<%BANNER%>

Aspects of the Invasion and Management of Japanese Climbing Fern (Lygodium japonicum) in Southeastern Forests

HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Abstract
 Literature review
 Assessment of site variables associated...
 Herbicide efficacy for management...
 Conclusions and recommendations...
 Appendix: Initial analysis tables...
 References
 Biographical sketch
 

PAGE 1

ASPECTS OF THE INVASION AND MANA GEMENT OF JAPANESE CLIMBING FERN ( Lygodium japonicum ) IN SOUTHEASTERN FORESTS By ANDREA N. VAN LOAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Andrea N. Van Loan

PAGE 3

To my mother and father, for leading by exampl e, and to my husband for his partnership

PAGE 4

iv ACKNOWLEDGMENTS I would like to express my sincere ap preciation to my graduate committee members, who have provided the support and freedom necessary for me to successfully balance my professional, academic, and personal commitments, all of which have been major components of a life-changing period of time spent as a graduate student at the University. First, I thank Dr. Jarek Nowak, who initially chaired my graduate committee, and who recognized th e real priorities in my life and allowed me time to devote to them. Dr. Eric Jokela, who began as a committee member, and finished as my committee chair; and gave me the necessary s upport to move me to completion at a very critical stage. Dr. Greg MacDonald, Agronomy Department representative, has consistently provided friendshi p, support, and availability at short notice. Dr. Doria Gordon, who provided review, comment, and eco logical perspective to my efforts, as well as a dose of reality when stumbling bl ocks revealed themselves. And finally, Dr Alan Long, who graciously agreed to join my committee very late in the process. Invaluable support was provide d by other personnel at the School of Forest Resources and Conservation, including Cherie Arias, and Dr. Wendell Cropper. This process would not have been possibl e without the personal and professional support and flexibility provided by my co-worke rs, especially those within the Florida Division of Forestrys Forest Health Section, and the Florida Division of Plant Industry in Gainesville. In particular, my supervisor Dr. Ed Barnard, who demonstrated how to

PAGE 5

v truly support staff advancement in consiste ntly approving my flexible work schedules, and leave requests made for field research and academic needs. The field component of this research would not have been possible without the interest and assistance of several individua ls and entities. Critical support for the comparative herbicide trials was provided by Jerry Wyrick, and Wyrick and Sons Pine Straw, and Mr. Elton Headings. Key suppor t for the comparative invasion assessment was provided by: Mr. Dan Miller; the Northw est Florida Water Management District with assistance from Mr. John Valenta; and the Florida Division of Forestry, Blackwater River State Forest with assistance from Mr Tom Cathey, and Mr. Henry Thompson. This work would not have been possible without support from two funding sources. Research on the impact of site environment on Japanese climbing fern invasion levels was supported by a grant from the Univers ity of Florida, Institute of Food and Agricultural Sciences, Center for Aquatic a nd Invasive Plants, Invasive Plant Working Group. Herbicide research was supported by a grant from the USDA Forest Service Special Technology Development Progr am (Project # 03-DG-11083112-040). The role and importance of my family cannot be understated as in the end, all other aspects of life paled in comparison over th e last few years. My husband and partner, Chris Van Loan, supported me both at home, and as my research assistant for a significant portion of my field work; providi ng much-needed labor and assistance despite extreme temperatures, swarming mosquitos, and the relative un-attractiveness of weekends spent in Calhoun County. My mother, Barbara Christman, showed me ultimate perseverance and grace throughout her lif e. That lesson in perseverance became a touchstone for the writing of this thesis. My father, Harry Christman, showed me the

PAGE 6

vi strength of love as a caregiver in so ma ny senses of the word; allowing me enough peace of mind to continue to make forward progress with this process. While I already see the intrinsic rewards of the academic and rese arch training I have received, these human lessons have been the most humbling and significant of this experience.

PAGE 7

vii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES..........................................................................................................xii ABSTRACT.....................................................................................................................xi ii CHAPTER 1 LITERATURE REVIEW: BIOLOGY, ECOLOGY, DISPERSAL, ECONOMIC IMPACT, AND MANAGEMENT OF JAPANESE CLIMBING FERN....................1 Introduction................................................................................................................... 1 Biology........................................................................................................................ .2 Rhizomes...............................................................................................................2 Leaves....................................................................................................................2 Spores....................................................................................................................5 Gametophytes/Mating System...............................................................................8 Sporophytes...........................................................................................................9 Ecology.......................................................................................................................1 0 Native Range.......................................................................................................10 Introduced Range.................................................................................................10 Environmental Requirements..............................................................................12 Impacts................................................................................................................14 Vines as Invasive Plants......................................................................................16 Ferns as Invasive Plants.......................................................................................17 Related Species in the United States...................................................................18 Recognition and Management in the United States....................................................20 Management Status.............................................................................................20 Regulatory Status.................................................................................................20 Economic Impact.................................................................................................21 Management Practices.........................................................................................23 Prevention.....................................................................................................23 Chemical control..........................................................................................24 Mechanical control.......................................................................................26 Biological control.........................................................................................26 Prescribed fire treatment..............................................................................28

PAGE 8

viii Conclusion..................................................................................................................29 2 ASSESSMENT OF SITE VARIABLES ASSOCIATED WITH JAPANESE CLIMBING FERN OCCURRENCE IN THREE NORTH FLORIDA FORESTS...32 Introduction.................................................................................................................32 Methods and Materials...............................................................................................34 Site Selection.......................................................................................................34 Plot Layout..........................................................................................................34 Site Descriptions..................................................................................................35 Macroplot Measurements....................................................................................35 Intensive Sampling Plot Measurements..............................................................36 Data Analysis.......................................................................................................37 Initial Analyses....................................................................................................39 Discriminant Analysis.........................................................................................40 Regression Analysis............................................................................................40 Neural Network Analysis....................................................................................41 Results........................................................................................................................ .41 Site-Level Occurrence.........................................................................................41 Plot-Level Occurrence.........................................................................................42 Presence/Absence................................................................................................42 Cover Class Levels..............................................................................................44 Discriminant Analysis.........................................................................................45 Regression Analysis............................................................................................46 Neural Network Analysis....................................................................................47 Discussion...................................................................................................................48 Site Overview......................................................................................................49 Discriminant Analysis.........................................................................................50 Plot environmental variables........................................................................50 Plot tree species............................................................................................51 Regression Analysis............................................................................................53 Neural Network Analysis...................................................................................56 Conclusions.................................................................................................................57 3 HERBICIDE EFFICACY FOR MANAGeMENT OF JAPANESE CLIMBING FERN IN SOUTHEASTERN FORESTS..................................................................71 Introduction.................................................................................................................71 Forestry and Invasive Plant Impacts....................................................................72 Management Techniques.....................................................................................73 Materials and Methods...............................................................................................74 Study Area...........................................................................................................74 Herbicide Selection.............................................................................................75 Experimental Design...........................................................................................75 Treatment and Assessment Schedule..................................................................76 Data Analysis.......................................................................................................76 Results and Discussion...............................................................................................77

PAGE 9

ix Phenology of Japanese Climbing Fern................................................................78 Multiple Herbicide Trial: 2002 and 2004............................................................79 Short Term Treatment Effects.............................................................................79 Timber Harvest Effects........................................................................................82 Long Term Treatment Effects.............................................................................82 2002 Glyphosate/Metsulfuron Trial....................................................................83 Conclusions and Recommendations for Land Managers...........................................84 4 CONCLUSIONS AND RECOMMENDATIO NS FOR FUTURE RESEARCH......94 Ecological Impacts......................................................................................................95 Distribution and Spread..............................................................................................96 Management Gaps......................................................................................................97 Regulatory Gaps.......................................................................................................100 APPENDIX INITIAL ANALYSIS TABLES FROM CHAPTER 2...........................102 LIST OF REFERENCES.................................................................................................108 BIOGRAPHICAL SKETCH...........................................................................................119

PAGE 10

x LIST OF TABLES Table page 2-1 Japanese climbing fern ( Lygodium japonicum ) occurrence on three North Florida forest sites, 2002 and 2004..........................................................................64 2-2 Japanese climbing fern ( Lygodium japonicum ) occurrence across eight natural community types in three North Florida forest sites, 2004. Twenty-five plots assessed per site........................................................................................................64 2-3 Comparison of tree importance values (IV) by species vegetative index and presence or absence of Japanese climbing fern (Lygodium japonicum )..................65 2-4 Relationships between Japanese climbing fern ( Lygodium japonicum ) percent cover levels and significant plot variables and tree importance values in three North Florida forest sites..........................................................................................66 2-5 Plot environmental variables that disc riminate between Japanese climbing fern ( Lygodium japonicum) presence or absence in three North Florida forest sites......67 2-6 Plot environmental variables that disc riminate between Japanese climbing fern ( Lygodium japonicum) percent cover class levels in three North Florida forest sites.......................................................................................................................... .67 2-7 Tree species importance values that discriminate between presence and absence of Japanese climbing fern ( Lygodium japonicum ) presence or absence in three North Florida forest sites..........................................................................................68 2-8 Tree species importance values that discriminate between Japanese climbing fern ( Lygodium japonicum) percent cover class levels on plots with fern present in three North Florida forest sites.............................................................................68 2-9 Effects of significant plot environmental and tree importance values in multiple logistic regression for rela tionship to presence or ab sence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forest sites................................69 2-10 Effects of significant plot environmental and tree importance values in multiple logistic regression for rela tionship to percent cover cl ass levels of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forest sites................70

PAGE 11

xi 3-1 Herbicide chemical families and mode of action used in a Japanese climbing fern control trial in a North Fl orida pine plantation, 2002 to 2005..........................89 3-2 The effect of post-emergent herbicides on Japanese climbing fern foliar cover in a North Florida slash pine plantation, 2002 and 2004..............................................91 3-3 Change in Japanese climbing fern folia r cover on untreated control plots from two herbicide trials in a North Florida slash pine plantation...................................92 3-4 Japanese climbing fern herbicide trea tment options for land managers, based on results of evaluation in a North Flor ida slash pine plantation, 2002 to 2005...........93 A-1 Abiotic plot variable with significant relationship to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests......103 A-2 Biotic plot variables w ith significant non-paramteric re lationships to presence or absence of Japanese climbing fern ( Lygodium japonicum) in three North Florida forests.....................................................................................................................104 A-3 Soil variables with significant parametric relationship to presence or absence of Japanese climbing fern ( Lygodium japonicum) in three North Florida forests......105 A-4 Tree species with importance values (IV ) significantly related to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests.....................................................................................................................106 A-5 Tree species with importance values (IV) not significantly related to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests.....................................................................................................................107

PAGE 12

xii LIST OF FIGURES Figure page 1-1 Timeline for Japanese climbing fern plant development from spores.....................30 1-2 Worldwide distribution of Japanese climbing fern ( Lygodium japonicum ) from literature review (June, 2006). A) Native ra nge indicated in grey. B) Introduced range indicated in black. Dotted line indi cates northern and s outhern extent of recorded Japanese climbing fern field populations..................................................30 1-3 County-level distribution of Japanese climbing fern in the Southeastern United States with reference to the Southeas tern Coastal Plain physiographic region. Records based on vouchered herbarium speci mens from Southeastern herbaria, and records from regional vascular plant databases................................................31 2-1 Location of study sites used in 2002 Japa nese climbing fern site establishment assessment, Florida, USA.........................................................................................59 2-2 Sampling design for assessment of Japa nese climbing fern distribution on three forest sites in North Florida. A) Lay out of major transect lines and intensive sampling plots. B) Layout of sampling points within each intensive sampling plot........................................................................................................................... .60 2-3 Plot layout indicating 2004 Japanese climbing fern cover classes and natural community types, over 2000 aerial imag ery, Blackwater River State Forest (BRSF), Santa Rosa County, Florida.......................................................................61 2-4 Plot layout indicating 2004 Japanese climbing fern cover classes and natural community types, over 2000 aerial imag ery, Miller property (Miller), Calhoun County, Florida.........................................................................................................62 2-5 Plot layout indicating 2002 Japanese climbing fern cover classes and natural community types, over 2004 aerial imager y, Florida River Island (FRI), Liberty County, Florida.........................................................................................................63 3-1 Average live foliar cover of Japanese climbing fern on untreated control plots established in October 2002 and November 2004 as part of a multiple herbicide trial in a North Florida pine plantati on. Approximate month after treatment (MAT) indicated.......................................................................................................92

PAGE 13

xiii Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science ASPECTS OF THE INVASION AND MANA GEMENT OF JAPANESE CLIMBING FERN ( Lygodium japonicum) IN SOUTHEASTERN FORESTS By Andrea N. Van Loan December 2006 Chair: Eric J. Jokela Major Department: Forest Resources and Conservation The continual introduction of new non-nati ve invasive species, and the expanding distributions of those already present in th e United States make the need for objective priority-based management increasingly important. Japanese climbing fern Lygodium japonicum (Thunb.)Sw.) is a non-native weed of mesic to hydric natural communities and sites in the Southeastern United States. Gr owth ranges from scattered creeping stems to tangled mats with dense canopy, effectivel y eliminating understory vegetation, and smothering seedlings of overstory tree species. Two field studies explored key aspects of Japanese climbing fern management in fore sted settings: an assessment of herbicide efficacy for treatment of existing populations and a review of s ite characteristics associated with varying levels of establishment. The assessment of site charac teristics associated with va rying levels of Japanese climbing fern establishment was conducted in th ree forested sites in North Florida. Sites were characterized through a comparison of Japanese climbing fern percent cover (dependent variable) with a series of inde pendent plot abiotic a nd biotic variables; including measurements of soil, and vegetation, assessed on a gr id of sampling plots. At

PAGE 14

xiv each level of analysis, some plot variables and some species in the plant community showed significant re lationships to Japanese climbing fern occurrence levels. The general group of significant va riables for presence or absenc e of Japanese climbing fern included: tree evenness index, plot percent flooded (each winter), soil calcium, and the occurrence of Carpinus caroliniana Comparison of plots with Japanese climbing fern present for varying cover class levels identi fied photosynthetically active radiation, basal area, soil aluminum, and the occurrence of Carpinus caroliniana and Quercus nigra Site variables showing the strongest relationship to Japanese c limbing fern occurrence have a unifying basis in site hydrology. The herbicide treatment study was conducted in a heavily infested forest stand North Florida, where fifteen different herbicid es and rate combinations were assessed for effects on Japanese climbing fern foliar cover, compared with an untreated control. Treatments including the herbicide activ e ingredients glyphosate, imazapyr, and metsulfuron methyl showed the greatest leve l of Japanese climbing fern control at ten months after treatment; though only glyphosate and imazapyr sustained greater than 70% control to twelve months after treatment.

PAGE 15

1 CHAPTER 1 LITERATURE REVIEW: BIOLOGY, ECOLOGY, DISPERSAL, ECONOMIC IMPACT, AND MANAGEMENT OF JAPANESE CLIMBING FERN Introduction The first time I saw a climbing fern I was shocked, as if Id opened a window and a trout had flown in (Hipps 1989). These words may describe the reactions of many observing the non-native invasive climbing ferns in the southeastern United States This has been particularly true for Old World climbing fern ( Lygodium microphyllum (Cav.) R. Br.), as the species has spread rapidly and dramatically across South and now Central Florida. Japanese climbing fern ( Lygodium japonicum (Thunb.)Sw.) (Family, Schizeaceae) however, has spread with perhaps equal rapidity, through the understory of mesic to hydric s ites across Florida. Currently, it is found in nine southeastern states. Many aspects of the life history of Japanese climbing fern indicate the potentia l for this species to be more broadly pervasive regionally; though pe rhaps slightly less aggressive on individual sites. In Florida, Japanese climbing fern esta blishes dense mats of vegetation reaching 100% groundcover in some areas. Infestations of the plant affect pinestraw management on sites throughout North Florida, and threaten native species and habitats throughout the State. Though increasingly recognized as a pe st plant within its established range in the Southeast, Japanese climbing fern may pose a lower threat to agricu lture and biodiversity than Old World climbing fern (Pemberton and Ferriter 1998). As a result, greater research focus has been placed on Old World climbing fern, and many gaps exist in the

PAGE 16

2 literature on ecology and management of Japa nese climbing fern. A review of the literature on Japanese climbing fern follows and subsequent chapters will address climbing fern management and site ecology. In many areas, references and descriptions of Old World climbing fern are utilized to provide a more co mprehensive view of general climbing fern biology and ecology. The known similarities between Japanese climbing fern and Old World climbing fern suggest th at a greater understand ing of both species can be gained by study of either. Biology Rhizomes Japanese climbing fern rhizomes are wide ly creeping, bearing le aves in two rows on the dorsal side at the rate of three to four every 1 to 2 cm (Mueller 1982). Rhizomes grow 1 to 3 cm below the soil surface and are densely covered with multicellular hairs (Mueller 1982) which may help to preven t water loss. Roots are numerous, arising endogenously on the lateral and ventral side s near the shoot ap ex (Mueller 1982). Rhizomes branch dichotomously at 1 to 2 cm intervals after producti on of three to four leaves (Mueller 1982). As seen in some fe rn species, activation of rhizoid elongation may be due to presence of key soil minera ls (Ko 2003). No examples of vegetative reproduction exist in the liter ature beyond mention of establ ishment following transport of intact fronds with rhizomes contained in soil. Rhizomes are easily divided and transplanted due to their creeping, branching nature (Hoshizaki and Moran 2001). Leaves Japanese climbing fern leaves are borne on the upper surface of slender, dark brown rhizomes. The rachis of new climbi ng fern leaves displays searcher shoot morphology, similar to many twining flowering plants, with rapid early rachis elongation

PAGE 17

3 and delayed leaflet expansion (Mueller 1982). Similar to climbing angiosperms, leaves maintain continual apical growth, e xhibit apical dominance (Punetha 1987), circumnutate, and have bud-like dormant me ristems on each leaflet. These dormant meristems are capable of growing out to repla ce a damaged leaf tip, or to strengthen the leaf (Mueller 1981, Mueller 1983). Japanese climbing fern plants remain ever green below the frost line in Central and South Florida (Ferriter 2001, Valenta, Zeller, and Leslie 2001, Zeller and Leslie 2004). Above the frost line, winter dieback of Japa nese climbing fern fronds occurs to varying extents through the majority of its range. In the spring, the plants will re-sprout from cold-tolerant subterranean rhizomes (new st ems observed the first week in March in North Florida), and often utilize freeze-damaged stems as ladders to grow back into the canopy. After emergence from the soil in March th rough May, initial leaf growth is slow (1.2 to 1.5 cm every 3 days), but steadily incr eases for at least 70 days (Punetha 2000). A series of eight to twelve primary leaves ar e produced in the first five to six months, before adult leaves are initiate d. Successively formed leaves increase in size and take a pinnate form (Mueller 1982). At peak expansi on, leaves grow at a rapid rate (3.4 to 6.5 cm per day) for several weeks, including ma in rachis extension and steady addition of 10 to 25 new alternate primary rachis branches (pinnae). New pinnae emerge from the crozier every 1.7 days (Mueller 1983). Rachis internodes can reach 9 to 10 cm in mature plants (Mueller 1983, Punetha 2000). Leaves ex hibit indeterminate growth and can reach lengths of 2 to 30 m (M ueller 1981, Ferriter 2001).

PAGE 18

4 Pinnae on lower rachis are sterile, while successively newer, upper pinnae are increasingly fertile as the growing season progresses. Pi nnules (leaflets) are initially tightly coiled, and take five to seven week s to fully expand to 23 to 25 cm (Mueller 1983). Fertile pinnules bear two rows of sporangia along the unders ide of the leaflet margin (Ferriter 2001). Leaf lets are highly dissected, w ith a lacy appearance and triangular outline, generally 10 to 20 cm long, and about as wide (Nauman 1993, Langeland and Burks 1998). Leaflet shape and appearance varies noticeably according to age, season, and site. Upon achieving a height of 20 to 40 cm, leaves circumnutate at the rate of to 1 rotations pe r day, typically in a countercl ockwise direction (Mueller 1983). Leaves will grow vertically, using trees, shrubs, or other structures for support, or horizontally along the ground (Ferriter 2001), an d the climbing habit is undoubtedly an adaptation to increase sunlight access (Holttum 1959). Growth of the stem and leaflets is retarded when the fern is grown either upside down or horizontally (Punetha 2000). Dormant apices on primary rachis branches wi ll proliferate seven days after detopping the frond or delaminating the secondary rach is branches (Punetha 1987, Punetha 2000). When grown under stress (modified habit), th e height and vigor of Japanese climbing fern leaves declined signif icantly (Punetha 2000). In areas with seasonal frost events, Japanese climbing fern foliage will decrea se in percent cover during the cooling temperatures preceding the frost event. After frost, Japanese climbing fern foliage turns a rusty-brown color due to stem and leaf death. New leaves emerge from the rhizomes the following March. This seasonal dieback in sub-temperate and temperate climates is one factor which has likely prevented formati on of the canopy-covering arboreal blankets seen with Old World climbing fern (Zelle r and Leslie 2004). However, the increasing

PAGE 19

5 distribution of Japanese climbi ng fern in south Florida may lead to the development of such canopy-smothering growth on infested sites. Spores Japanese climbing fern spores are tetrahed ral to globose, with a verrucose surface and trilete aperture that are rust or tan colored. Spores fo rm on the underside of fertile leaf margins which roll under to cover the sporangia, forming false indusia (Langeland and Burks 1998). Japanese climbing fern pl ants can produce spores year-round in frostfree areas as seen in south Florida (Lott 2001) The effects of seasonal frost and freeze damage can limit much spore production to June through November to December (Lott and Volin 2001) in areas further north. In No rth Florida, peak spore release apparently occurs in October, when a brown haze of spores is visible in heavily infested stands (Van Loan 2006). Fertile leaflets of Old Worl d climbing fern each produce an estimated 28,600 spores, or 38,000 spores per square inch of fertile leaf area, translating to as many as 30 billion spores on heavily infested si tes (Volin, Lott, Muss, and Owen 2004). The typically larger fertile leafle ts of Japanese climbing fern may produce a similar quantity of spores (Lockhart 2006). Spore viability. Once released from the spor angia, spores require moist environmental conditions for germination a nd survival of the prothallia (gametophyte stage). Spores of Lygodium species are thick walled, whic h extends their environmental viability, an important cons ideration for management (Ferriter 2001). However, Lloyd and Klekowski (1970) predict shor t spore viability due to the chlorophyllous nature of the spores. In laboratory conditions, spore germ ination began within six days of sowing (Lott, Volin, Pemberton, and Austin 2003). Japanese climbing fern spore germination is induced by red light through a phytochrome-medi ated system, and is greatly influenced

PAGE 20

6 by temperature. Germination rates increase fr om a low at 20 C as temperatures increase (Kagawa and Sugai 1991, Sahi and Singh 1994), i ndicating that cooler temperatures will limit both germination and growing season. Sugai, Takeno, and Furuya (1997) found 40% to 80% germination rates for Japanese c limbing fern spores af ter storage at room temperature for 3 years. Perez-Garcia, Orozco-Segovia, and Riba. (1994) found spore viability of Lygodium heterodoxum in Mexico drastically re duced following storage in dry conditions (1% germination af ter 1 year). In contrast, wh en stored fully imbibed in water germination averaged 40% after 1 year. Light quality was not significant to spore germination, explaining germina tion success under tree canopies. Spore dispersal. The spores of Japanese climbing fern are tiny structures, easily transported by wind, water, animals, and equipment (Langeland and Burks 1998), and may possibly travel several miles from their point of origin. The climbing growth form of Japanese climbing fern may also contribute to rapid dispersal by placing spores at the height of the prevailing winds (Lott et al 2003). Annual flooding in the Appalachicola River floodplain, as well as periodic flooding of other rivers and streams in Florida, distributes Japanese climbing fern spores a nd frond fragments into new areas (Valenta et al 2001). Land managers have not ed observations of increased distribution of the fern in forest stands in the years following the El Ni o-induced flooding events, which occurred across central and north Florida in the wint er of 1997/98. These observations may be based on increased recognition of the species, in creased distribution of the species due to either flood-based distribution of the spores, or increased soil moisture levels that create more conducive conditions for germination and su rvival of spores al ready present in the soil and leaf litter s ubstrate (Ferriter 2001).

PAGE 21

7 The apparent spread pattern of Old Worl d climbing fern in south Florida overlaps with the expected spread pattern if dispersa l was primarily due to the prevailing winds of spring, summer and autumn. Old World climbi ng fern spore production appears to occur year-round, and one measurement of spore load ing in the air captured 724 spores per m3 per hour within a Martin County infestati on (Pemberton and Ferriter 1998). Similar, though potentially diminished, spore loading and wind-driven disp ersal of Japanese climbing fern may exist given similarities in biology between the two species, although estimates of spore loading are rare for this sp ecies in Florida. So me concern exists over the long distance spread of Lygodium spores in the convective columns of prescribed or wildfires. These fire-generated wind conditi ons can disperse smoke great distance away from a fire (more than 100 miles in some cases) and may show the ability to disperse spores in the same way. In addition, soil dist urbances associated with forest management activities have been shown to facilitate establishment of the native fern Dennstaedtia punctilobula (Michx.) Moore. (Groninger and McCormick 1992), and evidence of a similar effect has been seen with establis hment of Japanese climbing fern in some harvested sites in northwest Florida. Spore banks. Natural and cultivated soils can contain a large number of fern spores in a spore bank, often without cl ose proximity of sporophytes (Hamilton 1988, Dyer and Lindsay 1992, Ranal 2003). When peak spore maturation periods occur, spores may fall to the ground in large, scattered masses; a portion of which percolate into the soil pore space with germination inhibited by lack of light, and others may remain ungerminated on the soil surface (Ko 2003). Unlike limitations on fungal spore germination, fern spore germination does not seem to be affected by soil microbiostasis

PAGE 22

8 (Ko 2003). Ko (2003) suggested that fern spores are nutritiona lly independent, and germinate successfully on substrates without exogenous nutrients. Soil spore banks may play an important role in reproductive and establishment success for many fern species, creating immediate opportunities for spore ge rmination and gametophyte establishment after any soil disturbance (Ranal 2003). Evidence of fern spore survival in soils for one year has been obtained in multiple studies, a nd at least one study verifies survival of approximately two-thirds of fern spores for two years (Dyer a nd Lindsay 1992). Dyer and Lindsay (1992) found that fern spore banks can even survive forest and heath fires. Bark spore banks have been assessed for two tropical species, verify ing the potential for such a spore bank to contribute to the popul ation dynamic of pteri dophyte species (Ranal 2004). Spores present on bark of live or downed trees, and even in bark products harvested from infested stands may germinate a nd contribute to dispersal of this species. When coupled with movement of Japanese climbing fern spores in pine straw bales, concerns arise over the possible role of tim ber and non-timber forest product movement in spore dispersal in the southeastern United States. Gametophytes/Mating System Japanese climbing fern is a homospor ous fern with free-living haploid gametophytes and diploid sporophytes. Sexuall y mature gametophytes appear five weeks after spore germination (Lott et al. 2003). In the gametophyte generation, the fern is capable of three types of sexual reproductio n: intragametophytic selfing (egg and sperm from same gametophyte), intergametophy tic selfing (egg and sperm from two gametophytes originating from the same s porophyte), and intergametophytic crossing (outcrossing of egg and sper m from two gametophytes, originating from two different sporophytes) (Lott et al. 2003). According to Lott et al. (2003) the populations of

PAGE 23

9 Japanese climbing fern in Florida may have originated from only a few plants, with selection favoring selfing variants rather than outcrossing. It has been proposed that Japanese climbing fern has evolved an theridiogens such as gibberellin A73-methyl ester (GA-73-Me) (Wynne, Mander, Goto, Yamane, and Omori 1998), and gibberellin A9 methyl ester (GA9-Me) (Takeno and Fur uya 1980) which may promote outcrossing to avoid the potential inbreeding associated with intragametophytic selfing (Ferriter 2001). However, Lott et al. (2003) found intragamet ophytic selfing to be the primary mode of reproduction for Japanese climbing fern from two populations in Florida. Over 90% of the gametophytes produced this way develope d sporophytes. Therefore, though some evidence of outcrossing exists, intragametophytic selfing may be a stable mating system in Florida. If so, this reproductive strate gy in Japanese climbing fern may facilitate colonization of a diversity of habitats a nd establishment of di stant populations from single spores (Lott et al. 2003). There are four pairs of Lygodium species which have some amount of intergrade. This may be due to environmental conditions or it could be due to hybridization. Japanese climbing fern ( L. japonicum ) has been documented to intergrade with Lygodium flexuosum (L.) Sw. (Holttum 1959). Sporophytes Under stressful and variable environmenta l conditions as seen with drought and flood cycles, gametophytes that reach sexual ma turity quickly have the best opportunity to produce sporophytes before dying (Lott et al. 2003). Once fertilized in the gametophyte stage, sporophyte production begi ns. Sporophyte production occurred from week five to week twelve (Lott et al. 2003). Adult plants with climbing leaves and the

PAGE 24

10 capability to reproduce have formed by five to six months after spore germination (Mueller 1982) (Figure 1-1). Ecology Native Range Japanese climbing fern is native to temper ate and tropical eastern Asia and the East Indies, including India, China, Japan, Kor ea, and many other countries in the region (Figure 1-2) (Holttum 1959, Singh and Pa nigrahi 1984, Wee and Chua 1988, Ferriter 2001, Louisiana State University 2006, Roya l Botanic Gardens Melbourne 2006, USDA 2006). In its native range, Japa nese climbing fern occurs in forest edges, open forests, and secondary forests at both lower and highe r (2550 m) elevations, and is described as never colonizing in any ar ea(Holttum 1959, Gias Uddin, Mahal, and Pasha 1997); but has also been considered weedy in Taiwan and the Phillippines (Singh and Panigrahi 1984, Langeland and Burks 1998, Ferriter 2001). Holttum (1959) described Japanese climbing fern as Only found native in regi ons with pronounced dry season, during which fronds perhaps die. Introduced Range Considered to have been introduced to the United States around 1900 for ornamental purposes, the first naturalized population of Japanese climbing fern was identified in Georgia in 1903 (Pemberton and Ferriter 1998), and in north Florida in 1932 (Gordon and Thomas 1997). Noted as a rare escape in Georgia in 1964 (Radford et al. 1964), it was described as aggressively spread ing in moist woods and fields throughout Florida and adjacent states by 1968 (Beckner 196 8). In 2005, Japanese climbing fern was identified as spreading at an alarming ra te in Georgia (Evans and Moorhead 2005). Florida county records for th is species have increased from 29 in 1995 to 45 in 1999

PAGE 25

11 (Ferriter 2001), and 53 in 2006 (Van Loan 2006) due to rapid and successful spore dispersal, and improved recognition and repor ting. In Florida, recent occurrences in Broward, Collier, and Dade counties as well as Brevard, Hardee, Highlands, Manatee, Polk, and Sarasota counties verify a range ove rlap with Old World climbing fern (Ferriter 2001, Ferriter and Pernas 2006, Van Loan 2006). This raises concerns that Japanese climbing fern may invade mesic sites and tr anscend into hydric sites occupied by Old World climbing fern; a situation which is begi nning to occur in centr al and south Florida (Ferriter and Pernas 2006, Lockhart 2006). As may be true with Old World climbing fern, multiple naturalization events are possible for Japanese climbing fern (Fe rriter 2001) though possibly from the same propagative parent plant ma terial (Lott et al. 2003) These two fern species were often confused in early horticultura l literature. A photograph and de scriptive statements of Old World climbing fern in the 1905 catalog from Ro yal Palm Nursery (a primary distributor of tropical and subtropical flor a in Florida and the United Stat es) actually appear to be Japanese climbing fern. If so, it may have been sold by the nursery for some part of the period (1888-1930) that the nursery marketed Old World climbing fern (Pemberton and Ferriter 1998). Promoted in th e horticultural trade as a ha rdy, attractive plant (Ferriter 2001), Japanese climbing fern is often misidentified as Lygodium scandens (L.)Sw., a synonym for Old World climbing fern (Hoshizaki and Moran 2001). In the United States, Japanese climbing fern now occurs in eleven states including: Florida, Georgia, Alabama, Mississippi, Loui siana, Arkansas, Texas, South Carolina, North Carolina (Figure 1-3), as well as Calif ornia, Hawaii and Puerto Rico (Figure 1-2) (Ferriter 2001, Van Loan 2006). County-level occurrence records in the Southeastern

PAGE 26

12 United States indicate a cl ose and expanding range overlap with the Coastal Plain physiographic region (Sundell 1986, Martin, B oyce and Echternacht 1993). The Coastal Plain boundary may well serve as a good predicto r of the region most rapidly occupied by Japanese climbing fern. Within that range, Japanese climbing fern occurs primarily in floodplain forests, marshes, wetlands, s econdary woods, moist pinelands (including flatwoods), and disturbed areas such as culverts, fencelines, road shoulders and rights-ofway (Clewell 1982, Nauman, 1993, Wunder lin 1998, Langeland and Burks 1998). Growth ranges from scattered creeping stems to tangled mats with dense canopy, effectively eliminating unde rstory vegetation (Nauman 1993) and smothering seedlings of overstory tree species (La ngeland and Burks 1998). In the Apalachicola river basin, Japanese climbing fern is de scribed as occurring at epid emic proportions in bottomland forests, floodplain forests, sloughs, and mesic flatwoods (Fer riter 2001). Floodplain swamps are comparatively uninfested due to lower elevations a nd resultant regular flooding and inundation (Ferriter 200 1). Infestations also occur in xeric sites, but do not appear to expand as rapidly as in more mesi c sites, possibly due to the infrequency of appropriate conditions for gametophyt e establishment or fertilization. Environmental Requirements The most successful invasive plants are us ually capable of displacing native species because they have effective reproduc tive and dispersal mechanisms, superior competitive ability, limited herbivores or pat hogens, ability to occupy a vacant niche, and ability to alter site resource availabilit y, disturbance regimes, or both (Gordon 1998). While the ecological requirements of Japane se climbing fern are poorly understood, its broad distribution indicates a tolerance for a range of environmental conditions. Observations of land managers indicate redu ced success in extremely dry or consistently

PAGE 27

13 flooded sites (Ferriter 2001) a nd a potential preference for so ils with circumneutral pH (Nauman 1993). Fluctuating precipitation levels may have limited impact on plant growth, but foresters in north Florida reported th at in the dry spring of 1999 Japanese climbing fern was not as serious of a problem as in previous wetter springs (Ferriter 2001). Japanese climbing fern is the most widely establis hed non-native, invasive pest plant on Suwannee River Water Management Dist rict lands, and is considered to pose the greatest threat to na tural systems on those lands (Fe rriter 2001). Twenty populations were evaluated on District lands, and found to occur in the following natural community types: mixed hardwood and pine (9), bottoml and hardwood (7), longleaf pine-turkey oak (5), upland hardwood hammock (2), wet hardwood hammock (1), swamp hardwoods (1), and North Florida flatwoods (1). Sixteen of the 20 populations occurred between 50 and 60 feet in elevation. Eighteen of the popul ations were associated with disturbance (potentially an artifact of the location of the surveys) (Ferriter 2001). Human disturbance does not a ppear to be a prerequisi te for successful invasion (Ferriter 2001), but in at least two cases in north Florida, removal of pine litter from the forest floor of pine plantations was follo wed by rapid and complete establishment of Japanese climbing fern in sites where it was pr eviously either absent or present in very small quantities (Van Loan 2001). Sahi and Si ngh (1994) assessed the effects of sulphite exposure on Japanese climbing fern spore germination and rhizoid growth, and found both to be negatively impacted. This finding indicated that the presence and quantity of Japanese climbing fern gametophytes may be a bioindicator of sulfur dioxide air

PAGE 28

14 pollution. This finding may have use in some areas in India, within the native range of Japanese climbing fern. Japanese climbing fern plants remain ever green below the frost line, as does Old World climbing fern in central and sout h Florida (Ferriter 2001, Valenta et al 2001, Zeller and Leslie 2004). Above the frost line, winter dieback of Ja panese climbing fern fronds occurs to varying extents through the ma jority of its range; plants in shady, moist protected areas like floodplain and bottomla nd forests, are able to persist through the winter in places. Freeze-damaged/killed fronds turn rusty brown in color and are easily recognized. In the spring, the plants will re-sprout from moderately cold-tolerant rhizomes (new stems observed the first week in March in North Flor ida), and often utilize freeze-damaged stems as ladders to grow back into the canopy. Japanese climbing fern is promoted as an ornamental plant as far north as Kansas due to cold tolerance of rhizomes, but the Southwestern Fern Society does r ecommend mulching the plants after the first frost for additional rhizome protection in far northern areas (Ferriter 2001). This seasonal dieback in sub-temperate and temperate climat es is one factor which has likely prevented formation of the dense arboreal blankets in tree canopies seen w ith Old World climbing fern (Zeller and Leslie 2004). However, the increasing distribution of Japanese climbing fern in south Florida may soon lead to the development of such canopy-smothering growth on infested sites. Impacts In the Apalachicola River basin in Flor ida, Japanese climbing fern occurs in floodplain forests as a thick mat of growth that smothers native groundcover vegetation, and can extend into the lower overstory canopy (Valenta et al. 2001). Japanese climbing fern is listed as one of the top contributing factors to the degradati on of the Apalachicola

PAGE 29

15 River system (Berquist, Pable, Killingswort h, and Silvestri 1995). In its native range, Japanese climbing fern can establish rela tively large mats; however, these are much smaller than those commonly found in Florid a (Ferriter 2001). The primary effects of Japanese climbing fern on nativ e flora and fauna stem from impacts on understory plant communities, including the habitat of the Florida endangered plant species Sideroxylon thornei (Cronq.) Penn., Aristolochia tomentosa Sims., and Polygonum meisnerianum Cham. & Schlecht. (Ferriter 2001). Observations of Old World climbing fern, in south Florida and I ndia indicate that the dense growth of fronds into the overst ory tree canopy can serve as ladder fuels to carry prescribed fire and wildfire into th e canopy, resulting in tree death. Additionally, the fronds serve as a fuel conduit to allow fi res to burn into the center of cypress sloughs normally left unburned due to water levels, and embers of the dense fern mats can serve as fuels to carry spot fires away from th e primary burn unit (Marth ani et al. 1986, Roberts 1996, Roberts 1997). Brandt and Black (2001) found the impact of Old World climbing fern invasion to be a change in vegetativ e composition of the na tive community, resulting in a decrease in native species and a loss in plant diversity and evenness. In a comparative study, Clark (2002) found significa nt reduction in native plant cover, richness and diversity from uninvaded areas to similar habitats with heavy Old World climbing fern infestation. Impacts of a sim ilar type, but possibly lower scale may be seen in sites infested with Japanese climbing fern. Little is known about the economic impact of Japanese climbing fern invasion. Forest managers in Florida have noted imp acts including reduction in annual financial return and harvest lease longevity for pine straw production, an indus try which generates

PAGE 30

16 $79 million in revenue for forest landowners in Florida annually (Hodges, Mulkey, Alavalapati, Carter, and Ki ker 2004), and reductions in pi ne seedling survival during reforestation of heavily infested sites. Vines as Invasive Plants An estimated 138 non-native shrub and tree sp ecies have invaded native U.S. forest and shrub ecosystems (Pimental, Lach, Zuniga, and Morrison 2004), a figure which does not include invasive grasses, forbs, or vi nes. However, vines represent a large and aggressive group of invasive plants in th e United States. Nineteen of the 133 plant species on the 2005 Florida Exotic Pest Plan t Council (FLEPPC) Invasive Plant list are vines, representing 14% of the most inva sive plants in Florida (FLEPPC 2005). Horovitz, Pascarella, McMann, Freedman, and Hofstettler (1 998) proposed six functional groups of invasive species, including vin e blankets. The groups were defined to classify recolonization of a site following hurri canes, but other severe disturbance may be considered (i.e. timber harvesting in pine plantations or natural stands). The vine blanket group would include both introduced Lygodium species in Florida with the proposed effects of shading natives, including: the juvenile/seedling layer, tree re-sprouts from roots and fallen stems, and regrowth of new branches from standing stems. Vine blanket species are usually aggressi ve and may negatively influence forest regeneration by diverting resources, and st rangling native species seedlings. Vine invasion may drive species composition toward ruderal species, particularly in sites affected by a significant disturbance even t (Gordon 1998). Invasion of fire-evolved systems by non-indigenous (and native) vines ca n alter or prevent fire movement due to increased fuel moisture, and increased site relative humidity. Post-disturbance invasion by non-indigenous vines has been noted in other forests with Lonicera japonica Thunb.

PAGE 31

17 Invasion of temperate deciduous forests after smaller-scale natural disturbances have resulted in crushing and strangling of nativ e shrubs and saplings, and contributed to increased root and soil mois ture competition (Gordon 1998). Japanese climbing fern effectively coloni zes both basic invasibl e substrates: those highly disturbed by humans, such as pine plan tations, agricultural fields, road shoulders, ditches, and yards; and those relatively free of anthropogenic disturbance such as wetlands, river banks, hardwood hammocks, and some pinelands. Key differences also exist between Japanese climbing fern and se veral other non-native vines in the Southeast that possess a twining growth form. The twin ing growth form of Japanese climbing fern contrasts with the bark-adhering growth form of native vines in this area. This twining form enables linkage of tree canopies, incr eases the likelihood of collapse of the supporting plants, and changes the plant community structure (Hardt 1986, Gordon 1998). Native vines in the Southeast are prim arily temperate species, compared with the predominantly tropical origin of the invaders. As a result, the tropical invaders may gain a competitive advantage in the subtropical climat es of the Southeast, particularly as seen in Florida (Gordon 1998). Ferns as Invasive Plants There are 34 exotic ferns and fern allies in Florida, representing one-third of Floridas fern species (Pemberton 2003). Six of the 133 species on the 2005 FLEPPC plant list are ferns, representing 5% of the most invasive species in Florida (FLEPPC 2005). Pemberton (2003) lists Japanese climbi ng fern as one of five severely invasive ferns in Florida along with Lygodium microphyllum, Nephrolepis cordifolia (L.) Presl. and N. multiflora (Roxb.) Jarrett ex Morton and Tectaria incisa Cav. all of which

PAGE 32

18 reproduce prolifically and form dense stands or mats of growth, displacing or outcompeting native species. Ferns are also occurring in increasing numbers in the Hawaiian Islands, where 32 alien fern species have become naturalized, including Japanese climbing fern. Much as in the Southeast, the species was not obser ved to have spread from the original 1936 location until 1998. Since then colonizat ion has expanded rapidly, as additional populations have arisen in multiple loca tions on Oahu and Hawaii (Wilson 2002). According to Wilson (2002) the Hawaiian ec osystem continues to be challenged by invasion of new species of pt eridophytes as well as the spread of already present naturalized species. As a corollary, Dicranopteris linearis (Burm.) Underw., a common native climbing fern of Hawaii, is a successful colonizer on open substr ates due to factors including: indeterminate (cl onal) growth, shallow rhizomes, and a mat-forming capacity (Russell et al. 1998). These same characteristics also contribute to the ability of Japanese climbing fern to colonize habi tats in its introduced range. Related Species in the United States From 26 to 40 species occur in the genus Lygodium (Alston and Holltum 1959, Ferriter 2001) and most are placed in the family Schizaeaceae (Prantl 1881). Two additional Lygodium species are established in the S outheastern United States, both of which overlap in range with and create mana gement implications for Japanese climbing fern. A native species, Lygodium palmatum (Bernh.) Sw. occurs in swamps, streambeds and ravines in Mississippi, Alabama, and Georgia, and its range extends northeast through New England (Nauman 1993, Miller 2003); while a second, highly aggressive non-native species, Lygodium microphyllum (Cav.) R. Brown (Old World climbing fern), occurs in Central and South Fl orida (Ferriter and Pe rnas 2006). Native to moist habitats

PAGE 33

19 in Africa, southeastern Asia, Australia, and Polynesia (Pemberton 1998, Hoshizaki and Moran 2001), Old World climbing fern has spread rapidly through hydric sites in southern Florida (Pemberton and Ferriter 1998) where the sub-tropic al climate, and the ferns life history strategies (climbing vine growth form, highly successful reproductive and long-distance dispersal strategies (i.e., intragametophytic selfing, frond growth into tree canopies, wind and water-dri ven dispersal), and rapid development of sporophytes within a growing season (Lott et al. 2003)), ha ve facilitated the tendency for invaded sites to be converted to climbing fern monocultu res. Old World climbing fern was first observed growing naturalized in Southeast Florida in the 1960s (Beckner 1968, Nauman and Austin 1978). Systematic aerial surveys of Old World climbing fern estimated gross infested area in 1993 at 11,200 ha in S outh Florida (Pemberton and Ferriter 1998). Similar surveys completed in 2005 increased th e estimate of infested area to 48,878 ha in South and Central Florida, a small portion of which may actually be Japanese climbing fern as the species are indistinguishable from the air, and now have overlapping ranges in this area (Ferriter and Pern as 2006). Though Japanese climbing fern is increasingly recognized as a pest plant within its esta blished range in the Southeast, Old World climbing fern has received far greater research emphasis. As a result, much of what is known or published about Japanese clim bing fern biology, ecology, impacts, and management is the product or by-product of research and assessment of Old World climbing fern. This relationship has almost certainly enhanced awareness of Japanese climbing fern, but has potentially served to comparatively diminish perceptions of its invasiveness and impact as well.

