C Figure 4-5. Mean ( S.E.) percent change in the pools of nutrients from the initial storage at every sample time for each plant species. A) Change in carbon pool. B) Change in nitrogen pool. C) Change in phosphorus pool. 107

PAGE 108

Figure 4-6. Mean ( S.E.) pools of nutrients at every sample time for each plant species. A) Carbon pool. B) Nitrogen pool. C) Phosphorus pool. 400 600 800 1000 1200 M. quinquenervia T. distichum P. elliottii 10 12 14 16 18 20 0 200 400 600 800 1000 06122652 A B C Weeks Carbon Pool mg Carbon Pool (mg) Nitrogen Pool mg Nitrogen Pool (mg) Phosphorus Pool g Phosphorus Pool (g) 108

PAGE 109

CHAPTER 5 RECOVERY OF PLANT COMMUNITY ST RUCTURE AFTER A SEASONAL FIRE Introduction Ecosystem invasion by exotic plant species poses a significant th reat to community diversity, function, and stability (Mack and D'Antonio 1998, Mack et al. 2000, Fenn et al. 2003, Kourtev et al. 2003). Although this issue is of global concern, the invasion of exotic plants has been a significant problem in the State of Florid a, especially South Florida, which is home to Everglades. The Everglades is an extensive forest and graminoid wetland community that once occupied 4,000 square miles of th e states lower peninsula, over twice its present day land area (Gunderson 1994). This complex ecosystem has pr oved an ideal habitat for many exotic plant species. The Florida Exotic Pest Plant C ouncil has compiled a list of Floridas 125 most invasive exotic plants and separated them into two categories (Doren 2002 ). Sixty-six plants received a category I designation, which indicates they are considered highly disruptive to native plant communities. Arguably one of the worst cate gory I invasive plants is Melaleuca quinquenervia (Cav.) Blake, otherwise known as the paper-bark tree, cajeput, punk tree, or wh ite bottlebrush tree (Doren 2002). Due to its fire adapte d nature and copious seed production, M. quinquenervia is able to out-compete and replace many na tive species. It is estimated that M. quinquenervia populations currently cover over 161, 874 hectares in the State of Florida (Anonymous 2007). This large evergreen tree (25-30 m tall) is native to coastal areas of eastern Australia, southern New Guinea, and New Caledonia where it occurs typically along freshwater streams and swamps (Boland et al. 1987). It was intr oduced into South Florida in 1886 (Dray et al. 2006), originally for sale as an ornamental, but later was used fo r erosion control, as a forestry crop, and as an agricultural windrow plant (Meskimen 1962, Stocke r and Sanders Sr. 1981, Bodel et al. 1994). 109

PAGE 110

This exotic tree has colonized and thrived in most of the natural areas of South Florida, including bayhead tree islands, sawgrass prairies, pine flatwoods, pastures, and cypress forests (Bodel et al. 1994). DiStefano and Fisher (1983) found that th e relative density, frequency, and dominance of several native plant species were significantly diminished in sites invaded by M. quinquenervia A management plan was coopera tively developed by local, state, and federal agencies to reduce the impact of M. quinquenervia using chemical, mechanic al, and biological methods (Ferriter et al. 2005). The Sout h Florida Water Management Dist rict (SFWMD) has sustained a multi-year campaign to chemically and mechanically control M. quinquenervia on public lands. In 2003 alone the SFWMD chemically treated 1,795 hectares of M. quinquenervia with ground application and 4,118 hectares usin g aerial application (Ferriter et al. 2005). In addition, the SFWMD partially funded a biological control project headed by the United States Department of Agriculture, Agricultural Research Service (ARS) (Ferriter et al. 2005). This project, begun in 1986 at the ARS Invasive Plant Res earch Lab in Fort Lauderdale, is responsible for the release of four biological control agents to reduce or eliminate the capacity of M. quinquenervia to invade (Ferriter et al. 2005). Two of the biological agents, Oxyops vitiosa Pascoe (Coleoptera: Curculionidae) and Boreioglycaspis melaleucae Moore (Hemiptera: Psyllidae), have successfully established and are suppressing M. quinquenervia reproduction, growth, and recruitment on a landscape scale (Tipping et al. 2009). Although management schemes aim to reduce or remove the competitive advantage of invasive plants in order for native plant communities to recover, litt le work has been done evaluating the impact of treatment methods on pl ant community structure, especially in the context of native disturbance regimes. Many of Floridas natural areas th at have been invaded by M. quinquenervia depend on regular disturbances like fires to maintain community structure 110

PAGE 111

and function. Native plant communities in these areas depend on seasonal fires to open canopies and trigger the seed release and germination of plant species, provide temporary pulses of soluble nutrients, and reduce plant competiti on (Wade et al. 1980, Neary et al. 1999). The invasion of M. quinquenervia and other exotic plants have altered native fire patterns which may have long-term consequences for ecosystem function (Wade et al. 1980, Serbesoff-King 2003). The objective of this work was to elucidate changes in plant community structure in a Pinus elliottii EnglemTaxodium distichum (L.) L.C. Rich var. nutans (Ait.) Sweet ecotone forest after the invasion and management of M. quinquenervia Two main hypotheses were tested: after a seasonal fire 1) plant community structure will not be different in the invaded and biologically controlled but will be different in the invaded and chemically treated site compared to the non-invaded site and 2) the re-invasion of M. quinquenervia will be most severe in the chemically treated site compared to the biolog ically controlled and non-invaded sites. Materials and Methods Experimental and Statistical Justification In order to assure the proper assignment of treatment causality in experiments several fundamental assumptions must be met includi ng: random assignment of treatments across experimental units and treatment replication (Beyers 1998). Random assignment of treatments reduces the impact of non-treatment factors that could confound results a nd replication reduces the likelihood that random variati on is the cause of measured differences between treatments. Although the most effective field studies have randomly assigned, repl icated experimental treatments many times limitations of money, labor, and time make these conditions impossible. Without a proper experimental design, the use of in ferential statistics may only reveal differences between un-replicated plots and not the desired tr eatment effect (Hurlbert 1984). In other words, 111

PAGE 112

the null hypothesis becomes that there is no diffe rence between plots NOT that the treatment has no effect (Hurlbert 1984). In ecosystems all over the world land manage rs are manipulating natural areas in an attempt to restore function, provide habitat, or mitigate anthropogenic disturbance. Often times these treatments are done on one large tract of land or single e xperimental unit. Scientific analysis of these areas can be complex because both of the assumptions mentioned above are violated. Still large-scale field studies investigating these treatm ents must be done in order to assess the impact of natural resource management d ecisions. There is nothing to be gained from limiting scientific investigation when manageme nt must continue. Insight gained from determining large-scale treatment effects can be used to guide further research and prevent undesirable consequences of management decisions. Site Description The study site was located in the Belle Meade Tract of the Picayune Strand State Forest in Collier County, Florida. This area consists of ne arly level, poorly drained, low fertility soils which are loamy, siliceous, hyperthermic Arenic Glassoqualfs. The soil series is Pineda-BocaHallandale which is characterized by moderately to poorly drained sands which overly limestone bedrock to a depth of approxi mately 1.4 m (USDA 1998). The wa ter table fluctuates annually between greater than 15 cm below the soil surface to approximatel y 25 cm above. In an average year, the area has a distinct wet season from ap proximately July to December and a dry season from January to June. Average annual rainfa ll in this region is approximately 1.36 m (SERC 2007). Historically, the vegeta tion in this area was a mixed T. distichum-P. elliottii forest with a hardwood under-story. Over the past several deca des, many areas in this landscape have been invaded with M. quinquenervia and are now comprised of spar se populations of mature trees 112

PAGE 113

with dense understories of seed lings and saplings that can exc eed densities of 100 plants per square meter. In early May of 2007 the Great Basal fire burned approximately 8, 000 hectares in southwest Florida and all of the established experimental plots. As the fire was not planned no direct measurements of fire intensity were taken. However, all of the plots were equally affected as well as the entire su rrounding landscape. Diversity measurem ents were taken 1 year after the fire. Diversity Plots On May 15, 2007, after a seasonal fire, twenty-five, 1 m2 plots were established along five transects in each of th ree contiguous study areas. The areas sampled were: area #1) reproductive M. quinquenervia treated with herbicide in the summer of 2003 (hereafter referred to as herbicide site), area #2) reproductive M. quinquenervia treated with biological control (hereafter referred to as biological s ite), and area #3) native forest with no M. quinquenervia (hereafter referred to as noninvaded site). The two biolog ical control agents mentioned above, O. vitiosa and B. melaleucae, were common in all areas. Th e herbicide site consisted of reproductive M. quinquenervia that was treated aerially with Velpar (Hexazinone, 3cyclohexyl-6-(dimethylamino )-1-methyl-1,3,5-triazine-2,4(1 H,3H)-dione) in 2003. The biological site consisted of reproductive M. quinquenervia which was not treated with herbicides and has been under attack by O. vitiosa since 1998 and B. melaleucae since 2002. The noninvaded site was a P. elliottii-T. distichum ecotone forest with no history of M. quinquenervia invasion. Non-woody diversity was measured in each plot (n = 25) on February 12-14, 2008. Four, 100 m2 plots were established between the five transects in each of the sites and woody diversity was measured in each plot (n = 4) on February 17-19, 2008. Diversity was expressed by Simpsons Index, 1/D = 1 / ( pi 2), and Shannons Index, H = pi (ln pi), where pi is the 113

PAGE 114

relative (decimal) cover of the ith species (Peet 1974). Freque ncy is measured as the number of plots within each site where the species occu rred divided by the tota l number of plots. Statistical Analyses Values of measured woody and non-woody diversity and species abundance were calculated as a mean for each plot. ANOVA and Tukey means se paration tests were used to detect any differences among the plots. Differences are reported as signif icant for tests with p values 0.05. All statistical analys es were preformed using JMP 7.0.1 software (SAS Institute, North Carolina, USA). Results Full model results for the main effects and in teractions are reported in Tables A-23 and A24. Based on the magnitude of the results the ma in effect of site was the most important determinant of the measured respon se variables. There were no cons istent transect or plot effects for any of the measured variables. A total of 51 plant species were found across the three study sites with 37 species both in the non-invaded and biological sites, and 28 species in the herbicide site (Table 5-1). There was considerable overlap in species between the bi ological site and non-invaded sites with 68% of non-woody species, 100% of woody species, and 70% of all plant species found in both sites. The herbicide site had fewer plants in common with the non-inva ded: 59% of non-woody species, 66% of woody species, and 57% of all plant species. Overall, the non-woody plant species ric hness was highest in the non-invaded and herbicide sites compared to the biological site (Table 5-2). Woody plant species richness was highest in the non-invaded site fo llowed by the biological site and then the herbicide site. Both the Simpsons and Shannons indices for non-woody vegetation were highest in the non-invaded compared to the herbicide and biological sites. There was no difference in the Simpsons index 114

