|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 EVALU ATION OF EUCALYPTUS INVASIVENESS IN FLORIDA AN D METHODS FOR DIRECT CONTROL By KIMBERLY A. LORENTZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR T HE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2 2013 Kimberly A. Lorentz
3 To all of my family who are most precious to me
4 ACKNOWLEDGMENTS I want to express my gratitude to my advisor, Dr. Patrick Minogue, for his enthusiasm, for sharing his wealth of knowledge and for the material and instructive support that he provided. I am also thankful to my committee members, Dr s. Kimberly Bohn, Donald Rockw ood and Jason Ferrell for the care with whi ch they reviewed this thesis and for their valuable guidance along the way Insightful advice given by Drs. Anna Osiecka, Luke Flory and Damian Adams has also been a great help in the development of this research. I extend many thanks to James Colee, IFAS statistics, and Dr. Dwight Lauer Silvics Analytic for their assistance with statistical analyses in this thesis The a ssistance with germination testing that Jim Aldrich provided was greatly appreciated. I also thank Luke Wright, Seth Wright and Daniela Chevasco for assisting with data collection and for their pleasant company that made long days in the field go by quickly I am forever grateful for the companionship of m y fellow graduate students and friends at the University of Florida as we moved thr ough this learning experience together It has been a pleasure getting to know and collaborating with the many great people in the School of Forest Resources and Conservation. Most of all, loving thank s to my parents, Rick and Kathy Lorentz and to Marie L orentz and Johnny Grebenc for their unwavering suppo rt and encouragement throughout this journey
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ............................ 4 LIST OF TABLES ................................ ................................ ................................ ...... 7 LIST OF ABBREVIATIONS ................................ ................................ ....................... 9 ABSTRACT ................................ ................................ ................................ ............. 11 CHAPTER 1 EUCALYPTUS AS BIOENERGY FEEDSTOCK IN THE SOUTEASTERN UNITED STATES: INVASION RISK ASSESSMENT, MANAGEMENT PRACTICES AND POLICIES ................................ ................................ ........... 13 Introduction ................................ ................................ ................................ ....... 13 Risk Management Practices for Eucalyptus ................................ ...................... 18 Prevention ................................ ................................ ................................ .. 18 Containment ................................ ................................ ............................... 20 Control ................................ ................................ ................................ ........ 24 Impact Management ................................ ................................ ................... 26 Implications of Policy Alternatives ................................ ................................ ..... 27 Conclusion and Needed Research ................................ ................................ ... 30 2 POTENTIAL EUCALYPTUS INVASIVENESS IN FLORIDA'S NATIVE AND MODIFIED PLANT COMMUNITIES ................................ ................................ 32 Introduction ................................ ................................ ................................ ....... 32 Mate rials and Methods ................................ ................................ ...................... 34 Study Areas ................................ ................................ ................................ 34 Site Surveys ................................ ................................ ............................... 35 Seed Addition S tudies ................................ ................................ ................ 35 Germination testing ................................ ................................ .............. 35 Experimental design ................................ ................................ ............ 37 Seedling as sessments ................................ ................................ ......... 38 Statistical analysis ................................ ................................ ................ 38 Results and Discussion ................................ ................................ ..................... 40 Site Surveys ................................ ................................ ............................... 40 Seed Addition Studies ................................ ................................ ................ 41 Seedling emergence in seeded subplots ................................ ............. 41 Seedling emergence in non seeded areas ................................ .......... 41 Seedling survival ................................ ................................ .................. 44 Conclusion ................................ ................................ ................................ ........ 45
6 3 COMPARISON OF AMINOCYCLOPYRACHLOR TO STANDARD HERBICIDES FOR CONTROL OF EUCALYPTUS ................................ .......... 54 Introduction ................................ ................................ ................................ ....... 54 Herbi cides for Eucalyptus Control ................................ .............................. 54 Aminocyclopyrachlor for Control of Woody Plants ................................ ...... 56 Materials and Methods ................................ ................................ ...................... 57 Study Areas ................................ ................................ ................................ 57 Basal Bark Treatments and Experimental Design ................................ ...... 58 Basal Frill Treatments an d Experimental Design ................................ ........ 59 Tree Assessments ................................ ................................ ...................... 60 Statistical Analysis ................................ ................................ ...................... 60 Stem live height and crown reduction ................................ .................. 60 Phytotoxicity symptoms ................................ ................................ ....... 61 Aminocyclopyrachlor rate response ................................ ..................... 61 Herbicide impacts on non target trees ................................ ................. 62 Results and Discussion ................................ ................................ ..................... 63 Eucalyptus Control using Basal Bark Treatments ................................ ...... 63 Crown reduction ................................ ................................ ................... 63 Stem live height ................................ ................................ ................... 64 Phy totoxicity symptoms ................................ ................................ ....... 64 Eucalyptus Control using Basal Frill Treatments ................................ ........ 66 Crown reduction ................................ ................................ ................... 66 Stem live height ................................ ................................ ................... 67 Phytotoxicity symptoms ................................ ................................ ....... 67 Aminocyclopyrachlor Rate Response Relative to Tree Diamete r at Breast Height ................................ ................................ ................................ ...... 68 Impacts to Non target Vegetation ................................ ............................... 70 Conclusion ................................ ................................ ................................ ........ 71 4 SUMMARY AND IMPLICATIONS FOR MANAGEMENT AND FUTURE RESEARCH ................................ ................................ ................................ ...... 79 APPENDIX A POTENTIAL INVASION RISK MANAGEMENT PRACTICES FOR EUCALYPTUS ................................ ................................ ................................ .. 82 B AMINOCYCLOPYRACHLOR FOR CONTROL OF WOODY PLANTS: LITERATURE REVIEW ................................ ................................ .................... 84 LITERATURE CITED ................................ ................................ .............................. 86 BIOGRAPHICAL SKETCH ................................ ................................ ...................... 97