PAGE 34

20 Recognition and Management in the United States Management Status Recognition of Japanese climbing fern has increased among public conservation land managers in the southern United St ates since the mid 1990s, and has aided in increased detection and reporting. As a result, Japanese climbing fern was designated as a Category I invasive plant by the Florida Exot ic Pest Plant Council in 1995, due to its impacts in altering native plant communitie s by displacing native species, changing community structures or ecological functions (Ferriter 2001, Florid a Exotic Pest Plant Council 2005). The State exotic pest plant councils of Alabama, Georgia, and South Carolina also designate Japanese climbing fern as a significa nt invasive plant (Miller, Chambliss and Bargeron 2004). While non-re gulatory, these designa tions often guide management on public conservation lands. Th e USDA Forest Service Southern Region has designated Japanese climbing fern as a Category 1 (exotic pl ant) species known to be invasive and persistent, guiding its management on Na tional Forest lands since 2001 (Miller et al. 2004). Recognition of Japanese climbing fern by private land managers has also increased since the late 1990s, but management efforts have not increased at a similar rate, due to lack of awareness of the non-native inva sive status of the plant, research-based best management practices, and financial incentives or support for management of private-lands infestations. Regulatory Status In 1999, Japanese climbing fern was designa ted as a noxious weed in Florida (Rule 5B-57.007, F.A.C.) by the Florida Department of Agriculture and Consumer Services (FDACS), making it unlawful to introduce, posse ss, move, or release any living stage of the fern without a permit. This design ation has raised aw areness among private

PAGE 35

21 forestland managers and members of the fore st products industry (e sp. pinestraw) in Florida, and for increasing the demand for e ffective management guidelines. In addition, in Florida, both Collier and Okaloosa c ounties have local ordinances regulating L. japonicum (K. Burks, personal communication, 12/ 2001). Currently, Alabama is the only other state which designates Japanese climbi ng fern as a noxious weed (Van Loan 2006). Japanese climbing fern is not designated as a federal noxious weed according to the Federal Noxious Weed Act of 1974 (7 U.S.C. 2809). A petition was submitted to the United States Department of Agriculture, An imal and Plant Health Inspection Services (APHIS) in 1999, requesting that both Japanese and Old World climbing ferns be given federal noxious weed status. The widespread distribution of Japa nese climbing fern excluded it from definition as a quarantine pest, making designation unlikely without further research to more accurately identify th e potential distribution in the United States (Ferriter 2001). One potential di stribution is shown in Figur e 1-3, indicating that much can be done to limit the conti nued spread of this plant in the United States. Federal noxious weed status would have major imp acts on anthropogenic spread of Japanese climbing fern, including: sale of Japanese cl imbing fern in the horti cultural industry, sale of spores in traditional medicine industry, inte rstate shipment and sale of contaminated forest landscape products. Economic Impact In addition to negative environmental im pacts, non-native species are causing major economic losses in forestry, agriculture, and other segments of the U.S. economy (Pimental 2004). Little is known about th e specific economic impact of Japanese climbing fern invasion. Some estimates pl ace annual economic impact of forest product loss due to invasive species as high as $2 b illion in the United States (Pimental et al.

PAGE 36

22 2000). Impacts of invasive plants on the fore stry industry in Florida have been estimated at $38 million per year ($15 million in weed control costs, and $23 million in yield losses) (Lee 2005). Economic costs. Ground-based foliar treatment of Japanese climbing fern in natural areas can range from $45 to $450 per acre, depending on the infestation density, cover and distribution, habitat type, site a ccess and location, and ot her factors. With multiple treatments required to achieve maintena nce control, or eradication if possible, management costs for an infested site can reach the hundreds or thousands of dollars. Presently, the Florida pine straw indu stry, which generates $79 million in landowner revenue (Hodges et al. 2004) is mo st directly impacted by Japanese climbing fern invasion. While quantitative assessm ent of the economic impact of Japanese climbing fern invasion on the pine straw industry in the southeast has not been conducted, pine straw producers in Florida have noted some impacts. In at least two cases, producers have had to abandon pine plan tations leased for straw harvest due to the dense and impenetrable nature of Japanese c limbing fern invasion on site and the quantity of economically unacceptable contaminant produ ced by harvesting infested plantations. Management of this plant in pine straw pl antation sites requir es additional production cost for the product reducing profits for many producers (Van Loan 2001). In addition, Japanese climbing fern may pos e a serious economic risk to pine plantations in the Southeast through spread and intensification of fire in infested areas (Ferriter 2001). Economic uses. Japanese climbing fern has historic ally been cultivated in Florida and may have been sold widely for a thir ty year period between 1888 and 1930 (Ferriter 2001). Currently it is not being commercially produced in the State (Betrock Information

PAGE 37

23 Systems 2002) due to its noxious weed status. However, sale of Japanese climbing fern continues in other states, facili tated by the internet as in the case of the Bushman Plant Farm of Cleveland Texa s (Ferriter 2001). Japanese climbing fern spores are mark eted for multiple herbal treatments under the names spora Lygodii and the Chinese term hai jin sha (hai = sea, jin = gold, sha = sand) (Ferriter 2001). Some traditional medi cinal uses of spores and spore tinctures include: promotion of diuresis, relief of stranguria (caused by urinary stones), and treatment of eczema (Fan 1996), as well as ki dney and urinary function, colds and fever and others (Ferriter 2001). These uses are not expected to be ec onomically significant, but may contribute to a degree of spread of this species. Management Practices Integrated pest management utilizes the best appropriate combination of treatments including: chemical, manual/mechanical, cultu ral, and biological control techniques. Current control technologies for the manage ment of Japanese climbing fern are in development for application in both natural ar eas and managed properties. At present, herbicide use appears to be the most effectiv e control technique, but further research is need on integration of techniques, and im pact of human management practices on anthropogenic spread of the species. Prevention With all invasive species, the optimal approach to successful management is prevention of new introductions and preven tion of continued spread from existing infestations. At the nation al level, sale of products containing viable reproductive Japanese climbing fern plant material should be halted, an action which requires establishment of regulatory noxious weed stat us at the federal level, and throughout the

PAGE 38

24 southeastern states. In add ition to designation as a noxious w eed at the state and federal levels, a strategy for effective enforcement must be identified or developed. The forest products (i.e., timber, pine straw, pine bark mu lch) often harvested from infested sites are not easily screened for the presence of Japanese climbing fern spores, but multiple cases in North Florida have verified that in at least some cases, these products are capable of vectoring reproductive material (Van Loan 2006). Further attention to this type of movement will be critical to the success of attempts to manage Japanese climbing fern. When managing sites infested with Japa nese climbing fern, spore vectoring on contaminated personnel, clothing, and equipmen t must be considered. Hutchinson and Langeland (2006) evaluated Old World clim bing fern gametophyte formation from spores collected on clothing and equipment of herbicide applicators after working in three heavily infested sites receiving an initial herbicide treatment, and one site with a very low infestation undergoing a re-treatment. Contamination of clothing and equipment occurred in the initial and re-treatment sites, affecting shirts, pants, sprayers, disposable suits, chainsaws, gloves, machetes and footware. Contamin ation was significantly greater from the heavily infested initial treatments s ites versus the re-treat ment sites. Similar results are likely in sites infested with Japa nese climbing fern, and applicators and other land managers working in these sites s hould utilize the following good spore hygiene recommendations: spray off equipment and brush off clothing and accessories before leaving site, wash all clothi ng daily, store disposable suit s in plastic bags, limit any vehicle entry into infested area s (Hutchinson and Langeland 2006). Chemical control In general, less herbicide testing has b een completed on Japanese climbing fern than Old World climbing fern. Initial work that has been completed indicates that

PAGE 39

25 glyphosate provides a high level of control. Questions still remain, however, regarding the efficacy of other herbicides, tank mixes, dur ation of control effect effects of repeated treatments, and integration of treatment t echniques. Preliminary assessments have resulted in a range of herbicide reco mmendations for Japanese climbing fern management with varying efficacy, includi ng foliar applications of: glyphosate, metsulfuron, triclopyr, and imazapyr (Valen ta et al. 2001, Miller 2003, Zeller and Leslie 2004). Valenta et al (2001) evaluated low, medium, and high rates of glyphosate, triclopyr ester and triclopyr amine each in a re plicated study in the panhandle of Florida. At 315 days after treatment (DAT), the glyphos ate treatments yielded greater than 70% control (reduction in live foliar cover), while the tric lopyr amine and triclopyr ester applications showed no control. In a pair of unreplicated demonstrat ion trials established in north Florida for pine straw producers, Ze ller and Leslie (2004) treated inter-rows in two pine plantations, compari ng broadcast and spot applica tion techniques, using high and low rates of glyphos ate, triclopyr ester, 2, 4-D+dicamba, hexazinone and metsulfuron. When assessed in the peri od from 300 to 400 DAT, glyphosate broadcast treatments yielded greater th an 80% control of Japanese climbing fern, but also caused notable damage to native understory plants. Metsulfuron broadcast treatments showed similar high control levels at 240 DAT, but by 370 DAT, control ranged from 21% at low rates to 62% at high rates. Dama ge to native understory plants was lower in metsulfuron treatment areas than observed with glyphosate. In an attempt to begin filling some of the knowledge gaps in Japanese climbing fern management, herbicide evaluations were conducted as a part of this research. The results of these trials,

PAGE 40

26 comparing fifteen common forestry and invasi ve plant herbicide treatments are reported in Chapter 3. Mechanical control Mechanical methods for treatment or c ontrol of Japanese climbing fern are considered only as components of an integr ated management approach, as demonstrated with Old World climbing fern. Observations on Old World climbing fern indicate that vines will re-grow after both cutting of stems and hand-pul ling of roots and rhizomes, and that mechanical techniques must be used in conjunction with herb icide application to achieve acceptable treatment efficacy. Stocke r, Ferriter, Thayer, Rock, and Smith(1997) assessed a combination of tr imming and flooding on Old Wo rld climbing fern but found no long-term control. Similar results are li kely for Japanese climbing fern. Cutting arboreal fern stems slightly above ground-leve l prior to herbicide application has been used to increase efficiency of Old Worl d climbing fern treatments in canopy-topping infestations. Heavy equipment has been used in some situ ations to remove the thick rachis mat formed by the ferns overlapping growth (Ferriter 2001). Use of heavy equipment in climbing fern infestations ra ises concerns about equipment-based spore vectoring, damage to native species, and e xposure of mineral soil aiding re-colonization of sites by Japanese climbing fern or othe r non-native invasive plants (Hutchinson, Ferriter, Serbesoff-King, Langeland, and Rodgers 2006). Biological control Literature searches by Pemberton (1998) re vealed very few known insect feeders of Lygodium ferns. However, insect biocontro l candidates have been identified for Old World climbing fern and may exist for Japanese climbing fern Ongoing research efforts are in place to identify biological control ag ents for Old World climbing fern with the

PAGE 41

27 first approved agent, a pyralid moth, Austromusotima camptozonale (Hampson), released in 2005 (Pemberton 2006). It is possible that biological cont rol agents released on Old World climbing fern will have some impact on Japanese climbing fern populations in Florida, particularly below the frost line wh ere populations of the two species overlap, and host preference for Japanese climbing fern is evaluated for all screened agents. The possibility of range ove rlap with the native L. palmatum affects consideration of Japanese climbing fern classical biol ogical control efforts. A foliar rust fungus ( Puccinia lygodii (Har.) Arth. (Uredinales)) has been isolated from Japanese climbing fern in Louisiana, and several locations in north and central Florida. The fungus, native to South America, causes severe damage to the leaflets, including necrosis, browning, and drying (Rayachhetry, Pemberton, Smith, and Leahy 2001), and has been observed to apparently incr ease in distribution and impact in north Florida climbing fern populations. The visual impact of this fungus on climbing fern in north Florida has been most noticeable in early winter (i.e., N ovember and December), when varying levels of damage have been observed on approximately 95% of foliage in some forest stands in north Florida. In preliminary observation, the primary impact of this fungus in north Florida appears to be a reduction in photosynt hetically active leaf surface area immediately precedi ng seasonal leaf dieback. Th e fungus does not appear to affect fern re-growth in the spring. Rust f ungi are considered difficult to propagate in laboratory conditions, possibly lim iting further consideration of P. lygoddii as a bioherbicide candidate. Strandbe rg (1999) evaluated the effect s of inoculation with the Colletotrichum acutatum H. Simmonds (fern anthracnose) fungus, but found disease damage on Japanese climbing fern to be mi nimal. Jones, Rayamajhi, Pratt, and Van

PAGE 42

28 (2003) evaluated the fungus Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. for pathogenicity on both invasive climbing fern s, finding some browning and wilting of the pinnules of Japanese climbing fern, and as much as 50% necrosis on Old World climbing fern plants. Prescribed fire treatment Foresters in north central Flor ida have reported that prescribed fire is not effective in controlling Japanese climbi ng fern in pine plantations (F erriter 2001). The effects of prescribed burning have not been directly evaluated on Japanese climbing fern; however, results might be expected to be similar to those in Old World climbing fern. Prescribed fire was evaluated both alone and in combin ation with herbicides for control of Old World climbing fern by Roberts and Richar dson (1994), and Stocker et al. (1997). Roberts and Richardson (1994) c oncluded that fire alone cann ot be used to manage Old World climbing fern based on measurements of percent cover and height and number of stems. Fire used in combination with a si ngle herbicide applicati on was ineffective as well. Stocker et al. (1997) compared combined treatments of trimming and flooding, contact herbicide and flooding, and burni ng and flooding, and found Old World climbing fern to successfully recover follo wing all treatments. Prescrib ed fire has been found to be an effective biomass reduction tool prior to he rbicide application, resu lting in a reduction in herbicide required to control Old World climbing fern (Ferriter 2001). Haywood et al. (2001) assessed the impacts of 37 years of seasonal prescribed burning on vegetation in a drymesic upland longleaf pine site in Louisiana. In measuring occurrence of herbaceous vegetati on on the study plots, Japanese climbing fern was found to be the dominant species on 3% of the plots burned biennially in March, and the second most common species on 3% of the plots burned biennially in July

PAGE 43

29 Japanese climbing fern was less common on pl ots burned biennially in May. Anecdotal observations on the study plots indicate that regardless of burn regime, Japanese climbing fern is increasing in presence on all plots. Similar increases in percent cover of Ja panese climbing fern on regularly burned sites have been observed in north and cen tral Florida (J. Wyrick, 6/15/2002, personal communication.). Climbing fern infestation ma y alter fire behavior on a site by serving as a ladder fuel to carry flames into the tr ee canopy. In addition, c oncern exists about the possible long-distance spread of climbing fe rn spores in the convective columns of prescribed fires or wildfire s (Ferriter 2001). These fire-g enerated wind conditions can disperse smoke great distances away from a fire (more than100 miles in some cases) and may show the ability to disperse spores in the same manner. Conclusion While some aspects of the biology, ecology, and management of Japanese climbing fern have been described, clear gaps do stil l exist. Questions remain about invasiveness in various habitats, interactions with na tive plant species, spore longevity in the environment, effects of forest management practices, anthropogenic spread, and many more. The chapters which follow include further exploration of habitat/invasion associations, and herbicide management in an attempt to address some of these questions.

PAGE 44

30 S p o r e g e r m i n a t i o n S p o r e r e l e a s e 012345678910111213141516171819202122232425Weeks Gametophyte development Rapid sporophyte production Continued sporophyte production Adult plants forming climbing leaves. Figure 1-1. Timeline for Japa nese climbing fern plant development from spores. Figure 1-2. Worldwide distributi on of Japanese climbing fern ( Lygodium japonicum ) from literature review (June, 2006). A) Native range indicated in grey. B) Introduced range indicated in black. Dotted line indicates northern and southern extent of recorded Japa nese climbing fern field populations.

PAGE 45

31 Figure 1-3. County-level distri bution of Japanese climbing fern in the Southeastern United States with reference to the S outheastern Coastal Plain physiographic region. Records based on vouchered herbarium specimens from Southeastern herbaria, and records from regiona l vascular plant databases.

PAGE 46

32 CHAPTER 2 ASSESSMENT OF SITE VARIABLES ASSOCI ATED WITH JAPANESE CLIMBING FERN OCCURRENCE IN THREE NORTH FLORIDA FORESTS Introduction A simplified view of plant invasions includes two essential stages: transport of organisms or propagules to new sites, and establishmen t and population expansion in the invaded sites (Shea and Chesson 2002). Both stages are aff ected at some level by the conditions of the invaded site. On any given site or region, invader growth rate is affected by resources, the physical environment, ecological adaptations of the invasive species, and natural enemies. Invasive potential may be determined by envi ronmental factors including soil pH and nutrient levels, site climatic conditions and native plant community (L angeland and Burks 1998; Davis, Grime, and Thompson. 2000; Shea and Chesson 2002). Thus, the mechanisms that determine site invasibility by non-native species, in va rious ecosystems, are often unknown (Lavorel, Prieur-Richard, and Grigulis 1999). Japanese climbing fern is one species for which the details establishment success are poorly quantified. Some fern species are dependent on specific site conditions for successful establishment and sprea d, including climate, soil substrate, soil pH, elevation, and moisture regimes (Peck, P eck and Farrar 1990; Whittier and Moyroud 1993; Stewart 2002). Establishment a nd persistence of gametophyte popul ations are strongly affected by environmental events (Peck et al. 1990). Mo st ferns in the Schizaeaceae are vigorous in edaphically low nutrient environmen ts (Page 2002). This toleran ce opens a variety of sites to establishment, including intrinsically nutrien t poor savannahs, post wildfire sites, erosion surfaces, and epiphytic habitats (Page 2002).

PAGE 47

33 While the ecological requirements of Japane se climbing fern are poorly understood, its broad distribution indicates a tolerance for a ra nge of environmental conditions. In its native range, Japanese climbing fern occurs in forest ed ges, open forests, and secondary forests at both lower and higher elevations (Singh and Panigr ahi 1984, Gias Uddin et al. 1997, Ferriter 2001). In the Southeast United States, Japanese clim bing fern occurs primarily in floodplain forests, marshes, wetlands, secondary woods, moist pinela nds (including flatwoods) and disturbed areas such as road shoulders and rights-of-w ay (Clewell 1982, Nauman, 1993, Wunderlin 1998, Langeland and Burks 1998). Infestations also occu r in xeric sites such as sandhills, but do not appear to expand as rapidly as in more mesic sites. Observations of land managers indicate redu ced success in extremely dry or consistently flooded sites (Ferriter 2001) and a potential preference for soils with a pH close to 7 (Nauman 1993). Most ferns flourish at pH between 6 and 7 (Hoshizaki and Moran 2001). Fluctuating water levels may have limited impact on plant grow th, but foresters in nort h Florida reported that in the dry spring of 1999, Japanese climbing fern wa s not as serious of a problem as in previous wetter springs (Ferriter 2001). Hu man disturbance does not appear to be a prerequisite for successful invasion (Ferriter 2001), but populations are often associat ed with ditches and culverts due to increased site moisture levels, and possibl y due to associated soil disturbance as well. The objectives of this study are to classify and describe biotic and abiotic variables associated with the presence or absence, and perc ent cover of Japanese climbing fern, as well as those associated with varying leve ls of percent cover. Identification of clear associations between site environmental factors and fern establishmen t success may aid prioritization of treatments for managing Japanese climbing fern invasions, and improve understanding of habitat vulnerability to fern invasion.

PAGE 48

34 Methods and Materials Site Selection Three forested study sites were selected in North Florida (Figure 2-1), each containing large areas infested with Japanese climbing fern, and no hi story of herbicide treatment in the past 10 years. Sites were selected in areas with visually heavy local regional infestation levels, with the intention of reducing variability in site spore loading, and ther efore in propagule-based establishment opportunity. Sites were also select ed to maximize natural community variability, with multiple natural community types represente d at each study area. Plot measurements were taken between June and September, 2002. Plot Layout On each site, a square macroplot 400 m by 400 m in size (16 ha) was identified as the study area and subjected to ecological sampling. Percent cover of Japanese climbing fern was determined through line intercept (Myers and Shelton 1980), with intensive sampling plots distributed in equal distances al ong transect lines (Figure 2-2). In each study site, a baseline was established along a road marking a stand boundary and linear disturbance, or edge. Five parallel major transect lines were established, originating from the baseline at 100 m intervals, and extending 400 m into the stand, pe rpendicular to the baseline. A grid of 25 intensive sampling plots (plots), each 10 m by 10 m in size, were established at 100 m spacing along the major transect lines. The center of each plot was permanently marked. On each plot, five minor transect lines were established, each 10 m long and spaced 2.5 m apart. A grid of 25 sampling points (point s) was established every 2.5 m along these minor transect lines (Figure 2-2 ). Each point was mark ed with a pin flag. Percent cover of Japanese climbing fern was determined for all major tran sect lines and minor tr ansect lines through line intercept measurements.

PAGE 49

35 Site Descriptions Site 1 Blackwater River State Forest, Santa Rosa County (BRSF) is a publicly-owned 78,000+ ha property managed by the Florida Divi sion of Forestry. The study area included natural, fire-managed upland sandhill community, with a temporal stream drainageway cutting curvilinearly through a portion of the site (Fi gure 2-3). A large portion of the drainageway is heavily infested with Chinese privet ( Ligustrum sinense Lour). Site 2. The Miller property, in Ca lhoun County (Miller) is a privately owned forestland with portions of the study area in the followi ng community types: mesic flatwood slash pine plantation (previously managed for pinestraw production), botto mland hardwoods, longleaf pine upland, mixed pine/hardwoods, and cypress wetlands (Figure 2-4). Site 3. Florida River Island, in Liberty County (F RI) is a publicly-owned floodplain forest island situated between the Apalachicola and Florida Rivers, and managed by the Northwest Florida Water Management District. The study area is primarily cypress swamp, and floodplain swamp hardwood forest, including areas of higher elevation on ri dges and natural levees with limited soil saturation, and low flat swales which remain inundated for much of the year (Figure 2-5) (Leitman, Sohn and Franklin. 1983). Macroplot Measurements Physical characteristics of each site (16 ha macroplot) were assessed, including locally significant natural features (bodies of water, roads, etc.), and natu ral community boundaries; delineated using field mapping and interpretatio n of aerial imagery (Figure 2-3, Figure 2-4, Figure 2-5). Percent cover of Ja panese climbing fern in the unde rstory was measured using the line-intercept method along the 400 meter-long major transect lines to generate an average percent coverage by site. The s ite-level values were compared to average plot-level values on each site for significant differences in measured percent cover. Lack of agreement between the

PAGE 50

36 two measurements would have indicated an in adequate sampling design. The percent cover of Japanese climbing fern was defined as: the tota l linear distance covered by the foliage over the entire length of the transect lines within a si te, divided by the total transect line length and multiplied by (Myers and Shelton 1980). The same measurements were taken along the 10 meter-long minor transect lines on each plot, and the same technique used to generate an average percent coverage for each intensive sampling plot. Intensive Sampling Plot Measurements The majority of the classifying measuremen ts were taken on the 10 m by 10 m intensive sampling plots. For each plot, the following featur es were measured on each of the 25 sampling points within the plot to genera te a composite average value for the plot. Plot-level percent cover of Japanese climbing fern was measured and was the dependent variable for subsequent analyses. The species and diameter at breast he ight (dbh), was measured for all trees (minimum 2 inch dbh) on the 10 m by 10 m intensive sampling pl ots. These measurements were used to aid in typification of site, and allow for calculation of plot basal area (m2 ha-1), species richness, species diversity, species evenness, and species importance values. Understory vegetation, and midstory vegetati on percent cover were recorded as ocular estimates of all plant material excluding Japanese climbing fern Overstory percent canopy cover was determined visually using a non-spherical dens iometer. Values were recorded on a scale with 100% being full canopy with no light pe netration, and 0% being no canopy overhead. Relative light penetration (mol m-2sec-1) was determined by measurements of photosynthetically active radiation (PAR) with a Li-Cor light mete r inside the stand and in the open. Average light intensity for each intensiv e sampling plot was obtained by measuring PAR just above the forest floor ve getation over the 25 sampling points for each plot, and averaging these values to obtain a plot average. These av erages were then corrected with a relative PAR

PAGE 51

37 factor measured under zero canopy conditions for the respective sampling period to determine the relative light penetration fo r each intensive sampling plot measured that day. Light readings were taken between 10:00 am and 2:00 pm on days with no cloud cover to ensure uniformly peak light conditions. Site hydrology was described with thr ee variables: Hydrol ogy class (Xeric=1, Xeric/Mesic=1.5, Mesic=2, Mesic/Hydric=2.5, Hydric=3) assigned according to site observations, percent of the plot flooded fr om winter 2002 through spring 2003, and the number of months each plot remained flooded. Plot topogr aphic relief was ocularly assessed as concave, convex, or flat. Plot percent slope was dete rmined using a hypsometer. Slope aspect was determined with a compass. A soil core (diameter 2 cm, depth of 15 cm) was collected with a soil probe at each of the 25 points per plot to generate a composite plot so il sample. Composite soil samples were allowed to air dry, and screened to pass a 2 mm sieve prio r to analysis for: macronutrients, organic matter content and soil pH. Mehlich-1 extractable P, K, Ca, Mg, Cu, Mn, Zn, Al, Na, and Fe were measured using the Mehlich1 extraction solution (0.0125 M H2SO4 and 0.05M HCl) to provide a soil solution ratio of 1:4. Units for all elem ents are recorded as mg per kg. Soil pH was measured using a 2:1 water to soil ratio. Soil organic matter percent was measured through Walkley-Black dichromate methodology. This method was originated to assi st in the selection of herbicide rates for mineral soil contai ning less than 6% organic matter. Data Analysis Plot environmental variables and calculated indices were usedto assess relationships between variables and the presence or absence of Japanese climbing fern across all plots. To improve interpretation, results of these analyses are presented in four functional groups: plot abiotic variables, plot soil va riables, plot vegetation variable s, and tree species importance

PAGE 52

38 values. Variables were further assessed for re lationships between three levels of Japanese climbing fern percent cover across plots w ith Japanese climbing fern present. Plot abiotic variables include : photosynthetically active radi ation, percent slope, percent exposed soil, plot hydrology class, plot percent flooded (2003), and months flooded (2003). Plot soil variables include: pH, P, Ca, Fe, Mg, Al, K, Zn, Na, Mn, Cu, and percent organic matter. Plot vegetation variables include: understory percent cover, mi dstory percent cover, and tree canopy percent cover. Plot tree inve ntory data were utilized for th e calculation of additional plot vegetation variables: tree basal area (m2/ha), tree species richness (spe cies per plot), Simpsons Reciprocal (1/D) Index (Magurran 1988) of tree sp ecies diversity, and Piel ous J Index of tree species evenness (Pielou 1975). Tree importance values were calculated for each species on each plot using the equations below: IMPORTANCE VALUE: Relative density + relative dominance + relative frequency. DENSITY: Total stems of species/total plot area. RELATIVE DENSITY: (Density of species/sum density of all species) X 100. DOMINANCE: Mean basal area per tree of species X number of trees in species. RELATIVE DOMINANCE: (Dominance of species/su m dominance of all species) X 100. FREQUENCY: (Plots with species / Total number of plots) X 100. RELATIVE FREQUENCY: (Frequency of species/sum frequency of all species) X 100. Data Transformation. Anderson-Darling test statis tics were calculated to assess normality of distributions for each environmen tal variable, across all study plots (n=75, alpha=0.01). Distributio ns of most variables were non-norma l. Non-normal data were then transformed and normality re -assessed. Natural log, log10, and log2 transformations were

PAGE 53

39 performed on all variables; a constant (value=1 ) was added to each variable value prior to log transformation to allow inclusi on of zero values. Additionall y, inverse, and square root transformations were performed on all variables, and arcsine transformations were performed on variables with percentage values. Transformed data showing the stronge st probability of normal distribution for each variable were utilized in su bsequent parametric analyses of environmental data. Tree importance values were assessed using non-parametric analyses. The data subset including only plots with Japanese climbing fe rn (n=29) was also assessed for normality of distribution. In this case most variab les had normal distribution of values. Initial Analyses Presence/absence. Initial analyses were conducte d to measure the significance of differences in plot environmental variables a nd tree importance values, on the presence or absence of Japanese climbing fern. Significant differences in variable mean values between plots with Japanese climbing fern present versus absent were assessed using the Kruskal-Wallis one-way rank analysis of varian ce test (alpha=0.05) for non-norm ally distributed variables. Kruskal-Wallis p-values are calculated using the chi-square approximation (Siegel 1988). Variance equality was assessed for normally dist ributed plot environmen tal variables using the folded F-statistic (alpha=0.05). Significant differe nces in normal variable mean values between plots with Japanese climbing fe rn present versus absent we re assessed using a pooled twosample t-statistic for variables with equal variance, and Satterthwaites two-sample t-statistic for those with unequal variances. Significant correla tions between variable mean values on plots with Japanese climbing fern present versus absent were assessed using the Spearman R correlation coefficient (alpha=0.05) for non-nor mal variables, and the Pearson correlation coefficient (alpha=0.05) for normal variables. Correlation strength ratings were assigned based on coefficient values: weak=0.2 to 0.49, modera te=0.5 to 0.79, strong=0.8 to 1.0 (Cohen 1988).

PAGE 54

40 Cover class. Separate analyses were conducted for all plots with Japanese climbing fern present (n=29) from all sites, to assess relatio nships between three Japanese climbing fern cover classes, and plot environmental variables and tree importance valu es. Normal distributions were achieved with raw data for most variables, allo wing use of parametric tests where appropriate without data transformation. Cove r classes utilize the following ra nks: 0= 0%, 1= >0 to 5%, 2= >5 to 25%, 3= 26 to 75%, 4= 76 to 95%, 5= 96 to 100% (Horvitz and Koop 2001). All plots fell in cover classes 1, 2, or 3. Parametric one-way analysis of variance te sts (alpha=0.05), and nonparametric Kruskal-Wallis analysis of varian ce tests (alpha=0.05) were performed on all environmental variable and tree impo rtance values by cover class. Discriminant Analysis Plot environmental variables and tree species found to vary significantly according to Japanese climbing fern presence or absence, and Japanese climbi ng fern cover class levels were evaluated in stepwise discrimina nt analyses of covariance. An alyses were conducted on plots at the study-level (n=75) and site-level (n=25 each). Plot variables and tr ee values were assessed separately. Significance was determined by th e results of the Kruskal-Wallis analysis of variance tests and two-sample t-tests. Power fo r variable entry or removal in the discrimnant analysis (alpha=0.15) was measured by the W ilks Lambda likelihood ratio criterion (SAS Version 9.1). Data were assessed at the study-level and the site-level. Regression Analysis Multiple logistic regression analysis was pe rformed on significant variables from the Japanese climbing fern presence/absence assessme nt, and the Japanese climbing fern cover class assessment. Spearman R correlation coefficients were calculated across all variable to assess levels of variable correlation. Principal components analysis was used to combine strongly

PAGE 55

41 correlated (correlation coefficient exceeds 0.8) environmental variable s into the component factors for use in the regression analyses. Neural Network Analysis Artificial neural network analysis was conducte d utilizing all plot va riables to develop a deterministic classification model to assess the abili ty of the data set to correctly classify plots according the presence or absence of Japanese climbing fern. The model was developed as a C program, and modified for use in Python. Pl ots were assigned or values according to Japanese climbing fern absence or presence resp ectively; these values served as the supervised network teacher. A non-linear ac tivation function in the model used a radial basis transfer function, which assumes a Gaussian curve. Netw ork inputs, hidden layers (the radial basis transfer function), and networ k outputs are connected by weight multipliers, summed by nodes. Results Results of respective analyses are presen ted in Tables 2-1 th rough 2-10, and appendix tables, and Figures 2-3, 2-4, and 2-5 at the end of this chapter. To improve interpretation, results of initial statistical tests are presented in four functional gr oups: plot abiotic variables, plot soil variables, plot vegetation variables, and tree species importance values. Site-Level Occurrence Percent cover of Japanese climbing fern on a ll sites is presented in Table 2-1, including comparisons between plot measurements taken in 2002, and re-assessed on two sites in 2004. Among the three sites, Florida River Island (FRI) ha d the highest site-level and plot-level percent cover of Japanese climbing fern in 2002 (13.7%), a nd the greatest number of plots infested (18), but it could not be re-assessed in 2004 due to a herbicide treat ment applied to the Japanese climbing fern on the site in 2003. The Miller site had the second highest si te-level percent cover (3.00%), with eight plots infested in 2002, while percent cover on the Bl ackwater River state

PAGE 56

42 forest site was lowest (1.3%), due to a small nu mber of infested plots (3) with lower individual percent cover levels. The number of plots with Japane se climbing fern present increased from 2002 to 2004 by one plot each on the Miller site a nd the Blackwater River State Forest (BRSF) site. Plot percent cover of Japanese clim bing fern increased (0.4%) from 2002 to 2004 on the Miller site, but decreased (0.6 %) on the BRSF site over the same time period. Two-sample ttests revealed that neither cha nge was significant (alpha=0.05). Plot-level Ja panese climbing fern presence and percent cover values from 200 2 are used in statistical analyses. Macroplot (site-level) cover was not re-assessed Japanese climbing fern cover classes are i llustrated in Figure 2-3 (BRSF), Figure 2-4 (Miller), and Figure 2-5 (FRI). Across all site s, plots on xeric pine uplands were uninfested, while mesic pine community types (pine plan tation, mesic flatwoods mixed pine/hardwood), and hydric community types ( hydric cypress wetland, hydric hardwood swamp, and hydric floodplain forest: ridge, swale, and slough) each averaged a 55% Japanese climbing fern occurrence level. Table 2-2 indicates Japanese climbing fern presence/absence levels according to natural community type. Plot-Level Occurrence Presence/Absence Plot environmental variables. Means of the following plot ab iotic and biotic variables varied significantly (alpha=0.05) according to Ja panese climbing fern pr esence (in order of significance): site, PAR, unders tory percent cover, canopy percent cover, Pielou tree evenness index, hydrology class, Simpsons reci procal tree diversity index, tree species richness, months flooded, percent flooded, and trees per plot. Unde rstory percent cover an d PAR were negatively correlated with Japanese climbing fern presence ; all other significant variables showed positive correlations. Correlation strengths were w eak (Table 2-4, Tables A-1 to A-4).

PAGE 57

43 In particular, photosynthetically activ e radiation ranged from 133 mol s -1m-2 on sites with Japanese climbing fern present to 455 mol s -1m-2 on sites without Japanese climbing fern. Hydrology class increased in moisture level w ith Japanese climbing fern presence. Flood duration averaged two months on sites with Japa nese climbing fern compared to one month on sites without Japanese climbing fern, and percent flooded increased from 15% to 27% with the presence of Japanese climbing fern. Means of the following plot soil variables varied significantly (alpha=0.05) in parametric tests of relationship to Japane se climbing fern presence: pH, phosphorus (P), calcium (Ca), iron (Fe), magnesium (Mg), aluminum (Al), potassium (K), zinc (Zn), and sodium (Na). Manganese (Mn), copper (Cu), and percent organic matter did not vary significantly (Table A-3). Phosphorus and pH were moderately correlated with fern presence, while other significant variables were weakly correlated. Iron and aluminum demonstrated negative relationships with fern presence, while other significant variables were positively correlated. Soil pH on sites with Japanese climbing fern present averaged 6.0, while sites without Japanese climbing fern were slightly more acidic, averaging 5.2. Soil alum inum was significantly higher on plots with Japanese climbing fern absent (332 mg kg-1) than plots with fern present (192 mg kg-1). Conversely, higher levels of soil calcium occurred in plots with Japanese climbing fern present (284 mg kg-1) than absent (101 mg kg-1). Plot tree species. Importance values were analysed by species, comparing mean values for plots with Japanese climbing fern present versus absent. Fortythree tree species were identified across all plots, of which twenty-one occurred on one plot only. Species with importance values that varied significantly acco rding to Japanese climbing fern presence or absence include: Liquidambar styraciflua, Pinus palust ris, Carya aquatica, Quercus lyrata,

PAGE 58

44 Ulmus americana, Carpinus caroliniana, Quercus nigra and Taxodium distichum (Table 2-7, Table A-4), while those not exhibiting a significan t relationship are listed in Table A-5. Mean importance values for the majority of significan t tree species demonstrate a positive relationship with Japanese climbing fern pr esence; the one exception being P. palustris which demonstrates a moderate negative relationship. Ot her significant species are, at best, weakly correlated with Japanese climbing fern presence. Tree species were assigned vegetative index classes according to the Florida Department of Environmental Protections Wetland Delineati on Index (Sec.62-340.450, F.A.C.) (FLDEP 2006). Analyses were conducted to assess relationshi ps between tree importance values within vegetative index classes and Japanese climbing fe rn presence or absence (Table 2-3). Tree importance values on plots with Japanese climbi ng fern present versus absent were compared, significant differences between Japanese clim bing fern presence and absence in importance values for facultative upland trees, and obligate wetland trees. Analysis of variance detected significant (P=0.005) variation in importance values between ve getative index classes on plots with Japanese climbing fern pres ent, as well as on plots with Japanese climbing fern absent (P<0.0001). Cover Class Levels Plots with Japanese climbing fern presen t were analyzed as a subgroup to assess relationships between thr ee Japanese climbing fern percent cover levels (raw data or cover class), and plot environmental variables and tree impor tance values. No signi ficant (alpha=0.05) relationships were detected betw een plot and tree variables an d Japanese climbing fern cover class in parametric one way analysis of variance. However, significant (alpha=0.05) nonparametric relationships were detected betwee n percent cover of Japanese climbing fern and three variables, including plot basal area which peaked (30 m2 ha-1) on moderately infested (class

PAGE 59

45 2) plots. Soil aluminum reached 249 mg kg-1 on moderately infested pl ots, an average of 70 mg kg-1 more than on slightly (class 1) or heav ily (class 3) infested plots. Occurrence of Carpinus caroliniana also showed significant re lationships to Japanese climbing fern cover class (Table 2-4). Significant (alpha=0.05) corr elations to Japanese climbing fe rn cover class were measured for basal area, plot percent flooded, photosynthe tically active radiation, and the occurrence of Carpinus caroliniana and Quercus nigra (Table 2-4). Discriminant Analysis Stepwise discriminant analysis of covarian ce was used to further explore relationships between plot variables and Japanese climbing fern occurrence, and cover class. Variables were measured for relationships at the study (n=75) le vel, and at each respective site (n=25). Across all study plots, soil calcium was the most si gnificant discriminator (P <0.0001) of Japanese climbing fern presence or absence (Table 2-5). Site level analyses revealed relationships between other variables including percent fl ooded (P=0.02), and months flooded (P=0.0001) at BRSF; while soil iron is signifi cant (P=0.002) at the Miller site, and tree species richness is significant (P=0.03) at FRI. (Tables 2-5 and 2-6) Of the variables that discriminated between Japanese climbing fern presence or absence (Table 2-5), percen t flooded is the only one also significant (P=0.09) as a discriminator (alpha=0. 10) between Japanese climbing fern cover class levels (Table 2-6). In addition, soil aluminum and plot basal ar ea were also both significant, discriminators of Japanese climbing fern cover cl ass at the study level. Site level analyses revealed the strongest relationshi p between variables and Japanese climbing fern cover class, to be soil aluminum at the Miller site (P=0.01), followed by per cent flooded at the Miller site (P=0.14). At FRI, basal area and soil aluminum were significant (alpha=0.15) discriminators of Japanese climbing fern cover class.

PAGE 60

46 Quercus nigra was a significant discriminator betw een Japanese climbing fern presence and absence at the study level (n=75, P<0.0001) a nd at BRSF (n=25, P<0.0001) (Table 2-7); as well as between cover class levels at Florida River Island (Table 2-8). Other significant tree species were discriminators of Japanese cl imbing fern presence or absence, such as: Pinus palustris (P<0.0001) and Liquidambar styraciflua (P=0.0005) at the study level, Ligustrum sinense at BRSF (P<0.0001), Pinus elliotii at the Miller site (P=0.02), and Nyssa aquatica (P=0.02) and Acer rubrum (P=0.08) at FRI. Carpinus caroliniana was the only significant discriminator of cover class (P=0.02) at the study level. Discriminant analysis was not possible for BRSF due to the small number of plots with Japane se climbing fern present; and no tree species were accepted into the analysis for the Miller site. Regression Analysis Multiple logistic regression analysis was performed on significant variables from the Japanese climbing fern presence/absence and co ver class assessments. Principal components analysis was used to combine strongly co rrelated (correlation coefficient exceeds 0.8) environmental variables into the following f actors: FLOOD=Hydrology cl ass + plot percent flooded + months flooded, SOIL1=pH (inverse) + phosphorus (inverse) + cal cium (natural log), SOIL2=Zn + K (natural log), TREE=Pielou spec ies evenness index + Simpsons reciprocal species diversity index + species richness. Prin cipal components analysis explained more than 80% of the variance among factor components for al l factors, indicating s ubstantial agreement, and validating use of factors in subseque nt multiple regression analysis. Variables generating significant effects in the regression analyses are listed in order of forward selection into the regre ssion models in Tables 2-9 (pre sence/absence) and 2-10 (cover class). The first entry into th e presence/absence model is an in tercept (value=0.4), followed by the principal component factor SOIL1, accounting for 40% of the variation in the model.