PAGE 115

for woody vegetation among the sites. Shannons index for woody vegetation was highest in the non-invaded site compared to the biological and herbicide sites. The abundance of M. quinquenervia live and dead seedlings was highest in the biological compared to the non-invaded and herbic ide sites (Table 5-3). Many of the M. quinquenervia seedlings measured in the biological were dead (37%) followe d by 5% in the non-invaded site and 0% in the herbicide site. The non -invaded site had highest abundance of T. distichum and P. elliottii mature trees compared to the herbicide and biological sites. Figure 1 shows the frequency occurrence of each spec ies in the three sites. Discussion Plant Community Structure Approaches to restoring communities after th e management of exotics can run the gamet from passive to active. The most passive appr oach relies on native plant communities recovering on their own, while more active methods involve directed efforts like replanting natives. Selecting the best approach de pends on several factors includi ng cost and the impact of the management itself on the native plant community. However, little work has been done evaluating the effects of current management programs on native community recovery in South Florida ecosystems. Smith et al. (2002) experimentally evaluated the concept of passive re storation as part of the Comprehensive Everglades Restoration Plan (CER P). The main focus of the CERP is to get the water right which would passively restor e the Everglades plant communities. In this study vegetation responses to increased hydroperiod length were measured to assess if the restoration of historic water flow patterns would result in cont rol of invasive plants and a rebound of native plant communities. Their result s indicated that restor ation of native plant communities will not be solely dependant on water flow patterns and, therefore, more active 115

PAGE 116

management strategies may be necessary in the future to restore historical plant communities (Smith et al. 2002). In contrast, MacDougall an d Turkington (2007) mana ged weeds using fire, cutting and raking, and weeding and noted reducti ons in several common, invasive agricultural weeds, coupled with increases in native plants. In this study, chemical a nd biological management of M. quinquenervia resulted in lower levels of plant diversity after a seasonal fire when compared to an unmanaged, non-invaded site (Table 5-2). Although the herbicide site containe d levels of species richness comparable to the non-invaded site, only 57% of the plant specie s were found in common compared to 70% of species in biological site. Si milarly, Mason and French (2007) found that despite significant reductions of Chrysanthemoides monilifera ssp. rotundata dune systems were unable to passively return to pre-invasion levels of native pl ant diversity. They attributed this deficit to collateral damage from aerial herbicide appl ications which prevented recovery of dune communities. Further work is needed to evaluate the long-term consequences of the treatment of M. quinquenervia on native plant recruitment. Plant Community Re-invasion The density of M. quinquenervia seedlings was highest in the biological site compared to the non-invaded and herbicide sites (Table 5-3). However, 37% of those seedlings were dead compared to only 5% in the non-invaded site a nd 0% in the herbicide site. The rate of M. quinquenervia re-invasion in the all sites is significantly lower than experienced in the same site prior to the introduction of th e biological control agents After a fire in 1998, 591 M. quinquenervia seedlings per square meter germinated co mpared to the 34 seedlings in this study (Tipping et al., unpublished data). The lower rate of post-fire M. quinquenervia seedling recruitment is the resu lt of the aforementioned biological co ntrol program. Tipping et al. (2008) 116

PAGE 117

found that herbivory for the two established biol ogical agents has reduced the seed production of M. quinquenervia by 99%. However, management of one exotic species may lead to its replacement by another. Ogden and Rejmanek (2005) used fire and herbicid es to significantly de crease the cover of the invasive fennel Foeniculum vulgare, only to have it replaced by non-native Mediterranean annual grasses. Another study found that riparian systems managed to control Impatiens glandulifera invasion were more readily invaded afte r management (Hulme and Bremner 2006). Although it was not one of the most a bundant species, the invasive grass Panicum repens or torpedo grass (ID #25) was found in both the herb icide and biological sites (Table 5-1, Figure 51). This species is a serious pest in the Stat e of Florida, has proven resistant to management programs, and may out-compete natives in the two treatment sites (David 1999). The potential for re-invasion following management reveals a need for a more active management approach that considers site specific environmental conditions (MacDougall and Turkington 2007). For example, a multi-habitat analysis of the control of the invasive plant Tamarix spp. identified several site characteristics that promoted greater cover and ric hness of natives (Bay and Sher 2008). Such analyses should extend to both abov eand belowground alterations to ecosystem function caused by exotics and th e tactics to manage them. Conclusion It is clear that plant invasions can result in devastating cha nges in natural systems, which justifies management operations designed to reduce or eliminate th eir effects (Mack and D'Antonio 1998, Mack et al. 2000, Fenn et al. 2003) Invasive species also pose significant direct and indirect challenges to ecosystem-wide re storation projects like CERP. Despite the fact that CERP will cost at least 30 billion dolla rs, require 30 years to complete, involve many agencies from all levels and jurisdictions, a nd affect 18,000 square miles over sixteen Florida 117

PAGE 118

counties, the management of exotics has been so mething of an afterthought. The United States Army Corps of Engineers, the managerial agen cy for CERP, stated th at once hydrology is restored to the Everglades, invasive exotic species, such as Melaleucawill continue to degrade the system by displacing na tive species (Anonymous 2004). While the ultimate goal of management programs is to restore ecosystem integrity, this work has shown that passive restoration may not be enough to restore plant community structure in this system. More detailed studies which incorporate ecosystems serv ices are needed to evaluate and value the role of native plant communities in overall ecosystem health, thereby guiding management decisions designed to protect and maintain them. 118

PAGE 119

Tables and Figures Table 5-1. Species list of all pl ants identified in the non-invade d, herbicide, and biologically controlled plots. ID Species Name Non-invaded Herbicide Biological -------------------non-woody-------------------1. unknown spp. 1 X X X 2. Pluchea rosea X X X 3. Andropogon glomeratus X X X 4. Leptochola spp. X X X 5. Cyperus elegans X X X 6. Cyperus haspan X X X 7. Scoparia dulcis. X X X 8. Eupatorium capillifolium X X X 9. Gamochaeta purpurea X X X 10. Diodia virginiana X X X 11. Symphyotrichum spp. X X X 12. Pluchea odorata X X X 13. Rhynchospora colorata X X X 14. Heloptroppium spp. X X X 15. Rhynchospora divergens X 16. Fuirena breviseta X 17. Mikania scandens X X X 18. Dichanthelium dichotomum X X X 19. Agalinis purpurea X X 20. Solanum americanum X X X 21. Ludwigia leptocarpa X X X 22. Hypericum fasciculatum X X 23. Erechtites hieracifolia X X X 24. Juncus scirpodes X 25. Panicum repens X X 26. Dichanthelium spp. X X 27. unknown spp. 2 X 119

PAGE 120

Table 5-1. continued. ID Species Name Non-invaded Herbicide Biological -------------------non-woody------------------28. Cyperus odoratus X 29. Eragrostis elliottii X 30. Gratiola ramose X X 31. Panicum hians X 32. Erechtites hieraciifolius X 33. Blechnum serrulatum X 34. Fuirena spp. X X 35. Cyperus polystachyos X 36. Oldenlandia uniflora X X 37. Violoa lanceolata X 38. Pteridium aquilinum X 39. Centella asiatica X 40. Hyptis alata X 41. unknown spp. 3 X 42. Sarcostemma clausum X 43. Conoclinium coelestinum X 44. Bigelowia nudata X 45. Mitreola petiolata X 46. unknown spp. 4 X 47. Xyris platylepis X X X 48. Cladium jamaicense X ----------------------woody---------------------49. Melaleuca quinquenervia X X X 50. Pinus elliottii X X X 51. Taxodium distichum X X 120

PAGE 121

Table 5-2. Diversity indices for the non-invaded, herbicide, and biologically controlled plots. V ariable Plant type Non-invaded Herbicide Biological P Richness Non-woody 9.40 0.63 a 9.83 0.54 a 7.21 0.55 b 0.004 Woody 3.00 0.00 a 1.00 0.00 c 2.25 0.25 b <0.0001 Simpson Non-woody 4.17 0.43 a 2.47 0.17 b 2.41 0.23 b 0.0001 Woody 1.77 0.16 a 1.00 0.00 b 1.07 0.10 b 0.0005 Shannon Non-woody 1.59 0.07 a 1.18 0.07 b 1.04 0.11 b <0.0001 Woody 0.68 0.09 a 0.00 0.00 b 0.16 0.02 b <0.0001 121

PAGE 122

Table 5-3. Woody species abundance for the non-i nvaded, herbicide, and biologically controlled plots. Species Plant Stage Non-invaded Herbicid e Biological P ------------------number of plants m-2-----------------Melaleuca quinquenervia Live Seedling (< 15cm) 5.24 0.95 b 2.58 0.47 b 22.83 13.62 a <0.0001 Dead Seedling (< 15cm) 0.28 0.15 b 0.00 0.00 b 13.63 5.87 a 0.007 Mature (>15cm) 1.78 0.99 2.27 0.29 3.79 0.94 0.24 Taxodium distichum Live Seedling (< 15cm) 0.20 0.08 a 0.00 0.00 b 0.00 0.00 b 0.004 Dead Seedling (< 15cm) 0.00 0.00 0.00 0.00 0.00 0.00 Mature (>15cm) 0.53 0.04 a 0.00 0.00 c 0.13 0.01 b <0.0001 Pinus elliottii Live Seedling (< 15cm) 0.20 0.13 0.04 0.04 0.17 0.10 0.47 Dead Seedling (< 15cm) 0.00 0.00 0.00 0.00 0.00 0.00 Mature (>15cm) 0.06 0.01 a 0.00 0.00 b 0.003 0.003 b <0.0001 122

PAGE 123

0 0.2 0.4 0.6 0.8 1 1357911131517192123252729313335373941434547a A 0 0.2 0.4 0.6 0.8 1 1357911131517192123252729313335373941434547b 0 0.2 0.4 0.6 0.8 1 1357911131517192123252729313335373941434547c B C Species ID Species Frequency Figure 5-1. Frequency data for each non-woody plan t species in the non-invaded, herbicide, and biological plots. A) Non-i nvaded site plant species fre quency. B) Herbicide site plant species frequency. C) Biologically controlled site plant species frequency. 123

PAGE 124

CHAPTER 6 SYNTHESIS Objective 1. An Investigation of Ecos ystem-Alteration after Management of Melaleuca quinquenervia. The work presented in Chapter 2 tested two main hypotheses: 1) herbivory from the biological control agents will lower M. quinquenervia litter quality and ra tes of litter production and 2) herbivore-induced changes in litter quality will lower soil nutrient storage and availability before and after a seasonal fire. This study identified an indirect mechanism whereby M. quinquenervia out-completes native plants and maintains a dominate position in low-resource ecosystems. When freed from the top-down regulation of herbivory, M. quinquenervia creates a positive feedback loop to growth and reproduction (Figur e 2-7). Initially, high aboveand below-ground biomass production allows M. quinquenervia to out-compete native plants fo r nutrient and light resources. Higher rates of nutrient uptake pr oduce greater amounts of high qua lity standing biomass, which eventually falls to the soil at a much higher volu me. This high quality litter maintains a larger population of soil microbial biomass that processes litter and soil nutrients at a faster rate. The resultant mineralized nutrients are quickly taken up by the extensive M. quinquenervia root biomass, further increasing the production of vegetative and re productive biomass. The no herbivory advantages formerly experienced by M. quinquenervia have now been removed with the introduction and establishment of two specialized herbivores. As a result, all ages of M. quinquenervia are now under continuous attack by insect herbivores, from the most recently recruited seedlings to the tallest and fully mature trees. This relentless herbivory slows the rate of above-ground biomass production and reduces the size of the root zone (Figure 2-7). In addition, herbivores remove a significant am ount of standing biomass, thereby preventing it from falling to the forest floor. The remaining litterfall has higher concentrations of resistant 124