7 LIST OF TABLES Table page 2 1 Expected germination in a controlled environment growth chamber ................. 47 2 2 Vegetation community characteristics at Gainesville, Florida ............................ 4 7 2 3 Vegetation community characteristics at Quincy, Florida ................................ .. 47 2 4 Mean percent Eucalyptus seedling emergence for disturbance, species and vegetation community variables ................................ ................................ ........ 48 2 5 Mean percent Eucalyptus seedling emergence for location, vegetation community and species treatment combinations ................................ ............... 49 2 6 Population means for Eucalyptus seedling emergen ce across both seeded and non seeded areas of 21 disturbed and 21 non disturbed treatment plots within each of the various vegetation communities at two Florida study locations. ................................ ................................ ................................ .......... 50 3 1 Treatments tested in the basal bark studies for Eucalyptus benthamii control in Florida ................................ ................................ ................................ ............ 73 3 2 Treatments tested in the basal frill studies for Eucal yptus benthamii control in Florida ................................ ................................ ................................ ........... 73 3 3 Crown reduction of Eucalyptus benthamii at 2, 6 and 12 months after treatment (MAT) with diameter specific basal bark herbicide applications at the eroded and non e roded study sites ................................ ............................ 74 3 4 Cr own reduction of Eucalyptus benthamii at 2, 6 and 12 months after treatment (MAT) with diameter specific basal frill herbicide applications at the non eroded study site ................................ ................................ ................. 75 3 5 Logistic regression model variables for aminocyclopyrachlor (AMCP) rate response in Eucalyptus benthamii mo rtality at two months after treatment ....... 75 3 6 Percentage of buffer (n = 411) and non treated (n = 48) Eucalyptus be nthamii trees that displayed symptoms of herbicide injury at 2, 6 and 12 months after treatment (MAT) in a study using aminocyclopyrachlor, imazapyr and triclopyr herbicides ................................ ................................ ............................ 75 A 1 Practices to evaluate and manage potential invasiveness of Eucalyptus in the Southeastern US ................................ ................................ ........................ 82 B 1 A list of reported research for the control of woody plant species using aminocyclopyrachlor (AMCP) herbicides ................................ .......................... 84
8 LIST OF FIGURES Figure page 2 1 Treatment plot dimensions and possible randomly assigned disturbance treatments within whole plots ................................ ................................ ............. 51 2 2 Seedling longevity for the Gainesville, Florida study location ............................ 52 2 3 Seed ling longevity for the Quincy, Florida study location ................................ .. 52 2 4 Time of Eucalyptus seedling survival by species for all seedlings at both Quincy and Gain esville study locations ................................ ............................. 53 3 1 Change in Eucalyptus benthamii live stem height from pre treatment values to 12 months after diameter specif ic basal bark treatments at the eroded (lighter bars) and non eroded (darker bars) study sites ................................ ..... 76 3 2 Change in Eucalyptus benthamii live stem height from pre treatment values to 12 months after diameter specific basal frill treatments at the non eroded study site ................................ ................................ ................................ ........... 77 3 3 Relationship between Eucalyptus benthamii stem diameter at breast height (DBH, measured at 137 cm height) and applied concentration of 120 g ae L 1 aminocyclopyrachlor (AMCP) i n methylated soybean oil carrier for the predicted likelihood of mortality at two months after diameter specific basal bark treatment ................................ ................................ ................................ ... 78 3 4 Relationship between Eucalyptus benthamii stem diameter at breast height (DBH, measured at 137 cm height) and applied concentration of 120 g ae L 1 aminocyclopyrachlor (AMCP) diluted in water fo r the predicted likelihood of mortality at two months after diameter specific basal frill treatment ................... 78
9 LIST OF ABBREVIATIONS a e Acid equivalent a i Active ingredient AMPC A minocyclopyrachlor ANOVA A nalysis of variance APHIS Animal and Plant Health Inspection Service BSD B asal stem diameter BTU British thermal unit C Celsius cm C entimeter d Day DBH D iamete r at breast height df D egrees of freedom FAC Florida Administrative Code FAO Food and Agriculture Organization of the United Nations ft Foot FTE Frost tolerant Eucalyptus g G ram gly Glyphosate ha H ectare HSD Honestly significant difference IFAS Universit y of Florida Institute of Food and Agricultural Sciences imaz I mazapyr L Liter M Mean
10 m M eter MAT M onths after treatment mg Milligram ml M illiliter mm Millimeter mo Month N North n S ample size NISC National Invasive Species Council NOAA National Oceanic and Atmospheric Administration NPS National Park Service REGWQ Ryan Einot Gabriel Welsch multiple range test RSB Roundtable on Sustainable Biofuels SE S tandard error SU Suflonylurea herbicides sp p S pecies triclo T riclopyr US United States USDA United Sta tes Department of Agriculture v/v V olume per volume ratio W West WRA Weed Risk A ssessment YAT Y ears after treatment
11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requiremen ts for the Degree of Master of Science EVALUATION OF EUCALYPTUS INVASIVENESS IN FLORIDA AND METHODS FOR DIRECT CONTROL By Kimberly A. Lorentz May 2013 Chair: Patrick Minogue Major: Forest Resources and Conservation Potential invasiveness is a concern for Eucalyptus species which are being planted in the southeastern US for landscaping mulch and possible bioenergy crops The purpose of this research was to evaluate the potential invasiveness of Eucalyptus sp ecies that are under consideration for large scale planting and to improve practices for control of Eucalyptus Surveys for natural recruitment within and proximate to seed bearing stands at two Florida locations found no Eucalyptus seedlings. Seed addition studies then examined the potential for see dling emergence and survival among E ucalyptus amplifolia E ucalyptus camaldulensis and E ucalyptus grandis relative to various disturbance levels, seeding densities and vegetation community types within and proximate to the mature Eucalyptus stands Greater survival was found for E. camaldulensis compared to the other species. Greater emergence was observed under disturbed conditions and within Eucalyptus communities. Overall, emergence of added seed was very low (0.0 to 0.32%) and no seedlings survived more than 13 weeks. Experiments were also conducted to compare aminocyclopyrachlor (AMCP) a new herbicide, to the standard imazapyr and triclopyr herbicide treatments for control of 29 month old Eucalyptus benthamii Basal bark applications of a 120 g ae L 1 AMCP
12 product formulation at 5% v/v in methylated soybean oil carrier resulted in 97 99% crown reduction of E benthamii and generally provided greater control than the standard 240 g ae L 1 imazapyr product formulation at 28.1% v/v or the 480 g ae L 1 tri clopyr ester product formulation at 75% v/v, when assessed at 6 and 12 mo after treatment. Basal frill applications of 120 g ae L 1 AMCP product formulation at 12.5% v/v in water resulted in 100% crown reduction of Eucalyptus and greater control than the s tandard 240 g ae L 1 imazapyr product formulation at 7.8% v/v or 360 g ae L 1 triclopyr am ine product formulation at 50% v/v treatments when evaluated 6 and 12 mo after treatment. These data indicate that under specific favorable conditions Eucalyptus spp. seedlings may establish within or proximate to planted stands in Florida, but the overall likelihood of establishment is low and complete control of unwanted Eucalyptus can be provided by AMCP herbicide at low rates.
13 CHAPTER 1 EUCALYPTUS AS BIOENERGY FE EDSTOCK IN THE SOUTEASTERN UNITED STATES: INVASION RISK ASSESSMENT, MANAGEMENT PRACTICES AND POLICIES Introduction Since the announcement of the US renewable energy initiative in 2006, fore warning of the high risk for biomass bioenergy crops to escape from cultivat ion and invade unmanaged areas ha s become prevalent throughout literature on renewable energy sustainability ( Raghu et al. 2006; Barney and Di T omaso 2008; Buddenhagen et al. 2009; Richardson and Blanchard 2011). Numerous authors advocate d that the precautionary principle should be applied in the adoption of bioenergy policies (Chimera et al. 2010; Davis et al. 2010; Witt 2010; McCormick and Howard 2013). These authors suggest ed that predictive risk assessment legal mechanisms sh ould b e key components of these policies in order to minimize negative environmental impacts of invasive species While a precautionary decision rule is attractive from an ecological perspective, historically invasive species policy in the US has allowed f or risk, in part because policy decisions must be made using scientific information that is inherently complex and difficult to attain. H ow to effectively manage potentially invasive species that are economically valuable and whose management is complicate d by many different stakeholders and government agencies remains uncertain Studies of conflicts over Acacia and Pinus invasions in South Africa and throughout the world contributed valuable insight into policy and management solutions for invasive forestr y species (van Wilgen et al. 2011, 2012; Wilson et al. 2011 a ). I n the southeastern US, potential invasiveness recently emerged as concern for Eucalyptus and several research efforts addressed invasion risk.