PAGE 61

47 Following general site classification by those soil pH correlates, the additional cumulative effect of importance values of Ligustrum sinense, Quercus nigra, Lindera benzoin Acer rubrum Quercus laevis and Nyssa aquatica accounting for a maximum of 68% of the model variation related to presence or absence of Japanese clim bing fern. When the presence/absence analysis was conducted for each respective site, few variables were entered into the model. At BRSF, the principal components analysis factor TREE wa s entered, accounting for a cumulative 49% of model variation. At the Mill er site, the SOIL1 factor en tered the model, followed by Pinus palustris accounting for 62% of cumulative model variatio n. At FRI, no variables were entered. A similar number of variables produced significan t effects in analysis of Japanese climbing fern cover class levels. The firs t entries into the cover class mode l were an intercept (value=9.2), followed by importance values for Carya aquatica, accounting for a cumulative 17% of model variation. A significant increase in model st rength came with additional of the principal components analysis factor FLOOD1, raising model accuracy to 38%. Six tree species: Catalpa bignoniodes, Platanus occi dentalis, Nyssa ogeechee, and Quercus laevis were accepted in the regression analysis, accounting for a total cumulativ e model variation of 75%. When the cover class analysis was conducted for each respective site, the only vari able entered into the model was FLOOD1 at the FRI site. No variables were entered for BRSF, or the Miller site. Neural Network Analysis The neural network analysis selected thirty plot variables and tree importance values for input into the model. Variab les positively related to Japane se climbing fern presence are indicated with a (+), while variables positively rela ted to absence are indict ed with (-). In order of rank in the model, they include: unde rstory percent cover (-), position (-), Pinus palustris importance (-), Quercus nigra importance (+), Ligustrum sinense importance (+), soil iron (-), percent exposed soil (-), basal area (+), Quercus falcata importance (-), Nyssa sylvatica

PAGE 62

48 importance (-), percent slope (+), soil phosphorus (+), per cent canopy cover (+ ), site, soil aluminum (-), soil calcium (+), hydrology class (+), soil magnesium, Bumelia lycoides importance (-), Quercus stellata importance (-), Liquidambar styraciflua importance (+), soil pH (+), Quercus laevis importance (-), Taxodium distichum importance (+), soil sodium (+), soil potassium (+), midstory percent cover (-), mont hs flooded (2003) (+), relief (-), and soil organic matter percent (+). Utilizing these variables, th e neural network correctly classified Japanese climbing fern presence for 22 of 29 plots, or 76 % of the time; Japanese climbing fern absence was correctly classified for 38 of 46 plots, or 83% of the time. Discussion In field observations, some trends in the pres ence and percent cover of Japanese climbing fern are observable. Identificati on and quantification of the specific causal agents of these trends is experimentally challenging, but some general re lationships are discernible from the results of this study. Some aspects of study design, primar ily a limited number of plots study-wide; and the macroplot sampling design restrict applic ation of most results beyond the study sites themselves. The following discussion begins wi th relationships within three main variable categories: site overview, plot environmental vari ables, and tree species. The final two sections, multiple logistic regression, and neural network analysis, focus on relationships across variable categories, seeking the most valid or predictive combination of variables. Twenty plot environmental variables and ei ght tree species dem onstrated significant variance in mean values related to the presence or absence of Japanese climbing fern (Tables: A1, A-2, A-3, A-4) in initial analyses. The major ity of these significant va riables and tree species either affect or indicate as pects of site hydrology (and asso ciated soil chemistry) and photosynthetic opportunity (e.g., hydrology class, percent flooded, photosynthetically active radiation). This reflects genera l fact that in Florida, hydrolog y, pH and fire exert overwhelming

PAGE 63

49 influence over site vegetation, mi nimizing effects of other factor s; including soil nutrient supply (Brown, Stone, and Carlisle 1990). The sensitivity of statistical tests used in each analysis should be considered in interpretation of all results, part icularly in consideration of forested sites outside the local region, or with different natural community types represented. Site Overview Percent cover of Japanese climbing fern vari ed significantly among site s (Table 2-1). The general grouping of natural commun ity types by site affects a broad range of variab les, including Japanese climbing fern occurrence. The primaril y hydric Florida River Is land (FRI) site had the highest site-level and plot-level percent cover of Japanese climb ing fern. Low areas and swales on this site flood annually with the elevation of water levels in the winter and spring like the majority of floodplain and bottomland hardwood fo rests along the Apalachicola River. Under these conditions, soil saturation is the prim ary limiting factor affecting climbing fern establishment and survival. Hi gher elevation areas in this fo rest type can become heavily infested with Japanese climbing fern as a resu lt of their generally mesic to hydric nature, and extensive flood-based dispersal of spores and on site via the annual fl oods (Figure 2-5). In addition to site characteristics, regional invasion history may also contribute to higher presence and percent cover of Japanese climbing fern for sites in the Apalachicola River basin (BRSF and Miller site), where the plant has been estab lished for several decades. The xeric to mesic Blackwater River State Forest had the lowest site-level percent c over of Japanese climbing fern (Table 2-1). This reduced site -level occurrence is primarily determined by the spatially limited nature of sites with appropriate hydrology (Fi gure 2-3), although Japanese climbing fern is increasing in occurrence in both longleaf pine ( Pinus palustris ) and slash pine ( Pinus elliottii ) forests on the site. The majority of Japanese climbing fern on the mesic to hydric Miller site

PAGE 64

50 occurred in the pine plantation area and ecotonal areas between upland and lower swamp hardwoods on the eastern portion of th e site (Figure 2-4, Table 2-2). The number of plots with Japanese climbi ng fern present increased by one on both the Blackwater River State Forest a nd Miller sites from 2002 to 2004. Foresters and land managers throughout Florida have also noted visually obvious increases in Japanese climbing fern cover and distribution throughout the no rth Florida region over the last decade (Ferriter 2001). This trend is likely to continue as local and regiona l propagule pressure from established plants and populations increases, particular ly given the limited nature of environmental restrictions on climbing fern occurrence. Plot-level percent co ver of Japanese climbing fern increased by 0.4% from 2002 to 2004 at the Miller site, but decreas ed by 0.6% at Blackwater River State Forest. Neither change was significant. The decrease in plot percent cover at BRSF may reflect the effect of visually high level of fo liar damage caused by the rust fungus Puccinia lygoddii in 2004, coupled with a small number of plots with hydrology appropriate fo r Japanese climbing fern. The increase in plot-level occurrence while also not significant, does document the continued expansion of this species on infested s ites, either through natura l dispersal, or possibly through human-vectored spore dispersal as a re sult of repeated entr y into the plots by researchers. Discriminant Analysis Plot environmental variables Stepwise discriminant analyses of plot variables (alpha=0.10 ) furthered support for the role of general site hydrology in determining occurren ce of Japanese climbing fe rn. Specifically, soil calcium, and soil iron, plot percent flooded, and plot understory percent cover discriminated between Japanese climbing fern presence and abse nce at the study level, maintaining support for the importance of site mesic to hydric hydrology in climbing fern establishment. Clawson,

PAGE 65

51 Lockaby and Rummer (2001) found soil calcium to increase as flooding levels increased on soils ranging from somewhat poorly drained to poorly dr ained within a floodplain forest. In addition to hydrologic variables, site-lev el analysis identified the tree evenness index as a significant indicator at Blackwater River Stat e Forest, probably due to the na turally low tree evenness in the xeric upland longleaf pine comm unity dominating this site (Fi gure 2-3). As a result, the significance of evenenss in this analysis proba bly reflects site hydrology and natural community type, rather than indicating a strong relationship between climbing fern occurrence and eveneness. This is further supported by the hi gh occurrence of Japanese climbing fern on the Pinus elliottii plots at Miller, which also had low eve nness values. Tree species richness was significant at Florida River Island, primarily due to the low richness measured on swale plots (Figure 2-4) which remain inundated for a large part of each year, thereby reducing vegetative establishment opportunity for many species including Japanese climbing fern. Discriminant analysis of plot s with Japanese climbing fern for relationships with varying cover class levels identified thr ee plot variables as discriminators Basal area was significant at the study level, driven by the pol ynomial relationship to Japanese climbing fern cover classes at Florida River Island, where basal area decreases significantly as Japanese climbing fern percent cover exceeds 25%. Plot percent flooded displa yed a negative discriminatory relationship with Japanese climbing fern cover, indicating that while adequate moisture is necessary for establishment, excessive moisture (and shortened growing season) will limit percent cover. Plot tree species Tree species showing significant relationships to Japanese cl imbing fern presence (Table A-1) are primarily classified as facultative we tland or obligate wetland species according to the vegetative index used for wetla nd delineation in Florida (Table 2-3 and Table A-4) (FLDEP 2006). Facultative upland species an d facultative species were signi ficantly less important than

PAGE 66

52 facultative wetland and obligate we tland species on sites with Japa nese climbing fern present, indicating an association with mesic to hydric site conditions. In comparison, no significant difference was measured between vegetative indices on plots with Japanese climbing fern absent, indicating that absence is affected by more than vegetativ e index of dominant species (e.g., establishment opportunity). Stepwise discriminant analysis of tree species importance values identified Quercus nigra, Liquidambar styraciflua, and Taxodium distichum as significant discriminators of Japanese climbing fern presence. Conversely, Nyssa sylvatica was a significant discriminator of Japanese climbing fern absence at the study level. Site-lev el analyses also identified the invasive plant Ligustrum sinense as a discriminator of Japanese climbi ng fern presence at Blackwater River State Forest. The relationship shown between th ese species indicates tw o key factors which are easily recognizeable in the field: the hydrologi c function of the exte nsive natural stream drainageways on the site provide a moisture re gime allowing for establishment of Japanese climbing fern and L. sinense across the site landscape; and movement of water through the system of drainageways on the site has proba bly played an important role in vectoring propagules of both species. The natural functi on and management of these sites are both complicated by the co-occurring popul ations of these species, as well as an additional non-native invasive plant, Lonicera japonica Removal of the often dens e canopy and/or midstory of L. sinense has resulted in rapid expans ion of Japanese climbing fern in multiple cases. In addition to the documented relationship at Blackwater River State Forest, the pine plantation plots at the Miller site also had co-occuring infest ations of Japanese climbing fern and L. sinense though plants were too small to census with plot trees

PAGE 67

53 Pinus elliottii is a significant discriminator for th e presence of Japanese climbing fern on the Miller site, as a result of the P. elliottii plantation covering the western portion of the study area (Figure 2-4). This site bias does reflect a significant regional relationship between P. elliottii plantations and Japanese climb ing fern. High levels of Ja panese climbing fern invasion are clearly visible throughout mesi c pine plantations in the Apalach icola River basin in Florida. One study documented a 22% Japanese climbing fern occurrence level in North Florida P. elliottii plantations, a forest type covering 5.1 milli on acres in the State (Van Loan 2006). Silvicultural management practices on these site s may be contributing to Japanese climbing fern dispersal and site occupancy levels as a re sult of canopy removal, soil disturbance, and equipment-based spore dispersal. Significant increases have been seen in fern coverage and density on flatwoods and mesic plantation fore st types following timber harvest, pinestraw harvest, and prescribed bu rning in particular. Discriminant analysis of plot s with Japanese climbing fern for relationships with varying cover class levels identified four tree species as discriminators. Increasing importance of Carpinus caroliniana was significantly related to increa sing occurrence of Japanese climbing fern. A common associate of Q. nigra Carpinus caroliniana is considered to demonstrate best growth and development on rich wet-mesic s ites (Burns and Honkala 1965), further supporting the observations of the importance of general si te hydrology in affecting Japanese climbing fern establishment and spread on infested sites. Regression Analysis The multiple regression analysis further s upported the importance of mesic to hydric conditions for Japanese climbing fern establishm ent and spread on a site. The SOIL1 factor (Tables 2-9 and 2-10), reflecting combined asso ciations between soil ph osphorus, soil calcium, and pH, varies in association with site hydrology, particularly flooding. Grouping the results of

PAGE 68

54 the discriminant and multiple regression analyses reveals five soil variables (calcium, iron, potassium, phosphorus, and pH) associ ated with presence or absen ce of Japanese climbing fern, and one soil variable (aluminum) associated with cover class on infe sted sites. Each of the soil variables associated with the presence or absence of Japanese climbing fern varies in response to multiple factors of parent material, hydrology, mois ture source, and habitat inputs. Taken as a group, calcium, phosphorus, pH, and potassium levels in this study were significantly higher on plots with Japanese climbing fern present, typically the mesic to hydric sites, as opposed to absent. In a simplified compar ison, iron and aluminum in this study were significantly lower overall on plots with Japanese climbing fern pr esent; however, analysis of fern presence or absence at the site level provide d an expanded perspective. For many soil variables, a separation existed be tween values measured at FRI versus the other two sites. Calcium, phosphorus, and pH were all significantly higher at FRI compared with BRSF and Miller, while iron was si gnificantly lower at FRI. Iron availability in solution is influenced by soil pH, aeration, and reactions with organic matter. In pa rticular, iron decreases as pH increases, reaching a minimum at pH 7.4 to 8.5. Additionally, poo r soil aeration (reduced oxygen levels) caused by flooding can prompt oxidation and/or reduction of iron when influenced by other conditions (Dixon and Schulze 2002). The saturation of soils on FRI reduces soil iron, increasing its solubility, and allowing its movement out of the surface soil layer (Dixon and Schulze 2002). This effect is different from the brief, infreque nt periods of saturation in the upland drainageway at BRSF, which produced higher iron concentrations, probably due to moisture-induced oxidation (Di xon and Schulze 2002). However, th e significantly higher sitelevel occurrence of Japanese climbing fern at FR I is probably linked more to the general site

PAGE 69

55 hydrology which prompts the soil variable trends, rather than being associated with the individual soil variables themselves. The significance of soil aluminum is distingu ishing between Japanese climbing fern cover classes on infested plots was an artifact of two situations. Al uminum values on the infested BRSF upland drainageway pl ots and the mesic Miller plots were significantly higher than FRI values, reflecting the effect of lower pH and ex treme weathering in the mo re sandy soils seen at these sites. Low pH values have been shown to prompt a peak in available soil aluminum as seen in sites which repeatedly flood, and then drain (Darke and Walbri dge 2000). The addition of the higher values from the Miller and BRSF plots creates the polynomial relationship observed between Japanese climbing fern cover class and soil aluminum on infested plots. Aspects of site hydrology heavily influence fl uctuation in soil condi tions and components. Aluminum and iron are known to be important in controlling the retention of dissolved inorganic phosphorus in wetland soils, providing further su pport for positive relationships between those three variables (Darke and Walbridge 2000). Da rke and Walbridge (2000), and Clawson et al. (2000) recorded increasing levels of soil calcium in soils from a gradient of sites ranging from upland to floodplain swale, explai ning the trends in calcium obse rved, particularly at the site level on Miller and BRSF, both sites including a mixture of pine dominated habitat and hardwood dominated habitat. Unlike FRI which is dominated by hardwoods and slightly acidic to neutral soils. Soil pH varied significantl y by site due primarily to hydrology and natural community; with higher pH values (5.5 to 7.2) on the hydric FRI hardwood floodplain forest site, and significantly lower pH values (4.6 to 5.9) on mesic Miller site, followed by the xeric BRSF sandhill site (4.6 to 5.0), a trend seen in these habitat types stat ewide (Londo, Kusha, and Carter 2006, Platt and Schwartz 1990).

PAGE 70

56 As in the discriminant analysis, the effects of Quercus nigra and Carya aquatica influenced Japanese climbing fern occurr ence. These two species, together with Ligustrum sinense may be used as general indicators of site conditions appropriate for Japanese climbing fern occurrence in northwest Florida. However, primary application of this result should be restricted to similar natural community types on th e same managed areas as utilized in the study. Several years of unquantified obs ervations support the finding that Japanese climbing fern presence may be strongly correlated with the presence of L. sinense on a landscape-scale in North Florida. This tendency toward co-l ocation may also extend into Georgia, where L. sinense is widely established in upla nd drainageways, bottomland hardw ood and floodplain forests, sites with higher levels of Japanese c limbing fern in North Florida. Neural Network Analysis Artificial neural networks are attempts to predict complex, non-linear relationships based on a loose analog with living nervous systems. These networks have been used to classify samples according to discrete classes (Mehrotr a, Mohan and Ranka 2000), in this case the presence or absence of Japanese climbing fern. Though the network performed relatively well in classification of plots with cl imbing fern present and absent, the number of variables included (30) indicates overall model weakness in a si milar sense to loss of degrees of freedom in regression analysis. However, the predictive pe rformance does allow for some interpretation if primarily for the managed areas on which study sites were located. While, the model primarily selected variables which had demonstrated significant relationships with Japanese clim bing fern presence or absence in previous analyses, it selected some new variables, including position, relief, mi dstory percent cover, and upland tree species such as Quercus stellata and Quercus falcata The addition of thes e factors in the model network may indicate non-linear relationships that standard t echniques cannot correctly model

PAGE 71

57 due to non-collinearity or other statistical assumptions in these models. It is also likely that some factors are only useful for this particular dataset, due the limited sample size. However, all of the data available were used in the fitting of this model, leaving none for model validation. The basic model output does pr ovide support for the initial study hypothesis that a combination of site variables may be iden tified that affect the pr esence or absence of Japanese climbing fern on a site. However, the model itself serves only as a hypothesis without independent validation. If future research is planned on classifi cation of sites vulnerable to Japanese climbing fern invasion, a recommendation for study design is to use a larger number of plots on which a smaller number of variables meas ured may enable use of more robust analyses, provide adequate data for model validation, and reduce some amount of correlation between variables. Data gathered in any future site cl assification studies may be useful in validation of this model, or construction of others. Conclusions The reproductive strategy of Japanese clim bing fern plays a significant role in the establishment and spread of this species. Baker s Rule states that species that are capable of uniparental reproduction are more likely to be successful i nvaders, (Richardson 2004). The results of Foxcroft, Richardson, Rouget and MacFa yden (2004) suggest that propagule pressure can act as a fundamental driver of invasi ons, significantly reducing the importance of environmental limitations (Volin et al. 2004). Th is situation may at leas t partially explain the lack of strong results from this study. However, some relationships between site characteristics and Ja panese climbing fern invasion do exist, the most influe ntial of which is general site hydrology. Hydrology has a major role in determination of natura l communities and subsequent site environment and vegetation. In particular, flooding extent and duration have several impacts on ve getative communities,

PAGE 72

58 affecting growing season duration and estab lishment opportunity, and soil biogeochemical processes. A similar assessment examined the correlation between the distribution of Old World climbing fern and abiotic factors in two forested wetlands in South Florida (Volin et al. 2004). In that study, Old World climbing fern occurred on 32% (n=43) of the sampling units, a similar ratio to this study in which Japa nese climbing fern occurred on 39% (n=29) of the sampling units (plots). Differences in the sampling design a nd availability of existing distribution data permitted use of ordination analyses not appropria te for this data set (Volin et al. 2004). However, general conclusions were similar to thos e in this study. Similar to Japanese climbing fern in North Florida, presence of Old Worl d climbing fern was found to be significantly correlated with site hydrology, specifically a wet but not permanently inundated condition. Future studies of Japanese climbing fern i nvasion establishment ecology should be focused in the northern extent of its range in the southeastern United Stat es. If propagule pressure is the primary determinant of the invasive potential of this species, the majority of mesic and many hydric sites in Florida may already be vulnerabl e to Japanese climbing fern invasion. Rather than focus heavily on site vulnerability, a critical next step in study of th e ecology of this species is to quantify the effects of i nvasion on habitat quality and functi on, and native species diversity. The answers to these questions will be key in understanding the impact of Japanese climbing fern invasions and accurately determining the true priority of aggressive prevention and control actions.

PAGE 73

59 Figure 2-1. Location of study si tes used in 2002 Japanese climb ing fern site establishment assessment, Florida, USA.

PAGE 74

60 Figure 2-2. Sampling design for assessment of Japanese climbing fern distribution on three forest sites in North Florida. A) Lay out of major transect lines and intensive sampling plots. B) Layout of sampling poi nts within each intensive sampling plot

PAGE 75

61 Figure 2-3. Plot layout indicat ing 2004 Japanese climbing fern cover classes and natural community types, over 2000 aerial imagery, Blackwater River State Forest (BRSF), Santa Rosa County, Florida.

PAGE 76

62 Figure 2-4. Plot layout indicat ing 2004 Japanese climbing fern cover classes and natural community types, over 2000 aerial imag ery, Miller property (Miller), Calhoun County, Florida.

PAGE 77

63 Figure 2-5. Plot layout indicat ing 2002 Japanese climbing fern cover classes and natural community types, over 2004 aerial imager y, Florida River Island (FRI), Liberty County, Florida.

PAGE 78

64 Table 2-1. Japanese climbing fern ( Lygodium japonicum ) occurrence on three North Florida forest sites, 2002 and 2004. Site = BRSF Site = Miller Site = FRI 2002 2004 2002 2004 2002 2004 Average site % Lygodium cover (major transects). 1% NAa 3% NA 14% NA Total number of plotsb with Lygodium present. 3 4 8 9 18 NA Percent of plots with Lygodium present 12% 16% 32% 36% 72% NA Average plot % Lygodium cover. 3% 2% 5% 6% 14% NA Change in Lygodium plot % cover 2002-2004. -0.6% +0.4% NA aNA indicates measurement not available. bTwenty-five plots per site. Table 2-2. Japanese climbing fern ( Lygodium japonicum ) occurrence across eight natural community types in three Nort h Florida forest sites, 2004. Twentyfive plots assessed per site. Site = BRSF Site = Miller Site = FRI Site = All Natural community Plot ratioa Plot %a Plot ratio Plot % Plot ratioa Plot %a Plot ratio Plot % Xeric pine uplands Xeric plot total 0/19 0% 0/19 0% Mesic pine plantation Mesic pine flatwoods Mesic pine/hardwood Mesic plot total 4/6 67% 8/11 0/5 73% 0% 12/22 55% Hydric floodplain ridge Hydric floodplain swale 10/10 6/10 100% 60% Hydric cypress wetland Hydric hardwood swamp Hydric plot total 0/4 1/5 0% 20% 2/5 40% 16/34 47% aPlot ratio and plot % indicat e the number of plots with Lygodium present to total number of plots of the community type.

PAGE 79

65Table 2-3. Comparison of tree importance values (IV) by species vegetative index a nd presence or absence of Japanese climbing fern (Lygodium japonicum ). Vegetative indexa Number of species Lygodium status Mean IV Std. error IV T-valuesb (present vs. absent) ANOVA statisticc Facultative upland 5 Present 7 3 4.2 (P<0.0001)* 5.9 (P=0.0005)* Facultative 13 Present 3 3 -1.3 (P=0.20) Facultative wetland 12 Present 10 4 -1.67 (P=0.1) Obligate wetland 13 Present 8 4 -2.56 (P=0.01)* Facultative upland 5 Absent 25 18 44.9 (P<0.0001)* Facultative 13 Absent 2 1 Facultative wetland 12 Absent 7 3 Obligate wetland 13 Absent 5 2 a Vegetative index of tree speci es hydrologic associations. b Test statistic and p-values for comparison of mean importan ce values by Lygodium status (p resent versus absent) for each vegetative index, Two sample t-test alpha = 0.05, SAS Version 9.1. Asteri sk* indicates significant value. c Test statistic and p-values for comparison of tree importan ce values between vegetative indices for each Lygodium status (present and absent), one-way analysis of variance, alpha =0.05, SAS Version 9.1. Asterisk* indicates significant value.

PAGE 80

66Table 2-4. Relationships betw een Japanese climbing fern ( Lygodium japonicum ) percent cover levels and significant plot variables and tree importance values in three North Florida forest sites. Variable Lygodium cover classa Number of plots Mean Std. error Kruskal-Wallis valuesb Correlation coefficientsc Correlation strengthd 1 10 20 6 9.62 (P=0.01)* -0.39 (P=0.04)* Weak 2 9 30 2 Basal area (m2 ha-1) 3 10 11 1 1 10 178 27 6.62 (P=0.04)* 2 9 249 44 Soil Al (mg kg-1) 3 10 154 16 1 10 2 0 6.12 (P=0.05)* 0.38 (P=0.04)* Weak Carpinus caroliniana 2 9 2 0 3 10 14 6 1 10 45 15 -0.42 (P=0.03)* Weak 2 9 29 14 Percent flooded 3 10 8 3 1 10 92 29 0.40 (P=0.03)* Weak 2 9 107 31 Photosynthetically active radiation (mol m-2 s-1) 3 10 199 93 1 10 10 3 0.39 (P=0.04)** Weak 2 9 17 6 Quercus nigra 3 10 28 7 aJapanese climbing fern cover class levels: 1= >0 to 5%, 2= 5.1 to 25%, 3= 25.1 to 75%. b Asterisk* indicates significant valu e Kruskal-Wallis analysis of variance test and p-values alpha = 0.05. c Asterisk* indicates significant valu e, Pearson's correlation coefficient and p-values, alpha=0.05, SAS Version 9.1. d Correlation strength: weak=0.20.49, moderate=0.5-0.79, strong=0.8-1.0

PAGE 81

67Table 2-5. Plot environmental variables that discriminate between Japanese climbing fern ( Lygodium japonicum) presence or absence in three North Florida forest sites. All plotsb Site = BRSFb Site = Millerb Site = FRIb Variablea R2 F-value P>F R2 F-value P>F R2 F-value P>F R2 F-value P>F Soil Ca (natural log) 0.31 32.72 P<0.00010.28 7.25 P=0.01 0.16 4.21 P=0.05 Percent flooded 0.05 2.47 P=0.07 0.23 5.93 P=0.02 0.13 3.30 P=0.08 Soil Fe (inverse) 0.06 4.72 P=0.03 0.35 12.48 P=0.002 Understory (% cover) 0.06 4.48 P=0.04 Tree evenness index 0.18 4.77 P=0.04 Hydrology class 0.17 4.21 P=0.05 Months flooded 0.45 15.34 P=0.001 Species richness 0.20 5.65 P=0.03 Soil K (natural log) 0.14 3.40 P=0.08a Data transformation indicated in parentheses if used. bTest statistics from stepwise di scriminant analysis of covarian ce of plot environmental variab les, SAS Version 9.1. Variables listed are significant (alpha=0.15) at either the study-level or site -level as indicated. Table 2-6. Plot environmental variables that discriminate between Japanese climbing fern ( Lygodium japonicum) percent cover class levels in three North Florida forest sites. All plotsa Site = BRSFb Site = Millera Site = FRIa Variable R2 F-value P>F R2 F-value P>F R2 F-value P>F Basal area 0.20 3.23 0.05 0.242.26 0.11 Soil Al 0.19 2.92 0.07 0.8211.36 0.01 0.252.52 0.14 Percent flooded 0.18 2.61 0.09 0.623.31 0.14 a Test statistics from stepwise discrimina nt analysis of covariance of plot e nvironmental variables, SAS Version 9.1. Variables listed are significant (a lpha=0.15) at either the study-le vel or site-level as indicated. b Sample size for site=BRSF too small to conduct analysis.

PAGE 82

68Table 2-7. Tree species importance values that discriminate between presence a nd absence of Japanese climbing fern ( Lygodium japonicum ) presence or absence in thr ee North Florida forest sites. All plotsa Site = BRSFa Site = Millera Site = FRIa Variable Partial R2 F value P>F Partial R2 F value P>F Partial R2 F value P>F Partial R2 F value P>F Pinus palustris 0.21 18.9 <0.0001 Quercus nigra 0.08 6.6 <0.0001 0.94 3.7 <0.0001 Liquidambar styraciflua 0.10 8.3 0.01 Taxodium distichum 0.07 5.6 0.02 Quercus hemisphaerica 0.05 3.3 0.07 Lindera benzoin 0.05 3.3 0.07 Nyssa sylvatica 0.05 3.3 0.07 Ligustrum sinense 0.91 2.2 <0.0001 Pinus elliottii 0.23 6.9 0.02 Acer rubrum 0.01 3.5 0.08 Nyssa aquatica 0.21 5.9 0.02 aTest statistics, stepwise discri minant analysis of covariance of tree importan ce values. Species listed are significant (alpha= 0.10). Table 2-8. Tree species importance values that discriminate between Japanese climbing fern ( Lygodium japonicum) percent cover class levels on pl ots with fern present in thr ee North Florida forest sites. All plotsa Site = BRSFb Site = Millerb Site = FRIa Variable R2 F-value P>F R2 F-value P>F R2 F-value P>F Carpinus caroliniana 0.18 2.9 0.07 Quercus nigra 0.293.07 0.08 Carya aquatica 0.323.02 0.08 a Test statistics from stepwise discrimina nt analysis of covariance of plot envi ronmental variables. Species listed are significant (alpha=0.10). b Sample size for BRSF too small to conduct an alysis, no species selected for Miller site..

PAGE 83

69 Table 2-9. Effects of signifi cant plot environmental and tree importance values in multiple logistic regression for relationship to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forest sites. Area Effect (in order of entry)a EstimateR-squared (cumulative) Chi square score P>Chi square All plots Intercept 0.40 SOIL1 1.79 0.40 24.0 P<0.0001 Ligustrum sinense 0.29 0.46 11.4 P=0.0008 Quercus nigra -0.09 0.53 5.7 P=0.02 Lindera benzoin -0.17 0.59 9.4 P=0.002 Acer rubrum 0.03 0.62 4.4 P=0.04 Quercus laevis 0.01 0.65 3.9 P=0.05 Nyssa aquatica 0.75 0.68 4.5 P=0.03 BRSF Intercept -1.46 TREE 1.80 0.49 7.85 P=0.005 Miller Intercept -6.39 SOIL1 -1.88 0.42 7.93 P=0.005 Pinus palustris 0.02 0.64 4.99 P=0.03 FRI No variables entered a Significant effects in model (SAS Versi on 9.1).. Principal components analysis factors: SOIL1= soil P, soil Ca, and soil pH; TREE=tree species richness, evenness, and diversity. Model validit y is questionable due to maximum likelihood estimate.

PAGE 84

70 Table 2-10. Effects of signifi cant plot environmental and tree importance values in multiple logistic regression for relations hip to percent cover class levels of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forest sites. Area Effect (in order of entry)a EstimateR-squared (cumulative)bChi square score P>Chi square All plots Intercept 9.2 Carya aquatica 3.7 0.17 4.3 P=0.04 FLOOD1 -1.1 0.38 6.6 P=0.01 Catalpa bignoniodes -0.1 0.46 5.2 P=0.02 Platanus occidentalis -0.1 0.52 4.1 P=0.04 Nyssa ogeechee -1.3 0.59 5.5 P=0.02 Quercus laevis 0.2 0.75 9.0 P=0.003 BRSF No variables entered Miller No variables entered FRI Intercept 3.21 FLOOD1 -1.37 0.50 8.1 P=0.005 a Significant effects in model (SAS Version 9.1 ). Principal compone nts analysis factor: FLOOD1=hydrology class, percent flooded, a nd months flooded. Model validity is questionable due to maximum likelihood estimate.

PAGE 85

71 CHAPTER 3 HERBICIDE EFFICACY FOR MANAGEMENT OF JAPANESE CLIMBING FERN IN SOUTHEASTERN FORESTS Introduction Japanese climbing fern ( Lygodium japonicum ), is a non-native invasive plant widely established in the southeastern Un ited States. In this range, Japa nese climbing fern invades mesic and temporally hydric areas, including: floodplain fo rests, bottomland hardwood forests, marshes, wetlands, secondary woods, moist pinelands (especi ally flatwoods), limestone outcroppings, and disturbed areas such as road shoulders and rights-of-way (Clewell 1982, Nauman, 1993, Langeland and Burks 1998). Japanese climbing fern does presently occur on xeric sites as well, but to a more limited extent. Heavily infested sites may sustain populations av eraging 60-80% cover over large areas, effectively reduce or eliminate native gr oundcover and understory vegetation, and smother seedlings of overstory tree species (Horovitz et al. 1998, Lott et al. 2003, Zeller and Leslie 2004). Despite these impacts, large information gaps ex ist on aspects of the biology and management of this plant. For both public and private land ma nagers, identification of efficient and effective control strategies for Japanese climbing fe rn has become increasingly important. Recognition of this plant has increased among public conservation land managers in the southern United States since the mid 1990s, aidi ng in increased detection and reporting. Recognition of Japanese climbing fern by private land managers has also increased since the late 1990s, but management efforts have not increased at a similar rate due to lack of awareness of the non-native invasive status of the plant, research -based best management practices, and financial incentives or support for management of private-lands infestations. Designation as a noxious weed

PAGE 86

72 in Florida has been an important part in the pr ocess of raising awarene ss among private forestland managers and members of the forest products in dustry (e.g., pinestraw) in Florida, and for increasing the demand for effective management guideli nes. Stop sale orders have been issued for shipments of Japanese climbing fern-contaminated pinestraw bales in multiple cases by the FDACS Division of Plant Industry (Van Loan 2006), and awareness of the regulatory requirements is affecting some buyer preferences as well. Forestry and Invasive Plant Impacts The impact of species invasions on sustainable forest management is increasing in scale and recognition. Plant invasions have a ffected biological diversity, fore st health and productivity, water and soil quality, and socieoeconomic value. Plant invasions modify forest ecosystems by affecting fire and hydrological regimes, food webs, and re cruitment of dominant tree species (Vitousek, DAntonio, Loope and Westbr ooks 1996, Chornesky, Bartuska, Aplet, OBritton, CummingsCarlson, Davis, Eskow, Gordon, Gottschalk, Haac k, Hansen, Mack, Rahel, Shannon, and Wainger 2005). In managed forests, establishment of invasive plants is increasingly facilitated by expanding human access, fragmentation, and so il or canopy disturbance often a ssociated with silvicultural practices (With 2002, Chornesky et al. 2005). Some estimates place annual economic impact of forest product loss due to invasive species as high as $2 billion in the United States (Pimental et al. 2000). Impacts of invasive plants on the forestry industry in Florida have been estimated at $38 million per year ($15 million in weed control cost s, and $23 million in yield losses) (Lee 2005). Little is known about the economic impact of Japanese climbing fern invasion. Forest managers in Florida have noted impacts including reduction in annual financial return and harvest lease longevity for pinestraw production, an indus try which generates $79 million in revenue for forest landowners in Florida annually (Hodges et al. 2004), and reductions in pine seedling survival during reforestation of heavily infested sites. In particular, mesic private fo restlands are moderately

PAGE 87

73 to heavily infested with Japanese climbing fern throughout much of the northwest Florida region. A 2002 survey (Barnard and Van Loan 2002) conducted across 280 pine pl antations in north and west Florida recorded presence of Japanese climbing fe rn in 22% of the stands on mesic sites, exceeding occurrence of four other regionally significant non-native inva sive plants including: cogon grass ( Imperata cylindrica (L.) Beauv.), Chinese tallow ( Triadica sebifera (L.) Small), kudzu ( Pueraria montana var. lobata (Willd.) Maesen & S. Almeida), and air potato ( Dioscorea bulbifera L.). With approximately 5.1 million acres of mesic pine plan tations alone in north Florida (Brown 1995), the regional scale of this invasion generates manageme nt problems including: need for coordinated and cost-effective management strategies, legal impli cations affecting harvest and sale of forest products from infested areas, and high probability of continual reinvasion of treated sites. Management Techniques Integrated pest management utilizes the most appropriate combination of treatments including: chemical, manual/mechanical, cultu ral, and biological control techniques. Mechanical and cultural (i.e., prescribed burning) treatments ha ve generally proven ineffective, often promoting Japanese climbing fern re-growt h, and possibly vectoring spores. To date, chemical treatments have been most effective, but little ex ploratory research has been conducted. A strategy that crosses propert y boundaries is required for su ccessful management of most invasive species (Chornesky et al. 2005). The re productive strategy (i.e. spore-dispersal, selffertilization) of Japanese climbing fern facilitate s rapid spread and establishment in remote areas (Lott 2003), and continual re-inva sion if all populations in an area are not addressed jointly. Identification of consistently effective treatment guidelines is an important component in developing a management commitment that cr osses property boundaries and is implemented by both public and private land managers Current knowledge indicates th at treatment with herbicides

PAGE 88

74 will be a required component of successful mana gement of Japanese climbing fern in the United States. This study evaluates those herbicides which have received preliminary assessment for Japanese climbing fern control, as well as others commonly used on both forestry and vegetation management. The objective of this study were to: (1 ) evaluate a range of he rbicides to be applied postemergence for control of Japanese climbing fern and (2) evaluate the efficacy of multiple rate combinations of glyphosate and metsulfuron, two herb icides expected to have the greatest efficacy against Japanese climbing fern. Materials and Methods Study Area Field experiments to evaluate the effect of postemergence herbicides applied to Japanese climbing fern were established in 2002 and replicat ed in 2004. All experiments were conducted in a privately-owned slash pine ( Pinus elliottii Engelm. var. elliottii ) plantation in Calhoun County, Florida, USA (3023N, 859W). The plantation was establis hed in 1979, managed for pinestraw harvest from 1997-1999, and underwent a third row th inning (removal of every third row of pine trees) in 2000. Japanese climbing fern was first recorded on the site in 1997, but may have been present earlier. By 2002, the stand was heavily in fested, with Japanese climbing fern dominating the understory and trellis ing into the midstory. For both of the 2002 trials, the soil was a Dothan sandy loam, 0-2 percent slope (fine-loamy, kaolin itic, thermic Plinthic Kandiudults) with 0.48% organic matter, and a soil pH of 5.3. The 2004 trial was conducte d within the same stand, but on slightly different soil; Robertsdale fine sandy loam (fine-loamy, siliceous, semiactive, thermic Plinthaquic Paleudults) with 1.87% organic matter and a soil pH of 5.2. The average daily temperature is 12o C in the winter and 33o C in the summer. Freezing temperatures occur on an average of 20 days each winter.

PAGE 89

75 Herbicide Selection Herbicides were selected based on their weed control spectrum, label-status for forestry applications, performance in initial field assessm ents by land managers, and commonality with best management practices for treatment of other non-na tive invasive plants regularly co-occurring with Japanese climbing fern in Florid a (Table 3-1). A non-ionic surfac tant Induce (Helena Chemical) was added at 0.25% v/v according to label recommendations. In response to land manager concerns and re quests, two formulations of glyphosate were evaluated; a traditional formulation (Accord ), and a generic (G lyphos) formulation, representing the many glyphosate products availa ble for purchase since the expiration of the Roundup patent. Land manager interest in this comparison was primarily due to the lower per unit cost of the generic formulations. Both form ulations were evaluated at a low application rate (2.24 kg ai per ha, or 2.5% in solution), and a high application rate (4.48 kg ai per ha, or 5% in solution). Experimental Design Two herbicide trials were established in October 2002. The Multiple Herbicide Trial included fifteen herbicide treatments and an untreat ed control (see Table3-2). Treatment plot size was 6.8 m by 3.4 m. The Glyphosat e/Metsulfuron Trial included tank-mixed combinations of four rates of glyphosate (0, 1.12, 2.24, 4.48 kg ai per ha) and metsulfuron (0, 0.03, 0.06, 0.12 kg ai per ha), and an untreated c ontrol, for a total of sixteen treatments. Treatment plot size for this trial was 3.4 m by 3.4 m and were smaller than desired due to constraints within the study site. Treatment plots were established in the thinned rows of the stand. All treatments were applied using a CO2-pressurized hand-held boom sprayer. Spra y volume was calibrated to deliver 225 liters per hectare (20 gallons per acre). Experiment s were conducted as randomized complete block designs, and all treatments were replicated four times for a total of sixty four plots per trial. The

PAGE 90

76 Multiple Herbicide Trial was replicated within the study site in 2004 with treatment plots sized 6.8 m by 3.4 m. The Glyphosate/Me tsulfuron trial was not repli cated in 2004 due to lack of adequate area with contiguous Japanese climbi ng fern coverage in the study site, and the expectation of greater utility of results from the Multiple Herbicide trial for land managers. Treatment and Assessment Schedule 2002 trials. Herbicides were applied on October 23, 2002. Percent cover of live Japanese climbing fern was rated at 1 month after treatment (MAT), 5, 6, 8, and 12 MAT. Upon returning to the study site for the 12 MAT rating, it was discover ed that the pine stand had been unexpectedly thinned in late September, 2003 (removal of sele cted trees from leave rows remaining after 2000 thinning, and skidding of logs thr ough all plots). Aboveground Japa nese climbing fern foliage was mechanically removed from all plots by the harvest operation (percent cover = 0), but plot corner stakes were adequately intact to re-define original plot boundaries. Follo wing harvest, percent cover of live Japanese climbing fern was rated at 15, 24, 29, 31, 35, and 36 MAT. 2004 trial The 2002 Multiple Herbicide Trial was repe ated within the original study site to remove the potential confounding effects due to logging. Treatment plot size was 3.4 by 6.8 m. Herbicides were applied on Nove mber 10, 2004. Percent cover of live Japanese climbing fern was rated at 1, 4, 6, 10, and 12 MAT. Data Analysis Percent cover of live (green folia r tissue) Japanese climbing fern was visually determined and recorded for each plot to rate effectiveness of the various treatments. All plots were rated on a scale from 0% (no climbing fern on plot) to 100% (plo t completely covered by climbing fern) in 5% increments one day prior to treatment to establis h a baseline. Following tr eatment, plots were rated using the same technique according to the assessme nt schedules indicated above; all ratings were were conducted by the same observer.

PAGE 91

77 For all plots, percent live Japa nese climbing fern cover was co nverted to percent control for analysis using the following equation: Percent control = Percent cover pre-treatment Percent cover at MAT X 100 Percent cover pre-treatment Use of percent control values accounts for di fferences in pre-treatment fern cover across treatment plots. The results from each trial were analyzed using the general linear model approach to detect significance of model effects among all treatments. Fishers least significance difference (LSD) test was used to make pairwise comparis ons among treatment means (SAS Institute 2005). For all tests, significance was asse ssed at the alpha=0.05 level. A two-sample t-test was performed to evaluate th e effect of plot size on trial results. Means from untreated control plots in the 2002 Multiple Herbicide trial and the 2002 Glyphosate/ Metsulfuron trial were compared, with the hypo thesis that no diffe rence would exist. Results and Discussion The reproductive strategy (i.e. spore-dispersal, self-fertilization) of Japanese climbing fern facilitates rapid spread and establishment in re mote areas (Lott et al. 2003), and continual reinvasion if all populations in an area are not addressed jointly. Identification of consistently effective treatment guidelines is an important co mponent in developing a management commitment that crosses property boundaries and is implemen ted by both public and private land managers. A strategy that crosses property bound aries is required for successful management of most invasive species (Chornesky et al. 2005). Cu rrent knowledge indicates that trea tment with herbicides will be a required component of successful management of Japanese climbing fern in the United States. The following results verify and expand the curren t knowledge base on herbic ide selection for this purpose.