PAGE 125

materials such as lignin which, in turn, increase s its turnover time. C oncomitantly smaller pools of soil microbial biomass are supported, which furthe r reduces the rate of nut rient turnover. This is clear evidence of how herbivory not only controls populations of M. quinquenervia directly by reducing plant growth and reproduction, but also i ndirectly by interruptin g its positive feedback growth cycle which would otherwise mainta in its dominance in the ecosystem. We predict that M. quinquenervia populations exposed to herb ivory will be less invasive after native disturbances such as fire. A post-fire census of the experimental plots revealed that 73% of the trees were killed in the herbivory pl ots compared to only 41% of the trees in the nonherbivory plots (Tipping, unpublished data). Surviving trees in the herbivory plots may have been weakened by herbivory perhaps resulting in less root biomass which should reduce their ability to efficiently scavenge the pulse of av ailable nutrients produced immediately after the fire. Native plants should benefit from the reduced competition for light, space, and nutrients. In contrast, without herbivory, populations of M. quinquenervia will likely benefit disproportionally from the increases in storage a nd availability of nutrients, compared to native plants. Although some plant mortality would be expected from fires, over the longer term populations may actually benefit from reduced in tra-specific competition. Thus M. quinquenervia would continue to invade and dominate hi gh and low resource plant communities. Objective 2. Assessing the Impact of a Se asonal Fire on Native, Invaded, and Managed Plots. The work presented in Chapter 3 tested two main hypotheses were tested: 1) M. quinquenervia invasion and treatment with an herbicide will reduce the quantity and availability of nutrients before and after a seasonal fire compared to an non-invaded site and 2) M. quinquenervia invasion and treatment with biological control agents wi ll not alter the quantity and availability of nutrients befo re and after a seasonal fire compared to an non-invaded site. 125

PAGE 126

Determining the best management practices fo r exotic species require s consideration of a broad array of factors an d their potential interacti ons, including future inte ractions with natural events like fires. Most efforts to date ha ve emphasized above ground factors like plant and animal diversity and richness, w ith little to no consideration of below ground factors like nutrient storage, nutrient cycling, and microbial comm unity diversity. This study clearly shows how these foundational ecosystem compon ents were affected by the management of exotics in the backdrop of a natural fire event. The data co llected on the storages of carbon, nitrogen, and phosphorus were compiled to give a broader ecosystem picture (Figures 3-4, 3-5, and 3-6). Before the fire, both the herbicid e and biological treatment areas stored less carbon and nitrogen compared to the native site. The observed differences were mostly due to the smaller litter nutrient storages. In contrast, post-fire the biological site stored th e most carbon and no longer had higher levels of nitrogen compared to the non-invaded site. Total storages of phosphorus were not different before the fire while every s ite experienced an increas e of phosphorus after the fire. Biological control of M. quinquenervia using insect herbivores has proven to be effective at controlling plant growth a nd reproduction (Tipping et al. 2 009). The results of this study suggest that this method had less of an impact on nutrient storage and cycling than herbicides. Additional questions remain including how both methods affect re-vegetation over the longer term. Although herbicides remain a valuable tool in the management of invasive species, more attention needs to be paid to the resulting consequences for ecosystem structure and function. Practices such as active revegetation with native plants may mitigate the deleterious impacts of the treatment and help to prevent future invasion. If evaluations of the below-ground side effects of exotic plant management remain rare, then advancing our understand ing of basic ecosystem 126

PAGE 127

structure and restoration will suffer and any alterations to foundational communities like microbes may permanently and negatively alter ecosystem function. Objective 3. An Analysis of Nati ve and Non-native Litter Quality. The work presented in Chapter 4 tested two main hypotheses were tested: 1) M. quinquenervia will have the slowest rate of decomposition and 2) M. quinquenervia litter will release least amount of carbon, nitr ogen, and phosphorus compared to T. distichum and P. elliottii litter. The invasion of M. quinquenervia has been shown to alter ecosystem structure and function (Myers 1983, 1984, Bodel et al. 1994, Martin et al. 2009). However, this work has shown that M. quinquenervia may not significantly alter the basic ecosystem processes of organic matter decomposition and nutrient turnover in invaded P. elliottii T. distichum ecotone forests. This indicates the need for ecosystem-s pecific studies to evaluate the impact of plant invasions. Although M. quinquenervia has colonized and th rived in most natural areas of South Florida, the consequences for ecosystem func tion may not be the same for each community (Bodel et al. 1994). Currently there is an inte grated plant management program to control M. quinquenervia in South Florida ecosystems. Mechanical, chemical, and biological control programs have contained the spread and eliminated the invasive potential of existing M. quinquenervia populations (Ferr iter et al. 2005, Tipping et al 2008, Tipping et al. 2009). However, live non-invasive M. quinquenervia trees remain part of ve getative landscape and are targets for future management. Treatment of remnant M. quinquenervia populations with chemical or mechanical methods may cause sign ificant collateral dama ge to native plant communities and may interrupt ecosystem function. Further work is needed to determine if communities would actually benefit from the removal of this exo tic, but now less invasive plant. 127

PAGE 128

Objective 4. Recovery of Plant Communi ty Structure after a Seasonal Fire The work presented in Chapter 5 will test two main hypotheses: after a seasonal fire 1) plant community structure will not be different in an invaded and biologica lly controlled site but will be different in an invaded and chemically controlled site compared to the non-invaded site and 2) the re-invasion of M. quinquenervia will be most severe in the chemically controlled site compared to the biologically controlled and non-invaded sites. It is clear that plant invasions can result in de vastating changes in natu ral systems. Several control methods exist that can he lp stop the spread of invasive species and minimize the impact of established populations. While the goal of management programs is to reduce invasive populations and restore ecosystem integrity, this work has shown that passive restoration may not be enough to restore plant community structure. Plant community di versity was highest in the non-invaded site compared to both the biol ogical and herbicides sites as measured by two diversity indices. In addition, the chemically treated site had the fewest number of mature native trees and the highest percentage of live M. quinquenervia seedlings. Further study is needed to evaluate both the long-term resilience and structure of managed ecosystems. Overall Conclusions Invasive species also pose signi ficant direct and indirect challenges to ecosystem-wide restoration projects like CERP. Despite the fact that CERP will cost at least 30 billion dollars, require 30 years to complete, invol ve many agencies from all levels and jurisdictions, and affect 18,000 square miles over sixteen Florida countie s, the management of exotics has been something of an afterthought. The United States Army Corps of Engineers, the managerial agency for CERP, stated that once hydrology is restored to th e Everglades, invasive exotic species, such as Melaleucawill continue to de grade the system by displacing native species (Anonymous 2004). While the ultimate goal of management programs is to restore ecosystem 128

PAGE 129

integrity, this work has shown that passive re storation may not be enough to restore plant community structure and function. The results of this study were complied into Figures 6-1, 6-2, 6-3, and 6-4 to illustrate the effects of the invasion and management of M. quinquenervia on basic ecosystem structure and function in a s ub-tropical South Florid a wetland site. More detailed studies of this kind are needed to evaluate and value the role of native plant communities in overall ecosystem health, thereby guiding mana gement decisions designed to protect and maintain them. 129

PAGE 130

Figures 90 g Microbial Biomass Carbon m-2 2221 g DW Litter Biomass m-2 1077 g C m-2 23 g N m-2 323 mg P m-2 1083 g C m-2 86 g N m-2 1930 mg P m-2 3550 g Soil Organic Matter m-2 3460 g DW Aboveground Biomass m-2* 287 g DW Belowground Biomass m-2* 125 g C m-2 yr-1 2.5 g N m-2 yr-1 48 mg P m-2 yr-1 258 g DW Litterfall Biomass m-2 year-1* Soil 0-15 cm Taxodium distichum 1.34 g SMN m-2 16.3 mg SMP m-2 Figure 6-1. Conceptual model of a fo rest ecosystem dominated by mature Taxodium distichum trees (*data from Martin et al. 2009). 130

PAGE 131

98 g Microbial Biomass Carbon m-2 683 g DW Litter Biomass m-2 295 g C m-2 6 g N m-2 150 mg P m-2 738 g C m-2 43 g N m-2 1840 mg P m-2 2555 g Soil Organic Matter m-2 4000 g DW Aboveground Biomass m-2* 494 g DW Belowground Biomass m-2 94 g C m-2 yr-1 2 g N m-2 yr-1 46 mg P m-2 yr-1 248 g DW Litterfall Biomass m-2 year-1 Soil 0-15 cm Melaleuca quinquenervia 0.52 g SMN m-2 7.4 mg SMP m-2 Figure 6-2. Conceptual model of a fore st ecosystem dominated by early stage Melaleuca quinquenervia trees (*data from Tipping, unpublished data). 131

PAGE 132

79 g Microbial Biomass Carbon m-2 650 g DW Litter Biomass m-2 320 g C m-2 6 g N m-2 100 mg P m-2 603 g C m-2 44 g N m-2 1569 mg P m-2 3254 g Soil Organic Matter m-2 818 g DW Aboveground Biomass m-2* 118 g DW Belowground Biomass m-2 10 g C m-2 yr-1 0.2 g N m-2 yr-1 3 mg P m-2 yr-1 54 g DW Litterfall Biomass m-2 year-1 0.5 g SMN m-2 23 mg SMP m-2 Melaleuca quinquenervia with biological control Soil 0-15 cm Figure 6-3. Conceptual model of a fore st ecosystem dominated by early stage Melaleuca quinquenervia trees controlled with biological agents (*data from Tipping, unpublished data). 132

PAGE 133

80 g Microbial Biomass Carbon m-2 710 g DW Litter Biomass m-2 367 g C m-2 6 g N m-2 97 mg P m-2 718 g C m-2 42 g N m-2 1600 mg P m-2 1920 g Soil Organic Matter m-2 ~0 g Aboveground Biomass m-2 ~0 g Belowground Biomass m-2 ~0 g C m-2 yr-1 ~0 g N m-2 yr-1 ~0 mg P m-2 yr-1 ~0 g Litterfall Biomass m-2 year-1 Soil 0-15 cm 0.59 g SMN m-2 11.8 mg SMP m-2 Melaleuca quinquenervia with chemical control Figure 6-4. Conceptual model of a fore st ecosystem dominated by early stage Melaleuca quinquenervia trees controlled with herbicides. 133

PAGE 134

APPENDIX A FULL MODEL STATISTICAL RESULTS Chapter 2 Model Results Table A-1. Full model results for main effects and interactions for Melaleuca quinquenervia (MQ) litterfall biomass and carbon, nitroge n, and phosphorus transfer in Chapter 2. Effects: site (herbivory and non-herb ivory) and transect (N, E, S, W). Variable Source DF Sums of Squares F ratio Prob > F MQ Litterfall Biomass Site 1 56110 170 <0.0001 Transect 3 480 0.49 0.70 Site*Transect 3 304 0.31 0.82 MQ Litterfall Carbon Transfer Site 1 11776 219 <0.0001 Transect 3 159 0.99 0.42 Site*Transect 3 88.4 0.55 0.66 MQ Litterfall Nitrogen Transfer Site 1 4.21 177 <0.0001 Transect 3 0.04 0.59 0.63 Site*Transect 3 0.01 0.19 0.90 MQ Litterfall Phosphorus Transfer Site 1 2757 93.2 <0.0001 Transect 3 34.6 0.39 0.76 Site*Transect 3 33.2 0.37 0.77 134