14 Members of the diverse Eucalyptus genus ( family Myrtaceae ) are now among the most common exotic trees throughout the world, although they are controversial wherever they occur outside of their native Australia. Commercial Eucalyptus planting s in the southeastern US are increasing in both area and biomas s production to provide f eed stock for an emerging bioenergy industry. Of the approximately 90 Eucalyptus species that have been introduced to North America (Paine et al. 2010), several exhibit the desired features for low cost delivered biomass relative to biomass productivity, rotation length, establishment and maintenance, harvesting and storage (Gonzale z et al. 2011). Eucalyptus can currently be produced in the southeastern US at a cost per BTU that is competitive with coal (Dougherty and Wright 2012) a nd research is under way to further improve the cellulosic conversion process and other aspects of the supply chain (Gonzalez et al. 2011) Additionally, t ransgenic and conventional breeding programs have developed cultivars with superior cold tolerance th at hold promise for plantings beyond the current frost limited northern range boundary of central Florida (Hinchee et al. 2011). While Eucalyptus species ha ve the potential to make a substantial contribution to bioenergy production in the southeastern US consideration must be given to reports of invasiveness elsewhere since history of invasiveness has been shown to correctly predict invasion probability in a new introduced range 90 percent of the time ( Panetta 1993). Naturalization (reproducing and maint aining populations without human help) has been reported for 40 Eucalyptus species including E. globulus in California and Hawaii and E. camaldulensis, E. robusta and E. s aligna in many regions ( Rejmnek and Richardson 2011 ). Further contributing to conce rn is the likeness between many
15 traits which are selected to maximize biofuel crop yield and the ecological traits of successful invasive species ( Raghu et al. 2006; Barney and DiTomaso 2008; NISC 2009) Many Eucalyptus species possesses these shared trait s, which include rapid accumulation of biomass, perennial growth form, prolific seed production low disease pest resistance, and tolerance to drought and low soil fertility (Booth 2012). An addition al risk factor is that an estimated 1.09 million ha of n ew short rotation woody crops may be planted in order to meet proposed Renewable Portfolio Standards in Florida alone (Hodges et al. 2010), resulting in high potential production and release of seed propagules. Little data have focused on the costs related to invasions of non native trees introduced through plantation forestry (Dodet and Collet 2012) Correspondingly, the eco nomic and ecological impacts of introductions of Eucalyptus in the sout h e a stern US have not been directly quantified Data are absent likely because it is inherent ly difficult to predict costs associated with an introduction ( Parker et al. 1999; Anderson et al 2004) although certain potential costs may be anticipated based on studies from other locations where Eucalyptus are invasive. F or example, i ncreased wildfire intensity and frequency is a chief concern for areas that are considering large scale planting of Eucalyptus because i gnition of Eucalyptus foliage significant ly increased the severity of a 1991 fire in the Oakland Berkeley h ills in California that cost over $1 billion (Pagni 1992). Concern has also been raised over the reports that Eucalyptus have cause d desiccati on of streams and groundwater such that nearby crop productivity was reduced in areas of east Africa (FAO 2010b; S enbeta et al. 2010). In South Africa, Eucalyptus invasions in two watersheds lead to 6.0 and 9.4% reductions in natural river flows and
16 the corresponding costs of control programs to prevent these losses were $US 4.1 and 6.6 million and would rise to $278 .0 and 11.1 million if areas became fully invaded (Le Maitre et al. 2002). Ecological impacts of Eucalyptus have been observed in California wh ere dense monocultures formed thereby altering ecosystem structure function and native faunal composition and density (Sax 2002). Although anecdotal evidence supports the potential invasiveness and high costs associated with Eucalyptus introduction in the southeastern US Eucalyptus ha ve been remarkably unobtrusive to date Eucalyptus ha ve been grown for mulch in central Florida for more than 40 years without significant problems, although four species ( E. camaldulensis E. grandis E. robusta and E. torelliana ) are reported to have established ( Wu n derlin and Hansen 2008 ). This pattern is consistent with a degree o f success at the global scale that is orders of magnitude less than that of other introduced trees (Rejmnek and Richardson 2011). Some authors suggest ed that Eucalyptus species are safe choice s for restoration projects in Brazil (da Silva et al. 2011) and are non invasive forestry alternative s to problematic Pinus and Acacia species (Dodet and Collet 2012). Some authors hypothesize d that Eucalyptus ha ve a relatively poor invasive ability because seed s are short lived and have low viability and because ther e is high seedling mortality due to herbivory ( Li et al. 2003; Bec e rra and Bustamente 2008; da Silvia et al. 2011) Other possible limiting factors include s usceptibility to pathogens and fungi (Rejmnek and Richardson 2011a) and intolerance to competition (Adams et al. 2003; Garau et al. 2009). R ecent studies conducted to improve herbicide site preparation or post planting options to combat high mortality in plantation establishment
17 (Blazier et al. 2012; Osiecka and Minogue 2011) offer supporting evidence for the sensitivity of Eucalyptus seedlings Eucalyptus ecology is still marked by many gaps in knowledge and contradictions meriting significant future research The goal of the following literature synthesis was to frame the steps that can be taken to a ddress potential environmental problems while this uncertainty remains E xisting and potential risk assessment and management practices for controlling invasive spread of Eucalyptus in the southeastern US are reviewed and implications of alternative polici es for non native bioenergy crops are discussed subsequently I nformation about effective management practices and policies for invasive forestry species was obtained from published literature identified by searching several scientific databases for keywor ds including: Eucalyptus Acacia Pinus invasive, bioenergy, W eed R isk A ssessment, management, policy and sustainability. Management practices for Eucalyptus are presented with respect to four commonly recognized phases of intervention: prevention, contai nment, control, and management and restoration of the affected ecosystem s (Table A 1). P ractices could Wilson et al. 2011a) or supply chain stages (e.g. McCormick and Howard 2013) However, a framework of intervention phases that generally corresponded with steps in the invasion process (i.e., introduction, establishment, spread and impact) was favorabl e for adequately convey ing the importance of certain strategies such as mitigatio n of invasion impacts which are not exclusively connected to a specific stage in Eucalyptus life cycle or in the supply chain.