PAGE 92

78 Phenology of Japanese Climbing Fern To fully understand the herbicide treatment effects for any species, it is important to consider other factors which may influence plant performa nce. On all plots in this study, three primary factors affected climbing fern foliar cover: he rbicide treatment (varied by treatment), seasonal dieback and growth (uniform across treatment s), and mechanical damage from the 2003 timber harvest (uniform across treatments). Japanese cl imbing fern plants remain evergreen below the frost line in central and south Fl orida (Ferriter 2001, Valenta et al 2001, Zeller and Leslie 2004). Above the frost line, winter dieback of Japanese climbing fern fronds occurs to varying extents through the majority of its range. In the spri ng, the plants will re-sprout from cold-tolerant subterranean rhizomes (new stems observed the fi rst week in March in No rth Florida), and often utilize freeze-damaged stems as ladders to grow back into the canopy. This seasonal dieback in sub-temperate and temperate climates is one fact or which has likely prevented formation of the canopy-covering arboreal blankets seen with Old World climbing fern ( Lygodium microphyllum ) (Zeller and Leslie 2004). However, the increasing distribution of Japanese climbing fern in south Florida may lead to the development of such canopy-smothering growth on infested sites. Japanese climbing fern seasonal foliar phenol ogy was recorded on the untreated control plots in each trial throughout the study period. At in ception of the 2002 and 2004 trials (October), untreated Japanese climbing fern plants had ach ieved maximum annual foliar coverage, followed by a seasonal frost event (December-February) which re duced plot live foliar cover to 11% on average (min. 5%, max 20%). Initial emer gence of new croziers (growing s hoots) occurred in mid-March, with measurable new growth of climbing fern pr esent by mid-April. Foliar cover continued to increase steadily through the summer and fall un til peaking again in October for both the 2002 and 2004 trials (Figure 3-1). Despite these seasonal fluctuations a nd the 2003 timber harvest event, Japanese climbing fern cover on th e control plots averaged a 3.1% a nnual increase over the initial

PAGE 93

79 percent cover (Table 3-3). Howeve r, analysis of variance did not id entify this change as significant (alpha = 0.05) Multiple Herbicide Trial: 2002 and 2004 Average initial climbing fern percent groundcover on the Multipl e Herbicide trial plots was 70% (min 30%, max 95%). A significant effect was measured for the interaction of treatment and year between 2002 and 2004, and as a result, the data from each year are presented separately. Discussion of short-term treatment effects obs erved through 12 MAT will focus on the 2004 trial results. This ensures continuity not possible in the 2002 plots due to the interruptive effect of the 2003 timber harvest. Results from the 2002 trials are used for discussion of long-term treatment effects observed at 24 and 36 MAT. The plot ratings from 5 and 29 MAT, and 15 MAT are not discussed as they represent effects of seasonal fr ost damage and timber harvest respectively, rather than herbicide treatment effect. Review of result s from this trial and those from Valenta et al. (2001), and Zeller and Leslie (2004) indicate that the best perfo rming (successful) treatments typically yield a minimum of 70% control at 12 MAT; a standard which will be utilized in the following discussion. Short Term Treatment Effects Short term treatment effects reported herein include effects of trea tments applied in 2004, from treatment through twelve months after treatment. At 1 MAT (December 2004), three treatments exhibited a level of damage to Japane se climbing fern foliage that was significantly (P<0.05) different than observed in the untreated control plots (Table 3-2). Hexazinone and 2,4D+dicamba showed the highest initial damage to Japanese climbing fern with 91% control (reduction in live foliar cover), followed by the dicam ba treatment with 80% control. This initial response is probably due to burnoff of the foliage rather than the damage to underground plant structures necessary to achieve long-term contro l. Similar effects were observed in early

PAGE 94

80 measurements of treatments by Valenta et al. ( 2001), and Zeller and Lesl ie (2004). Three other treatments yielded control levels greater than 70%: the high rate of glyphosate in both the traditional (77%) and generic (76%) formul ations, and the tank mix of glyphosate+imazapyr (72%). The tank mix of glyphosate+metsulfur on yielded 65% control, while the glyphosate treatments at low rate yielded 54% and 29% control, respectively, for the generic and traditional formulations. All other treatments yielded less than 50% cont rol. The untreated control plots yielded a 32% reduc tion in foliar cover, which may be explained by early winter decline in foliage density observed following the fall peak in foliar cover or as a result of foliar rust damage, and/or damage to fronds originat ing in other treatment plots and growing across adjacent control plots. AT 6 MAT (May 2005), treatments separated into three significant (P<0.05) groups (Table 32). The largest group included el even treatments exhibiting be tween 89% and 100% control of Japanese climbing fern. This gr oup included all of the herbicides and herbicide tank mixes in the trial with amino acid inhibition as the mode of action: glyphosate (both rates and formulations), imazapyr, metsulfuron, glyphosate+metsulfur on, glyphosate+imazapyr, imazapyr+metsulfuron, imazapic, and sulfometuron. In a statewide survey of climbing fern treatment regimes used by public land managers, many favored a six-month re turn interval for herbicide treatments (C. Lockhart, 5/04/2006, personal communication). With such a regime, all treatments in this group may be considered in managing Japanese climbi ng fern. A second group of treatments, including hexazinone, dicamba, and 2,4-D +dicamba, exhibited approximately 60% cont rol, while triclopyr ester dropped to 25%, indicating higher fern coverage than on the untreated control plots (35%). This control level observed in the untreated plots is primarily explained by lingering effects of winter climbing fern dieback.

PAGE 95

81 At 10 MAT (September 2005), two significan t (P<0.05) groups of treatments were observable. As found at 6 MAT, the group of trea tments exhibiting the highest level of control (69% to 94%) included all of the amino acid inhibitors, with the exception of sulfometuron where control levels were reduced to 0.1%. A second group of treatments including hexazinone, dicamba, 2,4-D+dicamba, sulfometuron and the untreated control plots e xhibited between 0.1% and 14% control; while triclopyr ester plots showed a nega tive control level, indicating that climbing fern cover had increased by 12% from th e pre-treatment levels. The untre ated control plot s still yielded 14% control at this assessment as climbing fe rn foliar cover continue d to increase through the remainder of the growing season (Figure 3-1). At 12 MAT (November 2005), thre e significant (P<0.05) treatme nt groups were observable (Table 3-2). The herbicide group exhibiting the hi ghest level of reduction in climbing fern included four treatments yielding greater th an 90% control (the t raditional and generic formulations of glyphosate at the high application rate, a nd tank mixes of glyphosate+imazapyr and glyphosate+metsulfuron); and three treatments exhibiting from 73% to 87% control (the traditional and generic formulations of glyphosat e at the low application rate, imazapyr, and the tank mix of metsulfuron+imazapyr.) At this post treatment date, the effects of metsulfuron and imazapic treatments had diminished to 41% a nd 27% control, respectively, while hexazinone exhibited only 3 percent control. For five treatments, percen t cover increased a bove pre-treatment levels, yielding a negative percent control value, including: dicamba (8%), sulfometuron (11%), triclopyr ester (12%), and 2, 4-D+dicamba ( 12%), and the untreated control plots (3%). Glyphosate comparison. Results from the 2004 Multiple Herbicide trials were consistent with initial expectati ons, in that the treatments that in cluded glyphosate were among the highest performing treatments. Direct comparison of the traditional and generic glyphosates at both

PAGE 96

82 high and low application rates yielded little diffe rence between the two formulations. At 1 MAT, significant difference (P<0.05) was de tected between the traditional and generic formulations at the low application rate, but at all subsequent measurements no difference was detected between the two formulations. Timber Harvest Effects The October 2003 timber harvest clearly affect ed all of the 2002 study plots by physically removing the aboveground fern growth, reducin g canopy cover across the stand, increasing the amount of downed woody debris, and possibly re-dis tributing Japanese climbing fern spores across the study site. The harvest effects appeared to be uniform across all plots based on visual assessment. The plots then underwent a period of post-harvest recovery which lasted for approximately one year, and during which climbing fern reclaimed the site understory.. Treatment effects became identifiable again at 24 MAT, a nd lasted through the final assessment at 36 MAT. Long Term Treatment Effects Long term treatment effects reported herein re fer to effects measured on both the 2002 and 2004 plots from 12 MAT to 36 MAT. Treatment eff ects which were beginning to appear in the 2002 plots at 8 MAT, and which also were clear in the 2004 plots at 12 MAT, reappeared in the 2002 plots at 24 MAT (Table 3-2). Analyses rev ealed many overlapping re lationships, but yielded two primary treatment groups. Nine treatments exhi bited significantly high er (P<0.05) levels of climbing fern control (38% to 70%) than seen in the untreated control plots, including: all treatments that included glyphosat e (traditional and generic formulations, low and high rates and tank mixes), and metsulfuron and imazapyr. Treatment effects remained similar at 36 MAT (Table 3-2). Six trea tments continued to exhibit significantly higher leve ls of climbing fern control (35% to 41%) than observed in the untreated control plots, including both rates of the traditional glyphosate formulation and the low

PAGE 97

83 application rate of the generic glyphos ate formulation, imazapyr, and tank mixes of glyphosate+imazapyr and metsulfuron+imazapyr. Se ven herbicide treatments exhibited between 29% control (reduction) and 24% incr ease in climbing fern cover, not significantly different than the untreated control plots. Climbing fern cover on the untreated plots increased above initial pretreatment levels by 6.25%. 2002 Glyphosate/Metsulfuron Trial The initial decision to conduct the Glyphosate/Me tsulfuron rate combination trial was made based on performance of these two herbicides in preliminary assessments (Valenta et al. 2001, Zeller and Leslie 2004). Metsulfuron showed poten tial for limiting damage to non-target plant species during invasive plant management. In add ition to the recovery of ecosystem value through retention or re-establishment of native species on sites affect ed by non-native species invasions, several studies have shown the importance of si te occupancy in preventing new or recurring site invasion by non-native species, pa rticularly following a disturba nce event such as chemical treatment (Burke and Grime 1996, Davis et al. 20 00, Prieur-Richard and Lavorel 2000, DAntonio and Myerson 2002). Identification of a control strategy which limits loss of non-target species may play a significant role in long term success of any Japanese climbing fern management strategy. Zeller and Leslie (2004) ob served glyphosate treatments to yield better long term control (70% to 80% reduction in fern c over at one year after treatment) than metsulfuron treatments, but more non-target damage occurred compared to mets ulfuron. In sites where herbicide applications are likely to be repeated within a single year this finding may indicate use of metsulfuron for retention of desirable no n-target species. However, conc erns exist among weed management specialists about rapid developmen t of resistance to sulfonylurea he rbicides such as metsulfuron (Tranel and Wright 2002), particularly with plants which reproduce as rapidly and prolifically as the climbing ferns (Lott et al. 2003). Treatments th at involve tank-mixing glyphosate and metsulfuron

PAGE 98

84 hold promise as a technique to reduce both the n on-target damage seen with glyphosate, and the possibility of herbicide-driven selection of metsulfuron-resistan t climbing fern populations (Diggle et al. 2003, Kudsk and Streibig 2003). This con cern prompted the evaluation of multiple rate combinations of glyphosate and metsulfuron in this study. Average initial climbing fern cover on the Gl yphosate/Metsulfuron pl ots was 66% (min 30%, max 90%). As with the 2002 Multiple Herbicid e plots, percent groundcover of live Japanese climbing fern was rated at 1, 5, 6, 9, and 12 MAT. Plots were affected by thinning before 12 MAT, at which time percent cover for all plots was mechan ically reduced to zero. In the period following harvest, plots were rated at 15, 24, 29, 31, 35, and 36 MAT. The plot size for the Glyphos ate/Metsulfuron trial was 3.4 m by 3.4 m, 50% smaller than the plot size for the Multiple Herbicide trials. This f actor was significant in determining the usefulness of the trial results (data not shown) At nearly all assessment dates, the percent control of Japanese climbing fern on the untreated control plots wa s not significantly different (P<0.05) than the herbicide treatment plots. The untreated contro l plot data from the Glyphosate/Metsulfuron trial and the 2002 Multiple Herbicide trial were compared to assess the effect of plot size. Significant differences (P<0.05) due to plot size were seen at 5, 6, 8, and 24 MAT, indicating that the treatment differences detected in this study represented a Type 1 error rather than actual treatment differences. A recommendation for future research is to limit th e minimum herbicide trial plot size to 3.4 m by 6.8 m (200 square feet). Further ev aluation of these treatments may be appropriate; however, in the interest of achieving long-term control of Japane se climbing fern a higher priority may be placed on research which evaluates the effect of repeated herbicide applications. Conclusions and Recommendations for Land Managers The typical goal of invasive plant manage ment on public conservation lands is, when possible, to eradicate the species from infested sites in such a way as to promote maximal recovery

PAGE 99

85 of native plant species following treatment. This may also be the management goal on some private forestlands, but for many others, preservation of native midstory and groundcover species is secondary to production of a profita ble and legally saleable forest product (Allen et al. 2005). In the case of the pinestraw industry, the vegetation mana gement goal is typically to remove understory vegetation from a harvest site that may impede th e pinestraw harvest process, or serve as a nonstraw contaminant in the pinestraw bales (Dur yea 1998, Taylor and Foster 2003). These varying goals will impact implementation of the results of this study by land managers in the field. The data of greatest interest to land managers are the results from the 2004 Multiple Herbicide trial in the 0 to 12 MAT period. Increasingly, a six month return interval is recommended or implemented in climbing fern management (Fe rriter 2001, Lockhart 2006). While managers often strive to revisit treatment sites multiple ti mes in a year, many operational invasive plant management programs on public and private lands in clude a functional minimum return interval of approximately one year between herbicide appli cations. However, budget and staff limitations can prevent mangers from achieving this minimum, resu lting in return intervals which may extend to two years or more. The long term (>12 MAT) results from this tria l illustrate that even when such extended return intervals occur, some treatment effects w ill remain, to the point of reducing the Japanese climbing fern coverage by 39% on average acro ss the good treatments at three years posttreatment (Table 2). While positive in some sens e, this extended suppressive effect should not prompt land managers to delay return to a site for management purposes, as the species sporebased reproductive strategy allows even small in festations to contribut e significantly to reinfestation of a site, and spread to other sites. In addition, the vacancies created in areas where herbicide treatments are successful, are most likely to be filled by either Ja panese climbing fern or

PAGE 100

86 another non-native invasive plant, providing add itional incentive for regular follow-up management visits (Davis et al. 2000, DAn tonio and Myerson 2002). Land managers will consider three main factor s when determining which approach to use in treatment of Japanese climbing fern: level of cont rol, duration of control, and per area treatment cost. As indicated earlier, a minimum level of control in successful treatments may be 70% reduction of Japanese climbing fern foliage (Table 3-4). However, in this study four treatments (glyphosate high rate traditional and gen eric, glyphosate + imazapyr, and glyphosate + metsulfuron) maintained greater th an 90% control for up to 12 MAT (T able 3-2). This highest level of control also limits spore production, a factor which is very important in the long term management of this plant. Three treatments from th is trial provided control for 6 (sulfometuron) to 10 (metsulfuron and imazapic) MAT. Managers may decide to treat Japanese climbing fern with one of these treatments as a secondary component of herbicide applicatio n for other management purposes, or in the case of metsulfuron, if reduced damage to non-target species is a management priority. Sulfometuron has been recommended for herbaceous weed control in hardwood plantations both preplanting and postplanting (Rhodenbaugh and Yeiser 1994, Ezell and Nelson 2001). Bottomland hardwood, floodplain forest, a nd mixed pine-hardwood stands are both readily invaded by Japanese climbing fern in Florida, and in some cases actively managed for hardwood production. In such sites, sulfometuron can be applied to dormant hardwood seedlings, and over conifer seedlings (Schuler, Robison and Quicke 2004). However, any damaging effects of these compounds on Japanese climbing fern will likely only last 6 months, and require attention to regular follow-up treatments. Based on level of control achieved and appr oximate prices in 2006 (Ferrell, Gray and MacDonald 2006), the glyphosate treatments were the most cost-effective, and efficacious at both

PAGE 101

87 low and high application rates using either the t raditional or generic formulations. Glyphosate is widely used in many aspects of forestry and invasive plant management, as well as in other vegetation management applications. Valenta et al. (2001) and Zeller and Leslie (2004) also found glyphosate treatments to be among the most effectiv e, achieving greater than 70% control at 0.07, 1.34, and 2.67 kg ai/ha at 315 days after treatment and 1.12 and 3.36 kg ai/ha at 400 days after treatment, respectively. This wide use facilita tes the ready expansion of treatment protocols established for other purposes to include Japanese climbing fern. Imazapyr is another herbicide widely used in pine plantation manageme nt (Lauer et al. 2002) and invasive plant management in the South. T hough currently less cost e ffective than glyphosate, imazapyr is effective against Japa nese climbing fern when applie d alone and in tank mixes with glyphosate and metsulfuron; residual effects were observable up to 36 MAT. Though effective, selection of imazapyr for Japanese climbing fern management should include consideration of soil activity and associated non-ta rget species damage (Tu, Hurd and Randall 2001). Clark (2002) found no significant difference in native plant cover, richness or dive rsity on sites infested with Old World climbing fern that were either treated with he rbicide or left untreated; illustrating that with heavy infestations, the damaging effects of chemical treatments may be no greater than the effects of heavy infestation. Regardless of which successful treatment is selected, managers must be vigilant in completion of regular follow-up treatments to reduce Japanese climbing fern infestations to control levels acceptable within their ma nagement priority framework. Spore contamination during treatment. When implementing any management program for Japanese climbing fern, spore ve ctoring on contaminated personnel, clothing, and equipment must be considered. Hutchinson and Langeland (2006 ) evaluated Old World climbing fern gametophyte

PAGE 102

88 (germinated spore) formation from spores co llected on clothing and equipment of herbicide applicators after working in three heavily infested sites receiving an initial herbicide treatment, and one site with a very low infestation undergoing a re-treatment. Contam ination of clothing and equipment occurred in the initial a nd re-treatment sites, affecting sh irts, pants, sprayers, disposable suits, chainsaws, gloves, machetes and footware. Contamination was significantly greater from the heavily infested initial treatments sites versus the re-treatment sites. Similar results are likely in sites infested with Japanese climbing fern, and ap plicators working in thes e sites should utilize the following good spore hygiene recommendations: sp ray off equipment and brush off clothing and accessories before leaving site, wash all clothing daily, store disposable suits in plastic bags, limit any vehicle entry into infested areas (Hutchinson and Langeland 2006.) This study has confirmed and expanded knowledge of the efficacy and treatment duration of several herbicide rates and comb inations for use in management of Japanese climbing fern, particularly for forest managers in the Southeast United States. However, additional research may clarify other aspects of the treatment of this sp ecies. Targets for future research may include: further assessment of the impacts of herbicide tr eatments on non-target native species, measurement of effects of repeated treatments, and assessmen t of mechanical, biological and other treatment techniques for efficacy in an integrated management approach. Finally, a comparison of infestation levels and treatment efficacies in disturbed vers us undisturbed sites may improve consideration and prioritization of Japanese climbing fern infestations in sites schedules for silvicultural management practices.

PAGE 103

89Table 3-1. Herbicide chemical families and mode of action used in a Japanese climbing fern control trial in a North Florida pine plantation, 2002 to 2005. Active ingredient Familya Mode of actiona Mechanism of actiona,b Absorptiona Accumulationa Plant deatha Soil half lifeb Imazapic Imidazolinone Amino acid inhibitor ALS inhibitor Leaves, stems, roots Underground tissues several weeks 120 daysc Imazapyr Imidazolinone Amino acid inhibitor ALS inhibitor Leaves, roots Meristematic regions 1to 2 weeks 25 to 142 days (3 to 24 month)c Sulfometuron Sulfonylurea Amino acid inhibitor ALS inhibitor Leaves, roots Meristematic regions 2 to 3 weeks 20 to 28 days Metsulfuron Sulfonylurea Amino acid inhibitor ALS inhibitor Leaves, roots Meristematic regions 1 to 2 weeks 1 to 6 weeks Metsulfuron + imazapyr Sulonylurea + imidazolinone Amino acid inhibitor ALS inhibitor Leaves, roots Meristematic regions 1 to 2 weeks see Imazapyr Glyphosate Glyphosate Amino acid inhibitor EPSP inhibitor Leaves, green stems Underground tissues, young leaves, meristems 4 to 7 days 47 days Glyphosate + imazapyr Glyphosate + imidazolinone Amino acid inhibitor EPSP + ALS inibitors Leaves, stems, roots Underground tissues, young leaves, meristems 1 to 2 weeks see Imazapyr Glyphosate + metsulfuron Glyphosate + sulfonylurea Amino acid inhibitor EPSP + ALS inibitors Leaves, stems, roots Underground tissues, young leaves, meristems 1 to 2 weeks see Glyphosate Dicamba Benzoic acid Growth regulator Synthetic auxin Leaves, stems, roots Growing points 3 to 5 weeks 14 to 28 days Triclopyr ester Picolinic acid or carboxylic acid Growth regulator Synthetic auxin Leaves, roots Growing points 3 to 5 weeks 10 to 46 days

PAGE 104

90Table 3-1. Continued Active ingredient Familya Mode of actiona Mechanism of actiona,b Absorptiona Accumulationa Plant deatha Soil half lifeb 2,4-D + dicamba Phenoxy acetic acid + benzoic acid Growth regulators Synthetic auxin Leaves, roots Growing points of root and shoot 3 to 5 weeks 1 to 4 weeks Hexazinone Triazine Photosynthetic inhibitor Binds to D1 protein in photosystem II Leaves, roots Leaves 90 days (soil:1 to 6 months, leaf litter: to 3 years) a Ahrens 1994 b ALS = Acetolactate synthase, EPSP = 5-e ndolpyruvyl -shikimate-3phosphate synthase c Tu et al. 2001

PAGE 105

91Table 3-2. The effect of post-emergent herbicides on Japanese clim bing fern foliar cover in a North Florida slash pine plantation, 2002 and 2004. 1 MATb 6 MAT 8 MAT10 MAT12 MAT 24 MAT 36 MATTreatmentc Rate 2002d 2004 2002 2004 2002 2004 2002e2004 2002 2002 kg ai/ha ------------------------------------------% reduction in Japanese climbing fern cover---------------------------------Untreated control 7 a 32 bc 38 a 35 a 25 a 14 b -3 a -10 a -7 ab Glyphosate "traditional" 2.24 10 a 29 abc 96 d 89 c 91 c 69 c 73 d 70 f 38 c Glyphosate "tradtional" 4.48 7 a 77 gh 94 d 100 c 91 c 91 cd 91 d 51 def 41 c Glyphosate "generic" 2.24 5 a 54 def 94 d 98 c 91 c 89 cd 87 d 56 ef 40 c Glyphosate "generic" 4.48 7 a 76 gh 92 d 97 c 93 c 93 d 92 d 38 bcdef 19 bc Metsulfuron 0.063 9 a 27 ab 98 d 100 c 95 c 74 cd 41 c 43 cdef 13 abc Imazapyr 0.84 10 a 48 cde 100 d 100 c 92 c 90 cd 77 d 47 cdef 35 c Imazapic 0.56 6 a 32 bc 100 d 100 c 100 c 81 cd 27 bc 29 abcdef 14 abc Sulfometuron 0.16 8 a 9 a 73 c 91 c 51 b 0.1 ab -11 a 15 abcde 0.4 abc Dicamba 3.36 18 ab 80 gh 53 b 57 b 26 a 2 ab -8 a 4 abc -24 a Triclopyr ester 1.12 28 bc 40 bcd 60 b 25 a 44 b -12 a -12 a 11 abcd 5 abc Hexazinone 3.36 87 d 91 h 52 b 62 b 37 ab 14 b 3 ab -1 ab 4 abc 2,4-D + dicamba 3.36 41 c 91 h 77 c 61 b 44 b 5 ab -12 a 23 abcde 15 abc Glyphosate + metsulfuron 2.24 + 0.063 14 ab 65 efg 97 d 100 c 92 c 93 d 91 d 52 def 29 bc Glyphosate + imazapyr 2.24 + 0.84 19 ab 72 fgh 100 d 100 c 100 c 94 d 92 d 57 ef 41 c Metsulfuron + imazapyr 0.063 + 0.84 7 a 42 bcd 100 d 100 c 95 c 91 cd 80 d 48 cdef 37 c Coefficient of variation 2.014 2.014 2.014 2.014 2.014 2.014 2.014 2.014 2.014 Least significant difference (0.05)16 21 12 21 17 22 25 44 41 a Means followed by the same letter are not significantly differen t at P<0.05. Negative values i ndicate increase in fern cover ab ove pre-treatment levels b Months after treatment (MAT). c For all treatments except the untreated contro l, "Induce" surfactant was added at the rate of 0.025% v/v (Helena Chemical). d Year of study initiation. e Plots were mechanically cleared by timber harvest event. Treatment differences remained obscured until 24 MAT.

PAGE 106

92 Table 3-3. Change in Japanese climbing fern foliar cover on untreated control plots from two herbicide trials in a North Florida slash pine plantation Trial year Initial % cover a Final % cover a Time period %Change/yearb 2002 77.5 (.6) 83.4 (.4) 3 years +2.5% NS 2004 70 (.7) 72.5 (.8) 1 year +3.6% NS Mean % increase in climbing fern foliar cover = +3.1% NS a Mean (n=4), standard error b % Change/year = [(Final%-Initial%)/Init ial%]/Time period. NS indicates nonsignificant change in cover (SAS Version 9.1, analysis of variance, alpha=0.05) Figure 3-1. Average live foliar cover of Ja panese climbing fern on untreated control plots established in October 2002 and November 2004 as part of a multiple herbicide trial in a North Florida pine plantation. Approximate month after treatment (MAT) indicated. 0 10 20 30 40 50 60 70 80 90 100 0 MAT Oct./Nov. 1 MAT Nov./Dec. 5 MAT Mar./Apr. 6 MAT May 8 MAT Jun.(02) 10 MAT Sep (04) 12 MAT Oct./Nov. 36 MAT Oct./Nov. Month 2002 trial plots (standard error bars shown) 2004 trial plots (standard error bars shown)% fern cove r

PAGE 107

93 Table 3-4. Japanese climbing fern herbicide treatment options for land managers, based on results of evaluation in a North Flor ida slash pine plantation, 2002 to 2005. Treatment effect durationa Recommended herbicides Solution rates 6 month 10 month 12 month 24 month Cost per hectare (acre)b Cost per unitb Glyphosate 2.50% Yes Yes Yes Yes $15 to $43 ($6 to $17) $12 to $35 per gallon Glyphosate 5% Yes Yes Yes Yesc $30 to $86 ($12 to $35) $12 to $35 per gallon Glyphosate + imazapyr 2.5% + 0.94% Yes Yes Yes Yesc $153 ($62) Glyphosate + metsulfuron 2.5% + 0.56 g/L (0.075 oz/gal) Yes Yes Yes Yesc $111 ($45) Imazapyr 0.94% Yes Yes Yes Yes $116 ($47) $250 per gallon Metsulfuron+ imazapyr 0.56 g/L (0.075 oz/gal) + 0.94% Yes Yes Yes Yes $190 ($77) Metsulfuron 0.56 g/L (0.075 oz/gal) Yes Yes No No $74 ($30) $20 per ounce Imazapic 1.25% Yes Yes No No $346 ($140) $560 per gallon Sulfometuron 1.12 g/L (0.15 oz/gallon) Yes No No No $229 ($93) $500 per pound a Yes indicates treatment sustained a minimum of 70% control for 6,10,&12 month periods, and a significant difference from control for 24 month period. b based on per unit costs from Ferrell et al. 2006 c Greater than 90% control.

PAGE 108

94 CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH The continual introduction of new species a nd the expanding distributions of those already present in the United States make the need for objective priority-based management increasingly important (Richardson 2004). However, for many species, the type and scale of economic and environmenta l impacts are poorly or loosely quantified, making it difficult to prioritize species for use of limited management dollars. The disparity in recognition and prioritization of management of the two non-native invasive climbing ferns in Florida provides an exampl e of the influence of documented or clear ecological impacts (as undeniably seen with Old World climbing fern), versus less obvious, or un-quantified impacts (as seen with Japanese climbing fern). Further pursuit of site cla ssification data for Japanese climbing fern is not strongly recommended. In general, the successf ul design and implementation of broad classification assessments is difficult, and id entification of meaningful relationships is often not feasible, or efficiently accomplished. In future consideration of a broad site classification study design, if a similar grid system is used, increasing the sampling intensity (i.e. establishing a greater number of transects and plots per study area) might capture a more normal distribution in plot va riable values, and Ja panese climbing fern cover, allowing for more robust and producti ve analyses, and possibly more definitive results. Another approach would be to narro w the plot selection to transition zones or ecotones between Japanese climbing fern presence and absence which might explain the basic limiting factors that exclude Japanese climbing fern from some areas. The

PAGE 109

95 apparent causes are linked to so il inundation. However, the re sults of the site assessment described in Chapter 2, and review of current distribution ecology in Chapter 1 indicate that it is likely that Japanese climbing fern will occupy a broad range of sites of mesic to hydric moisture regimes, with other site char acteristics playing minor and varying effects on establishment, or percent cover. Some aspects of management of Japanese climbing fern infestations do require further research review. Most invasive species research seeks to answer at least one of the following key questions: Where does it o ccur?, What are the ecological impacts?, How do we manage it and what are the cost s?, and Can we halt further spread, and how? Though the results repor ted herein do begin to addre ss parts of these questions, further work is needed in se veral areas. Research recomm endations to further answer these questions follow. Ecological Impacts Of highest research priority is an asse ssment of the impacts of Japanese climbing fern invasion at the site, sp ecies, and community levels. In the United States, Japanese climbing fern has already invaded a significan tly larger geographic area than Old World climbing fern, but without the consistent fo rmation of canopy-smothering infestations. Though not extending over the tops of oversto ry trees, ground-leve l Japanese climbing fern infestations can reach 70% to 100% gr oundcover in heavily infested sites. These heavy infestations can effectiv ely preclude recruitment of na tive plant species, in strata ranging from groundcover to overstory, and in addition may have direct competitive effects on mature plants from midstory to ground level. The impacts of infestations occurring at lower densities over many acres are not as clear, particularly since the ecological functions and importance of indivi dual native species are poorly understood in

PAGE 110

96 many cases. Research is needed to evaluate the true impacts of these infestations on native flora and fauna to help clarify the priority of prevention and management of this species on public and private lands. Specifi c research priorities include: impacts of Japanese climbing fern invasion on native plant communities at groundcover, understory, midstory, and overstory levels; and impacts on native arthropod communities. Distribution and Spread Efforts are currently in place through th e cooperative Central Florida Lygodium Strategy to evaluate the effect of the 2004 and 2005 hurricane seasons on both the range extension, and level of site occupancy of both climbing fe rns (Serbesoff-King 2006). As Japanese climbing fern continues to expand its distribution in central and south Florida, seasonal climatic restrictions wi ll be removed and the relative biological potential of the two species will be more clearly comparable. However, Japanese climbing fern also continues northward range exte nsion, under-recognized by many. The true extent of Japanese climbing fern s county-level range in Georgia has been under-reported by the large reporting databases, as verified by queries of individual herbaria in the Southeast. In particular, the curator of the Val dosta State University Herbarium added seven new Georgia county re cords (Figure 1-3) in a one-month period after simply increasing attention to Japanese climbing fern occurrence in his daily travels, in response to a query from the author (R. Clark, personal communi cation, 7/14/2006). A similar trend for under-reporting this species pr obably exists in Alabama, Mississippi, Arkansas, South Carolina, a nd East Texas. Well beyond si mple incursions into the southern regions of infested states, Japa nese climbing fern al ready extends throughout Louisiana, and reaches well in to central regions of Mississippi, Alabama, Georgia, and South Carolina.

PAGE 111

97 Recommendations for future work in this area include: active pursuit of an increase in recognition and reporting of Japanese climbing fern occurrence in these states to further aid in understanding the current and fu ture impacts of this species on mesic and hydric natural communities throughout the Southeas tern United States. This may require regional or state-level funding for surveys in target areas. In addition to location, descriptions of site should be reviewed for commonalities indicat ive of common vectors or dispersal mechanisms. Assessment of re gional spore loading levels in Northwest Florida and in the central and northern ranges of each infested state in the Southeastern United States (Figure 1-3) may help clarif y the infestation extent and eradication feasibility across the current range. Spore loading assessments must be able to differentiate spores of L. japonicum and Lygodium palmatum. Outside the United States, the present stat us and possible future distribution of Japanese climbing fern and Old World climbi ng fern in the Caribbean islands may also need review. These systems may be particularly vulnerable due to the fact that an empty niche may exist in Caribbean forests due to th e dearth of native vi nes in those systems (Horvitz et al. 1998), and proximity to the heavil y infested State of Florida (Figure 1-2). Management Gaps Rejmanek and Richardson (1996) suggest that non-native species become successful invaders based on small widely dispersed seeds becoming established in anthropogenically-disturbed ecosystems. Hor ovitz et al. (1998) propose that invasion success is determined by recruitment and persis tence at low densities in natural habitats, followed by rapid population expansion after a di sturbance event. Bo th of these invasion paradigms apply to Japanese climbing fern, indicating that the s cattered, lower density populations typical of many in fested sites are simply th e preliminary stages of

PAGE 112

98 infestations which await rele ase through site or canopy distur bance. In forested sites under silvicutural management, the unfortuna te result may be rapid conversion of scattered, incipient popu lations to site dominating near-monocultures following full or partial harvests. This conversion will complicate future management, including increasing the cost and reducing the success of reforestation e fforts, and altering the form and diversity of groundcover plant communities. This effect has already been observed following removal of pine needle litter from mesic pine planta tions in north Florida as a component of the pine straw industry (Van Loan 2001). In particular because of Japanese climb ing ferns reproductive strategy, proactive treatment at the invasion fr ont is a far better manageme nt strategy than waiting until regional populations and spor e loading prevent eradication from being a feasible management goal. Therefore, development of reporting programs and supporting training sessions on iden tification and treatment of Japanese climbing fern in the counties forming the northern extent of the species in the United States is recommended. State and county government involvement is critical to raise public awar eness, and regional experts will be required to verify sites as n ecessary. Public and private land managers should receive focused training and land mana gement funding support, concurrent with provision of the same to counties immediately adjacent to the north and south. This way, further northward spread of the species may be halted, allowing focused local education and treatment programs to begin moving th e range southward. This strategy is recommended for immediate implementation in the Southeastern United States. Eradication may be possible in some regions or states, leaving heavily infested states with pockets requiring ongoing main tenance-control treatments.

PAGE 113

99 Given that site-based rest rictions on establishment are potentially limited beyond general site hydrology, improving management of infested sites may be far more important targets for future research than further assessment of habitat preferences. Some recommended priorities include: assessment of the e ffect of repeated herbicide applications of higher-performing herbicide treatments within a single year, and over several years. A comparison of treatment ef ficacy and associated seasonal spore loading in stands treated in late spri ng versus fall may further assist in setting treatment timelines. Given the fact that one single treatment w ill not eradicate this species from a site, quantification of the extent a nd cost of an eradication-base d management strategy will aid land managers in long term planning. The eff ect of varying levels and timings of site disturbance should be assessed, including common silvicultura l treatments to facilitate adaptive management on managed lands. When establishing herbicide or other treatment plots in Japanese climbing fern infestati ons, a minimum plot size of 3.4 m by 6.8 m to prevent treatment effects from being carried into adjacent plots. A second approach would be to use buffers between plots This leads to the next recommendation; focused pursuit of a biological control agent for Japanese climbing fern. Presentl y, Japanese climbing fern receives some review as a result of the sear ch for an agent for Old World climbing fern. Range overlap with L. palmatum does reduce the likelihood that a hos t-specific agent will be found, but the possibility does still exist. Foreign explor ation for possible insect or fungal agents in various areas of the range of Japanese c limbing fern is recommended to address the feasibility of biological control of Japanese climbing fern.

PAGE 114

100 Regulatory Gaps Finally, increased recognition of this plant at a regulatory level may be the only path to halting its continued sp read. In particular, an in-sta te review of the noxious weed status of this species in Ge orgia, Texas, Mississippi, Arka nsas, Lousiana, South Carolina, and North Carolina is highly recommended, and may aid in generating adequate focus to effectively treat and manage the widely o ccurring incipient populations that reflect Florida infestation levels of decades past. Concurrently, the possible range in the United States should be reviewed more carefully, as this is a key question to support petition of this species for federal noxious weed status. If at minimu m, the Southeastern Coastal Plain represents the possibl e United States range, populat ions are not yet known in: Tennessee, Missouri, Oklahoma, or Virginia, as well as large areas of eastern Texas, and northern Mississippi (Figure 12). If however, the possible range is closer to that indicated by the current norther n (43 N) and southern (33 S) extent of the species, appropriate sites ranging from Boston, Massachusetts, USA to Santiago, Chile might be vulnerable (Figure 1-2). Both scenarios are possible at some level, and both support the need for designation of Japanese climbing fern as a federal noxious weed by the United States Department of Agricultu re (USDA). Development of a petition to this effect for submission to the USDA is recommended. The need for coordination of regulatory and management approaches is easily illustrated by forest product movement and the related management practices; and the potential for spread of Japanese climbing fe rn associated with each. The ability for Japanese climbing fern to be baled and sold in contaminated pinestraw bales has been shown repeatedly, indicating that despite some legal basis an effective preventative, or punative system has not yet been implemente d. The most comprehensive approach to

PAGE 115

101 preventing the continued vectoring of Japanese climbing fern this way is to halt the continued harvest or sale of any forest products from infested stands. Legal basis exists already for this action in Fl orida and Alabama. However, comprehensive implementation of such a requirement would require significant increases in funding for state or federal agricultural inspectors in rural, forested areas. One approach which may achieve practical success would include funding and tr aining a rapid response team in each state with known infestations to address further expansion, particularly northward. A second step would included providing cost-share funds for treatment of infestations on private forestlands. With participation in the co st-share program, landowners could provide forest products through a Fern-free certifica tion program, and successfully comply with state agricultural rules. As the gaps in knowledge continue to be filled with current and future research, a more complete understanding of Japanese clim bing ferns occurrence and impacts in the Southeastern United States will be possible. Communication of the r eal nature of these impacts will be key in succe ssful future management of this species, and the natural communities vulnerable to its invasion. The re sults reported in this document represent early steps in development of a regional land scape-based management approach for this species capable of effectively addressing its regional spread.

PAGE 116

APPENDIX INITIAL ANALYSIS TABLES FROM CHAPTER 2

PAGE 117

103 Table A-1. Abiotic plot vari able with significant relations hip to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests. Variablea Lygodium status Plots Mean Std. error Kruskal-Wallis valuesb Spearman R correlation coefficientsc Correlation strengthd Present 29 2.52 0.12 Moderate Site Absent 46 1.67 0.11 19.42 (P<0.0001)* 0.50 (P<0.0001)* Present 29 133.4934.75 Weak Plot PAR Absent 46 454.9853.59 15.00 (P<0.0001)* -0.45 (P<0.0001)* Present 29 2.59 0.12 Weak Hydrology class Absent 46 1.96 0.14 8.57 (P=0.003)* 0.34 (P=0.003)* Present 29 2.07 0.42 Weak Plot flood duration (months, 2003) Absent 46 1.03 0.26 4.74 (P=0.029)* 0.25 (P=0.029)* Present 29 0.27 0.07 Weak Percent flooded Absent 46 0.15 0.05 4.51 (P=0.034)* 0.25 (P=0.033)* a PAR=photosynthetically active radiation. b Asterisk* indicates significant value Kruskal-Wallis one way analysis of variance test p-value, alpha = 0.05, SAS Version 9.1 c Asterisk* indicates significant valu e Spearman's correlation coefficient a nd p-values, alpha=0.05, SAS Version 9.1 d Correlation strength: weak=0.20.49, moderate=0.5-0.79, strong=0.8-1.0

PAGE 118

104Table A-2. Biotic plot variable s with significant non-paramteric re lationships to presence or absence of Japanese climbing fer n ( Lygodium japonicum) in three North Florida forests. Variable a Lygodium status Plots Mean Standard error Kruskal-Wallis valuesb Spearman R correlation coefficientsc Correlation strengthd Present 29 29.14 4.49 Weak Plot % understory cover Absent 46 61.30 5.85 11.30 (P<0.0001)* -0.39 (P=0.0005)* Plot % canopy cover Present 29 60.04 2.56 10.48 (P<0.001)* 0.38 (P=0.0009)* Weak Absent 46 41.34 3.38 Present 29 0.65 0.07 Weak Plot tree evenness index Absent 44 0.37 0.06 8.71 (P=0.003)* 0.32 (P=0.005)* Present 29 3.07 0.30 Weak Plot tree species richness Absent 46 2.13 0.23 6.72 (P=0.010)* 0.30 (P=0.009)* Present 29 2.24 0.22 Weak Plot tree diversity index Absent 44 1.72 0.15 6.97 (P=0.008)* 0.31 (P=0.074)* Present 29 11.07 0.98 Weak Trees per plot Absent 46 8.37 0.94 4.25 (P=0.040)* 0.24 (P=0.038)* a Plot tree evenness index = Piel ou J', plot tree diversity inde x=Simpson's reciprocal 1/D. b Asterisk* indicates significant value Kruskal-Wallis one way an alysis of variance test p-value, alpha = 0.05, SAS Version 9. 1 c Asterisk* indicates significant value Spearman R correlation coeffi cient, alpha=0.05, SAS Version 9.1 d Correlation strength: weak=0.20.5, moderate=0.50.8, strong=0.8-1.0

PAGE 119

105Table A-3. Soil variables with significant parametric relati onship to presence or absence of Japanese climbing fern ( Lygodium japonicum) in three North Florida forests. Variable (data transformation)a Lygodium status N Mean Standard error T valuesb Pearson correlation coefficientsc Correlation strengthd Present 29 5.99 0.12 Moderate Soil pH (inverse) Absent 46 5.23 0.09 5.17 (P<0.0001)* 0.51 (P<0.0001)* Present 29 26.03 2.78 Moderate Soil P (inverse) Absent 46 9.41 1.96 5.24 (P<0.0001)* 0.51 (P<0.0001) Present 29 283.9335.51 Weak Soil Ca (natural log) Absent 46 101.2620.32 -5.72 (P<0.0001)* 0.49 (P<0.0001)* Present 29 13.52 1.14 Weak Soil Fe (inverse) Absent 46 31.50 4.08 -5.18 (P<0.0001)* -0.37 (P=0.001)* Present 29 34.03 3.61 Weak Soil Mg (inverse) Absent 46 17.96 2.81 5.04 (P<0.0001)* 0.38 (P=0.0007)* Present 29 191.7118.38 Weak Soil Al Absent 46 332.8928.41 4.17 (P<0.0001)* -0.39 (P=0.0005)* Present 29 40.76 7.02 Weak Soil K (natural log) Absent 46 21.27 2.89 -2.90 (P=0.005)* 0.32 (P=0.005)* Present 29 2.19 0.30 Weak Soil Zn Absent 46 1.33 0.16 -2.48 (P=0.017)* 0.30 (P=0.009)* Present 29 29.56 6.73 Weak Soil Na Absent 46 13.03 2.68 -2.28 (P=0.028)* 0.29 (P=0.011)* a Where indicated, data transformations utilized to im prove distribution normality for parametric analyses. b Asterisk* indicates significant value tw o sample T-test, alpha = 0.05, SAS Version 9.1 c Asterisk* indicates significant value Pearson's correlation coeffi cient, alpha=0.05, SAS Version 9.1 d Correlation strength: weak=0.20.5, moderate=0.50.8, strong=0.8-1.0

PAGE 120

106Table A-4. Tree species with importance values (IV) significantly related to presence or absenc e of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests. Species Veg. indexa Total plots Lygodium status Plots Mean IV Kruskal-Wallis valuesb Spearman R correlation coefficientsc Corr. Strengthd Present 10 25.4 17.94 (P=0.0001)* 0.34 (P=0.003)* Weak Liquidambar styraciflua L. FAC 10 Absent 0 5.0 Present 3 28.3 -15.88 (P<0.0001)* -0.45 (P<0.0001)* Weak Pinus palustris Mill. FACU 28 Absent 25 108.1 Present 12 40.2 11.88 (P=0.0006)* 0.37 (P=0.001)* Weak Carya aquatica (Michx. f.) Nutt. OBL 16 Absent 4 13.9 Present 7 13.1 6.28 (P=0.012)* 0.19 (P=0.10) Quercus lyrata Walt. OBL 9 Absent 2 6.6 Present 7 13.3 6.28 (P=0.012)* 0.19 (P=0.11) Ulmus americana L. FACW 9 Absent 2 7.4 Present 3 5.6 4.89 (P=0.027)* 0.22 (P=0.06) Weak Carpinus carolinian a Walt. FACW 3 Absent 0 2.0 Present 9 17.8 4.35 (P=0.037)* 0.29 (P=0.01)* Weak Quercus nigra L. FACW 14 Absent 5 9.7 Present 4 10.1 4.00 (P=0.046)* 0.27 (P=0.02)* Weak Taxodium distichum (L.) L. C. Rich OBL 5 Absent 1 3.3 a Vegetative index: FAC = facultative, FACU = facultative upland, OBL=oblig ate wetland, FACW=facultative wetland. b Asterisk* indicates significant value Kruskal-Wallis an alysis of variance p-values, alpha = 0.05, SAS Version 9.1 c Asterisk* indicates significant value Spearman R correlation coefficient, alpha=0.05, SAS Version 9.1 d Correlation strength: weak=0.20.5, moderate=0.50.8, strong=0.8-1.0

PAGE 121

107 Table A-5. Tree species with importance va lues (IV) not significantly related to presence or absence of Japanese climbing fern ( Lygodium japonicum ) in three North Florida forests. Species Vegetative indexa Total plots Lygodium status Plots Mean IV Std. dev. IV Present 8 58.2 83.2 Pinus elliottii Engelm. FACW 18 Absent 10 41.0 67.7 Present 5 20.0 40.6 Nyssa aquatica L. OBL 8 Absent 3 15.3 44.1 Present 1 4.8 9.7 Nyssa sylvatica Marsh. OBL 6 Absent 5 13.5 35.3 Present 3 7.0 16.2 Acer rubrum L. FACW 6 Absent 3 9.6 29.4 Present 4 8.8 15.6 Betula nigra L. OBL 5 Absent 1 4.3 8.9 Present 0 3.0 0 Quercus stellata Wangenh. FACU 5 Absent 5 8.7 20.3 Present 0 2.0 0 Quercus falcata Michx. FACU 4 Absent 4 5.5 14.3 Present 0 2.0 0 Magnolia grandiflora L. FAC 3 Absent 3 4.8 13.2 Present 1 5.2 17.5 Liriodendron tulipifera L. FACW 3 Absent 2 4.4 12.0 Present 1 2.6 3.3 Quercus laevis Walt. FACW 3 Absent 2 2.6 2.7 Present 0 1.0 0 Bumelia lycioides (L.) Pers. FAC 2 Absent 2 1.4 1.9 Present 0 1.0 0 Persea palustris (Raf.) Sarg. OBL 2 Absent 2 1.5 2.4 Present 0 1.0 0 Quercus hemisphaerica Bartr. ex Willd. FAC 2 Absent 2 2.6 7.8 Present 0 1.0 0 Taxodium ascendens Brongn. OBL 2 Absent 2 1.5 2.4 Present 1 1.3 1.9 Catalpa bignonioides Walt. FAC 1 Absent 0 1.0 0 Present 1 2.9 10.4 Celtis laevigata Willd. FACW 1 Absent 0 1.0 0 Present 1 4.1 16.9 Ligustrum sinense Lour. FAC 1 Absent 0 1.0 0 Present 1 3.6 14.1 Nyssa ogeche Bartr ex Marsh. OBL 1 Absent 0 1.0 0 Present 1 2.2 6.3 Planera aquatica Walt. ex Gmel. OBL 1 Absent 0 1.0 0 Present 1 2.2 6.5 Platanus occidentalis L. FACW 1 Absent 0 1.0 0

PAGE 122

108 Table A-5. Continued Species Vegetative indexa Total plots Lygodium status Plots Mean IV Std. dev. IV Present 1 1.3 1.9 Populus heterophylla L. OBL 1 Absent 0 1.0 0 Present 1 2.5 8.0 Quercus michauxii Nutt. FACW 1 Absent 0 1.0 0 Present 1 1.8 4.5 Viburnum nudum L. FACW 1 Absent 0 1.0 0 Present 0 1.0 0 Cornus florida L. FACU 1 Absent 1 1.3 2.1 Present 0 1.0 0 Cyrilla racemiflora L. FAC 1 Absent 1 1.5 3.7 Present 0 1.0 0 Ilex opaca Ait. FAC 1 Absent 1 1.0 0.3 Present 0 1.0 0 Diospyros virginiana L. FAC 1 Absent 1 1.6 4.3 Present 0 1.0 0 Lindera benzoin (L.) Blume FACW 1 Absent 1 1.5 3.1 Present 0 1.0 0.2 Myrica cerifera L FAC 1 Absent 1 1.2 1.0 Present 0 1.0 0 Persea borbonia (L.) Spreng. FAC 1 Absent 1 1.3 1.8 Present 0 1.0 0 Pinus taeda L. FAC 1 Absent 1 3.3 15.5 Present 0 1.0 0 Quercus incana Bartr. FACU 1 Absent 1 1.3 2.2 Present 0 1.0 0 Symplocos tinctoria (L.) L' Her. FAC 1 Absent 1 2.4 9.7 Present 1 1.7 3.5 Gleditsia aquatica Marsh. OBL 1 Absent 1 1.6 4.1 Present 0 3.7 14.3 Magnolia virginiana L. OBL 1 Absent 1 1.3 1.6 a Vegetative index: FAC = facultative speci es, FACU = facultative upland species, FACW=facultative wetland species, OBL=obligate wetland species.