PAGE 135

Table A-2. Full model results for main effects and interactions for lit ter % moisture, litter biomass, and % litter loss in Chapter 2. Eff ects: treatment (pre and post fire), site (herbivory and non-herbivory), a nd transect (N, E, S, W). Variable Source DF Sums of Squares F ratio Prob > F % Litter Moisture Treatment 1 3.79 1000 <0.0001 Transect 3 0.04 3.10 0.04 Site 1 0.0005 0.13 0.72 Transect*Site 3 0.006 0.56 0.64 Trt*Transect 3 0.004 0.35 0.79 Trt*Site 1 0.005 1.40 0.24 Trt*Trans*Site 3 0.002 0.15 0.92 Litter Biomass Treatment 1 1767 21.2 <0.0001 Transect 3 547 2.19 0.11 Site 1 1.62 0.02 0.89 Transect*Site 3 115 0.46 0.71 Trt*Transect 3 130 0.52 0.67 Trt*Site 1 12.0 0.14 0.71 Trt*Trans*Site 3 65.1 0.26 0.85 % Litter Loss Site 1 0.05 0.62 0.45 Transect 3 0.32 1.34 0.31 Site*Transect 3 0.09 0.36 0.78 135

PAGE 136

Table A-3. Full model results for main effects and interactions for lit ter carbon, nitrogen, and phosphorus concentration in Chapter 2. Effect s: treatment (pre and post fire), site (herbivory and non-herbivory), and transect (N, E, S, W). Variable Source DF Sums of Squares F ratio Prob > F Litter Carbon Concentration Treatment 1 314044 27.9 <0.0001 Transect 3 38216 1.13 0.35 Site 1 303 0.03 0.87 Transect*Site 3 19916 0.59 0.63 Trt*Transect 3 60583 1.79 0.17 Trt*Site 1 20076 1.78 0.19 Trt*Trans*Site 3 42893 1.27 0.30 Litter Nitrogen Concentration Treatment 1 52.2 3.66 0.06 Transect 3 72.4 1.69 0.19 Site 1 1.95 0.14 0.71 Transect*Site 3 33.4 0.78 0.51 Trt*Transect 3 46.2 1.08 0.37 Trt*Site 1 11.9 0.84 0.37 Trt*Trans*Site 3 39.4 0.92 0.44 Litter Phosphorus Concentration Treatment 1 752936 77.2 <0.0001 Transect 3 42237 1.44 0.25 Site 1 167401 17.2 0.0002 Transect*Site 3 102707 3.5 0.03 Trt*Transect 3 3376 0.12 0.95 Trt*Site 1 46122 4.73 0.04 Trt*Trans*Site 3 111402 3.81 0.02 136

PAGE 137

Table A-4. Full model results for main effects and interactions for lit ter carbon, nitrogen, and phosphorus storage in Chapter 2. Effects: treatment (pre and post fire), site (herbivory and non-herbivory), and tran sect (N, E, S, W). Variable Source DF Sums of Squares F ratio Prob > F Litter Carbon Storage Treatment 1 431115 13.1 0.001 Transect 3 152683 1.55 0.23 Site 1 637 0.02 0.89 Transect*Site 3 27282 0.28 0.84 Trt*Transect 3 87838 0.89 0.46 Trt*Site 1 5219 0.16 0.69 Trt*Trans*Site 3 26524 0.27 0.85 Litter Nitrogen Storage Treatment 1 75.0 5.48 0.03 Transect 3 94.9 2.31 0.10 Site 1 0.08 0.006 0.94 Transect*Site 3 16.7 0.41 0.75 Trt*Transect 3 43.7 1.06 0.38 Trt*Site 1 1.39 0.10 0.75 Trt*Trans*Site 3 22.9 0.56 0.65 Litter Phosphorus Storage Treatment 1 477 0.05 0.83 Transect 3 64501 2.06 0.13 Site 1 15121 1.45 0.24 Transect*Site 3 19291 0.62 0.61 Trt*Transect 3 10798 0.35 0.79 Trt*Site 1 6.18 0.0006 0.98 Trt*Trans*Site 3 39145 1.25 0.31 137

PAGE 138

Table A-5. Full model results for main effects and interactions for soil % moisture, bulk density, and organic matter in Chapter 2. Effects: treat ment (pre and post fi re), site (herbivory and non-herbivory), transect (N, E, S, W), and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F % Soil Moisture Treatment 1 0.05 369 <0.0001 Transect 3 0.0004 0.99 0.40 Site 1 0.0004 2.99 0.09 Depth 1 0.002 13.9 0.0004 Trt*Transect 3 0.00007 0.16 0.92 Trt*Site 1 0.0002 1.27 0.26 Trt*Depth 1 0.01 73.2 <0.0001 Transect*Site 3 0.0002 0.45 0.72 Transect*Depth 3 0.0003 0.72 0.54 Site*Depth 1 0.0005 3.58 0.06 Trt*Tran*Site 3 0.0003 0.66 0.58 Trt*Tran*Depth 3 0.00006 0.14 0.94 Trt*Site*Depth 1 0.000001 0.01 0.92 Bulk Density Treatment 1 0.004 0.23 0.64 Transect 3 0.22 3.91 0.01 Site 1 0.22 11.9 0.0009 Depth 1 3.05 163 < 0.0001 Trt*Transect 3 0.05 0.87 0.46 Trt*Site 1 0.02 1.08 0.30 Trt*Depth 1 0.17 9.21 0.003 Transect*Site 3 0.16 2.90 0.04 Transect*Depth 3 0.05 0.80 0.50 Site*Depth 1 0.02 1.31 0.26 Trt*Tran*Site 3 0.07 1.17 0.33 Trt*Tran*Depth 3 0.08 1.44 0.24 Trt*Site*Depth 1 0.004 0.22 0.64 Organic Matter Treatment 1 0.00004 0.13 0.72 Transect 3 0.0007 0.72 0.54 Site 1 0.002 4.75 0.03 Depth 1 0.002 5.29 0.02 Trt*Transect 3 0.002 2.54 0.06 Trt*Site 1 0.00007 0.23 0.64 Trt*Depth 1 0.002 7.01 0.01 Transect*Site 3 0.002 1.90 0.14 Transect*Depth 3 0.003 3.24 0.03 Site*Depth 1 0.000006 0.02 0.89 Trt*Tran*Site 3 0.002 2.51 0.07 Trt*Tran*Depth 3 0.001 1.03 0.38 Trt*Site*Depth 1 0.0006 1.82 0.18 138

PAGE 139

Table A-6. Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus concentration in Chapter 2. Effect s: treatment (pre and post fire), site (herbivory and non-herbivory), tr ansect (N, E, S, W), and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F Soil Carbon Concentration Treatment 1 86.4 7.45 0.008 Transect 3 61.1 1.76 0.16 Site 1 33.6 2.89 0.09 Depth 1 879 75.7 < 0.0001 Trt*Transect 3 18.0 0.52 0.67 Trt*Site 1 2.50 0.22 0.64 Trt*Depth 1 3.75 0.32 0.57 Transect*Site 3 23.3 0.67 0.57 Transect*Depth 3 41.3 1.19 0.32 Site*Depth 1 11.2 0.97 0.33 Trt*Tran*Site 3 21.3 0.61 0.61 Trt*Tran*Depth 3 11.2 0.32 0.81 Trt*Site*Depth 1 7.53 0.65 0.42 Soil Nitrogen Concentration Treatment 1 0.24 9.76 0.003 Transect 3 0.16 2.09 0.11 Site 1 0.16 6.20 0.02 Depth 1 2.50 99.6 < 0.0001 Trt*Transect 3 0.02 0.32 0.81 Trt*Site 1 0.02 0.63 0.43 Trt*Depth 1 0.13 5.15 0.03 Transect*Site 3 0.07 0.95 0.42 Transect*Depth 3 0.07 0.88 0.45 Site*Depth 1 0.07 2.64 0.11 Trt*Tran*Site 3 0.04 0.50 0.68 Trt*Tran*Depth 3 0.03 0.34 0.80 Trt*Site*Depth 1 0.02 0.64 0.42 Soil Phosphorus Concentration Treatment 1 0.46 33.9 <0.0001 Transect 3 0.06 1.58 0.20 Site 1 0.37 27.2 <0.0001 Depth 1 1.35 99.7 < 0.0001 Trt*Transect 3 0.006 0.14 0.93 Trt*Site 1 0.03 2.29 0.13 Trt*Depth 1 0.01 1.06 0.31 Transect*Site 3 0.05 1.15 0.33 Transect*Depth 3 0.10 2.35 0.08 Site*Depth 1 0.22 16.6 0.0001 Trt*Tran*Site 3 0.11 2.76 0.05 Trt*Tran*Depth 3 0.08 2.03 0.12 Trt*Site*Depth 1 0.12 9.14 0.004 139

PAGE 140

Table A-7. Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus storage in Chapter 2. Effects: treatment (pre and post fire), site (herbivory and non-herbivory), tr ansect (N, E, S, W), and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F Soil Carbon Storage Treatment 1 990057 19.3 <0.0001 Transect 3 273590 1.78 0.16 Site 1 332266 6.47 0.01 Depth 1 756964 14.7 0.0003 Trt*Transect 3 127783 0.83 0.48 Trt*Site 1 105413 2.05 0.16 Trt*Depth 1 4566 0.09 0.77 Transect*Site 3 240649 1.56 0.21 Transect*Depth 3 74137 0.48 0.70 Site*Depth 1 1817 0.04 0.85 Trt*Tran*Site 3 74813 0.49 0.69 Trt*Tran*Depth 3 37468 0.24 0.87 Trt*Site*Depth 1 96998 1.89 0.17 Soil Nitrogen Storage Treatment 1 1852 12.1 0.0008 Transect 3 478 1.04 0.38 Site 1 186 1.22 0.27 Depth 1 1102 7.21 0.009 Trt*Transect 3 113 0.25 0.86 Trt*Site 1 303 1.99 0.16 Trt*Depth 1 158 1.04 0.31 Transect*Site 3 463 1.01 0.39 Transect*Depth 3 82.2 0.18 0.91 Site*Depth 1 110 0.72 0.40 Trt*Tran*Site 3 54.7 0.12 0.95 Trt*Tran*Depth 3 41.5 0.09 0.97 Trt*Site*Depth 1 138 0.90 0.35 Soil Phosphorus Storage Treatment 1 5600965 37.2 <0.0001 Transect 3 173557 0.38 0.76 Site 1 563197 3.74 0.06 Depth 1 1234995 8.20 0.006 Trt*Transect 3 40432 0.09 0.97 Trt*Site 1 23821 0.16 0.69 Trt*Depth 1 1690 0.01 0.92 Transect*Site 3 711014 1.57 0.20 Transect*Depth 3 773409 1.71 0.17 Site*Depth 1 564850 3.75 0.06 Trt*Tran*Site 3 918276 2.03 0.12 Trt*Tran*Depth 3 975159 2.16 0.10 Trt*Site*Depth 1 648115 4.30 0.04 140

PAGE 141

Table A-8. Full model results for main effects and interactions for soil specifically mineralizable nitrogen and specifically mineralizable phos phorus in Chapter 2. Effects: treatment (pre and post fire), site (herbivory and nonherbivory), transect (N, E, S, W), and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F Specifically Mineralizable Nitrogen Treatment 1 21498 85.5 <0.0001 Transect 3 5756 7.63 0.0002 Site 1 919 3.65 0.06 Depth 1 116 0.46 0.50 Trt*Transect 3 3888 5.16 0.003 Trt*Site 1 1002 3.99 0.05 Trt*Depth 1 1909 7.59 0.007 Transect*Site 3 1411 1.87 0.14 Transect*Depth 3 1855 2.46 0.07 Site*Depth 1 1002 3.99 0.05 Trt*Tran*Site 3 708 0.94 0.43 Trt*Tran*Depth 3 1102 1.46 0.23 Trt*Site*Depth 1 231 0.92 0.34 Specifically Mineralizable Phosphorus Treatment 1 2.40 0.03 0.86 Site 1 445 6.22 0.02 Depth 1 16.5 0.23 0.64 Trt*Site 1 65.2 0.91 0.35 Trt*Depth 1 0.10 0.001 0.97 Site*Depth 1 5.61 0.08 0.78 Trt*Site*Depth 1 10.1 0.14 0.71 141