18 Risk M ana gement P ractices for Eucalyptus Prevention Prevention, which can be described as avoidance through appropriate risk ass essments and quarantine enforcement (McCormick and Howard 2013), is often considered the most cost effective approach for dealing with biological invasions (Leung et al. 2002). Accordingly, several authors have urged that tools such as the Australian Weed Risk Assessment (WRA) and quarantined experimental introductions be conducted to predict invasion risk posed by bioenergy crops (Barney and Di T omaso 2008; Chimera et al. 2010; Davis et al. 2010; Gordon et al. 2011; Flory et al. 2012) The WRA uses current weed status in other parts of the world, climate and environmental preferences, and biological attributes, to conclude if a species should be accepted, rejected or evaluated further for introduction (Pheloung et al 1999) (Table A 1) Conclusions from the WRA are used to identify low risk species for which a supply chain should be developed and high risk species that should not be considered as bioenergy feedstock candidates (Dodet and Collet 2012; Gordon et al. 2 012 ) It is an exceptionally valuable tool because it has been demonstrated to correctly identify invaders 90 percent of the time and non invaders 70 percent of the time (Gordon et al. 2008) and an assessment can typically be completed in five to eight hou rs with little cost (Daehler et al. 2004). Rejmnek and Richardson (2011) argue d that there is a particular need for tools such as the WRA to evaluate Eucalyptus and other groups that are widely planted but so far have proportionately low incidence of inv asiveness Accordingly, Eucalyptus invasion risk was recently evaluated for 38 of the most commercially important species using the WRA (Gordon et al. 2011, 2012) A study adjusted for the national scale
1 9 concluded that 15 of these Eucalyptus species have l ow risk of invasion, 14 have high risk and 9 require further information (Gordon et al. 2012). When the WRA was modified the invasion risk of three Eucalyptus species it concluded that E. grandis and E. camaldulensis posed high invasion risk and that further information was required for E. amplifolia (Gordon et al. 2011). While these results are useful, this approach to assessing Eucalyptus invasion risk is limited because assessment s need to be performed for each can didate genotype and cultivar individually because plant structure and function hence invasive potential, can vary widely within a species (Casler et al. 2004 ; Flory et al. 2012 ). Life history i nformation specific to new Freeze Tolerant Eucalyptus (FTE) l ines (ArborGen, Ridgeville, SC) is still largely unknown and evaluation of these cultivars using the WRA is not possible Additionally, concerns over assessor subjectivity and inconsistency (Davis et al. 2011, Gordon et al. 2008) and difficulties regarding the interpretation of the WRA questions for biomass crops (Barney and Di T omaso 2008, Flory et al. 2012) must be weighed Question s also remain about how to treat the nine Eucalyptus species that fall into the evaluate further category. While qualitativ e risk assessment by the WRA has been developed extensively existing empirical research to investigate Eucalyptus invasion risk is limited The only instance of empirical research in the southeastern US consisted of an observational study in which surveys for natural r ecruitment were conducted with in and proximate to Eucalyptus stands (Callaham et al. 2013). These authors detected limited establishment of E. amplifolia E. grandis and E. r obusta seedlings in modified land use types at latitudes south of 27 N Other examples of empirical research include t wo experiments
20 that found limited invasive potential of E. s aligna E. grandis and E. urograndis in Brazil, (da Silva et al. 2011; Emer and Fonseca 201 0 ) Notably, no studies reported the use of experiment s to test Eucalyptus invasiveness in the southeastern US While broad application experiments as part of a bioenergy crop screening protocol has been proposed as a way to inform proposed regulatory restrictions (Davis et al. 2010), experimental tests for a re still rare. This is possibly due to uncertainty about how to translate response variables such as seedling survival and growth into meaningful measures of risk that can be easily interpreted by policymakers (Table A 1) The absence of standardized proce dures or agency mandate for the high cost and oversight of experimental evaluations (Flory et al. 2012) have also limit ed the effectiveness of this approach. Containment Containment refers to management practices that decrease the likelihood of spread fro m a site (McCormick and Howard 2013). One potentially very effective containment approach for managing Eucalyptus invasion risk is trait selection during breeding (Table A 1) The likelihood of spread can be reduced by decreasing fecundity or by increasing the age to maturity, although the later method may negatively influence productivity (Gordon et al. 2012). This strategy was successfully implemented in other taxonomic groups, including sterile clones of Pinus species used in South Africa and triploid hy brid Leucaena in Hawaii (Richardson 1998). Likewise, elimination of seed production is thought to be a feasible goal in Eucalyptus (Gordon et al. 2012) and elimination of pollen production has already been accomplished in the transgenic hybrid, E. grandis x E. urophylla (AGEH427) (Hinchee et al. 2011). The ability to ensure containment of genetically modified trees through sterility is significant because it
21 eliminates the need for costly, imprecise and complex ecological research to understand and predict the impacts of spread (FAO 2010a). However, the major limitation to this approach is that the effectiveness of containment technology is uncertain due to relatively novel use in forestry (FAO 2010a). G enetic research is also costly (Wang and Brummer 2012) and p rogress in the commercial development of genetically modified organisms is slow under the current stringent regulatory structure (Strauss and Viswan a th 2011). Further consideration must also be given to the acceptability of genetically modified tree s (Dodet and Collet 2012) and the possibility of controversy over intellectual property concerns similar to those that have arisen for agriculture (FAO 2010a). Contrary to the use of trait selection to hinder potential for spread, a key possible risk mana gement step is disallowing the development of trees that may be harder to control due to genetic modifications that confer herbicide tolerance or insect resistance (Table A 1) Examples of hazardous traits exist in Australian developed E. camaldulensis tha t is resistant to chrysomelid beetles and toler ant to broad spectrum glufosinate ammonium herbicide (Harcourt et al. 2000) although this cultivar has not been introduced in the United States However the USDA Animal and Plant Health Inspection Service (A PHIS) has permitted field trials and received petition s to deregulate Eucalyptus cultivars that have altered cold tolerance, fertility, lignin biosynthesis, growth rate, or selectable markers (Harfouche et al. 2011). Whether cold tolerance and increased gr owth rate are traits that would increase vigor in a way that promotes invasiveness is a matter that requires more research.
22 In addition to containment practices using genetic technology, other practices th at create of temporal boundaries to dispersal may also decrease the movement of propagules. Harvesting trees before seed maturation could significantly reduce the possibility of spread (Table A 1) Harvesting within six months of floral onset is recommended for conventionally bred cold tolerant E. grandis which the WRA predicted to have high invasion risk (Flory et al. 2012). However, observance of this recommendation is not compulsory under existing regulations of biomass plantings ( FAC 2008 ) and the degree to which this recommended practice has been exe cuted is unknown. G rowers may oppose this management practice because flowering and seed maturation may occur earlier than the optimal harvest age. For example, E. grandis is known to flower as early as 2 to 3 years after germination (Hodgson 1976) but th e age of maximum sustained yield is 2.7 to 2.9 years with even longer optimum economic rotation lengths (Langholtz et al. 2005). Moreover, Eucalyptus are often managed to produce several coppice harvests within a rotation and information about the effect o f coppi ci ng on flowering and seed production is not readily available. H arvesting relative to unpredictable seed development would make it difficult for growers to make sound economic decisions, which require prior knowledge of the length of each stage of the rotation (Langholtz et al. 2005). The use of physical barriers to dispersal such as buffer zones around plant ings is generally recommended (Barney and DiTomaso 2008; Flory et al. 2012) and is required in the state of Florida under a unique biomass plan ting permitting system ( FAC 2008) (Table A 1) Two strategies exist regarding buffer zones: establish a stable vegetation community to impede seedling establishment, or maintain fallow areas by regular