PAGE 123

109 LIST OF REFERENCES Aherns, W. H. ed. 1994. Herbicide Handbook seventh edition. Champaign, IL: Weed Science Society of America. 352 p. Allen, H. L., T. R. Fox, and R. G. Campbe ll. 2005. What is ahead for intensive pine plantation silviculture in the sout h? South. J. Appl. For. 29:62. Alston, A. H., and R. E. Holtum. 1959. No tes on the taxonomy and nomenclature in the genus Lygodium (Schiziaceae). Reinwardtia 5:11. Barnard, E. L. and A. N. Van Loan. 2002. An assessment of fusiform rust and selected non-native invasive pest plants in longleaf and slash pine plantations established with Florida Division of Forestry seedli ngs. A report to the USDA Forest Service Forest Health Monitoring Program, Regi on 8, Atlanta, GA: Florida Division of Forestry. 17 p. Beckner, J. 1968. Lygodium microphyllum another fern escaped in Florida. Am. Fern J. 58:93. Berquist, G. T., A. M. Pable, M. C. Killings worth, and J. A. Silvestri. 1995. Apalachicola River and Bay ecosystem management pl an: environmental indicator system. Florida Center for Public Management, Ta llahassee, FL: Florida State University. 19 p. Betrock Information Systems. 2002. Betrocks PlantFinder: wholesale guide to foliage and ornamental plants. Hollywood, FL: Betr ocks Information Systems, Inc. 148 p. Brandt, L.A., and D.W. Black. 2001. Impacts of the introduced fern, Lygodium microphyllum on the native vegetation of tree is lands in the Arthur R. Marshall Loxahatchee National Wildlife Re fuge. Florida Sc ientist 64:191. Brown, M. J. 1995. Floridas Forests, 1995. Resour. Bull. SRS-48. Asheville, NC: U. S. Department of Agriculture, Forest Serv ice, Southern Research Station. 83 p. Brown, R. B., E. L. Stone, and V. W. Carlisle. 1990. Soils., In R. L. Myers, and J. J. Ewel eds. 1990 Ecosystems of Florida. Orlando, FL : University of Central Florida Press. Pp. 35. Brumback, W. E. 1985. Raising the clim bing fern from spores. Arnoldia 45:27.

PAGE 124

110 Burke, M. J., and J. P. Grime. 1996. An experimental study of plant community invisibility. Ecology 77:776. Chornesky, E. A., A. M. Bartuska, G. H. Ap let, K. O. Britton, J. Cummings-Carlson, F. W. Davis, J. Eskow, D. R. Gordon, K. W. Gottschalk, R. A. Haack, A. J. Hansen, R. N. Mack, F. J. Rahel, M. A. Shannon, L. A. Wainger, and T. B. Wigley. 2005. Science priorities for reducing the threat of invasive species to su stainable forestry. BioScience 55:335. Clawson, R. J., B. G. Lockaby, and B. Rummer. 2001. Changes in production and nutrient cycling across a wetness gradient within a floodplain forest. Ecosystems 4:126. Clewell, A. F. 1985 Guide to the Vascular Pl ants of the Florida Panhandle. Tallahassee, FL: University Press of Florida. 605 p. Cohen, J. 1988. Statistical power analysis for th e behavioral sciences (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Assoc. 500 p. DAntonio, C. D. and L. A. Meyerson. 2002. Exotic plant species as problems and solutions in ecological restoration: a synthesis. Restor ation Ecol. 10:703. Darke, A. K., and M. R. Walbridge. 2000. Al and Fe biogeochemi stry in a floodplain forest: implications for P reten tion. Biogeochemistry 51:1. Davis, M. A., J. P. Grime, and K. Thompson. 2000. Fluctuating resources in plant communities: a general theory of invisibility. J. Ecol. 88:528534. Diggle, A. J., P. B. Neve, and F. P. Smit h. 2003. Herbicides used in combination can reduce the probability of herbicide resistance in finite weed populations. Weed Res. 43:371. Dixon, J. B., and D. G. Schulze. 2002. Soil mine ralogy with environmental applications. Madison, WI: Soil Science Society of America Pp.1. Duryea, M. 1998. Pine straw management in Fl oridas forests, Circular 831. Gainesville, FL: Florida Cooperative Extension Service, IFAS, University of Florida. 12 p. Dyer, A. F., and S. Lindsay. 1992. Soil spore banks of temperate ferns. Am. Fern J. 82:89. Evans, C. and D. Moorhead. 2005. Georgia EP PC events highlight emerging threats. Wildland Weeds 8:4. Ezell, A. W., and L. Nelson. 2001. Weed cont rol and crop tolerance after pre-emergent and post-emergent applications of sulfometuron in oak ( Quercus spp.) plantations. Weed Tech. 15:585.

PAGE 125

111 [FLDEP] Florida Department of Environmen tal Protection. 2006. Flor ida Department of Environmental Protections Wetland Delin eation Vegetative Index : Web page: www.dep.state.fl.us/water/wetlands /delineation/vegindex/vegindex.htm Accessed: July 10, 2006. [FLEPPC] Florida Exotic Pest Plant Council. 2005. Exotic Pest Plant Database: Web page: www.fleppc.org/database/data_intro.htm Accessed: December, 1, 2005. [FLEPPC] Florida Exotic Pest Plant Council. 2005. Florida Exotic Pest Plant Councils 2005 List of Invasive Species: Web page: www.fleppc.org Accessed: December, 10, 2005. [FNAI] Florida Natural Areas Inventory. 2005. Florida Invasive Pl ants GeoDatabase June 2005, Web page: www.fnai.org/gis_data.cfm Accessed: December 10, 2005 Fan, W. 1996. A manual of Chinese herbal me dicine: principles and practice of easy reference. Boston, MA: Shambhala Publishing. p. 207. Ferrell, J. A., C. J. Gray, and G. E. M acDonald. 2006. Approximate Herbicide Pricing 2006, SS-AGR-16. Gainesville, FL: Florid a Cooperative Extension Service, University of Florida. 4 p. Ferriter, A ed 2001. Lygodium Management Plan for Florida, a report from the Florida Exotic Pest Plant Councils Lygodium Task Force. West Palm Beach, FL: South Florida Water Management District. 59 p. Ferriter, A. and Pernas, T. 2006. An explosi on in slow motion: tracking the spread of Lygodium microphyllum in Florida. Wildland Weeds 9:7. Foxcroft, L. C., D. M. Richardson, M. R ouget, and S. MacFayden. 2004. Reconstructing fifty years of Opuntia stricta invasion in the Kruger National Park: environmental determinants and propagule pres sure. Divers. Distrib. 10:427. Gias Uddin, M.D, M. Mahal, and M.K. Pasha. 1997. Taxonomic account of genus Lygodium Sw. (Lygodiaceae) in Bangladesh. Indian Fern Journal 14:337. Gordon, D.R. 1998. Effects of invasive, nonindigenous plant species on ecosystem processes: lessons from Fl orida. Ecol. Appl. 8:975. Hamilton, R. G.. 1988. The significance of s pore ibanks in natural populations of Athyrium pycnocarpon and A. thelypteroides. Am. Fern J. 78:96. Haywood, J.D., F.L. Harris, H.E. Grelen, and H.A. Pearson. 2001. Vegetative response to 37 years of seasonal burning on a Louisiana longleaf pine site. S. J. App. For. 25:122. Hipps, C. B. 1989. Japanese climbing fern: an ex otic vine for a shady fence or trellis. Horticulture, (August 1989):38.

PAGE 126

112 Hodges, A., D. Mulkey, J. Alavalapati, D. Carter, and C. Kiker. 2004. Executive summary: Economic impacts of the forest products industry in Florida. Florida Forests 3:12. Holttum, R. E. 1959. Flora Malesiana, Series II, Pteridophyta. Volume I. Leyden, The Netherlands: Noordhoff International Publishing. Pp. 44. Horvitz, C. C., and A. Koop. 2001. Removal of non-native vines and post-hurricane recruitment of tropical hardwood fore sts of Florida. Biotropica 33:268-281. Horovitz, C.C., J.B. Pascarella, S. McMann, A. Freedman, and R.H. Hofstettler. 1998. Functional roles of invasive, non-indi genous plants in hurricane-affected subtropical hardwood fore sts. Ecol. Appl. 8:947. Hoshizaki, B. J. and R. C. Moran. 2001. Fern Growers Manual. Portland, OR: Timber Press 624 p. Hutchinson, J. T., A. Ferriter, K. Serbes off-King, K. Langeland, and L. Rodgers. 2006. Old World climbing fen management plan fo r Florida: A report from the Florida Exotic Pest Plant Councils Lygodium Task Force. Second Edition. West Palm Beach, FL: Florida Exotic Pest Plan t Council Lygodium Task Force. 109 p. Hutchinson, J. T. and K. A. La ngeland. 2006. Potential spread of Lygodium microphyllum spores by herbicide applic ators. Wildland Weeds 9:13. Jones, K. A. M. B. Rayamajhi, P. D. Pra tt, and T. K. Van. 2003. First report of the pathogenicity of Colletotrichum gloeosporioides on invasive ferns, Lygodium microphyllum and L. japonicum in Florida. Plant Dis. 87:101. Kagawa, T. and M. Sugai. 1991. Involvemen t of gibberellic acid in phytochromemediated spore germination of the fern Lygodium japonicum. J. Plant Physiol. 138:299. Ko, W. 2003. Germination of fern spores in natural soils. Am. Fern. J. 93:70. Kudsk, P. and J. C. Streibig. 2003. Herbic ides: a two-edged sword. Weed Research 43:90. Langeland, K. A., and K. C. Burks. eds. 1998. Identification and Bi ology of Non-native Plants in Floridas Natural Areas. Gainesville, FL: University of Florida Press. Pp. 14. Lauer, D. K., H. E. Quicke, and P. J. Mi nogue. 2002. Early season site preparation with different imazapyr formulations Proc. South. Weed Sci. 55:83. Lavorel, S., A. H. Prieur-Richard, and K. Grigulis. 1999. Invasibility and diversity of plant communities: from patterns to processes. Divers. Distrib. 5:41.

PAGE 127

113 Lee, D. J.. 2005. The private cost of upland i nvasive plants in Flor ida, University of Florida, Institute of Food and Agricultura l Sciences, Florida's Invasive Species Working Group: Web page: http://www.iswgfla.org/ Accessed: July, 15, 2006. Leitman, H. M., J. E. Sohn, and M. A. Franklin. 1983. Wetland hydrology and tree distributions of the Apalachicola Rive r floodplain, Florida, Water Supply Paper 2196. Tallahassee, FL: U. S. Geological Survey. 52 p. Lloyd, R.M., and E.J. Klekowski. 1970. Spore ge rmination and viability in pteridophyta: evolutionary significance of chlorophyllous spores. Biotropica 2:129. Lockhart, C. 2006. The double-trouble ferns: status surveys of Lygodium treatment sites. Wildland Weeds 9:23. Londo, A. J., J. D. Kusha, and R. C. Carter 2006. Soil pH and tree sp ecies suitability in the South, SREF-FM-002. Athens, GA: Southe rn Regional Extension Service. 4 p. Lott, M. S., J. C. Volin, R. W. Pember ton, and D. F. Austin. 2003. The reproductive biology of the invasive ferns Lygodium microphyllum and L. japonicum (Schizaeaceae): implications for invasive potential Am. J. Bot. 90:1144. Louisiana State University. 2006. The Herb arium of Louisiana State University, accession database: Web page: http://www.herbarium.lsu.edu Accessed: May 27, 2006. Magurran, A. E.. 1988. Ecological diversity a nd measurement. Princeton, NJ: Princeton University Press. 179 p. Marthani, G. P., V.K. Bahuguna, and P. Lal. Effect of forest fires on the ground vegetation of the moist deciduous sa l forest. India Forester 112:646. Martin, W. H., S. G. Boyce, and A. C. Echternacht eds 1993. Biodiversity of the southeastern United States: lowland and terrestrial communities. New York, NY: John Wiley and Sons. 502 p. Mehrotra, K., C. K. Mohan, and S. Ranka. 2000. El ements of artificial neural networks. Boston, MA: MIT Press. 348 p. Miller, J.H. 2000. Exotic Pest Pl ants of Southeastern Forest s: a field guide for forest inventory and monitoring. Auburn: USDA Forest Service, Southern Research Station. p. 41. Miller, J. H. 2003. Nonnative Invasive Plants of Southern Forests: A field guide for forest inventory and monitoring, Gen. Tech. Re p. SRS-62. Asheville, NC: USDA Forest Service, Southern Research Station. 93 p.

PAGE 128

114 Miller, J. H., E. B. Chambliss and C. T. Ba rgeron. 2004. Invasive plan ts of the thirteen southern states: Web page: http://www.invasive.org/seweeds.cfm Accessed: May 22, 2006. Mueller, R. J. 1981. Developmental morphology of the shoot of Lygodium ( Schizaeaceae), Ph. D. dissertation. Berkel ey, CA: University of California, Berkeley 141 p. Mueller, R. J. 1982. Shoot morphology of the climbing fern Lygodium (Schizaeaceae): general organography, leaf initiation, and branchi ng. Bot. Gaz. 143:319. Mueller, R. J. 1983. Indeterminate growth of the climbing leaves of a fern. BioScience 33:586. Nauman, C. E. 1993. Lygodium C Presl. In, Flora of North America Editorial Committee, eds., Flora of North America: Volume 2, Pteridophytes and Gymnosperms. New York:, NY: Oxford University Press. Pp.114. Page, C. N. 2002. Ecological stra tegies in fern evolution: a neopteridological overview. Rev. Paleobot. Palynol. 119:1. Peck, J., C. Peck and D. Farrar. 1990. Influen ces of life history attr ibutes on formation of local and distant fern popul ations. Am. Fern J. 80:126. Pemberton, R. W. and A. P. Ferri ter. 1998. Old World climbing fern ( Lygodium microphyllum ), a dangerous invasive weed in Florida. Am. Fern J 88:165. Pemberton, R. W. 2006. Taking down clim bing fern: biological control of Lygodium microphyllum Wildland Weeds 9:10. Pemberton, R. W. 2003. Th e common staghorn fern, Platycerium bifurcatum naturalizes in southern Florida. Am. Fern. J. 93:203. Penrod, K. A., and L. H. McCormick. 1996. Abundance of viable hay-scented fern spores germinated from hardwood forest soils at various distances from a source. Am. Fern J. 86:69. Perez-Garcia, B. A. Orozco-Segovia, and R. Riba. 1994. The effects of white flourescent light, far-red light, darkness, and moisture on spore germination of Lygodium heterodoxum (Schizaeaceae). Am. J. Bot. 81:1367. Pezeshki, S. R. 2001. Wetland plant responses to soil flooding. Environ. Exp. Bot. 46:299. Pielou, E. C. 1975. Ecological diversity. New York, NY: Wiley and sons. Pp. 7.

PAGE 129

115 Pimental, D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic costs associated with no n-indigenous species in th e United States. BioScience 50:53. Platt, W. J., and M. W. Schwartz 1990. Temperate hardwood forests. In R. L. Myers and J. J. Ewel eds. Ecosystems of Florida. Orlando, FL: University of Central Florida Press. Pp. 194. Prantl, K. 1881. Die Schizaeaaceen, Mor phology der gefasskryptogamen volume 2. Leipzig: Verlag Von W ilhelm Engelmann. Pp. 7. Prieur-Richard, A. H., and S. Lavorel. 2000. I nvasions: the perspectiv e of diverse plant communities. Austral Ecol. 25:1. Punetha, N. 1987. Apical dominance and apical inhibition in the fronds of Lygodium japonicum (Thunb.)Sw.. Phytomorphology 37:299. Punetha, N. 2000. Phenology of the climbing leaf of Lygodium japonicum (Thunb.) Sw. (Schizaeaceae). Phytomorphology 50:75. Radford, A. E., H. E. Ahles, and C. R. Be ll. 1968. Manual of the Vascular Plants of the Carolinas. Chapel Hill, NC: Univers ity of North Carolina Press. 1183 p. Rainbolt, C. R., B. A. Sellers, J. A. Ferrell, and G. E. MacDonald 2005. Herbicide Resistant Weeds, SS-AGR-243. Gainesville FL: Florida Cooperative Extension Service, University of Florida. 8 p. Ranal, M. A. 2003. Soil spore bank of ferns in a gallery forest of the Ecological Station of Pangg, Uberlandia, MG, Brazil. Am. Fern J. 93:97. Ranal, M. A.. 2004. Bark spore bank of ferns in a gallery forest of the Ecological Station of Panga, Uberlandia-MG, Brazil. Am. Fern J. 69:57. Rayachhetry, M. B., R. W. Pemberton, L. L. Smith, and R. Leahy. 2001. Pathogenicity assessment of Puccinnia lygodii a potential biological control agent of Lygodium japonicum in southeastern United States. Plant Dis. 85:232. Rhodenbaugh, E. J. and J. L. Yeiser. 1994. Hard wood seedling tolerance to selected Oust treatments. Proc. South Weed Sci. Soc. 47:98. Roberts, D. 1996. Climbing fern wreaks wetland havoc. Florida Department of Environmental Protection Res ource Management Notes 8:13. Roberts, D. 1997. L. microphyllum research and mitigation at Jonathan Dickinson State Park. Florida Department of Environm ental Protection Resource Management Notes 9:30.

PAGE 130

116 Roberts, D. and D. Richardson. 1994. Exotic climbing fern. Florida Department of Environmental Protection Res ource Management Notes 6:16. Royal Botanic Gardens Melbourne. 2006. Aust ralias Virtual Herbarium: Web page: http://www.rbg.vic.gov.au/c gi-bin/avhpublic/avh.cgi Accessed: May 27, 2006.. Russell, A. E., J. W. Raich, and P. M. V itousek. 1998. The ecology of the climbing fern Dicranopteris linearis on windward Mauna Loa, Hawaii. J. Ecol. 86:765. Sahi, A. N. and S. K. Singh. 1994. Effect of sulfite of spore germination and rhizoid development in the tropical fern Lygodium japonicum (Filicales: Lygodiaceae). Rev. Biol. Trop. 42:53. [SAS] Statistical Analysis Systems 2005. SAS Version 9.1 Cary, NC: SAS Institute. Schuler, J. L., D. J. Robison, and H. E. Quicke. 2004. Assessing the use of Chopper herbicide for establishing hardwood planta tions in a cutover site. South. J. App. For. 28:163. Shea, K., and P. Chesson. 2002. Community ecology theory as a framework for biological invasions. Trends Ecol. Evol 17:170. Siegel, S. 1988. Non-parametric Statistics fo r the Behavioral Sciences (2nd Edition). New York, NY: Castellan Jr. Pp.174. Singh, S. and G. Panigrahi. 1984. Systematics of the genus Lygodium Sw. (Lygodiaceae) in India. P. Indian A.S. Plant S.C. 93:119. State of Florida. 2001. Rules of the Departme nt of Agriculture a nd Consumer Services, Division of Plant Industry, Chapter 5B-57, In troduction or Release of Plant Pests, Noxious Weeds, Arthropods and Bi ological Control Agents: Web page: http://doacs.state.fl.us/~pi/5b-57.htm Accessed: December 15, 2003. Strandberg, J. O. 1999. Pathogenticity of the fern anthracnose fungus, Colletotrichum acutatum on wild and cultivated ferns in Florida. Proceedings of the Florida State Horticultural Society 112:274. Stocker, R. K., A. Ferriter, D. Thayer, M. Rock, and S. Smith. 1997. Old World climbing fern hitting south Florida below the belt. Wildland Weeds 1:6. Sundell, E. 1986. Noteworthy vascular plants from Arkansas. Castanea 51:211. Sugai, M., K. Takeno, and M. Furuya. 1977. Di verse responses of s pores in the lightdependent germination of Lygodium ja ponicum. Plant Sci. Letters 8:333. Takeno, K., and M. Furuya. 1980. Sexual differe ntiation in populati on of prothallia in Lygodium japonicum Bot. Mag. Tokyo 93:67.

PAGE 131

117 Taylor, E. L. and C. D. Foster. 2003 Mana ging your East Texas forest for the production of pine straw, Texas Cooperative Exte nsion Service Publication 805-113. College Station, TX: Texas A&M University. 10 p. Tranel, P. J. and T. R. Wr ight. 2002. Resistance of weeds to ALS-inhibiting herbicides: what have we learned? (Review) Weed Sci. 50:700. Tu, M., C. Hurd and J. M. Randall. 2001. Weed Control Methods Handbook: The Nature Conservancy: Web page: http://tnc.ucdavis.edu Accessed: May 20, 2006. [USDA] U. S. Department of Agricultu re. 1980. Soil Survey of Santa Rosa County, Florida. USDA Soil Conservation Service. 134 p. [USDA] U. S. Department of Agricultu re. 2006a. Germplasm Resources Information Network, Lygodium japonicum : Web page: http://www.ars-grin.gov/cgibin/npgs/html/taxon.pl?403435 Accessed: May 28, 2006. [USDA] U. S. Department of Agriculture 2006b. The PLANTS Database, National Plant Data Center, Lygodium japonicum : Web page: http://plants.usda.gov Accessed: June 7, 2006. University of Mississippi. 2006. Pullen Herbarium Specimen Database: Web page: http://herbarium.olemiss.edu/projects.html Accessed: June 7, 2006. University of North Carolina. 2006. The University of North Carolina Herbarium, Distribution maps: Web page: http://www.herbarium.unc.edu/distribution.htm Accessed: June 7, 2006. University of South Carolina. 2006. South Carolina Plant Atlas: Web page: http://cricket.bi ol.sc.edu/herb/ Accessed: June 6, 2006. Valdosta State University. 2006. The Valdos ta State University Herbarium Accession Database, Lygodium japonicum. Valdosta, GA: Valdosta State University. Valenta, J. T., M. Zeller, and A. Leslie. 2001. Glyphosate controls Japanese climbing fern in experimental plots (Florida). Ecological Restoration 19: 118. Van Loan, A. 2001. Weeds in the Woods Florida Forests, (Fall 2001):27. Van Loan, A. N. 2006. Japanese climb ing fern: the insidious other Lygodium Wildland Weeds 9:25. Vitousek, P. M., C. M. DAntonio, L. L. Loope, and R. Westbrooks. 1996. Biological invasions as global environmen tal change. Am. Sci. 84:468. Volin, J. C., M. S. Lott, J. D. Muss, and D. Owen. 2004. Predicting rapid invasion of the Florida Everglades by Old World climbing fern ( Lygodium microphyllum ). Divers. Distrib. 10:439.

PAGE 132

118 With, K. A. 2002. The landscap e ecology of invasive sp read. Conserv. Biol. 16:1192 1203. Wilson, K. A. 2002. Continued pteridophyte in vasion of Hawaii. Am. Fern J. 92:179 183. Wunderlin, R. P. 1998. Guide to the Vascular Plants of Florida. Gainesville, FL: University Presses of Florida 806 p. Wunderlin, R. P., and B. F. Hansen. 2003. A tlas of Florida Vascular Plants, Florida Center for Community Design and Research, Institute for Systematic Botany, University of South Florida, Tampa: Web page: http://www.plantatlas.usf.edu/ Accessed: December 1, 2005. Wynne, G. M., L. N. Mander, N. Goto, H. Yamane, and T. Omori. 1998. Biosyntheic origin of the antheridiogen, gibberellin A73 methyl ester, in ferns of the Lygodium genus. Tetrahedron Lett. 39:3877. Zeller, M. and D. Leslie. 2004. Japanese climbi ng fern controls in planted pine. Wildland Weeds 7:6. Zimmerman, D. W. 1998. Invalida tion of parametric and nonpa rametric statistical tests by concurrent violation of two assump tions. J. of Experiment. Ed. 67:55.

PAGE 133

119 BIOGRAPHICAL SKETCH Andrea (Andi) Van Loan has gained her greate st inspiration in life from the natural world, fostered and supported by her mothers appreciation and love of the same. Andi graduated from the Illinois Mathematics a nd Science Academy in Aurora, Illinois in 1990. After completing her Bachelors degr ee in Resource Ecology and Management at the University of Michigans School of Na tural Resources and th e Environment in 1994, Andi spent a brief time as a forester in Texas before moving to Florida. Andi has worked for the Florida Division of Forestry since 1995, first buildi ng the Ecology Unit at the Withlacoochee State Forest, and then a dvancing the agencys statewide focus on nonnative plant species through work as a biolog ist in the Forest Health Section. Andi enjoys time spent with her family, incl uding: husband, Chris Van Loan, father Harry Christman, brother Eric Chri stman, sister-in-law Connie Ch ristman, and nephews Connor and Strider Christman. Andis mother, Ba rbara Christman, will always serve as her inspiration in love of the natural world, a nd the barometer by which Andis capacity to treat other people fairly and with open caring will be measured.


Permanent Link: http://ufdc.ufl.edu/UFE0016062/00001

Material Information

Title: Aspects of the Invasion and Management of Japanese Climbing Fern (Lygodium japonicum) in Southeastern Forests
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016062:00001

Permanent Link: http://ufdc.ufl.edu/UFE0016062/00001

Material Information

Title: Aspects of the Invasion and Management of Japanese Climbing Fern (Lygodium japonicum) in Southeastern Forests
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016062:00001


This item has the following downloads:


Table of Contents
    Title Page
        Page i
        Page ii
    Dedication
        Page iii
    Acknowledgement
        Page iv
        Page v
        Page vi
    Table of Contents
        Page vii
        Page viii
        Page ix
    List of Tables
        Page x
        Page xi
    List of Figures
        Page xii
    Abstract
        Page xiii
        Page xiv
    Literature review
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
    Assessment of site variables associated with Japanese climbing fern occurrence in three north Florida forests
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
    Herbicide efficacy for management of Japanese climbing fern in southeastern forests
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
    Conclusions and recommendations for future research
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
    Appendix: Initial analysis tables from Chapter 2
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
    References
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
    Biographical sketch
        Page 119
Full Text











ASPECTS OF THE INVASION AND MANAGEMENT OF JAPANESE CLIMBING
FERN (Lygodiumjaponicum) IN SOUTHEASTERN FORESTS














By

ANDREA N. VAN LOAN


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

Andrea N. Van Loan

































To my mother and father, for leading by example, and to my husband for his partnership















ACKNOWLEDGMENTS

I would like to express my sincere appreciation to my graduate committee

members, who have provided the support and freedom necessary for me to successfully

balance my professional, academic, and personal commitments, all of which have been

major components of a life-changing period of time spent as a graduate student at the

University. First, I thank Dr. Jarek Nowak, who initially chaired my graduate

committee, and who recognized the real priorities in my life and allowed me time to

devote to them. Dr. Eric Jokela, who began as a committee member, and finished as my

committee chair; and gave me the necessary support to move me to completion at a very

critical stage. Dr. Greg MacDonald, Agronomy Department representative, has

consistently provided friendship, support, and availability at short notice. Dr. Doria

Gordon, who provided review, comment, and ecological perspective to my efforts, as

well as a dose of reality when stumbling blocks revealed themselves. And finally, Dr

Alan Long, who graciously agreed to join my committee very late in the process.

Invaluable support was provided by other personnel at the School of Forest Resources

and Conservation, including Cherie Arias, and Dr. Wendell Cropper.

This process would not have been possible without the personal and professional

support and flexibility provided by my co-workers, especially those within the Florida

Division of Forestry's Forest Health Section, and the Florida Division of Plant Industry in

Gainesville. In particular, my supervisor, Dr. Ed Barnard, who demonstrated how to









truly support staff advancement in consistently approving my flexible work schedules,

and leave requests made for field research and academic needs.

The field component of this research would not have been possible without the

interest and assistance of several individuals and entities. Critical support for the

comparative herbicide trials was provided by Jerry Wyrick, and Wyrick and Sons Pine

Straw, and Mr. Elton Headings. Key support for the comparative invasion assessment

was provided by: Mr. Dan Miller; the Northwest Florida Water Management District

with assistance from Mr. John Valenta; and the Florida Division of Forestry, Blackwater

River State Forest with assistance from Mr. Tom Cathey, and Mr. Henry Thompson.

This work would not have been possible without support from two funding sources.

Research on the impact of site environment on Japanese climbing fern invasion levels

was supported by a grant from the University of Florida, Institute of Food and

Agricultural Sciences, Center for Aquatic and Invasive Plants, Invasive Plant Working

Group. Herbicide research was supported by a grant from the USDA Forest Service

Special Technology Development Program (Project # 03-DG-11083112-040).

The role and importance of my family cannot be understated as in the end, all other

aspects of life paled in comparison over the last few years. My husband and partner,

Chris Van Loan, supported me both at home, and as my research assistant for a

significant portion of my field work; providing much-needed labor and assistance despite

extreme temperatures, swarming mosquitos, and the relative un-attractiveness of

weekends spent in Calhoun County. My mother, Barbara Christman, showed me

ultimate perseverance and grace throughout her life. That lesson in perseverance became

a touchstone for the writing of this thesis. My father, Harry Christman, showed me the









strength of love as a caregiver in so many senses of the word; allowing me enough peace

of mind to continue to make forward progress with this process. While I already see the

intrinsic rewards of the academic and research training I have received, these human

lessons have been the most humbling and significant of this experience.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

L IST O F T A B L E S .. ............ ................................................... ............... x...... .... ..x

LIST OF FIGURES ............................. .. .......... .................................. xii

ABSTRACT ......................... ............... .................................. xiii

CHAPTER

1 LITERATURE REVIEW: BIOLOGY, ECOLOGY, DISPERSAL, ECONOMIC
IMPACT, AND MANAGEMENT OF JAPANESE CLIMBING FERN .................... 1

In tro d u ctio n .................................................................................................. ............... 1
B io lo g y ...................................................................................................... ........ .. 2
R hizom es .................................................................. ................................... 2
L e av e s ......................................................... ................................................ .. 2
S p o re s .......................................................... .............................................. . 5
G am etophytes/M ating System .......................................................... ...............8...
Sporophytes .............. ................................................ .. ............... ... . ... .9
E c o lo g y ....................................................................................................... ......... 10
N ativ e R an g e ....................................................................................................... 10
Introduced Range ............................................. 10
E nvironm ental R equirem ents ......................................................... ................ 12
Im p acts .. ............ .. ............................................................................. . 14
V ines as Invasive Plants ................. ......................................................... 16
F erns as Invasive P lants........................................ ....................... ............... 17
Related Species in the United States ..................................................18
Recognition and Management in the United States ..............................................20
M anagem ent Status ................ .............. ............................................ 20
R regulatory Status. .......................... .................. ................................................ 20
E conom ic Im pact ... .. ......................................... ........................ . ........... 2 1
M anagem ent P ractices......................................... ........................ ................ 23
P rev en tio n ..................................................................................................... 2 3
C hem ical control .. .. .................. .............................................. 24
M echanical control .... ..... .... .......... .............................................. 26
B biological control ............................ ............................................... 26
P described fi re treatm ent ......................................................... ................ 28









C o n c lu sio n .................................................................................................................. 2 9

2 ASSESSMENT OF SITE VARIABLES ASSOCIATED WITH JAPANESE
CLIMBING FERN OCCURRENCE IN THREE NORTH FLORIDA FORESTS ...32

In tro d u ctio n ................................................................................................................ 3 2
M ethods and M materials .. ..................................................................... ................ 34
S ite S ele ctio n ....................................................................................................... 3 4
P lo t L ay o u t .......................................................................................................... 3 4
Site Descriptions...................................................................... 35
Macroplot Measurements ................... ................ 35
Intensive Sampling Plot Measurements ......................................................... 36
D ata A n aly sis....................................................................................................... 3 7
Initial A naly ses ........................................................................................... .. 39
D iscrim inant A naly sis ......................................... ........................ ................ 40
R egression A analysis .............. ...... ............ .............................................. 40
N eural N etw ork A analysis ...................................... ...................... ................ 4 1
R e su lts....................................................................................................... ....... .. 4 1
Site-L evel O ccurrence ........................................ .. ..................... .. ............. 4 1
P lot-L evel O ccurrence......................................... ........................ ................ 42
Presence/A absence ... ................................................................... ... ............ 42
Cover Class Levels ............. ........................................ 44
D iscrim inant A naly sis ......................................... ........................ ................ 4 5
R egression A analysis ................ .............. ........................................... 46
N eural N etw ork A analysis ...................................... ...................... ................ 47
D isc u ssio n ................................................................................................................ .. 4 8
S ite O v erv iew ...................................................................................................... 4 9
D iscrim inant A naly sis ......................................... ........................ ................ 50
Plot environmental variables ...................................................50
P lot tree sp ecies ........................................................ ............... ............. 5 1
R egression A analysis ............. ................ .............................................. 53
Neural Network Analysis ........................................................56
C o n c lu sio n s............................................................................................................... .. 5 7

3 HERBICIDE EFFICACY FOR MANAGeMENT OF JAPANESE CLIMBING
FERN IN SOUTHEASTERN FORESTS ............................................................. 71

In tro d u c tio n .......................... ............................. ................................................ .. 7 1
Forestry and Invasive Plant Impacts............................................................... 72
M anagem ent T echniques....................................... ...................... ................ 73
M materials and M ethods .. ..................................................................... ................ 74
S tu d y A re a ........................................................................................................... 7 4
H erbicide Selection ............. ................ ............................................... 75
E x p erim ental D esig n ......................................................................... ............... 7 5
Treatment and Assessment Schedule ............................................................. 76
D ata A n aly sis....................................................................................................... 7 6
R results and D discussion ............. .. ............... .............................................. 77


viii









Phenology of Japanese Clim bing Fern........................................... ................ 78
M multiple H erbicide Trial: 2002 and 2004....................................... ................ 79
Short Term Treatm ent E effects ........................................................ ................ 79
T im ber H arvest E ffects........................................ ........................ ................ 82
L ong Term Treatm ent Effects ........................................................ ................ 82
2002 G lyphosate/M etsulfuron Trial ............................................... ................ 83
Conclusions and Recommendations for Land Managers ................ ..................... 84

4 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH ...... 94

E co lo g ical Im p acts ...................................................................................................... 9 5
D distribution and Spread .............. ...... ............ ................................. ...............96
Management Gaps ..................................... .............................97
R regulatory G aps ...... ... ...... ........ ...................................................... .... .... 100


APPENDIX INITIAL ANALYSIS TABLES FROM CHAPTER 2 ...........................102

LIST O F R EFEREN CE S .. .................................................................... ............... 108

BIO GRAPH ICAL SK ETCH .................. .............................................................1...... 19















LIST OF TABLES


Table page

2-1 Japanese climbing fern (Lygodiumjaponicum) occurrence on three North
Florida forest sites, 2002 and 2004 ..................................................... ................ 64

2-2 Japanese climbing fern (Lygodiumjaponicum) occurrence across eight natural
community types in three North Florida forest sites, 2004. Twenty-five plots
assessed p er site ........................................................................................................ 6 4

2-3 Comparison of tree importance values (IV) by species vegetative index and
presence or absence of Japanese climbing fern (Lygodiumjaponicum)................65

2-4 Relationships between Japanese climbing fern (Lygodiumjaponicum) percent
cover levels and significant plot variables and tree importance values in three
N orth F lorida forest sites ......................................... ......... ............................ .... 66

2-5 Plot environmental variables that discriminate between Japanese climbing fern
(Lygodiumjaponicum) presence or absence in three North Florida forest sites .....67

2-6 Plot environmental variables that discriminate between Japanese climbing fern
(Lygodiumjaponicum) percent cover class levels in three North Florida forest
site s ....................................................................................................... . ....... .. 6 7

2-7 Tree species importance values that discriminate between presence and absence
of Japanese climbing fern (Lygodiumjaponicum) presence or absence in three
N orth F lorida forest sites ........................................ ......................... ................ 68

2-8 Tree species importance values that discriminate between Japanese climbing
fern (Lygodiumjaponicum) percent cover class levels on plots with fern present
in three N orth Florida forest sites................... ................................................. 68

2-9 Effects of significant plot environmental and tree importance values in multiple
logistic regression for relationship to presence or absence of Japanese climbing
fern (Lygodiumjaponicum) in three North Florida forest sites.............................. 69

2-10 Effects of significant plot environmental and tree importance values in multiple
logistic regression for relationship to percent cover class levels of Japanese
climbing fern (Lygodiumjaponicum) in three North Florida forest sites. .............70









3-1 Herbicide chemical families and mode of action used in a Japanese climbing
fern control trial in a North Florida pine plantation, 2002 to 2005 .........................89

3-2 The effect of post-emergent herbicides on Japanese climbing fern foliar cover in
a North Florida slash pine plantation, 2002 and 2004......................... ................ 91

3-3 Change in Japanese climbing fern foliar cover on untreated control plots from
two herbicide trials in a North Florida slash pine plantation ...............................92

3-4 Japanese climbing fern herbicide treatment options for land managers, based on
results of evaluation in a North Florida slash pine plantation, 2002 to 2005...........93

A-i Abiotic plot variable with significant relationship to presence or absence of
Japanese climbing fern (Lygodiumjaponicum) in three North Florida forests...... 103

A-2 Biotic plot variables with significant non-paramteric relationships to presence or
absence of Japanese climbing fern (Lygodiumjaponicum) in three North Florida
fo re sts ................................................................. ............................................... ... 1 0 4

A-3 Soil variables with significant parametric relationship to presence or absence of
Japanese climbing fern (Lygodiumjaponicum) in three North Florida forests......105

A-4 Tree species with importance values (IV) significantly related to presence or
absence of Japanese climbing fern (Lygodiumjaponicum) in three North Florida
fo re sts ................................................................. ............................................... ... 1 0 6

A-5 Tree species with importance values (IV) not significantly related to presence or
absence of Japanese climbing fern (Lygodiumjaponicum) in three North Florida
fo re sts ................................................................. ............................................... ... 1 0 7















LIST OF FIGURES


Figure page

1-1 Timeline for Japanese climbing fern plant development from spores ..................30

1-2 Worldwide distribution of Japanese climbing fern (Lygodiumjaponicum) from
literature review (June, 2006). A) Native range indicated in grey. B) Introduced
range indicated in black. Dotted line indicates northern and southern extent of
recorded Japanese climbing fern field populations. ............................ ................ 30

1-3 County-level distribution of Japanese climbing fern in the Southeastern United
States with reference to the Southeastern Coastal Plain physiographic region.
Records based on vouchered herbarium specimens from Southeastern herbaria,
and records from regional vascular plant databases........................... ................ 31

2-1 Location of study sites used in 2002 Japanese climbing fern site establishment
assessm ent, F lorida, U SA ........................................ ........................ ................ 59

2-2 Sampling design for assessment of Japanese climbing fern distribution on three
forest sites in North Florida. A) Layout of major transect lines and intensive
sampling plots. B) Layout of sampling points within each intensive sampling
p lo t ....................................................................................................... . ....... .. 6 0

2-3 Plot layout indicating 2004 Japanese climbing fern cover classes and natural
community types, over 2000 aerial imagery, Blackwater River State Forest
(BR SF), Santa R osa County, Florida .................................................. ................ 61

2-4 Plot layout indicating 2004 Japanese climbing fern cover classes and natural
community types, over 2000 aerial imagery, Miller property (Miller), Calhoun
C ou n ty F lo rid a ......................................................................................................... 6 2

2-5 Plot layout indicating 2002 Japanese climbing fern cover classes and natural
community types, over 2004 aerial imagery, Florida River Island (FRI), Liberty
C o u n ty F lo rid a ...................................................................... ................................... 6 3

3-1 Average live foliar cover of Japanese climbing fern on untreated control plots
established in October 2002 and November 2004 as part of a multiple herbicide
trial in a North Florida pine plantation. Approximate month after treatment
(MAT) indicated ..................... ............ ............................. 92










Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

ASPECTS OF THE INVASION AND MANAGEMENT OF JAPANESE CLIMBING
FERN (Lygodiumjaponicum) IN SOUTHEASTERN FORESTS

By

Andrea N. Van Loan

December 2006

Chair: Eric J. Jokela
Major Department: Forest Resources and Conservation


The continual introduction of new non-native invasive species, and the expanding

distributions of those already present in the United States make the need for objective

priority-based management increasingly important. Japanese climbing fern Lygodium

japonicum (Thunb.)Sw.) is a non-native weed of mesic to hydric natural communities and

sites in the Southeastern United States. Growth ranges from scattered creeping stems to

tangled mats with dense canopy, effectively eliminating understory vegetation, and

smothering seedlings of overstory tree species. Two field studies explored key aspects of

Japanese climbing fern management in forested settings: an assessment of herbicide

efficacy for treatment of existing populations, and a review of site characteristics

associated with varying levels of establishment.