PAGE 142

Table A-9. Full model results for main effects a nd interactions for soil microbial biomass carbon in Chapter 2. Effects: treatment (pre and post fire), site (herbivory and nonherbivory), transect (N, E, S, W), and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F Microbial Biomass Carbon Treatment 1 3530 99.5 <0.0001 Transect 3 272 2.55 0.06 Site 1 164 4.63 0.03 Depth 1 968 27.3 <0.0001 Trt*Transect 3 327 3.07 0.03 Trt*Site 1 124 3.49 0.07 Trt*Depth 1 695 19.6 <0.0001 Transect*Site 3 62.4 0.58 0.63 Transect*Depth 3 15.9 0.15 0.93 Site*Depth 1 0.03 0.0008 0.98 Trt*Tran*Site 3 103 0.97 0.41 Trt*Tran*Depth 3 51.8 0.49 0.69 Trt*Site*Depth 1 7.18 0.20 0.65 Table A-10. Full model results for main effects a nd interaction for root biomass in Chapter 2. Effects: site (herbivory and non-herbivory) and depth (0-5 and 5-15cm). Variable Source DF Sums of Squares F ratio Prob > F Root Biomass Site 1 177893 13.6 0.002 Depth 1 1768 0.14 0.72 Site*Depth 1 112 0.009 0.93 142

PAGE 143

Chapter 3 Model Results Table A-11. Full model results for main effects and interactions for lit ter % moisture, litter biomass, and % litter loss in Chapter 3. Eff ects: treatment (pre and post fire), site (non-invaded, herbicide, and biol ogical), transect nested in site (1, 2, 3, 4, and 5), and plot nested in site (1, 2, 3, 4, and 5). Variable Source DF Sums of Squares F ratio Prob > F % Litter Moisture Treatment 1 3.45 679 <0.0001 Site 2 0.44 43.1 <0.0001 Transect [Site] 12 0.13 2.16 0.02 Plot [Site] 12 0.04 0.65 0.79 Trt*Transect [Site] 12 0.11 1.86 0.05 Trt*Plot [Site] 12 0.06 0.97 0.48 Trt*Site 2 0.13 13.0 <0.001 Litter Biomass Treatment 1 19.0 333 <0.0001 Site 2 5.18 45.5 <0.0001 Transect [Site] 12 2.15 3.15 0.0009 Plot [Site] 12 2.56 3.75 0.0001 Trt*Transect [Site] 12 0.97 1.42 0.17 Trt*Plot [Site] 12 1.35 1.98 0.04 Trt*Site 2 0.40 3.54 0.03 % Litter Loss Site 2 1232 4.88 0.01 Transect [Site] 12 676 0.45 0.93 Plot [Site] 12 1162 0.77 0.68 143

PAGE 144

Table A-12. Full model results for main effects and interactions for lit ter carbon, nitrogen, and phosphorus concentration in Chapter 3. Effect s: treatment (pre and post fire), site (non-invaded, herbicide, and biol ogical), transect nested in site (1, 2, 3, 4, and 5), and plot nested in site (1, 2, 3, 4, and 5). Variable Source DF Sums of Squares F ratio Prob > F Litter Carbon Concentration Treatment 1 979255 131 <0.0001 Site 2 27985 1.87 0.16 Transect [Site] 12 21510 2.40 0.01 Plot [Site] 12 49017 0.55 0.88 Trt*Transect [Site] 12 201359 2.24 0.02 Trt*Plot [Site] 12 66924 0.75 0.70 Trt*Site 2 89411 5.98 0.004 Litter Nitrogen Concentration Treatment 1 80.3 10.2 0.002 Site 2 221 14.1 <0.0001 Transect [Site] 12 298 3.16 0.001 Plot [Site] 12 99.2 1.05 0.41 Trt*Transect [Site] 12 266 2.82 0.003 Trt*Plot [Site] 12 51.6 0.55 0.88 Trt*Site 2 15.7 1.00 0.37 Litter Phosphorus Concentration Treatment 1 606802 111 <0.0001 Site 2 215633 19.7 <0.0001 Transect [Site] 12 63462 0.97 0.49 Plot [Site] 12 143380 2.18 0.02 Trt*Transect [Site] 12 75967 1.16 0.33 Trt*Plot [Site] 12 135775 2.07 0.03 Trt*Site 2 24430 2.23 0.11 144

PAGE 145

Table A-13. Full model results for main effects and interactions for lit ter carbon, nitrogen, and phosphorus storage in Chapter 3. Effects: treatment (pre and post fire), site (noninvaded, herbicide, and biological ), transect nested in site (1, 2, 3, 4, and 5), and plot nested in site (1, 2, 3, 4, and 5). Variable Source DF Sums of Squares F ratio Prob > F Litter Carbon Storage Treatment 1 6496 455 <0.0001 Site 2 2252 78.9 <0.0001 Transect [Site] 12 633 3.70 0.0002 Plot [Site] 12 456 2.72 0.005 Trt*Transect [Site] 12 323 1.89 0.05 Trt*Plot [Site] 12 323 1.89 0.05 Trt*Site 2 851 29.8 <0.0001 Litter Nitrogen Storage Treatment 1 108 289 <0.0001 Site 2 68.2 90.8 <0.0001 Transect [Site] 12 18.8 4.16 <0.0001 Plot [Site] 12 11.6 2.57 0.007 Trt*Transect [Site] 12 6.31 1.40 0.19 Trt*Plot [Site] 12 7.47 1.66 0.10 Trt*Site 2 12.2 16.2 <0.0001 Litter Phosphorus Storage Treatment 1 1187 148 <0.0001 Site 2 1073 66.8 <0.0001 Transect [Site] 12 410 4.25 <0.0001 Plot [Site] 12 234 2.43 0.01 Trt*Transect [Site] 12 115 1.2 0.30 Trt*Plot [Site] 12 114 1.19 0.31 Trt*Site 2 63.1 3.93 0.02 145

PAGE 146

Table A-14. Full model results for main effects and interactions for soil % moisture, bulk density, and organic matter in Chapter 3. E ffects: treatment (pre and post fire), site (non-invaded, herbicide, and biol ogical), transect nested in site (1, 2, 3, 4, and 5), plot nested in site (1, 2, 3, 4, and 5) and depth (0-5 and 5-15 cm). Variable Source DF Sums of Squares F ratio Prob > F % Soil Moisture Treatment 1 143 520 <0.0001 Transect [Site] 12 10.2 3.11 0.0004 Plot [Site] 12 8.17 2.48 0.005 Site 2 9.96 18.2 <0.0001 Depth 1 4.07 14.8 0.0002 Trt*Site 2 0.27 0.49 0.62 Trt*Depth 1 34.6 126 <0.0001 Site*Depth 2 2.65 4.83 0.009 Trt*Transect [Site] 12 3.28 1.00 0.45 Trt*Plot [Site] 12 3.41 1.04 0.42 Trt*Site*Depth 2 0.31 0.56 0.57 Bulk Density Treatment 1 0.47 15.5 0.0001 Transect [Site] 12 0.62 1.70 0.07 Plot [Site] 12 2.08 5.64 <0.0001 Site 2 2.05 33.4 <0.0001 Depth 1 10.5 343 <0.0001 Trt*Site 2 0.02 0.27 0.76 Trt*Depth 1 0.47 15.3 0.0001 Site*Depth 2 1.01 16.4 <0.0001 Trt*Transect [Site] 12 0.51 1.38 0.18 Trt*Plot [Site] 12 0.46 1.24 0.25 Trt*Site*Depth 2 0.003 0.05 0.95 Organic Matter Treatment 1 0.04 2.40 0.12 Transect [Site] 12 0.41 2.14 0.02 Plot [Site] 12 0.58 3.01 0.0007 Site 2 1.20 37.8 <0.0001 Depth 1 3.12 195 <0.0001 Trt*Site 2 0.09 2.72 0.07 Trt*Depth 1 0.08 4.84 0.03 Site*Depth 2 0.50 15.7 <0.0001 Trt*Transect [Site] 12 0.21 1.09 0.37 Trt*Plot [Site] 12 0.09 0.49 0.92 Trt*Site*Depth 2 0.04 1.25 0.29 146

PAGE 147

Table A-15. Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus concentration in Chapter 3. Effect s: treatment (pre and post fire), site (non-invaded, herbicide, and biol ogical), transect nested in site (1, 2, 3, 4, and 5), plot nested in site (1, 2, 3, 4, and 5) and depth (0-5 and 5-15 cm). Variable Source DF Sums of Squares F ratio Prob > F Soil Carbon Concentration Treatment 1 0.14 2.70 0.10 Transect [Site] 12 1.18 1.94 0.03 Plot [Site] 12 2.50 4.10 <0.0001 Site 2 2.07 20.4 <0.0001 Depth 1 13.6 266 <0.0001 Trt*Site 2 0.18 1.79 0.17 Trt*Depth 1 0.34 6.62 0.01 Site*Depth 2 0.73 7.14 0.001 Trt*Transect [Site] 12 0.38 0.63 0.81 Trt*Plot [Site] 12 0.54 0.89 0.56 Trt*Site*Depth 2 0.08 0.82 0.44 Soil Nitrogen Concentration Treatment 1 0.007 0.11 0.74 Transect [Site] 12 1.74 2.45 0.005 Plot [Site] 12 2.88 4.06 <0.0001 Site 2 3.15 26.6 <0.0001 Depth 1 7.41 125 <0.0001 Trt*Site 2 0.11 0.91 0.40 Trt*Depth 1 0.06 1.03 0.31 Site*Depth 2 0.79 6.71 0.002 Trt*Transect [Site] 12 0.53 0.75 0.70 Trt*Plot [Site] 12 0.64 0.90 0.55 Trt*Site*Depth 2 0.03 0.22 0.81 Soil Phosphorus Concentration Treatment 1 5.45 6.88 0.009 Transect [Site] 12 29.2 3.07 0.0005 Plot [Site] 12 46.1 4.85 <0.0001 Site 2 37.5 23.7 <0.0001 Depth 1 79.5 100 <0.0001 Trt*Site 2 0.10 0.06 0.94 Trt*Depth 1 0.46 0.59 0.45 Site*Depth 2 8.64 5.46 0.005 Trt*Transect [Site] 12 4.29 0.45 0.94 Trt*Plot [Site] 12 9.02 0.95 0.50 Trt*Site*Depth 2 1.03 0.65 0.52 147