23 burning or mechanical or chemical control to disturb seedlings and to enhance detection of small sized individuals through monitoring (Ledgard 2001). While maintenance of buffer zones is enforceable through random checks, uncertainty exists about how large the barrier needs to be in order to prevent seed dis persal into proximate susceptible communities (Flory et al. 2012). The Institute of Food and Agricultural Sciences (IFAS) of the University of Florida recommends that a 75 ft barrier be established, while Florida law requires only a 25 ft barrier ( FAC 2008 ; Flory et al. 2012 ) However, evidence regarding patterns of seed dispersal suggests that neither of these sizes may be adequate. Seed dispersal occurs primarily via wind and seeds are deposited within a radius of twice the tree or canopy height (Cremer 1 977), which is approximately 100 ft for E. grandis that are known to grow up to 49 ft in height by 3.5 years of age (Rockwood et al. 2006). The possibility of seed dispersal beyond this area may also be increased due hurricane force winds, which are theore tically able to transport the fine seeds of confamilial Melaleuca quinquenervia a maximum distance of 7.1 kilometers, contributing to increases in populations following hurricanes (Browder and Schroeder 1981). The e stablishment of production areas as monoc lonal plantings can also serve as a physical barrier to contain spread (Flory et al. 2012) (Table A 1) D ifferent cultivars should remain separate in order to decrease risk of hybridization which may have a positive effect on invasive ability through incr eased vigor (Lee 2002). Single large blocks of Eucalyptus would also have comparatively less plantation border than dispersed plantings, resulting in less adjoining area that could receive wind dispersed seed. Decreasing the amount of plantation border is also critical because these areas
24 have been hypothesized to be more vulnerable to colonization by volunteer seedlings (da Silva et al. 2011). Additionally, g rowers m ight embrace the practice of monoclonal plantings for Eucalyptus considering the numerous benefits that pure culture provides including simplicity of management, predictability of yield and potentially increased yield compared to seedling based plantings in which trees may have non uniform growth (DeBell and Harrington 1993; Zalesny et al. 2011 ). A significant disadvantage of this practice is increased risk of total loss due to disease when entire stands are composed of identical genetic stock (DeBell and Harrington 1993). Regular c leaning of seeds and plant material from harvesting equipment i s another generally recommended practice for limiting dispersal of bioenergy crops (NISC 2009), although it has not been argued for in the context of Eucalyptus (Table A 1) It is reasonable to predict that grower willingness to participate in a cumbersome management activity may be low, especially when budgetary limitations and other higher impact priorities of regulatory agencies make enforcement impractical. Another shortcoming of this practice is that it only addresses a minor dispersal vector and does not affect wind dispersal of seeds. Additionally, as more money and time is spent cleaning equipment, efficiency decreases and thus there are diminishing returns to cleaning (Leung et al. 2005). Control Additional m anagement practices that aim to decreas e the spread of an already established invader represent a shift from the previously discussed proactive practices, to a reactive approach. Removal of Eucalyptus has proven to be challenging due to their habit of mass sprouting from the base and roots in r esponse to injury (NPS 2006). There has been experimentation with a wide variety of control methods, many of which
25 remain complex, expensive and inefficient (NPS 2006). Mechanical control treatments which are generally considered the least damaging control method for surrounding environments (Bean and Russo 1989), were applied in management projects and experienced varying levels of success (Table A 1) Burning of stumps and coppice sprouts was an ineffective method of control (Bean and Russo 1989; Little a nd van den Berg 2006). Stump grinding in which stumps were ground down to two feet below the ground surface (Bean and Russo 1989, NPS 2006), was used in management projects alt hough regrowth sometimes occurred This method was also notably labor intensive and costly (NPS 2006). Stump light deprivation (tarping) in which thick plastic was stapled to the stump and the stump was then buried under mulch was also used to stop regrowth (NPS 2006) and m oderately increased mortality was observed when tarping was combined with herbicide treatment (Bean and Russo 1989). Whereas mechanical control is advantageous for some conservation objectives, c hemical control using herbicides ha ve generally been the most successful method for controlling Eucalyptus (Table A 1) R ecommended chemical methods include cut stump, basal bark, or basal frill applications of triclopyr, imazapyr or glyphosate herbicides though complete control of regrowth is not typically provided by single applications of herbicides using standard proto cols ( Bachelard et al. 1965; Bossard et al. 2000; Moore 2002; Little 2003; Little and van de n Berg 2006 ) Results of previous research and specific research needs for improving Eucalyptus control by herbicide treatments are further described in C hapter 3 o f this thesis Biological control (biocontrol), the control of invasive species populations through the introduction of predators or pathogens, is a management option that is typically
26 considered unacceptable when a species has economic value (Hoffmann et al. 2011), especially considering the costly and complicated research (Hobbs and Humphries 1995) and the lengthy regulatory process for approval of biocontrol agents (Montgomery 2011). While biocontrol was deemed an unlikely option for Eucalyptus (Rejmne k and Ric hardson 2011) and no published methods were found regarding the intentional release of biocontrol agents for Eucalyptus some authors have speculated that biocontrol may have essentially occurred through accidental introductions of Eucalyptus harm ing insect pests in California (Paine et al. 2011) (Table A 1) O ptions for biocontrol candidates exist among the 15 different Australian Eucalyptus feeding insect species from at least four different feeding guilds (2 borer species, 3 leaf eating beetle s pecies, 4 gall wasp species and at least 8 psyllid species) that have been introduced into California, Florida and Hawaii (Paine et al. 2011). If biocontrol is sought as a management strategy biocontrol programs for invasive Acacia in South Africa can ser ve as successful model s t hat balance social and ecological concerns. B iocontrol agent selection in South Africa focused on agents that only attack the flower buds, flowers or seed pods, to minimize the impact on commercial production but also reduce the c osts of follow up management and spread rates (Wilson et al. 2011 b ). Similarly, potential biocontrol efforts for Eucalyptus in the southeastern US might look to the galling wasp, Quadrastichodella nova which has been reported to infest seed capsules (Pain e et al. 2011). Impact M anagement Long term management practices for mitigating the ecological impacts of invaders have improved vastly in the past decade and are now viewed as useful and financially viable supplements to prevention and containment pract ices (Simberloff et al
27 2012) A number of practices may be used to minimize the impacts of Eucalyptus cultivation before and during commercial production Buffer zones around plantings mentioned previously for the purpose of limiting seed dispersal, can also provide surface water and wildfire protection by limiting the proximity of trees to waterways and by establishing a firebreak around the stand (Booth 2012; Flory et al. 2012) (Table A 1) The impacts of wildfires can also be minimized by reducing the density of the fuel load which accumulates below trees through the annual shedding of bark and limbs (NPS 2006) by harvesting trees before they begin to shed bark abundantly and by controlling the understory shrub layer (Goodrick and Santurf 2012). While addressing surface water depletion and fire hazard is relatively straightforward some ecosystem impacts that are consequential but are not readily detected (Simberloff et al. 2012) (Table A 1) Impacts such as reduced biodiversity and habitat loss which are not easily estimated through economic valuation and which are inherently incompatible with the presence of the large scale Eucalyptus plantings may have significant hidden costs. These costs are also known to rise as mitigation is delayed so it is important that the long term management of invasion impacts be viewed as the last option after unsuccessful prevention or containment ( Simberloff et al. 2012). Implications of Policy Alternatives There are many ways that Eucalyptus invasion risk can be ma naged, although no single option is ideal in a biological, economic and social sense and many require considerable further research. Challenge lies ahead in deciding what practices should be considered a priority (Flory et al. 2012). Consideration also nee ds to be given as to what kind of regulatory techniques should be used to implement policies regarding the management of non native biomass crops and who should assume the burden of