The assessment of site characteristics associated with varying levels of Japanese

climbing fern establishment was conducted in three forested sites in North Florida. Sites

were characterized through a comparison of Japanese climbing fern percent cover

(dependent variable) with a series of independent plot abiotic and biotic variables;

including measurements of soil, and vegetation, assessed on a grid of sampling plots. At









each level of analysis, some plot variables and some species in the plant community

showed significant relationships to Japanese climbing fern occurrence levels. The

general group of significant variables for presence or absence of Japanese climbing fern

included: tree evenness index, plot percent flooded (each winter), soil calcium, and the

occurrence of Carpinus caroliniana. Comparison of plots with Japanese climbing fern

present for varying cover class levels identified photosynthetically active radiation, basal

area, soil aluminum, and the occurrence of Carpinus caroliniana and Quercus nigra. Site

variables showing the strongest relationship to Japanese climbing fern occurrence have a

unifying basis in site hydrology.

The herbicide treatment study was conducted in a heavily infested forest stand

North Florida, where fifteen different herbicides and rate combinations were assessed for

effects on Japanese climbing fern foliar cover, compared with an untreated control.

Treatments including the herbicide active ingredients glyphosate, imazapyr, and

metsulfuron methyl showed the greatest level of Japanese climbing fern control at ten

months after treatment; though only glyphosate and imazapyr sustained greater than 70%

control to twelve months after treatment.














CHAPTER 1
LITERATURE REVIEW: BIOLOGY, ECOLOGY, DISPERSAL, ECONOMIC
IMPACT, AND MANAGEMENT OF JAPANESE CLIMBING FERN

Introduction

"The first time I saw a climbing fern I was shocked, as if I'd opened a window and

a trout had flown in" (Hipps 1989).

These words may describe the reactions of many observing the non-native invasive

climbing ferns in the southeastern United States. This has been particularly true for Old

World climbing fern (Lygodium microphyllum (Cav.) R. Br.), as the species has spread

rapidly and dramatically across South and now Central Florida. Japanese climbing fern

(Lygodiumjaponicum (Thunb.)Sw.) (Family, Schizeaceae), however, has spread with

perhaps equal rapidity, through the understory of mesic to hydric sites across Florida.

Currently, it is found in nine southeastern states. Many aspects of the life history of

Japanese climbing fern indicate the potential for this species to be more broadly

pervasive regionally; though perhaps slightly less aggressive on individual sites.

In Florida, Japanese climbing fern establishes dense mats of vegetation reaching

100% groundcover in some areas. Infestations of the plant affect pinestraw management

on sites throughout North Florida, and threaten native species and habitats throughout the

State. Though increasingly recognized as a pest plant within its established range in the

Southeast, Japanese climbing fern may pose a lower threat to agriculture and biodiversity

than Old World climbing fern (Pemberton and Ferriter 1998). As a result, greater

research focus has been placed on Old World climbing fern, and many gaps exist in the









literature on ecology and management of Japanese climbing fern. A review of the

literature on Japanese climbing fern follows, and subsequent chapters will address

climbing fern management and site ecology. In many areas, references and descriptions

of Old World climbing fern are utilized to provide a more comprehensive view of general

climbing fern biology and ecology. The known similarities between Japanese climbing

fern and Old World climbing fern suggest that a greater understanding of both species

can be gained by study of either.

Biology

Rhizomes

Japanese climbing fern rhizomes are widely creeping, bearing leaves in two rows

on the dorsal side at the rate of three to four every 1 to 2 cm (Mueller 1982). Rhizomes

grow 1 to 3 cm below the soil surface and are densely covered with multicellular hairs

(Mueller 1982) which may help to prevent water loss. Roots are numerous, arising

endogenously on the lateral and ventral sides near the shoot apex (Mueller 1982).

Rhizomes branch dichotomously at 1 to 2 cm intervals after production of three to four

leaves (Mueller 1982). As seen in some fern species, activation of rhizoid elongation

may be due to presence of key soil minerals (Ko 2003). No examples of vegetative

reproduction exist in the literature beyond mention of establishment following transport

of intact fronds with rhizomes contained in soil. Rhizomes are easily divided and

transplanted due to their creeping, branching nature (Hoshizaki and Moran 2001).

Leaves

Japanese climbing fern leaves are borne on the upper surface of slender, dark

brown rhizomes. The rachis of new climbing fern leaves displays "searcher shoot"

morphology, similar to many twining flowering plants, with rapid early rachis elongation









and delayed leaflet expansion (Mueller 1982). Similar to climbing angiosperms, leaves

maintain continual apical growth, exhibit apical dominance (Punetha 1987),

circumnutate, and have bud-like dormant meristems on each leaflet. These dormant

meristems are capable of growing out to replace a damaged leaf tip, or to strengthen the

leaf (Mueller 1981, Mueller 1983).

Japanese climbing fern plants remain evergreen below the frost line in Central and

South Florida (Ferriter 2001, Valenta, Zeller, and Leslie. 2001, Zeller and Leslie 2004).

Above the frost line, winter dieback of Japanese climbing fern fronds occurs to varying

extents through the majority of its range. In the spring, the plants will re-sprout from

cold-tolerant subterranean rhizomes (new stems observed the first week in March in

North Florida), and often utilize freeze-damaged stems as ladders to grow back into the

canopy.

After emergence from the soil in March through May, initial leaf growth is slow

(1.2 to 1.5 cm every 3 days), but steadily increases for at least 70 days (Punetha 2000). A

series of eight to twelve primary leaves are produced in the first five to six months,

before adult leaves are initiated. Successively formed leaves increase in size and take a

pinnate form (Mueller 1982). At peak expansion, leaves grow at a rapid rate (3.4 to 6.5

cm per day) for several weeks, including main rachis extension and steady addition of 10

to 25 new alternate primary rachis branches pinnaee). New pinnae emerge from the

crozier every 1.7 days (Mueller 1983). Rachis internodes can reach 9 to 10 cm in mature

plants (Mueller 1983, Punetha 2000). Leaves exhibit indeterminate growth and can reach

lengths of 2 to 30 m (Mueller 1981, Ferriter 2001).









Pinnae on lower rachis are sterile, while successively newer, upper pinnae are

increasingly fertile as the growing season progresses. Pinnules (leaflets) are initially

tightly coiled, and take five to seven weeks to fully expand to 23 to 25 cm (Mueller

1983). Fertile pinnules bear two rows of sporangia along the underside of the leaflet

margin (Ferriter 2001). Leaflets are highly dissected, with a lacy appearance and

triangular outline, generally 10 to 20 cm long, and about as wide (Nauman 1993,

Langeland and Burks 1998). Leaflet shape and appearance varies noticeably according to

age, season, and site. Upon achieving a height of 20 to 40 cm, leaves circumnutate at the

rate of 1/2 to 1%/4 rotations per day, typically in a counterclockwise direction (Mueller

1983). Leaves will grow vertically, using trees, shrubs, or other structures for support, or

horizontally along the ground (Ferriter 2001), and the climbing habit is undoubtedly an

adaptation to increase sunlight access (Holttum 1959). Growth of the stem and leaflets is

retarded when the fern is grown either upside down or horizontally (Punetha 2000).

Dormant apices on primary rachis branches will proliferate seven days after detopping

the frond or delaminating the secondary rachis branches (Punetha 1987, Punetha 2000).

When grown under stress (modified habit), the height and vigor of Japanese climbing

fern leaves declined significantly (Punetha 2000). In areas with seasonal frost events,

Japanese climbing fern foliage will decrease in percent cover during the cooling

temperatures preceding the frost event. After frost, Japanese climbing fern foliage turns a

rusty-brown color due to stem and leaf death. New leaves emerge from the rhizomes the

following March. This seasonal dieback in sub-temperate and temperate climates is one

factor which has likely prevented formation of the canopy-covering arboreal blankets

seen with Old World climbing fern (Zeller and Leslie 2004). However, the increasing









distribution of Japanese climbing fern in south Florida may lead to the development of

such canopy-smothering growth on infested sites.

Spores

Japanese climbing fern spores are tetrahedral to globose, with a verrucose surface

and trilete aperture that are rust or tan colored. Spores form on the underside of fertile

leaf margins which roll under to cover the sporangia, forming false indusia (Langeland

and Burks 1998). Japanese climbing fern plants can produce spores year-round in frost-

free areas as seen in south Florida (Lott 2001). The effects of seasonal frost and freeze

damage can limit much spore production to June through November to December (Lott

and Volin 2001) in areas further north. In North Florida, peak spore release apparently

occurs in October, when a brown haze of spores is visible in heavily infested stands (Van

Loan 2006). Fertile leaflets of Old World climbing fern each produce an estimated

28,600 spores, or 38,000 spores per square inch of fertile leaf area, translating to as many

as 30 billion spores on heavily infested sites (Volin, Lott, Muss, and Owen 2004). The

typically larger fertile leaflets of Japanese climbing fern may produce a similar quantity

of spores (Lockhart 2006).

Spore viability. Once released from the sporangia, spores require moist

environmental conditions for germination and survival of the prothallia gametophytee

stage). Spores of Lygodium species are thick walled, which extends their environmental

viability, an important consideration for management (Ferriter 2001). However, Lloyd

and Klekowski (1970) predict short spore viability due to the chlorophyllous nature of the

spores. In laboratory conditions, spore germination began within six days of sowing

(Lott, Volin, Pemberton, and Austin 2003). Japanese climbing fern spore germination is

induced by red light through a phytochrome-mediated system, and is greatly influenced









by temperature. Germination rates increase from a low at 200 C as temperatures increase

(Kagawa and Sugai 1991, Sahi and Singh 1994), indicating that cooler temperatures will

limit both germination and growing season. Sugai, Takeno, and Furuya (1997) found

40% to 80% germination rates for Japanese climbing fern spores after storage at room

temperature for 3 years. Perez-Garcia, Orozco-Segovia, and Riba. (1994) found spore

viability ofLygodium heterodoxum in Mexico drastically reduced following storage in

dry conditions (1% germination after 1 year). In contrast, when stored fully imbibed in

water germination averaged 40% after 1 year. Light quality was not significant to spore

germination, explaining germination success under tree canopies.

Spore dispersal. The spores of Japanese climbing fern are tiny structures, easily

transported by wind, water, animals, and equipment (Langeland and Burks 1998), and

may possibly travel several miles from their point of origin. The climbing growth form

of Japanese climbing fern may also contribute to rapid dispersal by placing spores at the

height of the prevailing winds (Lott et al. 2003). Annual flooding in the Appalachicola

River floodplain, as well as periodic flooding of other rivers and streams in Florida,

distributes Japanese climbing fern spores and frond fragments into new areas (Valenta et

al 2001). Land managers have noted observations of increased distribution of the fern in

forest stands in the years following the El Niho-induced flooding events, which occurred

across central and north Florida in the winter of 1997/98. These observations may be

based on increased recognition of the species, increased distribution of the species due to

either flood-based distribution of the spores, or increased soil moisture levels that create

more conducive conditions for germination and survival of spores already present in the

soil and leaf litter substrate (Ferriter 2001).









The apparent spread pattern of Old World climbing fern in south Florida overlaps

with the expected spread pattern if dispersal was primarily due to the prevailing winds of

spring, summer and autumn. Old World climbing fern spore production appears to occur

year-round, and one measurement of spore loading in the air captured 724 spores per m3

per hour within a Martin County infestation (Pemberton and Ferriter 1998). Similar,

though potentially diminished, spore loading and wind-driven dispersal of Japanese

climbing fern may exist given similarities in biology between the two species, although

estimates of spore loading are rare for this species in Florida. Some concern exists over

the long distance spread of Lygodium spores in the convective columns of prescribed or

wildfires. These fire-generated wind conditions can disperse smoke great distance away

from a fire (more than 100 miles in some cases) and may show the ability to disperse

spores in the same way. In addition, soil disturbances associated with forest management

activities have been shown to facilitate establishment of the native fern Dennstaedtia

punctilobula (Michx.) Moore. (Groninger and McCormick 1992), and evidence of a

similar effect has been seen with establishment of Japanese climbing fern in some

harvested sites in northwest Florida.

Spore banks. Natural and cultivated soils can contain a large number of fern

spores in a spore bank, often without close proximity of sporophytes (Hamilton 1988,

Dyer and Lindsay 1992, Ranal 2003). When peak spore maturation periods occur, spores

may fall to the ground in large, scattered masses; a portion of which percolate into the

soil pore space with germination inhibited by lack of light, and others may remain

ungerminated on the soil surface (Ko 2003). Unlike limitations on fungal spore

germination, fern spore germination does not seem to be affected by soil microbiostasis









(Ko 2003). Ko (2003) suggested that fern spores are nutritionally independent, and

germinate successfully on substrates without exogenous nutrients. Soil spore banks may

play an important role in reproductive and establishment success for many fern species,

creating immediate opportunities for spore germination and gametophyte establishment

after any soil disturbance (Ranal 2003). Evidence of fern spore survival in soils for one

year has been obtained in multiple studies, and at least one study verifies survival of

approximately two-thirds of fern spores for two years (Dyer and Lindsay 1992). Dyer

and Lindsay (1992) found that fern spore banks can even survive forest and heath fires.

Bark spore banks have been assessed for two tropical species, verifying the potential for

such a spore bank to contribute to the population dynamic of pteridophyte species (Ranal

2004). Spores present on bark of live or downed trees, and even in bark products

harvested from infested stands may germinate and contribute to dispersal of this species.

When coupled with movement of Japanese climbing fern spores in pine straw bales,

concerns arise over the possible role of timber and non-timber forest product movement

in spore dispersal in the southeastern United States.

Gametophytes/Mating System

Japanese climbing fern is a homosporous fern with free-living haploid

gametophytes and diploid sporophytes. Sexually mature gametophytes appear five weeks

after spore germination (Lott et al. 2003). In the gametophyte generation, the fern is

capable of three types of sexual reproduction: intragametophytic selfing (egg and sperm

from same gametophyte), intergametophytic selling (egg and sperm from two

gametophytes originating from the same sporophyte), and intergametophytic crossing

(outcrossing of egg and sperm from two gametophytes, originating from two different

sporophytes) (Lott et al. 2003). According to Lott et al. (2003) the populations of









Japanese climbing fern in Florida may have originated from only a few plants, with

selection favoring selfing variants rather than outcrossing. It has been proposed that

Japanese climbing fern has evolved antheridiogens such as gibberellin A73-methyl ester

(GA-73-Me) (Wynne, Mander, Goto, Yamane, and Omori 1998), and gibberellin A9

methyl ester (GA9-Me) (Takeno and Furuya 1980) which may promote outcrossing to

avoid the potential inbreeding associated with intragametophytic selfing (Ferriter 2001).

However, Lott et al. (2003) found intragametophytic selfing to be the primary mode of

reproduction for Japanese climbing fern from two populations in Florida. Over 90% of

the gametophytes produced this way developed sporophytes. Therefore, though some

evidence of outcrossing exists, intragametophytic selfing may be a stable mating system

in Florida. If so, this reproductive strategy in Japanese climbing fern may facilitate

colonization of a diversity of habitats and establishment of distant populations from

single spores (Lott et al. 2003).

There are four pairs of Lygodium species which have some amount of intergrade.

This may be due to environmental conditions, or it could be due to hybridization.

Japanese climbing fern (L. japonicum) has been documented to intergrade with Lygodium

flexuosum (L.) Sw. (Holttum 1959).

Sporophytes

Under stressful and variable environmental conditions as seen with drought and

flood cycles, gametophytes that reach sexual maturity quickly have the best opportunity

to produce sporophytes before dying (Lott et al. 2003). Once fertilized in the

gametophyte stage, sporophyte production begins. Sporophyte production occurred from

week five to week twelve (Lott et al. 2003). Adult plants with climbing leaves and the









capability to reproduce have formed by five to six months after spore germination

(Mueller 1982) (Figure 1-1).

Ecology

Native Range

Japanese climbing fern is native to temperate and tropical eastern Asia and the East

Indies, including India, China, Japan, Korea, and many other countries in the region

(Figure 1-2) (Holttum 1959, Singh and Panigrahi 1984, Wee and Chua 1988, Ferriter

2001, Louisiana State University 2006, Royal Botanic Gardens Melbourne 2006, USDA

2006). In its native range, Japanese climbing fern occurs in forest edges, open forests,

and secondary forests at both lower and higher (2550 m) elevations, and is described as

"never colonizing in any area"(Holttum 1959, Gias Uddin, Mahal, and Pasha 1997); but

has also been considered weedy in Taiwan and the Phillippines (Singh and Panigrahi

1984, Langeland and Burks 1998, Ferriter 2001). Holttum (1959) described Japanese

climbing fern as "Only found native in regions with pronounced dry season, during which

fronds perhaps die."

Introduced Range

Considered to have been introduced to the United States around 1900 for

ornamental purposes, the first naturalized population of Japanese climbing fern was

identified in Georgia in 1903 (Pemberton and Ferriter 1998), and in north Florida in 1932

(Gordon and Thomas 1997). Noted as a rare escape in Georgia in 1964 (Radford et al.

1964), it was described as aggressively spreading in moist woods and fields throughout

Florida and adjacent states by 1968 (Beckner 1968). In 2005, Japanese climbing fern

was identified as "spreading at an alarming rate" in Georgia (Evans and Moorhead 2005).

Florida county records for this species have increased from 29 in 1995 to 45 in 1999









(Ferriter 2001), and 53 in 2006 (Van Loan 2006) due to rapid and successful spore

dispersal, and improved recognition and reporting. In Florida, recent occurrences in

Broward, Collier, and Dade counties as well as Brevard, Hardee, Highlands, Manatee,

Polk, and Sarasota counties verify a range overlap with Old World climbing fern (Ferriter

2001, Ferriter and Pernas 2006, Van Loan 2006). This raises concerns that Japanese

climbing fern may invade mesic sites and transcend into hydric sites occupied by Old

World climbing fern; a situation which is beginning to occur in central and south Florida

(Ferriter and Pernas 2006, Lockhart 2006).

As may be true with Old World climbing fern, multiple naturalization events are

possible for Japanese climbing fern (Ferriter 2001) though possibly from the same

propagative parent plant material (Lott et al. 2003). These two fern species were often

confused in early horticultural literature. A photograph and descriptive statements of Old

World climbing fern in the 1905 catalog from Royal Palm Nursery (a primary distributor

of tropical and subtropical flora in Florida and the United States) actually appear to be

Japanese climbing fern. If so, it may have been sold by the nursery for some part of the

period (1888-1930) that the nursery marketed Old World climbing fern (Pemberton and

Ferriter 1998). Promoted in the horticultural trade as a hardy, attractive plant (Ferriter

2001), Japanese climbing fern is often misidentified as Lygodium scandens (L.)Sw., a

synonym for Old World climbing fern (Hoshizaki and Moran 2001).

In the United States, Japanese climbing fern now occurs in eleven states including:

Florida, Georgia, Alabama, Mississippi, Louisiana, Arkansas, Texas, South Carolina,

North Carolina (Figure 1-3), as well as California, Hawaii and Puerto Rico (Figure 1-2)

(Ferriter 2001, Van Loan 2006). County-level occurrence records in the Southeastern









United States indicate a close and expanding range overlap with the Coastal Plain

physiographic region (Sundell 1986, Martin, Boyce and Echternacht 1993). The Coastal

Plain boundary may well serve as a good predictor of the region most rapidly occupied

by Japanese climbing fern. Within that range, Japanese climbing fern occurs primarily in

floodplain forests, marshes, wetlands, secondary woods, moist pinelands (including

flatwoods), and disturbed areas such as culverts, fencelines, road shoulders and rights-of-

way (Clewell 1982, Nauman, 1993, Wunderlin 1998, Langeland and Burks 1998).

Growth ranges from scattered creeping stems to tangled mats with dense canopy,

effectively eliminating understory vegetation (Nauman 1993), and smothering seedlings

of overstory tree species (Langeland and Burks 1998). In the Apalachicola river basin,

Japanese climbing fern is described as occurring at epidemic proportions in bottomland

forests, floodplain forests, sloughs, and mesic flatwoods (Ferriter 2001). Floodplain

swamps are comparatively uninfested due to lower elevations and resultant regular

flooding and inundation (Ferriter 2001). Infestations also occur in xeric sites, but do not

appear to expand as rapidly as in more mesic sites, possibly due to the infrequency of

appropriate conditions for gametophyte establishment or fertilization.

Environmental Requirements

The most successful invasive plants are usually capable of displacing native species

because they have effective reproductive and dispersal mechanisms, superior

competitive ability, limited herbivores or pathogens, ability to occupy a vacant niche, and

ability to alter site resource availability, disturbance regimes, or both (Gordon 1998).

While the ecological requirements of Japanese climbing fern are poorly understood, its

broad distribution indicates a tolerance for a range of environmental conditions.

Observations of land managers indicate reduced success in extremely dry or consistently









flooded sites (Ferriter 2001) and a potential preference for soils with circumneutral pH

(Nauman 1993). Fluctuating precipitation levels may have limited impact on plant

growth, but foresters in north Florida reported that in the dry spring of 1999 Japanese

climbing fern was not as serious of a problem as in previous wetter springs (Ferriter

2001).

Japanese climbing fern is the most widely established non-native, invasive pest

plant on Suwannee River Water Management District lands, and is considered to pose the

greatest threat to natural systems on those lands (Ferriter 2001). Twenty populations

were evaluated on District lands, and found to occur in the following natural community

types: mixed hardwood and pine (9), bottomland hardwood (7), longleaf pine-turkey oak

(5), upland hardwood hammock (2), wet hardwood hammock (1), swamp hardwoods (1),

and North Florida flatwoods (1). Sixteen of the 20 populations occurred between 50 and

60 feet in elevation. Eighteen of the populations were associated with disturbance

(potentially an artifact of the location of the surveys) (Ferriter 2001).

Human disturbance does not appear to be a prerequisite for successful invasion

(Ferriter 2001), but in at least two cases in north Florida, removal of pine litter from the

forest floor of pine plantations was followed by rapid and complete establishment of

Japanese climbing fern in sites where it was previously either absent or present in very

small quantities (Van Loan 2001). Sahi and Singh (1994) assessed the effects of sulphite

exposure on Japanese climbing fern spore germination and rhizoid growth, and found

both to be negatively impacted. This finding indicated that the presence and quantity of

Japanese climbing fern gametophytes may be a bioindicator of sulfur dioxide air









pollution. This finding may have use in some areas in India, within the native range of

Japanese climbing fern.

Japanese climbing fern plants remain evergreen below the frost line, as does Old

World climbing fern in central and south Florida (Ferriter 2001, Valenta et al. 2001,

Zeller and Leslie 2004). Above the frost line, winter dieback of Japanese climbing fern

fronds occurs to varying extents through the majority of its range; plants in shady, moist

protected areas like floodplain and bottomland forests, are able to persist through the

winter in places. Freeze-damaged/killed fronds turn rusty brown in color and are easily

recognized. In the spring, the plants will re-sprout from moderately cold-tolerant

rhizomes (new stems observed the first week in March in North Florida), and often utilize

freeze-damaged stems as ladders to grow back into the canopy. Japanese climbing fern is

promoted as an ornamental plant as far north as Kansas due to cold tolerance of rhizomes,

but the Southwestern Fern Society does recommend mulching the plants after the first

frost for additional rhizome protection in far northern areas (Ferriter 2001). This seasonal

dieback in sub-temperate and temperate climates is one factor which has likely prevented

formation of the "dense arboreal blankets in tree canopies seen with Old World climbing

fern" (Zeller and Leslie 2004). However, the increasing distribution of Japanese climbing

fern in south Florida may soon lead to the development of such canopy-smothering

growth on infested sites.

Impacts

In the Apalachicola River basin in Florida, Japanese climbing fern occurs in

floodplain forests as a thick mat of growth that smothers native groundcover vegetation,

and can extend into the lower overstory canopy (Valenta et al. 2001). Japanese climbing

fern is listed as one of the top contributing factors to the degradation of the Apalachicola









River system (Berquist, Pable, Killingsworth, and Silvestri 1995). In its native range,

Japanese climbing fern can establish relatively large mats; however, these are much

smaller than those commonly found in Florida (Ferriter 2001). The primary effects of

Japanese climbing fern on native flora and fauna stem from impacts on understory plant

communities, including the habitat of the Florida endangered plant species Sideroxylon

thornei (Cronq.) Penn., Aristolochia tomentosa Sims., and Polygonum meisnerianum

Cham. & Schlecht. (Ferriter 2001).

Observations of Old World climbing fern, in south Florida and India indicate that

the dense growth of fronds into the overstory tree canopy can serve as ladder fuels to

carry prescribed fire and wildfire into the canopy, resulting in tree death. Additionally,

the fronds serve as a fuel conduit to allow fires to burn into the center of cypress sloughs

normally left unburned due to water levels, and embers of the dense fern mats can serve

as fuels to carry spot fires away from the primary bum unit (Marthani et al. 1986, Roberts

1996, Roberts 1997). Brandt and Black (2001) found the impact of Old World climbing

fern invasion to be a change in vegetative composition of the native community, resulting

in a decrease in native species and a loss in plant diversity and evenness. In a

comparative study, Clark (2002) found significant reduction in native plant cover,

richness and diversity from uninvaded areas to similar habitats with heavy Old World

climbing fern infestation. Impacts of a similar type, but possibly lower scale may be seen

in sites infested with Japanese climbing fern.

Little is known about the economic impact of Japanese climbing fern invasion.

Forest managers in Florida have noted impacts including reduction in annual financial

return and harvest lease longevity for pinestraw production, an industry which generates









$79 million in revenue for forest landowners in Florida annually (Hodges, Mulkey,

Alavalapati, Carter, and Kiker 2004), and reductions in pine seedling survival during

reforestation of heavily infested sites.

Vines as Invasive Plants

An estimated 138 non-native shrub and tree species have invaded native U.S. forest

and shrub ecosystems (Pimental, Lach, Zuniga, and Morrison 2004), a figure which does

not include invasive grasses, forbs, or vines. However, vines represent a large and

aggressive group of invasive plants in the United States. Nineteen of the 133 plant

species on the 2005 Florida Exotic Pest Plant Council (FLEPPC) Invasive Plant list are

vines, representing 14% of the most invasive plants in Florida (FLEPPC 2005).

Horovitz, Pascarella, McMann, Freedman, and Hofstettler (1998) proposed six functional

groups of invasive species, including "vine blankets". The groups were defined to

classify recolonization of a site following hurricanes, but other severe disturbance may be

considered (i.e. timber harvesting in pine plantations or natural stands). The "vine

blanket" group would include both introduced Lygodium species in Florida with the

proposed effects of shading natives, including: the juvenile/seedling layer, tree re-sprouts

from roots and fallen stems, and regrowth of new branches from standing stems.

Vine blanket species are usually aggressive and may negatively influence forest

regeneration by diverting resources, and strangling native species seedlings. Vine

invasion may drive species composition toward ruderal species, particularly in sites

affected by a significant disturbance event (Gordon 1998). Invasion of fire-evolved

systems by non-indigenous (and native) vines can alter or prevent fire movement due to

increased fuel moisture, and increased site relative humidity. Post-disturbance invasion

by non-indigenous vines has been noted in other forests with Lonicerajaponica Thunb.









Invasion of temperate deciduous forests after smaller-scale natural disturbances have

resulted in crushing and strangling of native shrubs and saplings, and contributed to

increased root and soil moisture competition (Gordon 1998).

Japanese climbing fern effectively colonizes both basic invasible substrates: those

highly disturbed by humans, such as pine plantations, agricultural fields, road shoulders,

ditches, and yards; and those relatively free of anthropogenic disturbance such as

wetlands, river banks, hardwood hammocks, and some pinelands. Key differences also

exist between Japanese climbing fern and several other non-native vines in the Southeast

that possess a twining growth form. The twining growth form of Japanese climbing fern

contrasts with the bark-adhering growth form of native vines in this area. This twining

form enables linkage of tree canopies, increases the likelihood of collapse of the

supporting plants, and changes the plant community structure (Hardt 1986, Gordon

1998). Native vines in the Southeast are primarily temperate species, compared with the

predominantly tropical origin of the invaders. As a result, the tropical invaders may gain

a competitive advantage in the subtropical climates of the Southeast, particularly as seen

in Florida (Gordon 1998).

Ferns as Invasive Plants

There are 34 exotic ferns and fern allies in Florida, representing one-third of

Florida's fern species (Pemberton 2003). Six of the 133 species on the 2005 FLEPPC

plant list are ferns, representing 5% of the most invasive species in Florida (FLEPPC

2005). Pemberton (2003) lists Japanese climbing fern as one of five severely invasive

ferns in Florida along with Lygodium microphyllum, Nephrolepis cordifolia (L.) Presl.

and N. multiflora (Roxb.) Jarrett ex Morton, and Tectaria incisa Cav., all of which









reproduce prolifically and form dense stands or mats of growth, displacing or out-

competing native species.

Ferns are also occurring in increasing numbers in the Hawaiian Islands, where 32

alien fern species have become naturalized, including Japanese climbing fern. Much as

in the Southeast, the species was not observed to have spread from the original 1936

location until 1998. Since then colonization has expanded rapidly, as additional

populations have arisen in multiple locations on O'ahu and Hawai'i (Wilson 2002).

According to Wilson (2002) "the Hawaiian ecosystem continues to be challenged by

invasion of new species of pteridophytes as well as the spread of already present

naturalized species." As a corollary, Dicranopteris linearis (Burm.) Underw., a common

native climbing fern of Hawaii, is a successful colonizer on open substrates due to factors

including: indeterminate clonall) growth, shallow rhizomes, and a mat-forming capacity

(Russell et al. 1998). These same characteristics also contribute to the ability of Japanese

climbing fern to colonize habitats in its introduced range.

Related Species in the United States

From 26 to 40 species occur in the genus Lygodium (Alston and Holltum 1959,

Ferriter 2001) and most are placed in the family Schizaeaceae (Prantl 1881). Two

additional Lygodium species are established in the Southeastern United States, both of

which overlap in range with and create management implications for Japanese climbing

fern. A native species, Lygodium palmatum (Bernh.) Sw. occurs in swamps, streambeds

and ravines in Mississippi, Alabama, and Georgia, and its range extends northeast

through New England (Nauman 1993, Miller 2003); while a second, highly aggressive

non-native species, Lygodium microphyllum (Cav.) R. Brown (Old World climbing fern),

occurs in Central and South Florida (Ferriter and Pernas 2006). Native to moist habitats









in Africa, southeastern Asia, Australia, and Polynesia (Pemberton 1998, Hoshizaki and

Moran 2001), Old World climbing fern has spread rapidly through hydric sites in

southern Florida (Pemberton and Ferriter 1998), where the sub-tropical climate, and the

fern's life history strategies (climbing vine growth form, highly successful reproductive

and long-distance dispersal strategies (i.e., intragametophytic selfing, frond growth into

tree canopies, wind and water-driven dispersal), and rapid development of sporophytes

within a growing season (Lott et al. 2003)), have facilitated the tendency for invaded sites

to be converted to climbing fern monocultures. Old World climbing fern was first

observed growing naturalized in Southeast Florida in the 1960s (Beckner 1968, Nauman

and Austin 1978). Systematic aerial surveys of Old World climbing fern estimated gross

infested area in 1993 at 11,200 ha in South Florida (Pemberton and Ferriter 1998).

Similar surveys completed in 2005 increased the estimate of infested area to 48,878 ha in

South and Central Florida, a small portion of which may actually be Japanese climbing

fern as the species are indistinguishable from the air, and now have overlapping ranges in

this area (Ferriter and Pernas 2006). Though Japanese climbing fern is increasingly

recognized as a pest plant within its established range in the Southeast, Old World

climbing fern has received far greater research emphasis. As a result, much of what is

known or published about Japanese climbing fern biology, ecology, impacts, and

management is the product or by-product of research and assessment of Old World

climbing fern. This relationship has almost certainly enhanced awareness of Japanese

climbing fern, but has potentially served to comparatively diminish perceptions of its

invasiveness and impact as well.









Recognition and Management in the United States

Management Status

Recognition of Japanese climbing fern has increased among public conservation

land managers in the southern United States since the mid 1990s, and has aided in

increased detection and reporting. As a result, Japanese climbing fern was designated as

a Category I invasive plant by the Florida Exotic Pest Plant Council in 1995, due to it's

impacts in "altering native plant communities by displacing native species, changing

community structures or ecological functions" (Ferriter 2001, Florida Exotic Pest Plant

Council 2005). The State exotic pest plant councils of Alabama, Georgia, and South

Carolina also designate Japanese climbing fern as a significant invasive plant (Miller,

Chambliss and Bargeron 2004). While non-regulatory, these designations often guide

management on public conservation lands. The USDA Forest Service Southern Region

has designated Japanese climbing fern as a "Category 1 (exotic plant) species known to

be invasive and persistent", guiding it's management on National Forest lands since 2001

(Miller et al. 2004). Recognition of Japanese climbing fern by private land managers has

also increased since the late 1990s, but management efforts have not increased at a

similar rate, due to lack of awareness of the non-native invasive status of the plant,

research-based best management practices, and financial incentives or support for

management of private-lands infestations.

Regulatory Status

In 1999, Japanese climbing fern was designated as a noxious weed in Florida (Rule

5B-57.007, F.A.C.) by the Florida Department of Agriculture and Consumer Services

(FDACS), making it unlawful to introduce, possess, move, or release any living stage of

the fern without a permit. This designation has raised awareness among private









forestland managers and members of the forest products industry (esp. pinestraw) in

Florida, and for increasing the demand for effective management guidelines. In addition,

in Florida, both Collier and Okaloosa counties have local ordinances regulating L.

japonicum (K. Burks, personal communication, 12/2001). Currently, Alabama is the only

other state which designates Japanese climbing fern as a noxious weed (Van Loan 2006).

Japanese climbing fern is not designated as a federal noxious weed according to the

Federal Noxious Weed Act of 1974 (7 U.S.C. 2809). A petition was submitted to the

United States Department of Agriculture, Animal and Plant Health Inspection Services

(APHIS) in 1999, requesting that both Japanese and Old World climbing ferns be given

federal noxious weed status. The widespread distribution of Japanese climbing fern

excluded it from definition as a quarantine pest, making designation unlikely without

further research to more accurately identify the potential distribution in the United States

(Ferriter 2001). One potential distribution is shown in Figure 1-3, indicating that much

can be done to limit the continued spread of this plant in the United States. Federal

noxious weed status would have major impacts on anthropogenic spread of Japanese

climbing fern, including: sale of Japanese climbing fern in the horticultural industry, sale

of spores in traditional medicine industry, interstate shipment and sale of contaminated

forest landscape products.

Economic Impact

In addition to negative environmental impacts, non-native species are causing

major economic losses in forestry, agriculture, and other segments of the U.S. economy

(Pimental 2004). Little is known about the specific economic impact of Japanese

climbing fern invasion. Some estimates place annual economic impact of forest product

loss due to invasive species as high as $2 billion in the United States (Pimental et al.









2000). Impacts of invasive plants on the forestry industry in Florida have been estimated

at $38 million per year ($15 million in weed control costs, and $23 million in yield

losses) (Lee 2005).

Economic costs. Ground-based foliar treatment of Japanese climbing fern in

natural areas can range from $45 to $450 per acre, depending on the infestation density,

cover and distribution, habitat type, site access and location, and other factors. With

multiple treatments required to achieve maintenance control, or eradication if possible,

management costs for an infested site can reach the hundreds or thousands of dollars.

Presently, the Florida pine straw industry, which generates $79 million in

landowner revenue (Hodges et al. 2004) is most directly impacted by Japanese climbing

fern invasion. While quantitative assessment of the economic impact of Japanese

climbing fern invasion on the pine straw industry in the southeast has not been

conducted, pine straw producers in Florida have noted some impacts. In at least two

cases, producers have had to abandon pine plantations leased for straw harvest due to the

dense and impenetrable nature of Japanese climbing fern invasion on site and the quantity

of economically unacceptable contaminant produced by harvesting infested plantations.

Management of this plant in pine straw plantation sites requires additional production

cost for the product reducing profits for many producers (Van Loan 2001). In addition,

Japanese climbing fern may pose a serious economic risk to pine plantations in the

Southeast through spread and intensification of fire in infested areas (Ferriter 2001).

Economic uses. Japanese climbing fern has historically been cultivated in Florida

and may have been sold widely for a thirty year period between 1888 and 1930 (Ferriter

2001). Currently it is not being commercially produced in the State (Betrock Information









Systems 2002) due to its noxious weed status. However, sale of Japanese climbing fern

continues in other states, facilitated by the internet as in the case of the Bushman Plant

Farm of Cleveland Texas (Ferriter 2001).

Japanese climbing fern spores are marketed for multiple herbal treatments under

the names spora Lygodii and the Chinese term "hai jin sha" (hai = sea, jin = gold, sha =

sand) (Ferriter 2001). Some traditional medicinal uses of spores and spore tinctures

include: promotion of diuresis, relief of stranguria (caused by urinary stones), and

treatment of eczema (Fan 1996), as well as kidney and urinary function, colds and fever

and others (Ferriter 2001). These uses are not expected to be economically significant,

but may contribute to a degree of spread of this species.

Management Practices

Integrated pest management utilizes the best appropriate combination of treatments

including: chemical, manual/mechanical, cultural, and biological control techniques.

Current control technologies for the management of Japanese climbing fern are in

development for application in both natural areas and managed properties. At present,

herbicide use appears to be the most effective control technique, but further research is

need on integration of techniques, and impact of human management practices on

anthropogenic spread of the species.

Prevention

With all invasive species, the optimal approach to successful management is

prevention of new introductions and prevention of continued spread from existing

infestations. At the national level, sale of products containing viable reproductive

Japanese climbing fern plant material should be halted, an action which requires

establishment of regulatory noxious weed status at the federal level, and throughout the









southeastern states. In addition to designation as a noxious weed at the state and federal

levels, a strategy for effective enforcement must be identified or developed. The forest

products (i.e., timber, pine straw, pine bark mulch) often harvested from infested sites are

not easily screened for the presence of Japanese climbing fern spores, but multiple cases

in North Florida have verified that in at least some cases, these products are capable of

vectoring reproductive material (Van Loan 2006). Further attention to this type of

movement will be critical to the success of attempts to manage Japanese climbing fern.

When managing sites infested with Japanese climbing fern, spore vectoring on

contaminated personnel, clothing, and equipment must be considered. Hutchinson and

Langeland (2006) evaluated Old World climbing fern gametophyte formation from

spores collected on clothing and equipment of herbicide applicators after working in three

heavily infested sites receiving an initial herbicide treatment, and one site with a very low

infestation undergoing a re-treatment. Contamination of clothing and equipment

occurred in the initial and re-treatment sites, affecting shirts, pants, sprayers, disposable

suits, chainsaws, gloves, machetes and footware. Contamination was significantly greater

from the heavily infested initial treatments sites versus the re-treatment sites. Similar

results are likely in sites infested with Japanese climbing fern, and applicators and other

land managers working in these sites should utilize the following "good spore hygiene

recommendations": spray off equipment and brush off clothing and accessories before

leaving site, wash all clothing daily, store disposable suits in plastic bags, limit any

vehicle entry into infested areas (Hutchinson and Langeland 2006).

Chemical control

In general, less herbicide testing has been completed on Japanese climbing fern

than Old World climbing fern. Initial work that has been completed indicates that









glyphosate provides a high level of control. Questions still remain, however, regarding

the efficacy of other herbicides, tank mixes, duration of control effect, effects of repeated

treatments, and integration of treatment techniques. Preliminary assessments have

resulted in a range of herbicide recommendations for Japanese climbing fern

management with varying efficacy, including foliar applications of: glyphosate,

metsulfuron, triclopyr, and imazapyr (Valenta et al. 2001, Miller 2003, Zeller and Leslie

2004). Valenta et al (2001) evaluated "low", "medium", and "high" rates of glyphosate,

triclopyr ester and triclopyr amine each in a replicated study in the panhandle of Florida.

At 315 days after treatment (DAT), the glyphosate treatments yielded greater than 70%

control (reduction in live foliar cover), while the triclopyr amine and triclopyr ester

applications showed no control. In a pair of unreplicated demonstration trials established

in north Florida for pine straw producers, Zeller and Leslie (2004) treated inter-rows in

two pine plantations, comparing broadcast and spot application techniques, using "high"

and "low" rates of glyphosate, triclopyr ester, 2,4-D+dicamba, hexazinone and

metsulfuron. When assessed in the period from 300 to 400 DAT, glyphosate broadcast

treatments yielded greater than 80% control of Japanese climbing fern, but also caused

notable damage to native understory plants. Metsulfuron broadcast treatments showed

similar high control levels at 240 DAT, but by 370 DAT, control ranged from 21% at

"low" rates to 62% at "high" rates. Damage to native understory plants was lower in

metsulfuron treatment areas than observed with glyphosate. In an attempt to begin filling

some of the knowledge gaps in Japanese climbing fern management, herbicide

evaluations were conducted as a part of this research. The results of these trials,









comparing fifteen common forestry and invasive plant herbicide treatments are reported

in Chapter 3.

Mechanical control

Mechanical methods for treatment or control of Japanese climbing fern are

considered only as components of an integrated management approach, as demonstrated

with Old World climbing fern. Observations on Old World climbing fern indicate that

vines will re-grow after both cutting of stems and hand-pulling of roots and rhizomes,

and that mechanical techniques must be used in conjunction with herbicide application to

achieve acceptable treatment efficacy. Stocker, Ferriter, Thayer, Rock, and Smith(1997)

assessed a combination of trimming and flooding on Old World climbing fern but found

no long-term control. Similar results are likely for Japanese climbing fern. Cutting

arboreal fern stems slightly above ground-level prior to herbicide application has been

used to increase efficiency of Old World climbing fern treatments in canopy-topping

infestations. Heavy equipment has been used in some situations to remove the thick

rachis mat formed by the fern's overlapping growth (Ferriter 2001). Use of heavy

equipment in climbing fern infestations raises concerns about equipment-based spore

vectoring, damage to native species, and exposure of mineral soil aiding re-colonization

of sites by Japanese climbing fern or other non-native invasive plants (Hutchinson,

Ferriter, Serbesoff-King, Langeland, and Rodgers 2006).

Biological control

Literature searches by Pemberton (1998) revealed very few known insect feeders of

Lygodium ferns. However, insect biocontrol candidates have been identified for Old

World climbing fern and may exist for Japanese climbing fern. Ongoing research efforts

are in place to identify biological control agents for Old World climbing fern with the









first approved agent, a pyralid moth, Austromusotima camptozonale (Hampson), released

in 2005 (Pemberton 2006). It is possible that biological control agents released on Old

World climbing fern will have some impact on Japanese climbing fern populations in

Florida, particularly below the frost line where populations of the two species overlap,

and host preference for Japanese climbing fern is evaluated for all screened agents. The

possibility of range overlap with the native L. palmatum affects consideration of Japanese

climbing fern classical biological control efforts.