PAGE 148

Table A-16. Full model results for main effects and interactions for soil carbon, nitrogen, and phosphorus storage in Chapter 3. Effects: treatment (pre and post fire), site (noninvaded, herbicide, and biological), transect nested in site (1, 2, 3, 4, and 5), plot nested in site (1, 2, 3, 4, and 5) and depth (0-5 and 5-15 cm). Variable Source DF Sums of Squares F ratio Prob > F Soil Carbon Storage Treatment 1 3816 0.06 0.80 Transect [Site] 12 2640419 3.73 <0.0001 Plot [Site] 12 2993650 4.23 <0.0001 Site 2 2021155 17.2 <0.0001 Depth 1 1178518 20.0 <0.0001 Trt*Site 2 306693 2.60 0.08 Trt*Depth 1 415874 7.06 0.009 Site*Depth 2 455101 3.86 0.02 Trt*Transect [Site] 12 582541 0.82 0.63 Trt*Plot [Site] 12 600287 0.85 0.60 Trt*Site*Depth 2 22031 0.19 0.83 Soil Nitrogen Storage Treatment 1 583 1.54 0.22 Transect [Site] 12 16472 3.64 <0.0001 Plot [Site] 12 21589 4.76 <0.0001 Site 2 24320 32.2 <0.0001 Depth 1 1.76 0.005 0.95 Trt*Site 2 668 0.88 0.41 Trt*Depth 1 522 1.38 0.24 Site*Depth 2 4019 5.32 0.006 Trt*Transect [Site] 12 2115 0.47 0.93 Trt*Plot [Site] 12 3137 0.69 0.76 Trt*Site*Depth 2 827 1.09 0.34 Soil Phosphorus Storage Treatment 1 4852024 25.2 <0.0001 Transect [Site] 12 9657604 4.19 <0.0001 Plot [Site] 12 9617087 4.17 <0.0001 Site 2 5228600 13.6 <0.0001 Depth 1 15766435 82.0 <0.0001 Trt*Site 2 503020 1.31 0.27 Trt*Depth 1 667172 3.47 0.06 Site*Depth 2 205794 0.54 0.59 Trt*Transect [Site] 12 916631 0.40 0.96 Trt*Plot [Site] 12 1603388 0.69 0.76 Trt*Site*Depth 2 868325 2.26 0.11 148

PAGE 149

Table A-17. Full model results for main eff ects and interactions for soil specifically mineralizable nitrogen and specifically mi neralizable phosphorus in Chapter 3. Effects: treatment (pre and post fire), s ite (non-invaded, herbic ide, and biological), transect nested in site (1, 2, 3, 4, and 5), plot nested in site (1, 2, 3, 4, and 5), and depth (0-5 and 5-15 cm). Variable Source DF Sums of Squares F ratio Prob > F Specifically Mineralizable Nitrogen Treatment 1 1782 8.37 0.004 Transect [Site] 12 4675 1.83 0.04 Plot [Site] 12 3426 1.34 0.20 Site 2 418 0.98 0.38 Depth 1 4251 20.0 <0.0001 Trt*Site 2 202 0.47 0.62 Trt*Depth 1 1187 5.57 0.02 Site*Depth 2 182 0.43 0.65 Trt*Transect [Site] 12 5835 2.28 0.009 Trt*Plot [Site] 12 1004 0.39 0.97 Trt*Site*Depth 2 638 1.5 0.23 Specifically Mineralizable Phosphorus Treatment 1 33.5 0.62 0.44 Site 2 347 3.20 0.05 Depth 1 197 3.64 0.06 Trt*Site 2 345 3.18 0.05 Trt*Depth 1 66.7 1.23 0.27 Site*Depth 2 563 5.19 0.01 Trt*Site*Depth 2 120 1.11 0.34 149

PAGE 150

Table A-18. Full model results for main effects and interactions for soil microbial biomass carbon in Chapter 3. Effects: treatment (pre and post fire), site (non-invaded, herbicide, and biological), transe ct nested in site (1, 2, 3, 4, and 5), plot nested in site (1, 2, 3, 4, and 5), and depth (0-5 and 5-15 cm). Variable Source DF Sums of Squares F ratio Prob > F Microbial Biomass Carbon Treatment 1 12.4 111 <0.0001 Transect [Site] 12 6.77 5.04 <0.0001 Plot [Site] 12 3.22 2.40 0.006 Site 2 1.85 8.28 0.0003 Depth 1 3.74 33.4 <0.0001 Trt*Site 2 1.55 6.91 0.001 Trt*Depth 1 0.16 1.43 0.23 Site*Depth 2 1.36 6.08 0.003 Trt*Transect [Site] 12 1.53 1.14 0.33 Trt*Plot [Site] 12 1.89 1.41 0.16 Trt*Site*Depth 2 0.93 4.13 0.02 150

PAGE 151

Chapter 4 Model Results Table A-19. Full model results for main effects a nd interactions for % mass loss, K value, and turnover time in Chapter 4. Effects: species ( Melaleuca quinquenervia, Pinus elliottii, and Taxodium distichum ), block (1, 2, and 3), and week (6, 12, 26, and 52). Variable Source DF Sums of Squares F ratio Prob > F % Mass Loss Species 2 11775 564 <0.0001 Block 3 6728 215 <0.0001 Week 2 389 18.6 <0.0001 Species*Block 6 616 9.84 <0.0001 Species*Week 4 197 4.73 0.002 Block*Week 6 277 4.43 0.008 Species*Blk*Wk 12 245 1.95 0.04 K Value Species 2 0.75 65.4 <0.0001 Block 2 0.09 8.31 0.003 Species*Block 4 0.004 0.19 0.94 Turnover Time Species 2 9.16 115 <0.0001 Block 2 1.47 18.5 <0.0001 Species*Block 4 0.28 1.76 0.18 151

PAGE 152

Table A-20. Full model results for main effect s and interactions for carbon, nitrogen, and phosphorus concentration in Chap ter 4. Effects: species ( Melaleuca quinquenervia, Pinus elliottii, and Taxodium distichum ), block (1, 2, and 3), and week (6, 12, 26, and 52). Variable Source DF Sums of Squares F ratio Prob > F Carbon Concentration Species 2 9789 54.4 <0.0001 Block 2 624 3.47 0.04 Week 3 767 2.84 0.04 Species*Block 4 338 0.94 0.45 Species*Week 6 1827 3.38 0.006 Block*Week 6 379 0.70 0.65 Species*Blk*Wk 12 1459 1.35 0.21 Nitrogen Concentration Species 2 827 626 <0.0001 Block 2 26.9 20.4 <0.0001 Week 3 79.8 40.2 <0.0001 Species*Block 4 5.56 2.10 0.09 Species*Week 6 33.0 8.33 <0.0001 Block*Week 6 17.6 4.43 0.0007 Species*Blk*Wk 12 11.9 1.51 0.14 Phosphorus Concentration Species 2 2358060 357 <0.0001 Block 2 161540 24.4 <0.0001 Week 3 265121 26.7 <0.0001 Species*Block 4 70364 5.32 0.0008 Species*Week 6 53100 2.68 0.02 Block*Week 6 250597 12.6 <0.0001 Species*Blk*Wk 12 75048 1.89 0.05 152

PAGE 153

Table A-21. Full model results for main effects a nd interactions for cha nge in carbon, nitrogen, and phosphorus storage in Chapter 4. Effects: species ( Melaleuca quinquenervia, Pinus elliottii, and Taxodium distichum ), block (1, 2, and 3), and week (6, 12, 26, and 52). Variable Source DF Sums of Squares F ratio Prob > F Change Carbon Storage Species 2 387 20.3 <0.0001 Block 2 199 10.5 0.0001 Week 3 13018 456 <0.0001 Species*Block 4 81.9 2.15 0.08 Species*Week 6 14420 252 <0.0001 Block*Week 6 113 1.97 0.08 Species*Blk*Wk 12 298 2.61 0.06 Change Nitrogen Storage Species 2 416 1.76 0.18 Block 2 634 2.68 0.08 Week 3 6735 19.0 <0.0001 Species*Block 4 188 0.40 0.81 Species*Week 6 11458 16.2 <0.0001 Block*Week 6 1087 1.53 0.18 Species*Blk*Wk 12 1734 1.22 0.28 Change Phosphorus Storage Species 2 14303 5.26 0.007 Block 2 35141 12.9 <0.0001 Week 3 73508 18.0 <0.0001 Species*Block 4 11054 2.03 0.10 Species*Week 6 49499 6.07 <0.0001 Block*Week 6 16327 2.00 0.08 Species*Blk*Wk 12 74049 4.54 <0.0001 153

PAGE 154

Table A-22. Full model results for main effects a nd interactions for litte r chemical composition in Chapter 4. Effects: species (Melaleuca quinquenervia, Pinus elliottii, and Taxodium distichum ), block (1, 2, and 3), and week (6, 12, 26, and 52). Variable Source DF Sums of Squares F ratio Prob > F % Soluble Fiber Species 2 5392 1326 <0.0001 Block 2 51.0 12.5 <0.0001 Week 3 1688 277 <0.0001 Species*Week 6 317 26.0 <0.0001 Block*Week 6 138 11.3 <0.0001 Block*Species 4 35.5 4.36 0.003 Species*Blk*Wk 12 107 4.36 <0.0001 % Lignin Species 2 3454 296 <0.0001 Block 2 44.0 3.77 0.03 Week 3 3784 216 <0.0001 Species*Week 6 572 16.3 <0.0001 Block*Week 6 141 4.03 0.002 Block*Species 4 56.8 2.43 0.06 Species*Blk*Wk 12 85.6 1.22 0.29 154

PAGE 155

Chapter 5 Model Results Table A-23. Full model results for main eff ects for non-woody plant species richness and diversity indices in Chapter 5. Effects: si te (non-invaded, herbicide, and biological), transect nested in site (1, 2, 3, 4, and 5), a nd plot nested in site (1, 2, 3, 4, and 5). Variable Source DF Sums of Squares F ratio Prob > F Non-woody Species Richness Site 2 99.9 7.63 0.001 Transect [Site] 12 71.2 0.91 0.55 Plot [Site] 12 192 2.45 0.01 Non-woody Shannon Index Site 2 4.08 10.3 0.0002 Transect [Site] 12 1.46 0.61 0.82 Plot [Site] 12 1.80 0.75 0.69 Non-woody Simpson Index Site 2 0.75 9.25 0.0004 Transect [Site] 12 0.54 1.10 0.38 Plot [Site[ 12 0.18 0.38 0.97 155

PAGE 156

Table A-24. Full model results for main effects for Melaleuca quinquenervia (MQ) live seedling, Melaleuca quinquenervia dead seedling, Pinus elliottii (PE) live seedling, and Taxodium distichum (TD) live seedling dens ities in Chapter 5. Effects: site (noninvaded, herbicide, and biological ), transect nested in site (1, 2, 3, 4, and 5), and plot nested in site (1, 2, 3, 4, and 5). Variable Source DF Sums of Squares F ratio Prob > F MQ Live Seedlings Site 2 6434 22.0 <0.0001 Transect [Site] 12 2136 1.22 0.30 Plot [Site] 12 3695 2.10 0.04 MQ Dead Seedlings Site 2 2924 7.30 0.002 Transect [Site] 12 5236 2.18 0.03 Plot [Site] 12 4368 1.82 0.07 PE Seedlings Site 2 0.33 0.89 0.42 Transect [Site] 12 3.96 1.75 0.09 Plot [Site] 12 3.56 1.58 0.13 TD Seedlings Site 2 0.66 4.85 0.01 Transect [Site] 12 0.00 0.00 1.00 Plot [Site] 12 0.80 0.98 0.48 156