28 management Various authors have suggested that importers, developers and growers who are responsible for introducing potentially invasive crops such as Eucalyptus should be responsible for damages to the environment (i.e. pays principle) rather than allowing that burden to be borne by tax payers or neighboring priva te landowners who are affected (Buddenhagen et al. 2009 ; Bradley et al. 2010 ; Chimera et al. 2010 ; Davis et al. 2010 ; Witt 2010; McCormick and Howard 2013) However Florida is the only state thus far that has adopted any legal authority governing the uses of non native species in biofuel produ ction (Environmental Law Institute 2010) and this novel regulation employs only containment and impact management approaches as they pertain to growers (FAC 2008) While the cost of field testing is assumed by the de veloper in the case of genetically modified crops such as FTE Eucalyptus regulated by APHIS, field testing to determine negative environmental impacts is not required for traditionally bred non native biomass crops including most of the Eucalyptus species currently being considered for feedstocks. Chimera et al. (2010) proposed that stringent protocols analogous to those for testing of genetically modified organisms should be developed for all biomass crops and that the expense of conducting such evaluation s be the responsibility of the developer However, stakeholders may be resistant to the adoption of more rules governing the cultivation of biomass crops, which may threaten the viability of a market that already has low profit margins. Additionally, bioma ss crop regulations have been unsuccessful elsewhere in the world. For example, laws that assigned growers of invasive Acacia species the responsibility of controlling seed spread failed because most landowners and growers had insufficient resources and th ere was a lack of commitment to prosecute offenders (van Wilgen et al. 2012). This
29 historical ineffectiveness of regulations may diminish support for the implementation of similar rules in the US at a time when many new growers are entering the market Re latively less discussion has emerged about the potential utility of a softer approach involving informal social control of non native bioenergy crops, although such an approach may be valuable considering the many previously reviewed uncertainties over the environmental risks associated with Eucalyptus culture Informal control through voluntary third party certification programs and promotion of sustainable forest management practices have long been important policy tools in the forestry sector (Ramenstein er and Simula 2002). Certification systems, which were developed to address public concerns related to deforestation and biodiversity loss, are well equipped to develop sustainability criteria for mitigating invasion risk of biomass crops, and to fulfill s uch criteria through proper enforcement and verification mechanisms (Lewandowski 2006). Indicative of this is a recent study finding that 10 of 17 existing agricultural or forestry certification schemes already included some criteria related to invasive sp ecies (Scarlat and Dallemand 2011). A criterion to prevent invasive bioenergy feedstock species from spreading outside the operation site is also included in the Roundtable on Sustainable Biofuels (RSB) certification system, which was launched in 2011 as t he first certification specifically for biofuel supply chains (RSB 2010). Inclusion of this criterion in certification schemes provides a way to persuad e landowners to plant cultivars that are predicted to present a low invasion risk, to maintain appropria te buffer zones, to control and monitor for escaped seedlings, and to carry out impact management practices. Certification schemes may also act as a vehicle for stakeholder education and knowledge sharing about how to improve the
30 economic and ecological ef ficiency of such protocols (Gootee et al. 2012). Most importantly, informal control is not bound by public officials who are constrained by the election cycle and local interests or by lengthy legal processes, which are cumbersome barriers to change for r egulations. Consequently informal control allows greater flexibility to respond to new developments in scientific information, which is critical for managing Eucalyptus in light of existing gaps in knowledge. Conclusion and Needed Research Eucalyptus spe cies are considered potentially invasive species and sources of environmental problems in various regions where they have been introduced, so recent concerns about potential invasiveness in the southeastern US demand attention. While producer are widely recommended, they are certainly not the only answer or even the best answer in an economic or social sense. Minimizing invasion risk could also be achieved through informal relationships, which place a lesser burden on the biomass producer and represent a flexible approach that is desirable under uncertain risk. As with the management of any invasive species, success is likely to require a landscape level response that integr ates a wide range of management practices implemented by different stakeholders at different stages of the invasion process. Studies made significant progress in the identification of risk factors and the specific level of risk posed by certain Eucalyptus species. However, further evidence is needed to improve the confidence of existing predictions. Management tools have also identified and include genetic, physical and temporal containment methods as well as herbicide and biocontrol techniques, although re search is also needed to improve the effectiveness of many of these methods. The purpose of the research that is described in the subsequent
31 chapters of this thesis was to further evaluate the potential invasiveness of Eucalyptus species being considered f or large scale planting in Florida and to improve practices for control of Eucalyptus using a new herbicide.
32 CHAPTER 2 POTENTIAL EUCALYPTUS INVASIVENESS IN FLORIDA'S NATIVE AND MODIFIED PLANT COMMUNITIES Introduction Invasion risk is shaped not only by species life history traits but also by s tochastic processes and local interactions with the biotic and abiotic features of ecosystems in the new range that ; Minton and Mack 2010). These complex p rocesses and relationships are often non linear and affected by temporal lag effects and positive feedbacks (Hulme 2011) Therefore it is difficult to predict invasions based on qualitative information using literature based tools like the WRA. While the W RA offers a useful starting point, more quantitative tests are needed to evaluate the many species and taxa of biofuel crops that are being proposed for introduction (Barney and DiTomaso 2008 ; Chimera et al 2010 ; Davis et al 2010 ; Hulme 2011). Several st udies have used experiments to assess factors contributing to invasion risk (Myers 1983; Minton and Mack 2010; Davis et al. 2011) Recent work synthesize d these past experimental tests of invasion potential and summarize d the key elements of a standardized protocol for testing invasion risk (Flory et al. 2012). Key recommendations were that controlled introductions should be performed at multiple sites that represent the communities that are susceptible to invasion (Ewel et al 1999 ; Parker and Kareiva 1996 ) and that factors important for establishment and performance such as disturbance, founder population size and timing of introduction be assessed as experimental variables (Flory et al. 2012). Previous observ ational and literature based studies (e.g., Ca llaham et al. 2013; Gordon et al. 2012) yielded valuable in sights about what Eucalyptus species have a high risk of be coming invasive and what areas might be at risk of colonization by
33 Eucalyptus seedlings. However, e xperimental evaluation of invasion risk w as needed to understand invasion risk relative to complex ecological processes and to resolve differing conclusions from observational and literature based work about the likelihood of introduced Eucalyptus species becoming invasive. Our study heeded the recent recommendations for experimental tests of invasion risk (Flory et al. 2012) and aimed to identify the invasive potential of three Eucalyptus species and to determine conditions conducive to natur alization in different vegetation communities. Of par ticular interest for Eucalyptus were the effects of seed density and site disturbance on seedling emergence and survival Disturbances such as vegetation removal were important to take into account because they have long been recognized as possible import ant drivers of invasion (Elton 1958 ) and have been suggested to be a prerequisite for Eucalyptus establishment (Wevill and Re a d 2010). High propagule has also been observed as one of the mechanisms responsible for tipping the bal ance to invasion (Gordon 20 11; Hobbs & Humphries 1995) and it is thought to be an especially important factor for Eucalyptus population performance Eucalyptus propagule pressure can potentially be very high considering that seed rain from mature Eucalyptus can be up to 4,000 seeds m 2 (Richardson and Rejmnek 2011) The over arching objective of this study was to evaluate the potential invasiveness of three Eucalyptus species being considered for large scale planting in Florida to supply bioenergy and fiber through two quantitative a pproaches. 1. Site surveys were conducted to determine the abundance and height distribution of Eucalyptus seedling recruitment within seed bearing stands and in the proximate plant communities where seed dispersal may occur. Plant communities included upland hardwood forest, non grazed pasture, intensively site prepared forest land and abandoned forest road.