A foliar rust fungus (Puccinia lygodii (Har.) Arth. (Uredinales)) has been isolated

from Japanese climbing fern in Louisiana, and several locations in north and central

Florida. The fungus, native to South America, causes severe damage to the leaflets,

including necrosis, browning, and drying (Rayachhetry, Pemberton, Smith, and Leahy

2001), and has been observed to apparently increase in distribution and impact in north

Florida climbing fern populations. The visual impact of this fungus on climbing fern in

north Florida has been most noticeable in early winter (i.e., November and December),

when varying levels of damage have been observed on approximately 95% of foliage in

some forest stands in north Florida. In preliminary observation, the primary impact of

this fungus in north Florida appears to be a reduction in photosynthetically active leaf

surface area immediately preceding seasonal leaf dieback. The fungus does not appear to

affect fern re-growth in the spring. Rust fungi are considered difficult to propagate in

laboratory conditions, possibly limiting further consideration of P. lygoddii as a

bioherbicide candidate. Strandberg (1999) evaluated the effects of inoculation with the

Colletotrichum acutatum H. Simmonds (fern anthracnose) fungus, but found disease

damage on Japanese climbing fern to be minimal. Jones, Rayamajhi, Pratt, and Van









(2003) evaluated the fungus Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. for

pathogenicity on both invasive climbing ferns, finding some browning and wilting of the

pinnules of Japanese climbing fern, and as much as 50% necrosis on Old World climbing

fern plants.

Prescribed fire treatment

Foresters in north central Florida have reported that prescribed fire is not effective

in controlling Japanese climbing fern in pine plantations (Ferriter 2001). The effects of

prescribed burning have not been directly evaluated on Japanese climbing fern; however,

results might be expected to be similar to those in Old World climbing fern. Prescribed

fire was evaluated both alone and in combination with herbicides for control of Old

World climbing fern by Roberts and Richardson (1994), and Stocker et al. (1997).

Roberts and Richardson (1994) concluded that fire alone cannot be used to manage Old

World climbing fern based on measurements of percent cover and height and number of

stems. Fire used in combination with a single herbicide application was ineffective as

well. Stocker et al. (1997) compared combined treatments of trimming and flooding,

contact herbicide and flooding, and burning and flooding, and found Old World climbing

fern to successfully recover following all treatments. Prescribed fire has been found to be

an effective biomass reduction tool prior to herbicide application, resulting in a reduction

in herbicide required to control Old World climbing fern (Ferriter 2001).

Haywood et al. (2001) assessed the impacts of 37 years of seasonal prescribed

burning on vegetation in a dry-mesic upland longleaf pine site in Louisiana. In

measuring occurrence of herbaceous vegetation on the study plots, Japanese climbing

fern was found to be the dominant species on 3% of the plots burned biennially in March,

and the second most common species on 3% of the plots burned biennially in July.









Japanese climbing fern was less common on plots burned biennially in May. Anecdotal

observations on the study plots indicate that regardless of bum regime, Japanese climbing

fern is increasing in presence on all plots.

Similar increases in percent cover of Japanese climbing fern on regularly burned

sites have been observed in north and central Florida (J. Wyrick, 6/15/2002, personal

communication.). Climbing fern infestation may alter fire behavior on a site by serving

as a ladder fuel to carry flames into the tree canopy. In addition, concern exists about the

possible long-distance spread of climbing fern spores in the convective columns of

prescribed fires or wildfires (Ferriter 2001). These fire-generated wind conditions can

disperse smoke great distances away from a fire (more thanl00 miles in some cases) and

may show the ability to disperse spores in the same manner.

Conclusion

While some aspects of the biology, ecology, and management of Japanese climbing

fern have been described, clear gaps do still exist. Questions remain about invasiveness

in various habitats, interactions with native plant species, spore longevity in the

environment, effects of forest management practices, anthropogenic spread, and many

more. The chapters which follow include further exploration of habitat/invasion

associations, and herbicide management in an attempt to address some of these questions.
















0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Weeks
Figure 1-1. Timeline for Japanese climbing fern plant development from spores.






Northern extent: latitude 43 degrees North j .- -"
--, ----- -.-_ --- "


\ / / ,.- .-_- '? range
B. Introduced .. "-* .-, _
range


| Southern extent: latitude 33 degrees South. Z






Figure 1-2. Worldwide distribution of Japanese climbing fern (Lygodiumjaponicum)
from literature review (June, 2006). A) Native range indicated in grey. B)
Introduced range indicated in black. Dotted line indicates northern and
southern extent of recorded Japanese climbing fern field populations.











Recorded distribution of Japanese climbing fern
In the Southeast United States, July14, 2006.
M County record = No record
.- -. Southeastern Coastal Plain




















s
I4-









Figure 1-3. County-level distribution of Japanese climbing fern in the Southeastern
United States with reference to the Southeastern Coastal Plain physiographic
region. Records based on vouchered herbarium specimens from Southeastern
herbaria, and records from regional vascular plant databases.














CHAPTER 2
ASSESSMENT OF SITE VARIABLES ASSOCIATED WITH JAPANESE CLIMBING FERN
OCCURRENCE IN THREE NORTH FLORIDA FORESTS

Introduction

A simplified view of plant invasions includes two essential stages: transport of organisms

or propagules to new sites, and establishment and population expansion in the invaded sites

(Shea and Chesson 2002). Both stages are affected at some level by the conditions of the

invaded site. On any given site or region, invader growth rate is affected by resources, the

physical environment, ecological adaptations of the invasive species, and natural enemies.

Invasive potential may be determined by environmental factors including soil pH and nutrient

levels, site climatic conditions and native plant community (Langeland and Burks 1998; Davis,

Grime, and Thompson. 2000; Shea and Chesson 2002). Thus, the mechanisms that determine

site invasibility by non-native species, in various ecosystems, are often unknown (Lavorel,

Prieur-Richard, and Grigulis 1999). Japanese climbing fern is one species for which the details

establishment success are poorly quantified. Some fern species are dependent on specific site

conditions for successful establishment and spread, including climate, soil substrate, soil pH,

elevation, and moisture regimes (Peck, Peck and Farrar 1990; Whittier and Moyroud 1993;

Stewart 2002). Establishment and persistence of gametophyte populations are strongly affected

by environmental events (Peck et al. 1990). Most ferns in the Schizaeaceae are vigorous in

edaphically low nutrient environments (Page 2002). This tolerance opens a variety of sites to

establishment, including intrinsically nutrient poor savannahs, post wildfire sites, erosion

surfaces, and epiphytic habitats (Page 2002).









While the ecological requirements of Japanese climbing fern are poorly understood, its

broad distribution indicates a tolerance for a range of environmental conditions. In its native

range, Japanese climbing fern occurs in forest edges, open forests, and secondary forests at both

lower and higher elevations (Singh and Panigrahi 1984, Gias Uddin et al. 1997, Ferriter 2001).

In the Southeast United States, Japanese climbing fern occurs primarily in floodplain forests,

marshes, wetlands, secondary woods, moist pinelands (including flatwoods), and disturbed areas

such as road shoulders and rights-of-way (Clewell 1982, Nauman, 1993, Wunderlin 1998,

Langeland and Burks 1998). Infestations also occur in xeric sites such as sandhills, but do not

appear to expand as rapidly as in more mesic sites.

Observations of land managers indicate reduced success in extremely dry or consistently

flooded sites (Ferriter 2001) and a potential preference for soils with a pH close to 7 (Nauman

1993). Most ferns flourish at pH between 6 and 7 (Hoshizaki and Moran 2001). Fluctuating

water levels may have limited impact on plant growth, but foresters in north Florida reported that

in the dry spring of 1999, Japanese climbing fern was not as serious of a problem as in previous

wetter springs (Ferriter 2001). Human disturbance does not appear to be a prerequisite for

successful invasion (Ferriter 2001), but populations are often associated with ditches and culverts

due to increased site moisture levels, and possibly due to associated soil disturbance as well. .

The objectives of this study are to classify and describe biotic and abiotic variables

associated with the presence or absence, and percent cover of Japanese climbing fern, as well as

those associated with varying levels of percent cover. Identification of clear associations between

site environmental factors and fern establishment success may aid prioritization of treatments for

managing Japanese climbing fern invasions, and improve understanding of habitat vulnerability

to fern invasion.









Methods and Materials

Site Selection

Three forested study sites were selected in North Florida (Figure 2-1), each containing

large areas infested with Japanese climbing fern, and no history of herbicide treatment in the past

10 years. Sites were selected in areas with visually "heavy" local regional infestation levels,

with the intention of reducing variability in site spore loading, and therefore in propagule-based

establishment opportunity. Sites were also selected to maximize natural community variability,

with multiple natural community types represented at each study area. Plot measurements were

taken between June and September, 2002.

Plot Layout

On each site, a square macroplot 400 m by 400 m in size (16 ha) was identified as the

study area and subjected to ecological sampling. Percent cover of Japanese climbing fern was

determined through line intercept (Myers and Shelton 1980), with intensive sampling plots

distributed in equal distances along transect lines (Figure 2-2). In each study site, a baseline was

established along a road marking a stand boundary and linear disturbance, or "edge". Five

parallel major transect lines were established, originating from the baseline at 100 m intervals,

and extending 400 m into the stand, perpendicular to the baseline.

A grid of 25 intensive sampling plots (plots), each 10 m by 10 m in size, were established

at 100 m spacing along the major transect lines. The center of each plot was permanently

marked. On each plot, five minor transect lines were established, each 10 m long and spaced 2.5

m apart. A grid of 25 sampling points (points) was established every 2.5 m along these minor

transect lines (Figure 2-2 ). Each point was marked with a pin flag. Percent cover of Japanese

climbing fern was determined for all major transect lines and minor transect lines through line

intercept measurements.









Site Descriptions

Site 1. Blackwater River State Forest, Santa Rosa County (BRSF) is a publicly-owned

78,000+ ha property managed by the Florida Division of Forestry. The study area included

natural, fire-managed upland sandhill community, with a temporal stream drainageway cutting

curvilinearly through a portion of the site (Figure 2-3). A large portion of the drainageway is

heavily infested with Chinese privet (Ligustrum sinense Lour).

Site 2. The Miller property, in Calhoun County (Miller) is a privately owned forestland

with portions of the study area in the following community types: mesic flatwood slash pine

plantation (previously managed for pinestraw production), bottomland hardwoods, longleaf pine

upland, mixed pine/hardwoods, and cypress wetlands (Figure 2-4).

Site 3. Florida River Island, in Liberty County (FRI) is a publicly-owned floodplain forest

island situated between the Apalachicola and Florida Rivers, and managed by the Northwest

Florida Water Management District. The study area is primarily cypress swamp, and floodplain

swamp hardwood forest, including areas of higher elevation on ridges and natural levees with

limited soil saturation, and low flat swales which remain inundated for much of the year (Figure

2-5) (Leitman, Sohn and Franklin. 1983).

Macroplot Measurements

Physical characteristics of each site (16 ha macroplot) were assessed, including locally

significant natural features (bodies of water, roads, etc.), and natural community boundaries;

delineated using field mapping and interpretation of aerial imagery (Figure 2-3, Figure 2-4,

Figure 2-5). Percent cover of Japanese climbing fern in the understory was measured using the

line-intercept method along the 400 meter-long major transect lines to generate an average

percent coverage by site. The site-level values were compared to average plot-level values on

each site for significant differences in measured percent cover. Lack of agreement between the









two measurements would have indicated an inadequate sampling design. The percent cover of

Japanese climbing fern was defined as: the total linear distance covered by the foliage over the

entire length of the transect lines within a site, divided by the total transect line length and

multiplied by "100" (Myers and Shelton 1980). The same measurements were taken along the

10 meter-long minor transect lines on each plot, and the same technique used to generate an

average percent coverage for each intensive sampling plot.

Intensive Sampling Plot Measurements

The majority of the classifying measurements were taken on the 10 m by 10 m intensive

sampling plots. For each plot, the following features were measured on each of the 25 sampling

points within the plot to generate a composite average value for the plot. Plot-level percent

cover of Japanese climbing fern was measured and was the dependent variable for subsequent

analyses. The species and diameter at breast height (dbh), was measured for all trees (minimum

2 inch dbh) on the 10 m by 10 m intensive sampling plots. These measurements were used to aid

in typification of site, and allow for calculation of plot basal area (m2 ha-1), species richness,

species diversity, species evenness, and species importance values.

Understory vegetation, and midstory vegetation percent cover were recorded as ocular

estimates of all plant material excluding Japanese climbing fern. Overstory percent canopy cover

was determined visually using a non-spherical densiometer. Values were recorded on a scale

with 100% being full canopy with no light penetration, and 0% being no canopy overhead.

Relative light penetration ([tmol m-2sec-1) was determined by measurements of

photosynthetically active radiation (PAR) with a Li-Cor light meter inside the stand and in the

open. Average light intensity for each intensive sampling plot was obtained by measuring PAR

just above the forest floor vegetation over the 25 sampling points for each plot, and averaging

these values to obtain a plot average. These averages were then corrected with a relative PAR









factor measured under zero canopy conditions for the respective sampling period to determine

the relative light penetration for each intensive sampling plot measured that day. Light readings

were taken between 10:00 am and 2:00 pm on days with no cloud cover to ensure uniformly

peak light conditions.

Site hydrology was described with three variables: Hydrology class (Xeric=l,

Xeric/Mesic=1.5, Mesic=2, Mesic/Hydric=2.5, Hydric=3) assigned according to site

observations, percent of the plot flooded from winter 2002 through spring 2003, and the number

of months each plot remained flooded. Plot topographic relief was ocularly assessed as concave,

convex, or flat. Plot percent slope was determined using a hypsometer. Slope aspect was

determined with a compass.

A soil core (diameter 2 cm, depth of 15 cm) was collected with a soil probe at each of the

25 points per plot to generate a composite plot soil sample. Composite soil samples were allowed

to air dry, and screened to pass a 2 mm sieve prior to analysis for: macronutrients, organic matter

content and soil pH. Mehlich-1 extractable P, K, Ca, Mg, Cu, Mn, Zn, Al, Na, and Fe were

measured using the Mehlich-1 extraction solution (0.0125 M H2SO4 and 0.05M HC1) to provide

a soil solution ratio of 1:4. Units for all elements are recorded as mg per kg. Soil pH was

measured using a 2:1 water to soil ratio. Soil organic matter percent was measured through

Walkley-Black dichromate methodology. This method was originated to assist in the selection

of herbicide rates for mineral soil containing less than 6% organic matter.

Data Analysis

Plot environmental variables and calculated indices were usedto assess relationships

between variables and the presence or absence of Japanese climbing fern across all plots. To

improve interpretation, results of these analyses are presented in four functional groups: plot

abiotic variables, plot soil variables, plot vegetation variables, and tree species importance









values. Variables were further assessed for relationships between three levels of Japanese

climbing fern percent cover across plots with Japanese climbing fern present.

Plot abiotic variables include: photosynthetically active radiation, percent slope, percent

exposed soil, plot hydrology class, plot percent flooded (2003), and months flooded (2003). Plot

soil variables include: pH, P, Ca, Fe, Mg, Al, K, Zn, Na, Mn, Cu, and percent organic matter.

Plot vegetation variables include: understory percent cover, midstory percent cover, and tree

canopy percent cover. Plot tree inventory data were utilized for the calculation of additional plot

vegetation variables: tree basal area (m2/ha), tree species richness (species per plot), Simpson's

Reciprocal (1/D) Index (Magurran 1988) of tree species diversity, and Pielou's J' Index of tree

species evenness (Pielou 1975).

Tree importance values were calculated for each species on each plot using the equations

below:

IMPORTANCE VALUE: Relative density + relative dominance + relative frequency.

DENSITY: Total stems of "species"/total plot area.

RELATIVE DENSITY: (Density of "species"/sum density of all species) X 100.

DOMINANCE: Mean basal area per tree of "species" X number of trees in species.

RELATIVE DOMINANCE: (Dominance of"species"/sum dominance of all species)
X 100.

FREQUENCY: (Plots with "species" / Total number of plots) X 100.

RELATIVE FREQUENCY: (Frequency of "species"/sum frequency of all species) X
100.

Data Transformation. Anderson-Darling test statistics were calculated to assess

normality of distributions for each environmental variable, across all study plots (n=75,

alpha=0.01). Distributions of most variables were non-normal. Non-normal data were then

transformed and normality re-assessed. Natural log, logo, and log2 transformations were









performed on all variables; a constant (value=l) was added to each variable value prior to log

transformation to allow inclusion of zero values. Additionally, inverse, and square root

transformations were performed on all variables, and arcsine transformations were performed on

variables with percentage values. Transformed data showing the strongest probability of normal

distribution for each variable were utilized in subsequent parametric analyses of environmental

data. Tree importance values were assessed using non-parametric analyses. The data subset

including only plots with Japanese climbing fern (n=29) was also assessed for normality of

distribution. In this case most variables had normal distribution of values.

Initial Analyses

Presence/absence. Initial analyses were conducted to measure the significance of

differences in plot environmental variables and tree importance values, on the presence or

absence of Japanese climbing fern. Significant differences in variable mean values between

plots with Japanese climbing fern present versus absent were assessed using the Kruskal-Wallis

one-way rank analysis of variance test (alpha=0.05) for non-normally distributed variables.

Kruskal-Wallis p-values are calculated using the chi-square approximation (Siegel 1988).

Variance equality was assessed for normally distributed plot environmental variables using the

folded F-statistic (alpha=0.05). Significant differences in normal variable mean values between

plots with Japanese climbing fern present versus absent were assessed using a pooled two-

sample t-statistic for variables with equal variance, and Satterthwaite's two-sample t-statistic for

those with unequal variances. Significant correlations between variable mean values on plots

with Japanese climbing fern present versus absent were assessed using the Spearman R

correlation coefficient (alpha=0.05) for non-normal variables, and the Pearson correlation

coefficient (alpha=0.05) for normal variables. Correlation strength ratings were assigned based

on coefficient values: weak=0.2 to 0.49, moderate=0.5 to 0.79, strong=0.8 to 1.0 (Cohen 1988).









Cover class. Separate analyses were conducted for all plots with Japanese climbing fern

present (n=29) from all sites, to assess relationships between three Japanese climbing fern cover

classes, and plot environmental variables and tree importance values. Normal distributions were

achieved with raw data for most variables, allowing use of parametric tests where appropriate

without data transformation. Cover classes utilize the following ranks: 0= 0%, 1= >0 to 5%, 2=

>5 to 25%, 3= 26 to 75%, 4= 76 to 95%, 5= 96 to 100% (Horvitz and Koop 2001). All plots fell

in cover classes 1, 2, or 3. Parametric one-way analysis of variance tests (alpha=0.05), and non-

parametric Kruskal-Wallis analysis of variance tests (alpha=0.05) were performed on all

environmental variable and tree importance values by cover class.

Discriminant Analysis

Plot environmental variables and tree species found to vary significantly according to

Japanese climbing fern presence or absence, and Japanese climbing fern cover class levels were

evaluated in stepwise discriminant analyses of covariance. Analyses were conducted on plots at

the study-level (n=75) and site-level (n=25 each). Plot variables and tree values were assessed

separately. Significance was determined by the results of the Kruskal-Wallis analysis of

variance tests and two-sample t-tests. Power for variable entry or removal in the discrimnant

analysis (alpha=0.15) was measured by the Wilk's Lambda likelihood ratio criterion (SAS

Version 9.1). Data were assessed at the study-level and the site-level.

Regression Analysis

Multiple logistic regression analysis was performed on significant variables from the

Japanese climbing fern presence/absence assessment, and the Japanese climbing fern cover class

assessment. Spearman R correlation coefficients were calculated across all variable to assess

levels of variable correlation. Principal components analysis was used to combine strongly









correlated (correlation coefficient exceeds 0.8) environmental variables into the component

factors for use in the regression analyses.

Neural Network Analysis

Artificial neural network analysis was conducted utilizing all plot variables to develop a

deterministic classification model to assess the ability of the data set to correctly classify plots

according the presence or absence of Japanese climbing fern. The model was developed as a

"C" program, and modified for use in Python. Plots were assigned "0" or "1" values according

to Japanese climbing fern absence or presence respectively; these values served as the supervised

network teacher. A non-linear activation function in the model used a radial basis transfer

function, which assumes a Gaussian curve. Network inputs, hidden layers (the radial basis

transfer function), and network outputs are connected by weight multipliers, summed by nodes.

Results

Results of respective analyses are presented in Tables 2-1 through 2-10, and appendix

tables, and Figures 2-3, 2-4, and 2-5 at the end of this chapter. To improve interpretation, results

of initial statistical tests are presented in four functional groups: plot abiotic variables, plot soil

variables, plot vegetation variables, and tree species importance values.

Site-Level Occurrence

Percent cover of Japanese climbing fern on all sites is presented in Table 2-1, including

comparisons between plot measurements taken in 2002, and re-assessed on two sites in 2004.

Among the three sites, Florida River Island (FRI) had the highest site-level and plot-level percent

cover of Japanese climbing fern in 2002 (13.7%), and the greatest number of plots infested (18),

but it could not be re-assessed in 2004 due to a herbicide treatment applied to the Japanese

climbing fern on the site in 2003. The Miller site had the second highest site-level percent cover

(3.00%), with eight plots infested in 2002, while percent cover on the Blackwater River state









forest site was lowest (1.3%), due to a small number of infested plots (3) with lower individual

percent cover levels. The number of plots with Japanese climbing fern present increased from

2002 to 2004 by one plot each on the Miller site and the Blackwater River State Forest (BRSF)

site. Plot percent cover of Japanese climbing fern increased (0.4%) from 2002 to 2004 on the

Miller site, but decreased (0.6%) on the BRSF site over the same time period. Two-sample t-

tests revealed that neither change was significant (alpha=0.05). Plot-level Japanese climbing

fern presence and percent cover values from 2002 are used in statistical analyses. Macroplot

(site-level) cover was not re-assessed

Japanese climbing fern cover classes are illustrated in Figure 2-3 (BRSF), Figure 2-4

(Miller), and Figure 2-5 (FRI). Across all sites, plots on xeric pine uplands were uninfested,

while mesic pine community types (pine plantation, mesic flatwoods, mixed pine/hardwood),

and hydric community types hydricc cypress wetland, hydric hardwood swamp, and hydric

floodplain forest: ridge, swale, and slough) each averaged a 55% Japanese climbing fern

occurrence level. Table 2-2 indicates Japanese climbing fern presence/absence levels according

to natural community type.

Plot-Level Occurrence

Presence/Absence

Plot environmental variables. Means of the following plot abiotic and biotic variables

varied significantly (alpha=0.05) according to Japanese climbing fern presence (in order of

significance): site, PAR, understory percent cover, canopy percent cover, Pielou tree evenness

index, hydrology class, Simpson's reciprocal tree diversity index, tree species richness, months

flooded, percent flooded, and trees per plot. Understory percent cover and PAR were negatively

correlated with Japanese climbing fern presence; all other significant variables showed positive

correlations. Correlation strengths were weak (Table 2-4, Tables A-i to A-4).









In particular, photosynthetically active radiation ranged from 133 kmol s -m-2 on sites with

Japanese climbing fern present to 455 kmol s -m-2 on sites without Japanese climbing fern.

Hydrology class increased in moisture level with Japanese climbing fern presence. Flood

duration averaged two months on sites with Japanese climbing fern compared to one month on

sites without Japanese climbing fern, and percent flooded increased from 15% to 27% with the

presence of Japanese climbing fern.

Means of the following plot soil variables varied significantly (alpha=0.05) in parametric

tests of relationship to Japanese climbing fern presence: pH, phosphorus (P), calcium (Ca), iron

(Fe), magnesium (Mg), aluminum (Al), potassium (K), zinc (Zn), and sodium (Na). Manganese

(Mn), copper (Cu), and percent organic matter did not vary significantly (Table A-3).

Phosphorus and pH were moderately correlated with fern presence, while other significant

variables were weakly correlated. Iron and aluminum demonstrated negative relationships with

fern presence, while other significant variables were positively correlated. Soil pH on sites with

Japanese climbing fern present averaged 6.0, while sites without Japanese climbing fern were

slightly more acidic, averaging 5.2. Soil aluminum was significantly higher on plots with

Japanese climbing fern absent (332 mg kg-1) than plots with fern present (192 mg kg-1).

Conversely, higher levels of soil calcium occurred in plots with Japanese climbing fern present

(284 mg kg-1) than absent (101 mg kg-1).

Plot tree species. Importance values were analysed by species, comparing mean values

for plots with Japanese climbing fern present versus absent. Forty-three tree species were

identified across all plots, of which twenty-one occurred on one plot only. Species with

importance values that varied significantly according to Japanese climbing fern presence or

absence include: Liquidambar styraciflua, Pinus palustris, Carya aquatica, Quercus lyrata,









Ulmus americana, Carpinus caroliniana, Quercus nigra, and Taxodium distichum (Table 2-7,

Table A-4), while those not exhibiting a significant relationship are listed in Table A-5. Mean

importance values for the majority of significant tree species demonstrate a positive relationship

with Japanese climbing fern presence; the one exception being P. palustris which demonstrates a

moderate negative relationship. Other significant species are, at best, weakly correlated with

Japanese climbing fern presence.

Tree species were assigned vegetative index classes according to the Florida Department of

Environmental Protection's Wetland Delineation Index (Sec.62-340.450, F.A.C.) (FLDEP 2006).

Analyses were conducted to assess relationships between tree importance values within

vegetative index classes and Japanese climbing fern presence or absence (Table 2-3). Tree

importance values on plots with Japanese climbing fern present versus absent were compared,

significant differences between Japanese climbing fern presence and absence in importance

values for facultative upland trees, and obligate wetland trees. Analysis of variance detected

significant (P=0.005) variation in importance values between vegetative index classes on plots

with Japanese climbing fern present, as well as on plots with Japanese climbing fern absent

(P<0.0001).

Cover Class Levels

Plots with Japanese climbing fern present were analyzed as a subgroup to assess

relationships between three Japanese climbing fern percent cover levels (raw data or cover class),

and plot environmental variables and tree importance values. No significant (alpha=0.05)

relationships were detected between plot and tree variables and Japanese climbing fern cover

class in parametric one way analysis of variance. However, significant (alpha=0.05)

nonparametric relationships were detected between percent cover of Japanese climbing fern and

three variables, including plot basal area which peaked (30 m2 ha-1) on moderately infested (class









2) plots. Soil aluminum reached 249 mg kg-1 on moderately infested plots, an average of 70 mg

kg-1 more than on slightly (class 1) or heavily (class 3) infested plots. Occurrence of Carpinus

caroliniana also showed significant relationships to Japanese climbing fern cover class. (Table

2-4). Significant (alpha=0.05) correlations to Japanese climbing fern cover class were measured

for basal area, plot percent flooded, photosynthetically active radiation, and the occurrence of

Carpinus caroliniana and Quercus nigra (Table 2-4).

Discriminant Analysis

Stepwise discriminant analysis of covariance was used to further explore relationships

between plot variables and Japanese climbing fern occurrence, and cover class. Variables were

measured for relationships at the study (n=75) level, and at each respective site (n=25). Across

all study plots, soil calcium was the most significant discriminator (P<0.0001) of Japanese

climbing fern presence or absence (Table 2-5). Site level analyses revealed relationships

between other variables including percent flooded (P=0.02), and months flooded (P=0.0001) at

BRSF; while soil iron is significant (P=0.002) at the Miller site, and tree species richness is

significant (P=0.03) at FRI. (Tables 2-5 and 2-6) Of the variables that discriminated between

Japanese climbing fern presence or absence (Table 2-5), percent flooded is the only one also

significant (P=0.09) as a discriminator (alpha=0.10) between Japanese climbing fern cover class

levels (Table 2-6). In addition, soil aluminum and plot basal area were also both significant,

discriminators of Japanese climbing fern cover class at the study level. Site level analyses

revealed the strongest relationship between variables and Japanese climbing fern cover class, to

be soil aluminum at the Miller site (P=0.01), followed by percent flooded at the Miller site

(P=0.14). At FRI, basal area and soil aluminum were significant (alpha=0.15) discriminators of

Japanese climbing fern cover class.









Quercus nigra was a significant discriminator between Japanese climbing fern presence

and absence at the study level (n=75, P<0.0001) and at BRSF (n=25, P<0.0001) (Table 2-7); as

well as between cover class levels at Florida River Island (Table 2-8). Other significant tree

species were discriminators of Japanese climbing fern presence or absence, such as: Pinus

palustris (P<0.0001) and Liquidambar styraciflua (P=0.0005) at the study level, Ligustrum

sinense at BRSF (P<0.0001), Pinus elliotii at the Miller site (P=0.02), and Nyssa aquatica

(P=0.02) and Acer rubrum (P=0.08) at FRI. Carpinus caroliniana was the only significant

discriminator of cover class (P=0.02) at the study level. Discriminant analysis was not possible

for BRSF due to the small number of plots with Japanese climbing fern present; and no tree

species were accepted into the analysis for the Miller site.

Regression Analysis

Multiple logistic regression analysis was performed on significant variables from the

Japanese climbing fern presence/absence and cover class assessments. Principal components

analysis was used to combine strongly correlated (correlation coefficient exceeds 0.8)

environmental variables into the following factors: FLOOD=Hydrology class + plot percent

flooded + months flooded, SOIL1=pH (inverse) + phosphorus (inverse) + calcium (natural log),

SOIL2=Zn + K (natural log), TREE=Pielou species evenness index + Simpson's reciprocal

species diversity index + species richness. Principal components analysis explained more than

80% of the variance among factor components for all factors, indicating substantial agreement,

and validating use of factors in subsequent multiple regression analysis.

Variables generating significant effects in the regression analyses are listed in order of

forward selection into the regression models in Tables 2-9 (presence/absence) and 2-10 (cover

class). The first entry into the presence/absence model is an intercept (value=0.4), followed by

the principal component factor SOIL1, accounting for 40% of the variation in the model.









Following general site classification by those soil pH correlates, the additional cumulative effect

of importance values of Ligustrum sinense, Quercus nigra, Lindera benzoin, Acer rubrum,

Quercus laevis, and Nyssa aquatica accounting for a maximum of 68% of the model variation

related to presence or absence of Japanese climbing fern. When the presence/absence analysis

was conducted for each respective site, few variables were entered into the model. At BRSF, the

principal components analysis factor TREE was entered, accounting for a cumulative 49% of

model variation. At the Miller site, the SOIL1 factor entered the model, followed by Pinus

palustris, accounting for 62% of cumulative model variation. At FRI, no variables were entered.

A similar number of variables produced significant effects in analysis of Japanese climbing

fern cover class levels. The first entries into the cover class model were an intercept (value=9.2),

followed by importance values for Carya aquatica, accounting for a cumulative 17% of model

variation. A significant increase in model strength came with additional of the principal

components analysis factor FLOOD1, raising model accuracy to 38%. Six tree species: Catalpa

bignoniodes, Platanus occidentalis, Nyssa ogeechee, and Quercus laevis were accepted in the

regression analysis, accounting for a total cumulative model variation of 75%. When the cover

class analysis was conducted for each respective site, the only variable entered into the model

was FLOOD 1 at the FRI site. No variables were entered for BRSF, or the Miller site.

Neural Network Analysis

The neural network analysis selected thirty plot variables and tree importance values for

input into the model. Variables positively related to Japanese climbing fern presence are

indicated with a (+), while variables positively related to absence are indicted with (-). In order

of rank in the model, they include: understory percent cover (-), position (-), Pinuspalustris

importance (-), Quercus nigra importance (+), Ligustrum sinense importance (+), soil iron (-),

percent exposed soil (-), basal area (+), Quercusfalcata importance (-), Nyssa sylvatica









importance (-), percent slope (+), soil phosphorus (+), percent canopy cover (+), site, soil

aluminum (-), soil calcium (+), hydrology class (+), soil magnesium, Bumelia lycoides

importance (-), Quercus stellata importance (-), Liquidambar styraciflua importance (+), soil pH

(+), Quercus laevis importance (-), Taxodium distichum importance (+), soil sodium (+), soil

potassium (+), midstory percent cover (-), months flooded (2003) (+), relief (-), and soil organic

matter percent (+). Utilizing these variables, the neural network correctly classified Japanese

climbing fern presence for 22 of 29 plots, or 76% of the time; Japanese climbing fern absence

was correctly classified for 38 of 46 plots, or 83% of the time.

Discussion

In field observations, some trends in the presence and percent cover of Japanese climbing

fern are observable. Identification and quantification of the specific causal agents of these trends

is experimentally challenging, but some general relationships are discernible from the results of

this study. Some aspects of study design, primarily a limited number of plots study-wide; and

the macroplot sampling design restrict application of most results beyond the study sites

themselves. The following discussion begins with relationships within three main variable

categories: site overview, plot environmental variables, and tree species. The final two sections,

multiple logistic regression, and neural network analysis, focus on relationships across variable

categories, seeking the most valid or predictive combination of variables.

Twenty plot environmental variables and eight tree species demonstrated significant

variance in mean values related to the presence or absence of Japanese climbing fern (Tables: A-

1, A-2, A-3, A-4) in initial analyses. The majority of these significant variables and tree species

either affect or indicate aspects of site hydrology (and associated soil chemistry) and

photosynthetic opportunity (e.g., hydrology class, percent flooded, photosynthetically active

radiation). This reflects general fact that in Florida, hydrology, pH and fire exert overwhelming









influence over site vegetation, minimizing effects of other factors; including soil nutrient supply

(Brown, Stone, and Carlisle 1990). The sensitivity of statistical tests used in each analysis should

be considered in interpretation of all results, particularly in consideration of forested sites outside

the local region, or with different natural community types represented.

Site Overview

Percent cover of Japanese climbing fern varied significantly among sites (Table 2-1). The

general grouping of natural community types by site affects a broad range of variables, including

Japanese climbing fern occurrence. The primarily hydric Florida River Island (FRI) site had the

highest site-level and plot-level percent cover of Japanese climbing fern. Low areas and swales

on this site flood annually with the elevation of water levels in the winter and spring like the

majority of floodplain and bottomland hardwood forests along the Apalachicola River. Under

these conditions, soil saturation is the primary limiting factor affecting climbing fern

establishment and survival. Higher elevation areas in this forest type can become heavily

infested with Japanese climbing fern as a result of their generally mesic to hydric nature, and

extensive flood-based dispersal of spores and on site via the annual floods (Figure 2-5). In

addition to site characteristics, regional invasion history may also contribute to higher presence

and percent cover of Japanese climbing fern for sites in the Apalachicola River basin (BRSF and

Miller site), where the plant has been established for several decades. The xeric to mesic

Blackwater River State Forest had the lowest site-level percent cover of Japanese climbing fern

(Table 2-1). This reduced site-level occurrence is primarily determined by the spatially limited

nature of sites with appropriate hydrology (Figure 2-3), although Japanese climbing fern is

increasing in occurrence in both longleaf pine (Pinuspalustris) and slash pine (Pinus elliottii)

forests on the site. The majority of Japanese climbing fern on the mesic to hydric Miller site









occurred in the pine plantation area and ecotonal areas between upland and lower swamp

hardwoods on the eastern portion of the site (Figure 2-4, Table 2-2).

The number of plots with Japanese climbing fern present increased by one on both the

Blackwater River State Forest and Miller sites from 2002 to 2004. Foresters and land managers

throughout Florida have also noted visually obvious increases in Japanese climbing fern cover

and distribution throughout the north Florida region over the last decade (Ferriter 2001). This

trend is likely to continue as local and regional propagule pressure from established plants and

populations increases, particularly given the limited nature of environmental restrictions on

climbing fern occurrence. Plot-level percent cover of Japanese climbing fern increased by 0.4%

from 2002 to 2004 at the Miller site, but decreased by 0.6% at Blackwater River State Forest.

Neither change was significant. The decrease in plot percent cover at BRSF may reflect the

effect of visually high level of foliar damage caused by the rust fungus Puccinia lygoddii in

2004, coupled with a small number of plots with hydrology appropriate for Japanese climbing

fern. The increase in plot-level occurrence while also not significant, does document the

continued expansion of this species on infested sites, either through natural dispersal, or possibly

through human-vectored spore dispersal as a result of repeated entry into the plots by

researchers.

Discriminant Analysis

Plot environmental variables

Stepwise discriminant analyses of plot variables (alpha=0.10) furthered support for the role

of general site hydrology in determining occurrence of Japanese climbing fern. Specifically, soil

calcium, and soil iron, plot percent flooded, and plot understory percent cover discriminated

between Japanese climbing fern presence and absence at the study level, maintaining support for

the importance of site mesic to hydric hydrology in climbing fern establishment. Clawson,









Lockaby and Rummer (2001) found soil calcium to increase as flooding levels increased on soils

ranging from somewhat poorly drained to poorly drained within a floodplain forest. In addition

to hydrologic variables, site-level analysis identified the tree evenness index as a significant

indicator at Blackwater River State Forest, probably due to the naturally low tree evenness in the

xeric upland longleaf pine community dominating this site (Figure 2-3). As a result, the

significance of evenenss in this analysis probably reflects site hydrology and natural community

type, rather than indicating a strong relationship between climbing fern occurrence and

eveneness. This is further supported by the high occurrence of Japanese climbing fern on the

Pinus elliottii plots at Miller, which also had low evenness values. Tree species richness was

significant at Florida River Island, primarily due to the low richness measured on swale plots

(Figure 2-4) which remain inundated for a large part of each year, thereby reducing vegetative

establishment opportunity for many species, including Japanese climbing fern.

Discriminant analysis of plots with Japanese climbing fern for relationships with varying

cover class levels identified three plot variables as discriminators. Basal area was significant at

the study level, driven by the polynomial relationship to Japanese climbing fern cover classes at

Florida River Island, where basal area decreases significantly as Japanese climbing fern percent

cover exceeds 25%. Plot percent flooded displayed a negative discriminatory relationship with

Japanese climbing fern cover, indicating that while adequate moisture is necessary for

establishment, excessive moisture (and shortened growing season) will limit percent cover.

Plot tree species

Tree species showing significant relationships to Japanese climbing fern presence (Table

A-1) are primarily classified as facultative wetland or obligate wetland species according to the

vegetative index used for wetland delineation in Florida (Table 2-3 and Table A-4) (FLDEP

2006). Facultative upland species and facultative species were significantly less important than









facultative wetland and obligate wetland species on sites with Japanese climbing fern present,

indicating an association with mesic to hydric site conditions. In comparison, no significant

difference was measured between vegetative indices on plots with Japanese climbing fern absent,

indicating that absence is affected by more than vegetative index of dominant species (e.g.,

establishment opportunity).

Stepwise discriminant analysis of tree species importance values identified Quercus nigra,

Liquidambar styraciflua, and Taxodium distichum, as significant discriminators of Japanese

climbing fern presence. Conversely, Nyssa sylvatica was a significant discriminator of Japanese

climbing fern absence at the study level. Site-level analyses also identified the invasive plant

Ligustrum sinense as a discriminator of Japanese climbing fern presence at Blackwater River

State Forest. The relationship shown between these species indicates two key factors which are

easily recognizable in the field: the hydrologic function of the extensive natural stream

drainageways on the site provide a moisture regime allowing for establishment of Japanese

climbing fern and L. sinense across the site landscape; and movement of water through the

system of drainageways on the site has probably played an important role in vectoring

propagules of both species. The natural function and management of these sites are both

complicated by the co-occurring populations of these species, as well as an additional non-native

invasive plant, Lonicerajaponica. Removal of the often dense canopy and/or midstory of L.

sinense has resulted in rapid expansion of Japanese climbing fern in multiple cases. In addition

to the documented relationship at Blackwater River State Forest, the pine plantation plots at the

Miller site also had co-occuring infestations of Japanese climbing fern and L. sinense, though

plants were too small to census with plot trees









Pinus elliottii is a significant discriminator for the presence of Japanese climbing fern on

the Miller site, as a result of the P. elliottii plantation covering the western portion of the study

area (Figure 2-4). This site bias does reflect a significant regional relationship between P.

elliottii plantations and Japanese climbing fern. High levels of Japanese climbing fern invasion

are clearly visible throughout mesic pine plantations in the Apalachicola River basin in Florida.

One study documented a 22% Japanese climbing fern occurrence level in North Florida P.

elliottii plantations, a forest type covering 5.1 million acres in the State (Van Loan 2006).

Silvicultural management practices on these sites may be contributing to Japanese climbing fern

dispersal and site occupancy levels as a result of canopy removal, soil disturbance, and

equipment-based spore dispersal. Significant increases have been seen in fern coverage and

density on flatwoods and mesic plantation forest types following timber harvest, pinestraw

harvest, and prescribed burning in particular.

Discriminant analysis of plots with Japanese climbing fern for relationships with varying

cover class levels identified four tree species as discriminators. Increasing importance of

Carpinus caroliniana was significantly related to increasing occurrence of Japanese climbing

fern. A common associate of Q. nigra, Carpinus caroliniana is considered to demonstrate best

growth and development on rich wet-mesic sites (Bums and Honkala 1965), further supporting

the observations of the importance of general site hydrology in affecting Japanese climbing fern

establishment and spread on infested sites.

Regression Analysis

The multiple regression analysis further supported the importance of mesic to hydric

conditions for Japanese climbing fern establishment and spread on a site. The SOIL1 factor

(Tables 2-9 and 2-10), reflecting combined associations between soil phosphorus, soil calcium,

and pH, varies in association with site hydrology, particularly flooding. Grouping the results of









the discriminant and multiple regression analyses reveals five soil variables (calcium, iron,

potassium, phosphorus, and pH) associated with presence or absence of Japanese climbing fern,

and one soil variable (aluminum) associated with cover class on infested sites. Each of the soil

variables associated with the presence or absence of Japanese climbing fern varies in response to

multiple factors of parent material, hydrology, moisture source, and habitat inputs. Taken as a

group, calcium, phosphorus, pH, and potassium levels in this study were significantly higher on

plots with Japanese climbing fern present, typically the mesic to hydric sites, as opposed to

absent. In a simplified comparison, iron and aluminum in this study were significantly lower

overall on plots with Japanese climbing fern present; however, analysis of fern presence or

absence at the site level provided an expanded perspective.

For many soil variables, a separation existed between values measured at FRI versus the

other two sites. Calcium, phosphorus, and pH were all significantly higher at FRI compared with

BRSF and Miller, while iron was significantly lower at FRI. Iron availability in solution is

influenced by soil pH, aeration, and reactions with organic matter. In particular, iron decreases

as pH increases, reaching a minimum at pH 7.4 to 8.5. Additionally, poor soil aeration (reduced

oxygen levels) caused by flooding can prompt oxidation and/or reduction of iron when

influenced by other conditions (Dixon and Schulze 2002). The saturation of soils on FRI reduces

soil iron, increasing its solubility, and allowing its movement out of the surface soil layer (Dixon

and Schulze 2002). This effect is different from the brief, infrequent periods of saturation in the

upland drainageway at BRSF, which produced higher iron concentrations, probably due to

moisture-induced oxidation (Dixon and Schulze 2002). However, the significantly higher site-

level occurrence of Japanese climbing fern at FRI is probably linked more to the general site









hydrology which prompts the soil variable trends, rather than being associated with the

individual soil variables themselves.

The significance of soil aluminum is distinguishing between Japanese climbing fern cover

classes on infested plots was an artifact of two situations. Aluminum values on the infested

BRSF upland drainageway plots and the mesic Miller plots were significantly higher than FRI

values, reflecting the effect of lower pH and extreme weathering in the more sandy soils seen at

these sites. Low pH values have been shown to prompt a peak in available soil aluminum as

seen in sites which repeatedly flood, and then drain (Darke and Walbridge 2000). The addition

of the higher values from the Miller and BRSF plots creates the polynomial relationship

observed between Japanese climbing fern cover class and soil aluminum on infested plots.

Aspects of site hydrology heavily influence fluctuation in soil conditions and components.

Aluminum and iron are known to be important in controlling the retention of dissolved inorganic

phosphorus in wetland soils, providing further support for positive relationships between those

three variables (Darke and Walbridge 2000). Darke and Walbridge (2000), and Clawson et al.

(2000) recorded increasing levels of soil calcium in soils from a gradient of sites ranging from

upland to floodplain swale, explaining the trends in calcium observed, particularly at the site

level on Miller and BRSF, both sites including a mixture of pine dominated habitat and

hardwood dominated habitat. Unlike FRI which is dominated by hardwoods and slightly acidic

to neutral soils. Soil pH varied significantly by site due primarily to hydrology and natural

community; with higher pH values (5.5 to 7.2) on the hydric FRI hardwood floodplain forest site,

and significantly lower pH values (4.6 to 5.9) on mesic Miller site, followed by the xeric BRSF

sandhill site (4.6 to 5.0), a trend seen in these habitat types statewide (Londo, Kusha, and Carter

2006, Platt and Schwartz 1990).