PAGE 157

LIST OF REFERENCES Andersen, J. M. 1976. Ignition method for determination of total phosphorus in lake sediments. Water Research 10:329-331. Anonymous. 1999. Ten years of successful me laleuca management in Florida: 1988 1998. Florida Exotic Pest Plant Council. Anonymous. 2004. Project management plan: Melale uca eradication and ot her exotic plants. United States Army Corps of Engineers and the South Florid a Water Management District. Anonymous. 2007. Weed Alert: Melaleuca ( Melaeuca quinquenervia ). Department of Environmental Protection, Bureau of Invasive Plant Management. Balciunas, J. K., D. W. Burrows, and M. F. Pu rcell. 1994. Field and laboratory host ranges of the Australian weevil, Oxyops vitiosa (Coleoptera, Curculionid ae), a potential biological control agent for the paperbark tree, Melaleuca quinquenervia Biological Control 4 :351360. Bardgett, R. D., and D. A. Wardle. 2003. He rbivore-mediated linkages between aboveground and belowground communities. Ecology 84:2258-2268. Baruch, Z., and G. Goldstein. 1999. Leaf construc tion cost, nutrient concentration, and net CO2 assimilation of native and invasive species in Hawaii. Oecologia 121:183-192. Bay, R. F., and A. A. Sher. 2008. Success of active revegetation after tamarix removal in riparian ecosytems of the southwestern United St ates: A quantitative assessment of past restoration projects. Restoration Ecology 16:113-128. Berg, B. 2000. Litter decomposition and organic matter turnover in northern forest soils. Forest Ecology and Management 133:13-22. Beyers, D. W. 1998. Causal inference in enviro nmental impact studies. Journal of the North American Benthological Society 17 :367-373. Blumenthal, D. M. 2006. Interactions between res ource availability and en emy release in plant invasion. Ecology Letters 9:887-895. Bodel, M. J., A. Ferriter, and D. D. Thayer. 1994. The biology, distribution, and ecological consequences of Melaleuca quinquenervia in the Everglades. Pages 341-355 in S. M. Davis and J. C. Ogden, editors. Everglades, The Ecosystem and Its Restoration. St. Lucie Press, Delray Beach, Florida. Boland, D. J., M. I. H. Brooker, C. M. Chippendale, N. Hall, B. P. M. Hyland, and R. D. Johnson. 1987. Forest trees of Australia. Ne lson Wadsworth Publisher, Melbourne, Australia. 157

PAGE 158

Burrows, D. W., and J. K. Balciuna s. 1999. Host-range and distribution of Eucerocoris suspectus (Hemiptera : Miridae), a potential biologi cal control agent for the paperbark tree Melaleuca quinquenervia (Myrtaceae). Environmental Entomology 28:290-299. Carson, C. F., K. A. Hammer, and T. V. Riley. 2006. Melaleuca alternifolia (tea tree) oil: a review of antimicrobial and other medicina l properties. Clinical Microbiology Reviews 19:50-59. Carter, M. C., and C. D. Foster. 2004. Prescrib ed burning and productivity in southern pine forests: a review. Forest Ecology and Management 191:93-109. Center, T. D., F. A. Dray, G. P. Jubinsky, and M. J. Grodowitz. 1999. Biological control of water hyacinth under conditions of maintenance management: Can herbicides and insects be integrated? Environmental Management 23:241-256. Center, T. D., and M. P. Hill. 2002. Field efficacy and predicted host range of the pickerelweed borer, Bellura densa a potential biological control ag ent of water hyacinth. Biocontrol 47:231-243. Center, T. D., P. D. Pratt, P. W. Tipping, M. B. Rayamajhi, T. K. Van, S. A. Wineriter, F. A. Dray, and M. Purcell. 2006. Field colonizat ion, population growt h, and dispersal of Boreioglycaspis melaleucae Moore, a biological contro l agent of the invasive tree Melaleuca quinquenervia (Cav.) Blake. Biological Control 39 :363-374. Certini, G. 2005. Effects of fire on propert ies of forest soils: a review. Oecologia 143 :1-10. Chapin, F. S., P. A. Matson, and H. A. Mooney. 2002. Principles of ecosystem ecology. Springer-Verlag New York, Inc., New York, New York. Chapman, S. K., S. C. Hart, N. S. Cobb, T. G. Whitham, and G. W. Koch. 2003. Insect herbivory increases litter quality and decomposition: An extension of the acceleration hypothesis. Ecology 84:2867-2876. Chapman, S. K., J. A. Schweitzer, and T. G. Whitham. 2006. Herbivory differentially alters plant litter dynamics of evergreen and deciduous trees. Oikos 114:566-574. Costello, S. L., P. D. Pratt, M. B. Rayamajhi, and T. D. Center. 2003. Arthropods associated with above-ground portions of the invasive tree, Melaleuca quinquenervia in South Florida, USA. Florida Entomologist 86:300-322. Couteaux, M. M., P. Bottner, and B. Berg. 1995. L itter decomposition, climate and litter quality. Trends in Ecology & Evolution 10 :63-66. D'Antonio, C., and L. A. Meyerson. 2002. Exotic plant species as problems and solutions in ecological restoration: A s ynthesis. Restoration Ecology 10:703-713. 158

PAGE 159

Davet, P. 2004. Microbial ecology of the soil and pl ant growth. Science Publishers, Inc., Enfield, New Hampshire. David, P. G. 1999. Response of exotics to rest ored hydroperiod at Dupuis Reserve, Florida. Restoration Ecology 7:407-410. Davis, M. A., J. P. Grime, and K. Thompson. 2000. Fluctuating resources in plant communities: a general theory of invasi bility. Journal of Ecology 88:528-534. Denslow, J. S., and C. M. D'Antonio. 2005. After biocontrol: Assessing indire ct effects of insect releases. Biological Control 35:307-318. DiStefano, J. F., and R. F. Fisher. 1983. Invasion potential of Melaleuca quinquenervia in Southern Florida, USA. Forest Ecology and Management 7:133-141. Doren, R. F. 2002. Weed's Won't Wait; Part Thre e, Executive summary of the assessment. A report to the South Florida Ecosystem Re storation Task force and Working Group produced through the support a nd partnership of the Agency members of the Noxious Exotic Weed Task Team. Dray, F. A., B. C. Bennett, and T. D. Center. 2006. I nvasion history of Melaleuca quinquenervia (Cav.) S.T. Blake in Florida. Castanea 71:210-225. Ehrenfeld, J. G. 2003. Effects of exotic plant invasions on so il nutrient cycling processes. Ecosystems 6:503-523. Eviner, V. T., and F. S. Chapin. 2003. Biogeoc hemical interaction a nd biodiversity. Pages 151173 in J. M. Melillo, C. B. Field, and B. Mo ldan, editors. Interac tions of the major biogeochemical cycles, global change and human impacts. Island Press, Washington, D.C. Ewel, J. J. 1986. Invasibility: Lessons from South Florida. in H. A. Mooney and J. A. Drake, editors. Ecology of Biological Invasions of North America and Hawaii. Springer-Verlag New York, Inc., New York, New York. Fenn, M. E., J. S. Baron, E. B. Allen, H. M. Ruet h, K. R. Nydick, L. Geiser, W. D. Bowman, J. O. Sickman, T. Meixner, D. W. Johnson, and P. Neitlich. 2003. Ecological effects of nitrogen deposition in the west ern United States. Bioscience 53:404-420. Ferriter, A., D. Thayer, C. Goodyear, R. Doren, K. Langeland, and J. Lane. 2005. Chapter 9: Invasive exotic species in the South Fl orida environment. South Florida Water Management's Annual Report. Findlay, S., P. Groffman, and S. Dye. 2003. Effects of Phragmites australis removal on marsh nutrient cycling. Wetlands Ecology and Management 11:157-165. 159

PAGE 160

Fine, P. V. A., I. Mesones, and P. D. Cole y. 2004. Herbivores promote habitat specialization by trees in Amazonian forests. Science 305 :663-665. Franks, S. J., A. M. Kral, and P. D. Pratt. 2006. Herbivory by introduced insects reduces growth and survival of Melaleuca quinquenervia seedlings. Environmental Entomology 35:366372. Franzluebbers, A. J. 1999. Potential C and N mine ralization and microbial biomass from intact and increasingly disturbed soils of vary ing texture. Soil Biology & Biochemistry 31:1083-1090. Frost, C. J., and M. D. Hunter. 2007. Recycling of nitrogen in herbivore feces: plant recovery, herbivore assimilation, soil retention, and leaching losses. Oecologia 151 :42-53. Gholz, H. L., C. S. Perry, W. P. Cropper, and L. C. Hendry. 1985. Litte rfall, decomposition, and nitrogen and phosphorus dynamics in a chronosequence of slash pine ( Pinus elliottii ) plantations. Forest Science 31:463-478. Giardina, C. P., and C. C. Rhoades. 2001. Cl ear cutting and burning affect nitrogen supply, phosphorus fractions and seedling growth in soils from a Wyoming lodgepole pine forest. Forest Ecology and Management 140:19-28. Greenway, M. 1994. Litter accession and accumulation in a Melaleuca quinquenervia (Cav) Blake, S.T. wetland in southeastern Queensland. Australian Journal of Marine and Freshwater Research 45 :1509-1519. Grierson, P. F., N. B. Comerford, and E. J. Jokela. 1999. Phosphorus mineralization and microbial biomass in a Florida Spodosol: e ffects of water potential, temperature and fertilizer application. Biology and Fertility of Soils 28:244-252. Gunderson, M. 1994. Vegetation of the everglades, determinants of community composition. Pages 323-340 in S. M. Davis and J. C. Ogden, ed itors. Everglades, The Ecosystem and Its Restoration. St. Lucie Press, Delray Beach, Florida. Heard, T. A., and S. L. Winterton. 2000. Interactions between nutrient status and weevil herbivory in the biological control of water hyacinth. Journal of Applied Ecology 37:117127. Hughes, R. F., J. B. Kauffman, and D. L. Cummings. 2000. Fire in the Brazilian Amazon 3: Dynamics of biomass, C, and nutrient pools in regenerating forests. Oecologia 124 :574588. Hulme, P. E., and E. T. Bremner. 2006. Assessing the impact of Impatiens glandulifera on riparian habitats: partitioning diversity com ponents following species removal. Journal of Applied Ecology 43:43-50. 160

PAGE 161

Hurlbert, S. H. 1984. Pseudoreplication and th e design of ecological field experiments. Ecological Monographs 54:187-211. Jackson, R. B., J. L. Banner, E. G. Jobbagy, W. T. Pockman, and D. H. Wall. 2002. Ecosystem carbon loss with woody plant inva sion of grasslands. Nature 418:623-626. Katterer, T., M. Reichstein, O. Andren, and A. Lomander. 1998. Temperature dependence of organic matter decomposition: a critical re view using literature data analyzed with different models. Biology and Fertility of Soils 27 :258-262. Kauffman, J. B., R. L. Sanford, D. L. Cummi ngs, I. H. Salcedo, and E. Sampaio. 1993. Biomass and nutrient dynamics associated with slash fires in neotropical dry forests. Ecology 74:140-151. Kaufman, S. R., and P. E. Smouse. 2001. Compar ing indigenous and introduced populations of Melaleuca quinquenervia (Cav.) Blake: response of s eedlings to water and pH levels. Oecologia 127 :487-494. Keane, R. M., and M. J. Crawley. 2002. Exotic pl ant invasions and the enemy release hypothesis. Trends in Ecology & Evolution 17 :164-170. Kohli, R. K., K. S. Dogra, D. R. Batish, and H. P. Singh. 2004. Impact of invasive plants on the structure and composition of natural vegetati on of northwestern Indian Himalayas. Weed Technology 18:1296-1300. Kourtev, P. S., J. G. Ehrenfeld, and M. Haggblom 2003. Experimental analysis of the effect of exotic and native plant species on the st ructure and function of soil microbial communities. Soil Biology & Biochemistry 35:895-905. LaRoche, F. B., and A. P. Ferriter. 1992. The rate of expansion of melaleuca in South Florida. Journal of Aquatic Plant Management 30:62-65. Lopez-Zamora, I., N. B. Comerford, and R. M. Muchovej. 2004. Root development and competitive ability of the invasive species Melaleuca quinquenervia (Cav.) ST Blake in the South Florida flatwoods. Plant and Soil 263:239-247. Luczak, C., M. A. Janquin, and A. Kupka. 1997. Si mple standard procedure for the routine determination of organic matter in marine sediment. Hydrobiologia 345 :87-94. MacDougall, A. S., and R. Turkington. 2007. Does th e type of disturbance matter when restoring disturbance-dependent gra sslands? Restoration Ecology 15:263-272. Mack, M. C., and C. M. D'Antonio. 1998. Impacts of biological invasions on disturbance regimes. Trends in Ecology & Evolution 13:195-198. 161