34 2. Seed addition studies were conducted to determine the relative potential for seedling emergence and survival among E. amplifolia E. camaldulensis and E. grandis sown at two seed densities and for disturbed and non disturbed conditions in the understory of reproductively mature Eucalyptus and in each of the aforementioned plant communities proximate to the Eucalyptus stands. Materials and Methods Study Ar eas Separate studies were conducted in the proximity of reproductively mature Eucalyptus stands at two locations in Florida. The Gainesville location was an E. amplifolia seed orchard on the University of Florida campus (2937'36" N, 8221'32" W) at appro ximately 23 m elevation. Eucalyptus were planted in this 0.7 ha stand at various times from 1992 to 1997. There had not been any vegetation management in the two years prior to initiating this study. The stand is adjacent to native upland hardwood forest a nd an abandoned forest road. The Quincy location is a 0.9 ha E. amplifolia progeny test planted in 1999 at the North Florida Research and Education Center, south of the city of Quincy (3032'32" N, 8435'25" W) at approximately 7 3 m elevation. This stand i s adjacent to non grazed pasture and intensively site prepared forest land. The intensively site prepared forest land previously supported a 13 year old eastern cottonwood ( Populus deltoides, W. Bartram ex Marshall) clone test harvested in January 2012 and was site prepared for planting Eucalyptus in March 2012 by the rake/pile/burn/disk method (Lowery and Gjerstad 1991). Eucalyptus amplifolia stands at both locations have relatively wide spacing between trees and some direct sunlight in the understory, pot entially fostering seedling establishment (Booth 2013). The Gainesville location has a temperate climate with highest temperatures in July (mean 27 C), lowest temperatures in January (mean 13 C), an average annual extreme minimum temperature of 6.7 to 3. 9 C (USDA hardiness zone 9a) and 125 cm average annual
35 precipitation (NOAA 2002; USDA 2012a). The Quincy location has highest temperatures in July (mean 27 C), lowest temperatures in January (mean 10 C), an annual average extreme minimum temperature of 9. 4 to 6.7 C (USDA hardiness zone 8b) and 143 cm average annual precipitation (NOAA 2002; USDA 2012a). Site Survey s In May 2012 line transect sampling was used to identify natural recruitment of E. amplifolia seedlings from the mature Eucalyptus trees At each study location, 1 m 2 plots (72 Gainesville, 238 Quincy) were evaluated using a sampling frame placed every 10 m on line transects established 20 m apart across the narrow dimension of the stand. Line the adjacent communities. Eucalyptus seeds are typically dispersed by wind within an estimated radius equal to twice the canopy or tree height (Cremer 1977), and 60 m is approximately twice the canopy height of the se E. amplifolia trees in these stands S eed A ddition S tudies Germination testing Germination capacity of Eucalyptus species is highly variable, ranging from 11 % to 98% (USDA 2008) Prior to initiating the seed addition experiment, the expected g ermination E. amplifolia E. camaldulensis and E. g randis seed stocks were determined by measuring the mean number of germinating seed per gram for each species, so that the amounts seed added into plots had an equal number of seeds that were expected to germinate (Table 2 1) Eucalyptus seeds are typicall y 1 3 mm long and weigh less than 0.5 mg up to 2.0 mg depending on species (Rejmnek and Richardson 2011). S eeds are also mixed with chaff (inert material) in seed capsules and the proportion of chaff by weight can range from 5:1 to 30:1 (USDA 2008). There fore,
36 the use of weight specific expected germination number was preferable because the small size of Eucalyptus seeds and the intermixed chaff made it difficult to count individual seeds accurately Using a protocol adapted from previously established gui delines for Eucalyptus germination testing (Boland 1986; USDA 2008), expected germination was evaluated in a controlled environment growth chamber. Growth chamber conditions corresponded to average minimum (20.1 C) and maximum (32.5 C) temperature conditio ns (NOAA 2002) and 14 hr photoperiod for June in north central Florida (Naval Meteorology and Naval Command 2012). Light was provided with the higher temperature for 14 hr followed by the lower temperature without light for 10 hr. While some alpine Eucalyp tus species require cold moist stratification to break dormancy (Boland 1986), the three species used in this study are not known to be among these. Eight 0.05 g samples of e ach seed lot were tested for each species. All materials that were exposed to seeds during preparation of the samples were autoclaved to reduce the likelihood of microbial growth. Samples were placed in 9 cm glass Petri dishes containing two thicknesses of filter paper as a substrate. Three ml of sterile distilled water was applied to th e filter paper using a glass pipette. Seeds were dispersed evenly over the filter paper and the Petri dishes were sealed with paraffin film to prevent samples from desiccating. A germination count was made after 5 d and every 2 d thereafter. During each c ount, normal and abnormal germinated seeds were counted and removed from the dish. A normal Eucalyptus seedling had a healthy radicle, hypocotyl and cotyledons. It was recommended that germination tests for E. grandis last 14 d, and whil e there was not spe cific informa tion available about the test duration for E. amplifolia and E. camaldulensis 10 to 21 d is generally suitable for most
37 species (Boland 1986 ). Thus, th e final count was made at 14 d and the total number of normal germinated seeds in each samp le was used to calculate the average number of germinating seed per g for each seedlot Experimental design The potential for Eucalyptus seedling establishment and survival in common native or modified vegetation communities was evaluated through a seed a ddition approach similar to da Silva et al. (2011). S ampling plots (Figure 2 1) were established within the understory of the E. amplifolia stand and in t wo proximate communities at both study locations, resulting in 63 plots at each location At the Gaine sville location, proximate communities included an upland native hardwood forest and an abandoned forest road. At the Quincy location, proximate communities included a non grazed pasture and an intensively prepared site for Eucalyptus planting, as describe d above. The characteristics of the communities studied are presented in Tables 2 2 and 2 3 In each vegetation community seeds from E. amplifolia E. camaldulensis and E. grandis were placed on separate nested paired plots to examine the effects of dist urbance and seeding density for each species (Figure 2 1). Adjacent 0.75 m 2 paired plots were randomly left non disturbed or were disturbed through removal of the vegetation layer and soil scarification with a rake. In the center of each of the paired plot s, two nested 0.25 m x 0.25 m paired subplots were randomly assigned to addition of amounts of seed expected to result in either 500 or 1000 germinating seeds m 2 These levels were chosen to reflect the hypothesized intermediate and maximum levels that oc cur naturally based on estimates of actual seed rain in the literature (Virtue and Melland 2003). The 0.625 m 2 non seeded areas outside of the smaller nested paired seeded subplots within the disturbed or non disturbed treatments were used to observe
38 Eucal yptus seedling establishment from mature trees Randomly distributed paired plots were replicated eight times in each vegetation community type for E amplifolia and E. grandis, and only five times for E. camaldulensis due to a shortage of seed. Seeds wer e sown on June 7, 2012 in Quincy and on June 8, 2012 in Gainesville, following 2.0 cm and 13.1 cm of rain (tropical storm event) in the previous 24 hours at each location, respectively (IFAS 2012). Seedling assessments A census was taken at 2, 4, 6, 8, 1 1, 14, 17, 21 and 25 weeks after treatment (WAT) with seed to determine emergence and survival of Eucalyptus seedlings. Newly emerged seedlings at each census were counted and marked with uniquely colored toothpicks to distinguish them from new seedlings i n subsequent censuses. Surviving seedlings of each emergence date group were counted at each census so that the approximate length of time that seedlings survived could be determined. Statistical analysis Seedling emergence Statistical analyses were perf ormed using SAS v. 9.3 (SAS Institute, Cary, NC). Because c umulative Eucalyptus seedling counts from the periodic plot assessments were generally too low to evaluate the effect of seeding density treatments, emergence data were combined for high and low de nsity seeded subplots within the disturbed or non disturbed treatments. Emergence data were analyzed as the proportion of seeds that emerged relative to the expected number of germinating seeds m 2 ( e.g., 62 seeds expected to germinate in high density subp lot + 31 seed s expected to germinate in low density subplot = 93 seeds expected to germinate in paired subplots ). A generalized linear model (PROC GLIMMIX) with a binomial distribution and a logit link function was used to determine the effect of