As in the discriminant analysis, the effects of Quercus nigra and Carya aquatica

influenced Japanese climbing fern occurrence. These two species, together with Ligustrum

sinense may be used as general indicators of site conditions appropriate for Japanese climbing

fern occurrence in northwest Florida. However, primary application of this result should be

restricted to similar natural community types on the same managed areas as utilized in the study.

Several years of unquantified observations support the finding that Japanese climbing fern

presence may be strongly correlated with the presence of L. sinense on a landscape-scale in

North Florida. This tendency toward co-location may also extend into Georgia, where L. sinense

is widely established in upland drainageways, bottomland hardwood and floodplain forests, sites

with higher levels of Japanese climbing fern in North Florida.

Neural Network Analysis

Artificial neural networks are attempts to predict complex, non-linear relationships based

on a loose analog with living nervous systems. These networks have been used to classify

samples according to discrete classes (Mehrotra, Mohan and Ranka 2000), in this case the

presence or absence of Japanese climbing fern. Though the network performed relatively well in

classification of plots with climbing fern present and absent, the number of variables included

(30) indicates overall model weakness in a similar sense to loss of degrees of freedom in

regression analysis. However, the predictive performance does allow for some interpretation if

primarily for the managed areas on which study sites were located.

While, the model primarily selected variables which had demonstrated significant

relationships with Japanese climbing fern presence or absence in previous analyses, it selected

some new variables, including position, relief, midstory percent cover, and upland tree species

such as Quercus stellata and Quercusfalcata. The addition of these factors in the model

network may indicate non-linear relationships that standard techniques cannot correctly model









due to non-collinearity or other statistical assumptions in these models. It is also likely that some

factors are only useful for this particular dataset, due the limited sample size.

However, all of the data available were used in the fitting of this model, leaving none for

model validation. The basic model output does provide support for the initial study hypothesis

that a combination of site variables may be identified that affect the presence or absence of

Japanese climbing fern on a site. However, the model itself serves only as a hypothesis without

independent validation. If future research is planned on classification of sites vulnerable to

Japanese climbing fern invasion, a recommendation for study design is to use a larger number of

plots on which a smaller number of variables measured may enable use of more robust analyses,

provide adequate data for model validation, and reduce some amount of correlation between

variables. Data gathered in any future site classification studies may be useful in validation of

this model, or construction of others.

Conclusions

The reproductive strategy of Japanese climbing fern plays a significant role in the

establishment and spread of this species. Baker's Rule states that species that are capable of

uniparental reproduction are more likely to be successful invaders, (Richardson 2004). The

results of Foxcroft, Richardson, Rouget and MacFayden (2004) suggest that propagule pressure

can act as a fundamental driver of invasions, significantly reducing the importance of

environmental limitations (Volin et al. 2004). This situation may at least partially explain the

lack of strong results from this study.

However, some relationships between site characteristics and Japanese climbing fern

invasion do exist, the most influential of which is general site hydrology. Hydrology has a major

role in determination of natural communities and subsequent site environment and vegetation. In

particular, flooding extent and duration have several impacts on vegetative communities,









affecting growing season duration and establishment opportunity, and soil biogeochemical

processes.

A similar assessment examined the correlation between the distribution of Old World

climbing fern and abiotic factors in two forested wetlands in South Florida (Volin et al. 2004). In

that study, Old World climbing fern occurred on 32% (n=43) of the sampling units, a similar

ratio to this study in which Japanese climbing fern occurred on 39% (n=29) of the sampling units

(plots). Differences in the sampling design and availability of existing distribution data

permitted use of ordination analyses not appropriate for this data set (Volin et al. 2004).

However, general conclusions were similar to those in this study. Similar to Japanese climbing

fern in North Florida, presence of Old World climbing fern was found to be significantly

correlated with site hydrology, specifically a wet but not permanently inundated condition.

Future studies of Japanese climbing fern invasion establishment ecology should be focused

in the northern extent of it's range in the southeastern United States. If propagule pressure is the

primary determinant of the invasive potential of this species, the majority of mesic and many

hydric sites in Florida may already be vulnerable to Japanese climbing fern invasion. Rather

than focus heavily on site vulnerability, a critical next step in study of the ecology of this species

is to quantify the effects of invasion on habitat quality and function, and native species diversity.

The answers to these questions will be key in understanding the impact of Japanese climbing

fern invasions and accurately determining the true priority of aggressive prevention and control

actions.




































Legend
Major rivers
Counties
A Study siles

N
0 37.5 75 150 225 300 W


Figure 2-1. Location of study sites used in 2002 Japanese climbing fern site establishment
assessment, Florida, USA.










I-400 meters


Baseline (road)


A
Figure 2-2. Sampling design for assessment of Japanese climbing fern distribution on three
forest sites in North Florida. A) Layout of major transect lines and intensive
sampling plots. B) Layout of sampling points within each intensive sampling plot












BlacKwater River state Forest, Santa Rosa county, FL


I


Legend: plot
Lygaolum cover
L1i No Lygoium
1W 1-5% cover
[16-25% cover
[1 26-75% cover


Figure 2-3. Plot layout indicating 2004 Japanese climbing fern cover classes and natural
community types, over 2000 aerial imagery, Blackwater River State Forest (BRSF),
Santa Rosa County, Florida.




























Legend: plot
Lyn dowum cover
PiJ No Lygodium
Il 1-5% cover
16-25% cover
[] 26-75% cover


Figure 2-4. Plot layout indicating 2004 Japanese climbi
community types, over 2000 aerial imagery,
County, Florida.


property (Miller), Calhoun






























C t Legend: plot
Lygpduum cover
1 al] 0 No Lygomum
l 1-s% cover
a 6.-25% cover
-3 [ 26-75% cover

Figure 2-5. Plot layout indicating 2002 Japanese climbing fern cover classes and natural
community types, over 2004 aerial imagery, Florida River Island (FRI), Liberty
County, Florida.










Table 2-1. Japanese climbing fern (Lygodiumjaponicum) occurrence on three North
Florida forest sites, 2002 and 2004.
Site = BRSF Site = Miller Site = FRI
2002 2004 2002 2004 2002 2004
Average site % Lygodium 1% NAa 3% NA 14% NA
cover (major transects).
Total number of plotsb with 3 4 8 9 18 NA
Lygodium present.
Percent of plots with 12% 16% 32% 36% 72% NA
Lygodium present
Average plot % Lygodium 3% 2% 5% 6% 14% NA
cover.
Change in Lygodium plot % -0.6% +0.4% NA
cover 2002-2004.
aNA indicates measurement not available.
bTwenty-five plots per site.


Table 2-2. Japanese climbing fern (Lygodiumjaponicum) occurrence across eight
natural community types in three North Florida forest sites, 2004. Twenty-
five plots assessed per site.
Site = BRSF Site = Miller Site = FRI Site = All
Plot Plot Plot Plot Plot Plot Plot Plot
Natural community ratio %a ratio % ratio %a ratio %
Xeric Dine uplands 0/19 0%


Xeric plot total
Mesic pine plantation
Mesic pine flatwoods
Mesic pine/hardwood
Mesic plot total
Hydric floodplain ridge
Hydric floodplain swale
Hydric cypress wetland
Hydric hardwood swamp
Hydric plot total


0/19 0%


8/11 73%
0/5 0%


4/6 67%


12/22 55%


0%
20%


10/10
6/10
2/5


100%
60%
40%


16/34 47%


aplot ratio and plot % indicate the number of plots with Lygodium present to total number of
plots of the community type.









Table 2-3. Comparison of tree importance values (IV) by species vegetative index and presence or absence of Japanese
climbing fern (Lygodium japonicum).


Vegetative index
Facultative upland
Facultative
Facultative wetland
Obligate wetland
Facultative upland
Facultative
Facultative wetland
Obligate wetland


Number of
species
5
13
12
13
5
13
12
13


Lygodium
status
Present
Present
Present
Present
Absent
Absent
Absent
Absent


Mean
IV
7
3
10
8
25
2
7
5


Std. error
IV
3
3
4
4
18


T-valuesb
(present vs. absent)
4.2 (P<0.0001)*
-1.3 (P=0.20)
-1.67 (P=0.1)
-2.56 (P=0.01)*


ANOVA
statistic
5.9 (P=0.0005)*




44.9 (P<0.0001)*


a Vegetative index of tree species hydrologic associations.
b Test statistic and p-values for comparison of mean importance values by Lygodium status (present versus absent) for each
vegetative index, Two sample t-test, alpha = 0.05, SAS Version 9.1. Asterisk* indicates significant value.
c Test statistic and p-values for comparison of tree importance values between vegetative indices for each Lygodium status
(present and absent), one-way analysis of variance, alpha=0.05, SAS Version 9.1. Asterisk* indicates significant value.









Table 2-4. Relationships between Japanese climbing fern (Lygodiumjaponicum) percent cover levels and significant
plot variables and tree importance values in three North Florida forest sites.
Lygodium Number Std. Kruskal-Wallis Correlation Correlation
Variable cover class of plots Mean error values coefficients strengthd
Basal area (m2 ha-) 1 10 20 6 9.62 (P=0.01)* -0.39 (P=0.04)* Weak
2 9 30 2
3 10 11 1
Soil Al (mg kg-1) 1 10 178 27 6.62 (P=0.04)*
2 9 249 44
3 10 154 16
Carpinus 1 10 2 0 6.12 (P=0.05)* 0.38 (P=0.04)* Weak
caroliniana 2 9 2 0
3 10 14 6
Percent flooded 1 10 45 15 -0.42 (P=0.03)* Weak
2 9 29 14
3 10 8 3
Photosynthetically 1 10 92 29 0.40 (P=0.03)* Weak
active radiation 2 9 107 31
([tmol m-2 S-1) 3 10 199 93
Quercusnigra 1 10 10 3 0.39 (P=0.04)** Weak
2 9 17 6
3 10 28 7
ajapanese climbing fern cover class levels: 1=>0 to 5%, 2= 5.1 to 25%, 3= 25.1 to 75%.
b Asterisk* indicates significant value Kruskal-Wallis analysis of variance test and p-values alpha = 0.05.
' Asterisk* indicates significant value, Pearson's correlation coefficient and p-values, alpha=0.05, SAS Version 9.1.
d Correlation strength: weak=0.2-0.49, moderate=0.5-0.79, strong=0.8-1.0









Table 2-5. Plot environmental variables that discriminate between Japanese climbing fern (Lygodiumjaponicum) presence or
absence in three North Florida forest sites.
All plotsb Site = BRSFb Site = Millerb Site FRIb
Variablea R2 F-value P>F R2 F-value P>F R2 F-value P>F R2 F-value P>F
Soil Ca 0.31 32.72 P<0.0001 0.28 7.25 P=0.01 0.16 4.21 P=0.05
(natural log)
Percent flooded 0.05 2.47 P=0.07 0.23 5.93 P=0.02 0.13 3.30 P=0.08
Soil Fe (inverse) 0.06 4.72 P=0.03 0.35 12.48 P=0.002
Understory 0.06 4.48 P=0.04
(% cover)
Tree evenness 0.18 4.77 P=0.04
index
Hydrology class 0.17 4.21 P=0.05
Months flooded 0.45 15.34 P=0.001
Species richness 0.20 5.65 P=0.03
Soil K 0.14 3.40 P=0.08
(natural log)
aData transformation indicated in parentheses if used.
bTest statistics from stepwise discriminant analysis of covariance of plot environmental variables, SAS Version 9.1. Variables
listed are significant (alpha=0.15) at either the study-level or site-level as indicated.

Table 2-6. Plot environmental variables that discriminate between Japanese climbing fern (Lygodiumjaponicum)
percent cover class levels in three North Florida forest sites.
All plots Site =BRSFb Site = Millera Site = FRIa
Variable R2 F-value P>F R2 F-value P>F R2 F-value P>F


Basal area
Soil Al
Percent flooded


0.20
0.19
0.18


3.23
2.92
2.61


0.05
0.07
0.09


0.82 11.36
0.62 3.31


a Test statistics from stepwise discriminant analysis of covariance of plot environmental variables, SAS Version 9.1.
Variables listed are significant (alpha=0.15) at either the study-level or site-level as indicated.
b Sample size for site=BRSF too small to conduct analysis.


0.01
0.14


0.24 2.26
0.25 2.52


0.11
0.14









Table 2-7. Tree species importance values that discriminate between presence and absence of Japanese climbing fern (Lygodium
japonicum) presence or absence in three North Florida forest sites.
All plotsa Site = BRSFa Site = Millera Site = FRIa
Partial F Partial F Partial F Partial F
Variable R2 value P>F R2 value P>F R2 value P>F R2 value P>F
Pinus palustris 0.21 18.9 <0.0001
Quercus nigra 0.08 6.6 <0.0001 0.94 3.7 <0.0001
Liquidambar 0.10 8.3 0.01
styraciflua
Taxodium distichum 0.07 5.6 0.02
Quercus 0.05 3.3 0.07
hemisphaerica
Lindera benzoin 0.05 3.3 0.07
Nyssa sylvatica 0.05 3.3 0.07
Ligustrum sinense 0.91 2.2 <0.0001
Pinus elliottii 0.23 6.9 0.02
Acer rubrum 0.01 3.5 0.08
Nyssa aquatica 0.21 5.9 0.02
aTest statistics, stepwise discriminant analysis of covariance of tree importance values. Species listed are significant (alpha=0.10).

Table 2-8. Tree species importance values that discriminate between Japanese climbing fern (Lygodiumjaponicum)
percent cover class levels on plots with fern present in three North Florida forest sites.
All plots Site = BRSFb Site = Millerb Site = FRIa
Variable R2 F-value P>F R2 F-value P>F R2 F-value P>F
Carpinus caroliniana 0.18 2.9 0.07
Quercus nigra 0.29 3.07 0.08
Carya aquatica 0.32 3.02 0.08
a Test statistics from stepwise discriminant analysis of covariance of plot environmental variables. Species listed are
significant (alpha=0.10).
b Sample size for BRSF too small to conduct analysis, no species selected for Miller site..









Table 2-9. Effects of significant plot environmental and tree importance values in
multiple logistic regression for relationship to presence or absence of
Japanese climbing fern (Lygodiumjaponicum) in three North Florida
forest sites.
Area Effect Estimate R-squared Chi square P>Chi
(in order of entry)a (cumulative) score square
All plots Intercept 0.40
SOIL1 1.79 0.40 24.0 P<0.0001
Ligustrum sinense 0.29 0.46 11.4 P=0.0008
Quercus nigra -0.09 0.53 5.7 P=0.02
Lindera benzoin -0.17 0.59 9.4 P=0.002
Acer rubrum 0.03 0.62 4.4 P=0.04
Quercus laevis 0.01 0.65 3.9 P=0.05
Nyssa aquatica 0.75 0.68 4.5 P=0.03

BRSF Intercept -1.46
TREE 1.80 0.49 7.85 P=0.005

Miller Intercept -6.39
SOIL1 -1.88 0.42 7.93 P=0.005
Pinus palustris 0.02 0.64 4.99 P=0.03

FRI No variables entered

a Significant effects in model (SAS Version 9.1).. Principal components analysis
factors: SOIL1= soil P, soil Ca, and soil pH; TREE=tree species richness, evenness,
and diversity. Model validity is questionable due to maximum likelihood estimate.









Table 2-10. Effects of significant plot environmental and tree importance values in
multiple logistic regression for relationship to percent cover class levels of
Japanese climbing fern (Lygodium japonicum) in three North Florida forest


sites.
Area Effect
(in order of entry)a
All plots Intercept
Carya aquatica
FLOOD
Catalpa bignoniodes
Platanus occidentalis
Nyssa ogeechee
Quercus laevis

BRSF No variables entered

Miller No variables entered


FRI


Intercept
FLOOD


Estimate R-squared Chi square
(cumulative)b score P>Chi square


9.2
3.7
-1.1
-0.1
-0.1
-1.3
0.2


3.21
-1.37


0.17
0.38
0.46
0.52
0.59
0.75


=0.04
=0.01
=0.02
=0.04
=0.02
=0.003


0.50


P=0.005


a Significant effects in model (SAS Version 9.1). Principal components analysis factor:
FLOOD l=hydrology class, percent flooded, and months flooded. Model validity is
questionable due to maximum likelihood estimate.














CHAPTER 3
HERBICIDE EFFICACY FOR MANAGEMENT OF JAPANESE CLIMBING FERN IN
SOUTHEASTERN FORESTS

Introduction

Japanese climbing fern (Lygodium japonicum), is a non-native invasive plant widely

established in the southeastern United States. In this range, Japanese climbing fern invades mesic

and temporally hydric areas, including: floodplain forests, bottomland hardwood forests, marshes,

wetlands, secondary woods, moist pinelands (especially flatwoods), limestone outcroppings, and

disturbed areas such as road shoulders and rights-of-way (Clewell 1982, Nauman, 1993, Langeland

and Burks 1998). Japanese climbing fern does presently occur on xeric sites as well, but to a more

limited extent. Heavily infested sites may sustain populations averaging 60-80% cover over large

areas, effectively reduce or eliminate native groundcover and understory vegetation, and smother

seedlings of overstory tree species (Horovitz et al. 1998, Lott et al. 2003, Zeller and Leslie 2004).

Despite these impacts, large information gaps exist on aspects of the biology and management of

this plant. For both public and private land managers, identification of efficient and effective

control strategies for Japanese climbing fern has become increasingly important.

Recognition of this plant has increased among public conservation land managers in the

southern United States since the mid 1990s, aiding in increased detection and reporting.

Recognition of Japanese climbing fern by private land managers has also increased since the late

1990s, but management efforts have not increased at a similar rate due to lack of awareness of the

non-native invasive status of the plant, research-based best management practices, and financial

incentives or support for management of private-lands infestations. Designation as a noxious weed









in Florida has been an important part in the process of raising awareness among private forestland

managers and members of the forest products industry (e.g., pinestraw) in Florida, and for

increasing the demand for effective management guidelines. Stop sale orders have been issued for

shipments of Japanese climbing fern-contaminated pinestraw bales in multiple cases by the FDACS

Division of Plant Industry (Van Loan 2006), and awareness of the regulatory requirements is

affecting some buyer preferences as well.

Forestry and Invasive Plant Impacts

The impact of species' invasions on sustainable forest management is increasing in scale and

recognition. Plant invasions have affected biological diversity, forest health and productivity, water

and soil quality, and socieoeconomic value. Plant invasions modify forest ecosystems by affecting

fire and hydrological regimes, food webs, and recruitment of dominant tree species (Vitousek,

D'Antonio, Loope and Westbrooks 1996, Chornesky, Bartuska, Aplet, O'Britton, Cummings-

Carlson, Davis, Eskow, Gordon, Gottschalk, Haack, Hansen, Mack, Rahel, Shannon, and Wainger

2005). In managed forests, establishment of invasive plants is increasingly facilitated by expanding

human access, fragmentation, and soil or canopy disturbance often associated with silvicultural

practices (With 2002, Chornesky et al. 2005). Some estimates place annual economic impact of

forest product loss due to invasive species as high as $2 billion in the United States (Pimental et al.

2000). Impacts of invasive plants on the forestry industry in Florida have been estimated at $38

million per year ($15 million in weed control costs, and $23 million in yield losses) (Lee 2005).

Little is known about the economic impact of Japanese climbing fern invasion. Forest

managers in Florida have noted impacts including reduction in annual financial return and harvest

lease longevity for pinestraw production, an industry which generates $79 million in revenue for

forest landowners in Florida annually (Hodges et al. 2004), and reductions in pine seedling survival

during reforestation of heavily infested sites. In particular, mesic private forestlands are moderately









to heavily infested with Japanese climbing fern throughout much of the northwest Florida region. A

2002 survey (Barnard and Van Loan 2002) conducted across 280 pine plantations in north and west

Florida recorded presence of Japanese climbing fern in 22% of the stands on mesic sites, exceeding

occurrence of four other regionally significant non-native invasive plants including: cogon grass

(Imperata cylindrica (L.) Beauv.), Chinese tallow (Triadica sebifera (L.) Small), kudzu (Pueraria

montana var. lobata (Willd.) Maesen & S. Almeida), and air potato (Dioscorea bulbifera L.). With

approximately 5.1 million acres of mesic pine plantations alone in north Florida (Brown 1995), the

regional scale of this invasion generates management problems including: need for coordinated and

cost-effective management strategies, legal implications affecting harvest and sale of forest

products from infested areas, and high probability of continual reinvasion of treated sites.

Management Techniques

Integrated pest management utilizes the most appropriate combination of treatments

including: chemical, manual/mechanical, cultural, and biological control techniques. Mechanical

and cultural (i.e., prescribed burning) treatments have generally proven ineffective, often promoting

Japanese climbing fern re-growth, and possibly vectoring spores. To date, chemical treatments

have been most effective, but little exploratory research has been conducted.

A strategy that crosses property boundaries is required for successful management of most

invasive species (Chornesky et al. 2005). The reproductive strategy (i.e. spore-dispersal, self-

fertilization) of Japanese climbing fern facilitates rapid spread and establishment in remote areas

(Lott 2003), and continual re-invasion if all populations in an area are not addressed jointly.

Identification of consistently effective treatment guidelines is an important component in

developing a management commitment that crosses property boundaries and is implemented by

both public and private land managers. Current knowledge indicates that treatment with herbicides









will be a required component of successful management of Japanese climbing fern in the United

States.

This study evaluates those herbicides which have received preliminary assessment for

Japanese climbing fern control, as well as others commonly used on both forestry and vegetation

management. The objective of this study were to: (1) evaluate a range of herbicides to be applied

postemergence for control of Japanese climbing fern and (2) evaluate the efficacy of multiple rate

combinations of glyphosate and metsulfuron, two herbicides expected to have the greatest efficacy

against Japanese climbing fern.

Materials and Methods

Study Area

Field experiments to evaluate the effect of postemergence herbicides applied to Japanese

climbing fern were established in 2002 and replicated in 2004. All experiments were conducted in a

privately-owned slash pine (Pinus elliottii Engelm. var. elliottii) plantation in Calhoun County,

Florida, USA (30 23'N, 85 9'W). The plantation was established in 1979, managed for pinestraw

harvest from 1997-1999, and underwent a third row thinning (removal of every third row of pine

trees) in 2000. Japanese climbing fern was first recorded on the site in 1997, but may have been

present earlier. By 2002, the stand was heavily infested, with Japanese climbing fern dominating

the understory and trellising into the midstory. For both of the 2002 trials, the soil was a Dothan

sandy loam, 0-2 percent slope (fine-loamy, kaolinitic, thermic Plinthic Kandiudults) with 0.48%

organic matter, and a soil pH of 5.3. The 2004 trial was conducted within the same stand, but on

slightly different soil; Robertsdale fine sandy loam (fine-loamy, siliceous, semiactive, thermic

Plinthaquic Paleudults) with 1.87% organic matter, and a soil pH of 5.2. The average daily

temperature is 12 C in the winter and 33 C in the summer. Freezing temperatures occur on an

average of 20 days each winter.









Herbicide Selection

Herbicides were selected based on their weed control spectrum, label-status for forestry

applications, performance in initial field assessments by land managers, and commonality with best

management practices for treatment of other non-native invasive plants regularly co-occurring with

Japanese climbing fern in Florida (Table 3-1). A non-ionic surfactant "Induce" (Helena Chemical)

was added at 0.25% v/v according to label recommendations.

In response to land manager concerns and requests, two formulations of glyphosate were

evaluated; a "traditional" formulation (Accord TM), and a "generic" (GlyphosTM) formulation,

representing the many glyphosate products available for purchase since the expiration of the

RoundupTM patent. Land manager interest in this comparison was primarily due to the lower per

unit cost of the "generic" formulations. Both formulations were evaluated at a low application rate

(2.24 kg ai per ha, or 2.5% in solution), and a high application rate (4.48 kg ai per ha, or 5% in

solution).

Experimental Design

Two herbicide trials were established in October 2002. The "Multiple Herbicide Trial"

included fifteen herbicide treatments and an untreated control (see Table3-2). Treatment plot size

was 6.8 m by 3.4 m. The "Glyphosate/Metsulfuron Trial" included tank-mixed combinations of

four rates of glyphosate (0, 1.12, 2.24, 4.48 kg ai per ha) and metsulfuron (0, 0.03, 0.06, 0.12 kg ai

per ha), and an untreated control, for a total of sixteen treatments. Treatment plot size for this trial

was 3.4 m by 3.4 m and were smaller than desired due to constraints within the study site.

Treatment plots were established in the thinned rows of the stand. All treatments were applied

using a C02-pressurized hand-held boom sprayer. Spray volume was calibrated to deliver 225 liters

per hectare (20 gallons per acre). Experiments were conducted as randomized complete block

designs, and all treatments were replicated four times, for a total of sixty four plots per trial. The









"Multiple Herbicide Trial" was replicated within the study site in 2004 with treatment plots sized

6.8 m by 3.4 m. The "Glyphosate/Metsulfuron" trial was not replicated in 2004 due to lack of

adequate area with contiguous Japanese climbing fern coverage in the study site, and the

expectation of greater utility of results from the "Multiple Herbicide" trial for land managers.

Treatment and Assessment Schedule

2002 trials. Herbicides were applied on October 23, 2002. Percent cover of live Japanese

climbing fern was rated at 1 month after treatment (MAT), 5, 6, 8, and 12 MAT. Upon returning to

the study site for the 12 MAT rating, it was discovered that the pine stand had been unexpectedly

thinned in late September, 2003 (removal of selected trees from "leave" rows remaining after 2000

thinning, and skidding of logs through all plots). Aboveground Japanese climbing fern foliage was

mechanically removed from all plots by the harvest operation (percent cover = 0), but plot comer

stakes were adequately intact to re-define original plot boundaries. Following harvest, percent

cover of live Japanese climbing fern was rated at 15, 24, 29, 31, 35, and 36 MAT.

2004 trial. The 2002 Multiple Herbicide Trial was repeated within the original study site to

remove the potential confounding effects due to logging. Treatment plot size was 3.4 by 6.8 m.

Herbicides were applied on November 10, 2004. Percent cover of live Japanese climbing fern was

rated at 1, 4, 6, 10, and 12 MAT.

Data Analysis

Percent cover of live (green foliar tissue) Japanese climbing fern was visually determined and

recorded for each plot to rate effectiveness of the various treatments. All plots were rated on a scale

from 0% (no climbing fern on plot) to 100% (plot completely covered by climbing fern) in 5%

increments one day prior to treatment to establish a baseline. Following treatment, plots were rated

using the same technique according to the assessment schedules indicated above; all ratings were

were conducted by the same observer.









For all plots, percent live Japanese climbing fern cover was converted to percent control for

analysis using the following equation:

Percent control = Percent cover pre-treatment Percent cover at MATf X 100
C,._ Percent cover pre-treatment 1


Use of percent control values accounts for differences in pre-treatment fern cover across

treatment plots. The results from each trial were analyzed using the general linear model approach

to detect significance of model effects among all treatments. Fisher's least significance difference

(LSD) test was used to make pairwise comparisons among treatment means (SAS Institute 2005).

For all tests, significance was assessed at the alpha=0.05 level.

A two-sample t-test was performed to evaluate the effect of plot size on trial results. Means

from untreated control plots in the 2002 Multiple Herbicide trial and the 2002 Glyphosate/

Metsulfuron trial were compared, with the hypothesis that no difference would exist.

Results and Discussion

The reproductive strategy (i.e. spore-dispersal, self-fertilization) of Japanese climbing fern

facilitates rapid spread and establishment in remote areas (Lott et al. 2003), and continual re-

invasion if all populations in an area are not addressed jointly. Identification of consistently

effective treatment guidelines is an important component in developing a management commitment

that crosses property boundaries and is implemented by both public and private land managers. A

strategy that crosses property boundaries is required for successful management of most invasive

species (Chornesky et al. 2005). Current knowledge indicates that treatment with herbicides will be

a required component of successful management of Japanese climbing fern in the United States.

The following results verify and expand the current knowledge base on herbicide selection for this

purpose.









Phenology of Japanese Climbing Fern

To fully understand the herbicide treatment effects for any species, it is important to consider

other factors which may influence plant performance. On all plots in this study, three primary

factors affected climbing fern foliar cover: herbicide treatment (varied by treatment), seasonal

dieback and growth (uniform across treatments), and mechanical damage from the 2003 timber

harvest (uniform across treatments). Japanese climbing fern plants remain evergreen below the

frost line in central and south Florida (Ferriter 2001, Valenta et al. 2001, Zeller and Leslie 2004).

Above the frost line, winter dieback of Japanese climbing fern fronds occurs to varying extents

through the majority of its range. In the spring, the plants will re-sprout from cold-tolerant

subterranean rhizomes (new stems observed the first week in March in North Florida), and often

utilize freeze-damaged stems as ladders to grow back into the canopy. This seasonal dieback in

sub-temperate and temperate climates is one factor which has likely prevented formation of the

canopy-covering arboreal blankets seen with Old World climbing fern (Lygodium microphyllum)

(Zeller and Leslie 2004). However, the increasing distribution of Japanese climbing fern in south

Florida may lead to the development of such canopy-smothering growth on infested sites.

Japanese climbing fern seasonal foliar phenology was recorded on the untreated control plots

in each trial throughout the study period. At inception of the 2002 and 2004 trials (October),

untreated Japanese climbing fern plants had achieved maximum annual foliar coverage, followed by

a seasonal frost event (December-February) which reduced plot live foliar cover to 11% on average

(min. 5%, max 20%). Initial emergence of new croziers (growing shoots) occurred in mid-March,

with measurable new growth of climbing fern present by mid-April. Foliar cover continued to

increase steadily through the summer and fall until peaking again in October for both the 2002 and

2004 trials (Figure 3-1). Despite these seasonal fluctuations and the 2003 timber harvest event,

Japanese climbing fern cover on the control plots averaged a 3.1% annual increase over the initial









percent cover (Table 3-3). However, analysis of variance did not identify this change as significant

(alpha = 0.05)

Multiple Herbicide Trial: 2002 and 2004

Average initial climbing fern percent groundcover on the Multiple Herbicide trial plots was

70% (min 30%, max 95%). A significant effect was measured for the interaction of treatment and

year between 2002 and 2004, and as a result, the data from each year are presented separately.

Discussion of "short-term" treatment effects observed through 12 MAT will focus on the 2004 trial

results. This ensures continuity not possible in the 2002 plots due to the interruptive effect of the

2003 timber harvest. Results from the 2002 trials are used for discussion of "long-term" treatment

effects observed at 24 and 36 MAT. The plot ratings from 5 and 29 MAT, and 15 MAT are not

discussed as they represent effects of seasonal frost damage and timber harvest respectively, rather

than herbicide treatment effect. Review of results from this trial and those from Valenta et al.

(2001), and Zeller and Leslie (2004) indicate that the best performing ("successful") treatments

typically yield a minimum of 70% control at 12 MAT; a standard which will be utilized in the

following discussion.

Short Term Treatment Effects

Short term treatment effects reported herein include effects of treatments applied in 2004,

from treatment through twelve months after treatment. At 1 MAT (December 2004), three

treatments exhibited a level of damage to Japanese climbing fern foliage that was significantly

(P<0.05) different than observed in the untreated control plots (Table 3-2). Hexazinone and 2,4-

D+dicamba showed the highest initial damage to Japanese climbing fern with 91% control

(reduction in live foliar cover), followed by the dicamba treatment with 80% control. This initial

response is probably due to "burn-off" of the foliage rather than the damage to underground plant

structures necessary to achieve long-term control. Similar effects were observed in early









measurements of treatments by Valenta et al. (2001), and Zeller and Leslie (2004). Three other

treatments yielded control levels greater than 70%: the "high" rate of glyphosate in both the

"traditional" (77%) and "generic" (76%) formulations, and the tank mix of glyphosate+imazapyr

(72%). The tank mix of glyphosate+metsulfuron yielded 65% control, while the glyphosate

treatments at "low" rate yielded 54% and 29% control, respectively, for the "generic" and

"traditional" formulations. All other treatments yielded less than 50% control. The untreated

control plots yielded a 32% reduction in foliar cover, which may be explained by early winter

decline in foliage density observed following the fall "peak" in foliar cover or as a result of foliar

rust damage, and/or damage to fronds originating in other treatment plots and growing across

adjacent control plots.

AT 6 MAT (May 2005), treatments separated into three significant (P<0.05) groups (Table 3-

2). The largest group included eleven treatments exhibiting between 89% and 100% control of

Japanese climbing fern. This group included all of the herbicides and herbicide tank mixes in the

trial with amino acid inhibition as the mode of action: glyphosate (both rates and formulations),

imazapyr, metsulfuron, glyphosate+metsulfuron, glyphosate+imazapyr, imazapyr+metsulfuron,

imazapic, and sulfometuron. In a statewide survey of climbing fern treatment regimes used by

public land managers, many favored a six-month return interval for herbicide treatments (C.

Lockhart, 5/04/2006, personal communication). With such a regime, all treatments in this group

may be considered in managing Japanese climbing fern. A second group of treatments, including

hexazinone, dicamba, and 2,4-D +dicamba, exhibited approximately 60% control, while triclopyr

ester dropped to 25%, indicating higher fern coverage than on the untreated control plots (35%).

This control level observed in the untreated plots is primarily explained by lingering effects of

winter climbing fern dieback.









At 10 MAT (September 2005), two significant (P<0.05) groups of treatments were

observable. As found at 6 MAT, the group of treatments exhibiting the highest level of control

(69% to 94%) included all of the amino acid inhibitors, with the exception of sulfometuron where

control levels were reduced to 0.1%. A second group of treatments including hexazinone, dicamba,

2,4-D+dicamba, sulfometuron and the untreated control plots exhibited between 0.1% and 14%

control; while triclopyr ester plots showed a negative control level, indicating that climbing fern

cover had increased by 12% from the pre-treatment levels. The untreated control plots still yielded

14% control at this assessment as climbing fern foliar cover continued to increase through the

remainder of the growing season (Figure 3-1).

At 12 MAT (November 2005), three significant (P<0.05) treatment groups were observable

(Table 3-2). The herbicide group exhibiting the highest level of reduction in climbing fern included

four treatments yielding greater than 90% control (the "traditional" and "generic" formulations of

glyphosate at the high application rate, and tank mixes of glyphosate+imazapyr and

glyphosate+metsulfuron); and three treatments exhibiting from 73% to 87% control (the

"traditional" and "generic" formulations of glyphosate at the low application rate, imazapyr, and the

tank mix of metsulfuron+imazapyr.) At this post treatment date, the effects of metsulfuron and

imazapic treatments had diminished to 41% and 27% control, respectively, while hexazinone

exhibited only 3 percent control. For five treatments, percent cover increased above pre-treatment

levels, yielding a negative percent control value, including: dicamba (8%), sulfometuron (11%),

triclopyr ester (12%), and 2, 4-D+dicamba (12%), and the untreated control plots (3%).

Glyphosate comparison. Results from the 2004 Multiple Herbicide trials were consistent

with initial expectations, in that the treatments that included glyphosate were among the highest

performing treatments. Direct comparison of the "traditional" and "generic" glyphosates at both









high and low application rates yielded little difference between the two formulations. At 1 MAT,

significant difference (P<0.05) was detected between the "traditional" and "generic" formulations at

the low application rate, but at all subsequent measurements no difference was detected between the

two formulations.

Timber Harvest Effects

The October 2003 timber harvest clearly affected all of the 2002 study plots by physically

removing the aboveground fern growth, reducing canopy cover across the stand, increasing the

amount of downed woody debris, and possibly re-distributing Japanese climbing fern spores across

the study site. The harvest effects appeared to be uniform across all plots based on visual

assessment. The plots then underwent a period of post-harvest "recovery" which lasted for

approximately one year, and during which climbing fern reclaimed the site understory.. Treatment

effects became identifiable again at 24 MAT, and lasted through the final assessment at 36 MAT.

Long Term Treatment Effects

Long term treatment effects reported herein refer to effects measured on both the 2002 and

2004 plots from 12 MAT to 36 MAT. Treatment effects which were beginning to appear in the

2002 plots at 8 MAT, and which also were clear in the 2004 plots at 12 MAT, reappeared in the

2002 plots at 24 MAT (Table 3-2). Analyses revealed many overlapping relationships, but yielded

two primary treatment groups. Nine treatments exhibited significantly higher (P<0.05) levels of

climbing fern control (38% to 70%) than seen in the untreated control plots, including: all

treatments that included glyphosate ("traditional" and "generic" formulations, low and high rates

and tank mixes), and metsulfuron and imazapyr.

Treatment effects remained similar at 36 MAT (Table 3-2). Six treatments continued to

exhibit significantly higher levels of climbing fern control (35% to 41%) than observed in the

untreated control plots, including both rates of the "traditional" glyphosate formulation and the low









application rate of the "generic" glyphosate formulation, imazapyr, and tank mixes of

glyphosate+imazapyr and metsulfuron+imazapyr. Seven herbicide treatments exhibited between

29% control (reduction) and 24% increase in climbing fern cover, not significantly different than

the untreated control plots. Climbing fern cover on the untreated plots increased above initial pre-

treatment levels by 6.25%.

2002 Glyphosate/Metsulfuron Trial

The initial decision to conduct the Glyphosate/Metsulfuron rate combination trial was made

based on performance of these two herbicides in preliminary assessments (Valenta et al. 2001,

Zeller and Leslie 2004). Metsulfuron showed potential for limiting damage to non-target plant

species during invasive plant management. In addition to the recovery of ecosystem value through

retention or re-establishment of native species on sites affected by non-native species invasions,

several studies have shown the importance of site occupancy in preventing new or recurring site

invasion by non-native species, particularly following a disturbance event such as chemical

treatment (Burke and Grime 1996, Davis et al. 2000, Prieur-Richard and Lavorel 2000, D'Antonio

and Myerson 2002). Identification of a control strategy which limits loss of non-target species

may play a significant role in long term success of any Japanese climbing fern management

strategy. Zeller and Leslie (2004) observed glyphosate treatments to yield better long term control

(70% to 80% reduction in fern cover at one year after treatment) than metsulfuron treatments, but

more non-target damage occurred compared to metsulfuron. In sites where herbicide applications

are likely to be repeated within a single year, this finding may indicate use of metsulfuron for

retention of desirable non-target species. However, concerns exist among weed management

specialists about rapid development of resistance to sulfonylurea herbicides such as metsulfuron

(Tranel and Wright 2002), particularly with plants which reproduce as rapidly and prolifically as the

climbing ferns (Lott et al. 2003). Treatments that involve tank-mixing glyphosate and metsulfuron









hold promise as a technique to reduce both the non-target damage seen with glyphosate, and the

possibility of herbicide-driven selection of metsulfuron-resistant climbing fern populations (Diggle

et al. 2003, Kudsk and Streibig 2003). This concern prompted the evaluation of multiple rate

combinations of glyphosate and metsulfuron in this study.

Average initial climbing fern cover on the Glyphosate/Metsulfuron plots was 66% (min 30%,

max 90%). As with the 2002 Multiple Herbicide plots, percent groundcover of live Japanese

climbing fern was rated at 1, 5, 6, 9, and 12 MAT. Plots were affected by thinning before 12 MAT,

at which time percent cover for all plots was mechanically reduced to zero. In the period following

harvest, plots were rated at 15, 24, 29, 31, 35, and 36 MAT.

The plot size for the Glyphosate/Metsulfuron trial was 3.4 m by 3.4 m, 50% smaller than the

plot size for the Multiple Herbicide trials. This factor was significant in determining the usefulness

of the trial results (data not shown). At nearly all assessment dates, the percent control of Japanese

climbing fern on the untreated control plots was not significantly different (P<0.05) than the

herbicide treatment plots. The untreated control plot data from the Glyphosate/Metsulfuron trial

and the 2002 Multiple Herbicide trial were compared to assess the effect of plot size. Significant

differences (P<0.05) due to plot size were seen at 5, 6, 8, and 24 MAT, indicating that the treatment

differences detected in this study represented a Type 1 error rather than actual treatment differences.

A recommendation for future research is to limit the minimum herbicide trial plot size to 3.4 m by

6.8 m (200 square feet). Further evaluation of these treatments may be appropriate; however, in the

interest of achieving long-term control of Japanese climbing fern a higher priority may be placed on

research which evaluates the effect of repeated herbicide applications.

Conclusions and Recommendations for Land Managers

The typical goal of invasive plant management on public conservation lands is, when

possible, to eradicate the species from infested sites in such a way as to promote maximal recovery









of native plant species following treatment. This may also be the management goal on some private

forestlands, but for many others, preservation of native midstory and groundcover species is

secondary to production of a profitable and legally saleable forest product (Allen et al. 2005). In the

case of the pinestraw industry, the vegetation management goal is typically to remove understory

vegetation from a harvest site that may impede the pinestraw harvest process, or serve as a non-

straw contaminant in the pinestraw bales (Duryea 1998, Taylor and Foster 2003). These varying

goals will impact implementation of the results of this study by land managers in the field.

The data of greatest interest to land managers are the results from the 2004 Multiple Herbicide

trial in the 0 to 12 MAT period. Increasingly, a six month return interval is recommended or

implemented in climbing fern management (Ferriter 2001, Lockhart 2006). While managers often

strive to revisit treatment sites multiple times in a year, many operational invasive plant

management programs on public and private lands include a functional minimum return interval of

approximately one year between herbicide applications. However, budget and staff limitations can

prevent mangers from achieving this minimum, resulting in return intervals which may extend to

two years or more.

The long term (>12 MAT) results from this trial illustrate that even when such extended

return intervals occur, some treatment effects will remain, to the point of reducing the Japanese

climbing fern coverage by 39% on average across the "good" treatments at three years post-

treatment (Table 2). While positive in some sense, this extended suppressive effect should not

prompt land managers to delay return to a site for management purposes, as the species' spore-

based reproductive strategy allows even small infestations to contribute significantly to re-

infestation of a site, and spread to other sites. In addition, the vacancies created in areas where

herbicide treatments are successful, are most likely to be filled by either Japanese climbing fern or









another non-native invasive plant, providing additional incentive for regular follow-up management

visits (Davis et al. 2000, D'Antonio and Myerson 2002).

Land managers will consider three main factors when determining which approach to use in

treatment of Japanese climbing fern: level of control, duration of control, and per area treatment

cost. As indicated earlier, a minimum level of control in successful treatments may be 70%

reduction of Japanese climbing fern foliage (Table 3-4). However, in this study four treatments

(glyphosate high rate "traditional" and "generic", glyphosate + imazapyr, and glyphosate +

metsulfuron) maintained greater than 90% control for up to 12 MAT (Table 3-2). This highest level

of control also limits spore production, a factor which is very important in the long term

management of this plant. Three treatments from this trial provided control for 6 (sulfometuron) to

10 (metsulfuron and imazapic) MAT. Managers may decide to treat Japanese climbing fern with

one of these treatments as a secondary component of herbicide application for other management

purposes, or in the case of metsulfuron, if reduced damage to non-target species is a management

priority. Sulfometuron has been recommended for herbaceous weed control in hardwood

plantations both preplanting and postplanting (Rhodenbaugh and Yeiser 1994, Ezell and Nelson

2001). Bottomland hardwood, floodplain forest, and mixed pine-hardwood stands are both readily

invaded by Japanese climbing fern in Florida, and in some cases actively managed for hardwood

production. In such sites, sulfometuron can be applied to dormant hardwood seedlings, and over

conifer seedlings (Schuler, Robison and Quicke 2004). However, any damaging effects of these

compounds on Japanese climbing fern will likely only last 6 months, and require attention to regular

follow-up treatments.

Based on level of control achieved and approximate prices in 2006 (Ferrell, Gray and

MacDonald 2006), the glyphosate treatments were the most cost-effective, and efficacious at both