PAGE 162

Mack, M. C., and C. M. D'Antonio. 2003. The effect s of exotic grasses on litter decomposition in a Hawaiian woodland: The importance of indirect effects. Ecosystems 6:723-738. Mack, R. N., D. Simberloff, W. M. Lonsdale H. Evans, M. Clout, and F. A. Bazzaz. 2000. Biotic invasions: Causes, epidemiology, globa l consequences, and control. Ecological Applications 10:689-710. Marrs, R. H., K. Galtress, C. Tong, E. S. Cox, S. J. Blackbird, T. J. Heyes, R. J. Pakeman, and M. G. Le Duc. 2007. Competing conservation go als, biodiversity or ecosystem services: Element losses and species recruitment in a managed moorland-bracken model system. Journal of Environmental Management 85:1034-1047. Martin, M. R., P. W. Tipping, a nd J. O. Sickman. 2009. Invasion by an exotic tree alters above and belowground ecosystem compon ents. Biological Invasions DOI 10.1007/s10530008-9366-3. Mason, T. J., and K. French. 2007. Management regi mes for a plant invader differentially impact resident communities. Biological Conservation 136:246-259. Mazzotti, F. J., T. D. Center, F. A. Dray, a nd D. D. Thayer. 1997. Ecological consequences of invasion by Melaleuca quinquenervia in South Florida wetlands: Paradise damaged, not lost. University of Florida, Institute of Food and Agricu ltural Sciences, Cooperative Extension Service Bulletin SS-WEC-123. Mazzotti, F. J., W. Ostrenko, and A. T. Smith. 1981. Effects of the exotic plants Melaleuca quinquenervia and Casuarina equisetifolia on small mammal populations in the Eastern Florida Everglades. Florida Scientist 44:65-71. Meskimen, G. F. 1962. A silvicultural study of the me laleuca tree in South Florida. Univeristy of Florida, Gainesville, Florida. Mitchell, C. E., and A. G. Power. 2003. Releas e of invasive plants from fungal and viral pathogens. Nature 421:625-627. Myers, R. L. 1983. Site susceptibility to invasion by the exotic tree Melaleuca quinquenervia in southern Florida. Journal of Applied Ecology 20:645-658. Myers, R. L. 1984. Ecological compression of Taxodium distichum var. nutans by Melaleuca quinquenervia in southern Florida. Pages 358-364 in K. C. Ewel and H. T. Odum, editors. Cypress swamps. University Presse s of Florida, Gain esville, Florida. Neary, D. G., C. C. Klopatek, L. F. De Bano, and P. F. Folliott. 1999. Fire effects on belowground sustainability: a review and synthesis. Forest Ecology and Management 122:51-71. 162

PAGE 163

Nessel, J. K., and E. Bayley. 1984. Distributi on and dynamics of organic matter and phosphorus in a sewage-enriched cypress swamp. Pages 262-278 in K. C. Ewel and H. T. Odum, editors. Cypress swamps. University Presse s of Florida, Gain esville, Florida. Ogden, J. A. E., and M. Rejmanek. 2005. Recovery of native plant commun ities after the control of a dominant invasive plant species, Foeniculum vulgare : Implications for management. Biological Conservation 125:427-439. Peet, R. K. 1974. The measurement of species diversity. Annual Review of Ecology and Systematics 5:285-307. Pidgeon, R. W. J., and S. C. Cairns. 1981. Deco mposition and colonization by invertebrates of native and exotic leaf material in a sm all stream in New-England (Australia). Hydrobiologia 77:113-127. Prieto-Fernandez, A., M. J. Acea, and T. Carb allas. 1998. Soil microbial and extractable C and N after wildfire. Biology and Fertility of Soils 27:132-142. Purcell, M. F., and J. K. Balciunas. 1994. Life history and distribution of the Australian weevil Oxyops vitiosa (Coleoptera, Curculionidae), a pot ential biological control agent for Melaleuca quinquenervia (Myrtaceae). Annals of the Entomological Society of America 87:867-873. Purcell, M. F., J. K. Balciunas, a nd P. Jones. 1997. Biology and host-range of Boreioglycaspis melaleucae (Hemiptera: Psyllidae), potent ial biological control agent for Melaleuca quinquenervia (Myrtaceae). Environmental Entomology 26:366-372. Ramey, V., and B. Hassell. 2005. Management in Fl orida waters. University of Florida and the Florida Department of Environmental Protecti on, editors. University Presses of Florida, Gainesville, Florida. Rayamajhi, M. B., T. K. Van, T. D. Center, J. A. Goolsby, P. D. Pratt, and A. Racelis. 2002. Biological attributes of the canopy-held mela leuca seeds in Australia and Florida, US. Journal of Aquatic Plant Management 40:87-91. Room, P. M., K. L. S. Harley, I. W. Forno, and D. P. A. Sands. 1981. Successful biologicalcontrol of the floating weed salvinia. Nature 294:78-80. Ross, D. J., T. W. Speir, K. R. Tate, and C. W. Feltham. 1997. Burning in a New Zealand snowtussock grassland: Effects on soil microbi al biomass and nitrogen and phosphorus availability. New Zeal and Journal of Ecology 21:63-71. Rothstein, D. E., P. M. Vitousek, and B. L. Simmons. 2004. An exotic tree alters decomposition and nutrient cycling in a Hawaiia n montane forest. Ecosystems 7:805-814. 163

PAGE 164

Rowland, A. P., and J. D. Roberts. 1994. Ligni n and cellulose fractionation in decomposition studies using acid-detergent fiber methods. Communications in So il Science and Plant Analysis 25:269-277. Serbesoff-King, K. 2003. Melaleuca in Florid a: A literature review on the taxonomy, distribution, biology, ecology, economic importa nce and control measures. Journal of Aquatic Plant Management 41:98-112. SERC, S. R. C. C. 2007. Naples Florida (086078) Period of Reco rd Monthly Climate Summary. Shea, K., and P. Chesson. 2002. Community ecology theory as a framework for biological invasions. Trends in Ecology & Evolution 17:170-176. Smith, S. M., P. V. McCormick, J. A. Leeds, and P. B. Garrett. 2002. Constraints of seed bank species composition and water de pth for restoring vegetation in the Florida Everglades, USA. Restoration Ecology 10:138-145. Standish, R. J., P. A. Williams, A. W. Robert son, N. A. Scott, and D. I. Hedderley. 2004. Invasion by a perennial herb in creases decomposition rate and alters nutrient availability in warm temperate lowland forest remnants. Biological Invasions 6:71-81. Stocker, R. K., and D. R. Sande rs Sr. 1981. Chemical control of Melaleuca quinquenervia Pages 129-134 in Proceedings of Melaleuca Symposium, Florida Department of Agriculture and Consumer Services Division of Forestry, Tallahassee, Florida. Tipping, P. W., M. R. Martin, K. R. Nimmo, R. M. Pierce, M. D. Smart, E. White, P. T. Madeira, and T. D. Center. 2009. Invasion of a West Everglades wetland by Melaleuca quinquenervia countered by classical biological control. Biological Control 48:73-78. Tipping, P. W., M. R. Martin, P. D. Pratt, T. D. Center, and M. B. Rayamajhi. 2008. Suppression of growth and reproduction of an exotic invasive tree by two introduced insects. Biological Control 44 :235-241. USDA, N. 1998. Soil Survey of Collier County Area, Florida. United State Department of Agriculture, Natural Res ource Conservation Service. Van, T. K., M. B. Rayachhetry, and T. D. Center. 1998. Reproductive ecology of melaleuca ( Melaleuca quinquenervia ) in south Florida. Weed Science Society of America 38:23. Van, T. K., M. B. Rayachhetry, T. D. Cent er, and P. D. Pratt. 2002. Litter dynamics and phenology of Melaleuca quinquenervia in South Florida. Jo urnal of Aquatic Plant Management 40:22-27. Vance, E. D., P. C. Brookes, and D. S. Je nkinson. 1987. An extraction method for measuring soil microbial biomass-C. Soil Biology & Biochemistry 19:703-707. 164

PAGE 165

Wade, D., J. Ewel, and R. Hofstetter. 1980. Fire in South Florida ecosystems. Southeastern Florida Experimental Station. Wan, S. Q., D. F. Hui, and Y. Q. Luo. 2001. Fi re effects on nitrogen pools and dynamics in terrestrial ecosystems: A meta-analysis. Ecological Applications 11:1349-1365. Wanthongchai, K., J. Bauhus, and J. G. Goldam mer. 2008. Nutrient losses through prescribed burning of aboveground litter and un derstorey in dry dipterocarp forests of different fire history. Catena 74:321-332. Wardle, D. A., G. W. Yeates, W. M. Williamson, K. I. Bonner, and G. M. Barker. 2004. Linking aboveground and belowground communities: the indirect influence of aphid species identity and diversity on a three trophic le vel soil food web. Oikos 107:283-294. Wheeler, G. S. 2003. Minimal increase in larval and adult performance of the biological control agent Oxyops vitiosa when fed Melaleuca quinquenervia leaves of different nitrogen levels. Biological Control 26:109-116. Wheeler, G. S., L. M. Massey, and I. A. Sout hwell. 2002. Antipredator defense of biological control agent Oxyops vitiosa is mediated by plant volatile s sequestered from the host plant Melaleuca quinquenervia Journal of Chemical Ecology 28:297-315. White, J. R., and K. R. Reddy. 2000. Influence of phosphorus loading on organic nitrogen mineralization of everglades soils. Soil Science Society of America Journal 64:15251534. Wienand, K. T., and W. D. St ock. 1995. Long-term phosphorus fertilization effects on the litter dynamics of an age sequence of Pinus elliottii plantations in the s outhern cape of South Africa. Forest Ecology and Management 75:135-146. Wineriter, S. A., G. R. Buckingham, and J. H. Frank. 2003. Host range of Boreioglycaspis melaleucae Moore (Hemiptera : Psyllidae), a potential biocontrol agent of Melaleuca quinquenervia (Cav.) S.T. Blake (Myrtaceae), under quarantine. Biological Control 27:273-292. Yurkonis, K. A., S. J. Meiners, and B. E. Wachholder. 2005. Invasion impacts diversity through altered community dynamics. Journal of Ecology 93:1053-1061. 165

PAGE 166

BIOGRAPHICAL SKETCH Melissa Rosemary Martin was born and ra ised in South Bend, Indiana where she developed a love of nature at an early age. As an undergraduate at the University of Notre Dame, Melissa participated in research on the plan t community structure and biogeochemistry of wetlands. After graduation in 2002, Melissa acce pted an internship through the Student Conservation Association at the USDA-ARS I nvasive Plant Research Laboratory in Fort Lauderdale, FL. Through this internship, she was introduced to research on the management and control of invasive exotic plan ts. Melissa continued her studies at the University of Florida, investigating ecosystem-level eff ects of the invasion of exotic pl ants in order to develop more effective management and restoration techniqu es. She received her M.S. degree from the department of Soil and Water Science in May of 2006 and her Ph.D. from the same department in May of 2009. She looks forward to working in a group that uses scientific studies to aid in the active management of natural systems. 166