39 treatment s on percent emergence and to estimate mean percent emergence for treatment combinations using data from both study locations. This approach has been used for determining effects of treatments on very low proportion germination data in similar experimenta l designs (Humber and Hermanutz 2011; Carillo Gaviln et al. 2012). Vegetation community, species, disturbance, disturbance by species and species by vegetation community were considered fixed effects. The three way interaction was dropped from the model b ecause it was not significant. The random effect of plot within each community could not be estimated and was dropped from the model due to nonconvergence when this effect was included The significance level for all tests was = 0.05. An additional analy sis was performed to capture the effect of disturbance, community and location experimental variables on Eucalyptus seedling emergence due to recruitment by the mature Eucalyptus Since seedling emergence was very close to zero in seed addition subplots an d because seed rain from the Eucalyptus canopy could not be excluded from these parts of the plot, total emergence counts for seeded and non seeded areas were combined. Because v ariation across plots could not be estimated analyses were performed on the t otal number of emerged seedlings summed across seeded and non seeded areas of the 21 0.75 m 2 disturbance treatment plots (total 15.75 m 2 ) within each of the various vegetation communities at the two locations. In order for the model to converge, v egetation communities that had zero emerged seedlings were not included in this analysis. A generalized linear model (PROC GLIMMIX) with a Poisson distribution and a log link was used because the total number of seeds in plots was unknown. Vegetation community (nes ted within location) and
40 disturbance were treated as fixed effects. Contrasts were used to compare combinations of factor level means at = 0.05. Seedling survival Data for the time that each seedling was first observed to when it was last observed were used to calculate the minimum and maximum possible time of survival for each of the 62 seedlings that germinated. The middle value of this range was assigned to each seedling to approximate time of survival. The significance of species, disturbance and ve getation community for time of survival were each evaluated for survival values classified as a categorical response (< 1 month, 1 to 2 squared tests under the Proc Freq procedure in SAS. Analyses were performed for the Gainesville and Quincy locations pooled together because the difference in time of survival between locations was not significant. The time of survival by species analysis were performed to compare seeding treatments in subplots and in whole plots to c apture species effects considering the possibility of movement of the added seed. Results and Discussion Site Surveys The 1 m 2 sampling plots in the Eucalyptus stands and surrounding communities included various types of ground cover including grasses, vi nes, forbs, shrubs, trees, bare mineral soil and leaf litter (Tables 2 4 and 2 5 ). No Eucalyptus seedlings were found in any of the 72 plots at the Gainesville study location or in the 238 plots at the Quincy location. The lack of natural Eucalyptus recrui tment observed in these surveys is consistent with results of similar studies in which natural recruitment of Eucalyptus seedlings were found in only 4 out of 16 surveys within and proximate to Eucalyptus stands in the Florida (Callaham et al. 2013). Simil ar surveys in Brazil found recruitment
41 with in Eucalyptus stands, but not in adjacent pine communities (Emer and Fonseca 2010). Seed Addition Studies Seedling emergence in seeded subplots The probability of Eucalyptus seedling emergence in seed addition su bplots was very low overall (0.0 0.32%) (Tables 2 4 and 2 5 ). Mean percent emergence was generally higher within disturbed (0.0001%) than in non disturbed areas (0.000001%) (Table 2 4 ). The highest percent emergence for the three tested Eucalyptus species was observed for E. camaldulensis (0.00028%) (Table 2 4 ). The Gainesville Eucalyptus stand had the highest percent emergence (0.0005%) among the various vegetation communities studied (Table 2 4 ). Correspondingly, the vegetation community by species combi nation that yielded the highest percentage of emerged seedlings was E. camaldulensis in the Gainesville Eucalyptus stand (0.32%) (Table 2 5 ). However, differences in the probability of emergence for disturbance, species and vegetation community factors wer e not significant. Da Silva et al. (2011) reported similarly low levels of Eucalyptus emergence in seed addition studies in Brazilian plant communities. W hen approximately 5,000 total seeds were added, only 111 seedlings were observed across 5 periodic cen suses (2.22% emergence). In their study, new seedlings were not distinguished from survivors, so the actual number of seedlings that emerged may have been lower if they inadvertently counted survivors as additional seedlings in subsequent censuses. Seedli ng emergence in non seeded areas Eucalyptus seedlings were observed in non seeded subplots beginning in early August, approximately eight weeks after the studies were initiated. The random
42 distribution of seedlings throughout the plots and the presence of open Eucalyptus seed capsules on the ground in plots at this same time suggested seedlings likely occurred in non seeded areas and seeded subplots as a result of recruitment from the mature Eucalyptus trees. While significant treatment effects were not det ected for emergence in seed addition subplots, significant effects were detected in the analysis that summed the total number of emerged Eucalyptus seedlings across seeded and non seeded areas to capture treatment effects on recruitment by the mature trees Location and Community No seedlings ever emerged in the native hardwood forest in Gainesville or in the intensively prepared land or non grazed pasture communities in Quincy. The total number of emerged Eucalyptus seedlings in the remaining vegetation c ommunities varied significantly by community (F 2,2 = 22.55, P = 0.043). Contrasts revealed that Eucalyptus seedling emergence was significantly greater in plots in the Eucalyptus stand in Gainesville (52 seedlings) compared to the stand in Quincy (7 seedl ings) (F 1, 2 = 24.81, P = 0.038). Both communities had more seedling emergence than forest road. Research by others has also shown greater recruitment for Eucalyptus seedlings within Eucalyptus stands compared to other vegetation communities. Emer and Fonse ca (2010) found that transplanted seedlings survived for an average of 9.3 months within a Eucalyptus stand, but only survived 2.7 to 5.3 months in other communities. These authors conducted s urveys for natural recruitment of Eucalyptus seedlings that also revealed the same trend. Emer and Fonseca (2010) reported that Eucalyptus seedling recruitment was not found in surveys of native pine forests and pine plantations, but did occur within mature Eucalyptus stands. Likewise, in their surveys adjacent to Euca lyptus stands, Callaham et al. (2013)
43 found 65 seedlings in Eucalyptus stands, but only 13 in neighboring wetland, four in pasture, two in pine, one in open forest and zero elsewhere (lawn, roadside, disturbed soil etc.). While these experiments cannot ful ly explain the mechanism driving the decreased success outside of Eucalyptus stands other authors hypothesize d that seed rain is not a limiting factor and that seedling mortality in native communities is related to plant community richness, plant abundanc e, soil fertility (Emer and Fonseca 2010) and also light availability (Booth 2013). In our experiment, mean total emergence including natural recruitment in plots was high as 28.3 thousand seedlings ha 1 for one combination of factor levels (Table 2 6 ). Th ese data support the hypothesis seed rain is not a limiting factor for Eucalyptus establishment. While this experiment did not provide information about the specific drivers of greater Eucalyptus seedling emergence in the Gainesville Eucalyptus stand compa red to the Quincy stand, possible influential factors include soil composition, canopy cover and weather near the beginning of the experiments. While both Gainesville and Quincy had similar mean temperatures (24.9 C and 24.7 C) throughout the first month o f the seed addition studies, a tropical storm event caused 13.3 cm of rain fall in the week following seeding treatments in Gainesville, while only 1.8 cm of rainfall occurred in Quincy. Although there was not conclusive evidence to evaluate the specific dr ivers of location differences for this study, the results of Callaham et al. (2013) indicated that invasiveness might be related to latitude. In their surveys in and around Eucalyptus stands throughout the southeastern US, they reported that no Eucalyptus seedling recruitment was seen north of 27 N which is approximately the southern limit of the range where freezing conditions are expected to occur annually (USDA 2012a). The
44 results of these experiments, which were located at 29 N (Gainesville) and 30 N (Quincy), support their conclusion that 27 N may represent a latitudinal threshold for establishment. Disturbance The effect of disturbance on Eucalyptus seedling emergence due to recruitment by mature Eucalyptus was also significant (F 1,2 =24.77, P = 0.038), with greater emergence occurring under disturbed conditions. The effect of disturbance is evident when Eucalyptus seedling emergence was compared for disturbance treatments within communities on a per hectare basis. The generalized linear model es timated that d isturbed treatments in the Gainesville Eucalyptus stand had a mean seedling emergence equivalent to 15.0 to 53.5 thousand seedlings ha 1 (95 % confidence interval), while emergence in this same community without disturbance was 1.1 to 20.2 th ousand seedlings ha 1 (Table 2 6 ). Likewise in Quincy, mean seedling emergence was estimated to be between 738 and 19.7 thousand seedlings ha 1 under disturbed conditions but was notably lower under undisturbed conditions (635 to 5.2 thousand seedlings ha 1 ).This trend was expected considering that disturbance improved germination of Eucalyptus seedlings in other studies (da Silva et al. 2011). This result supports the suggestion of Wevill and Read (2010) that rare disturbances such as fire or flooding whic h suppress competition are necessary for Eucalyptus to establish. Seedling s urvival Most seedlings survived less than one month, although a few lived up to approximately 3 months (Figures 2 2 and 2 squared tests found no significant effe ct s for plant community or disturbance on the time of survival. Species was not a significant factor when evaluated for seeded subplots. However, when the
45 effect of species on survival was evaluated for whole plots at both locations together, species emerg ed as a significant factor affecting survival ( 2 = 9.72, 1 df, n = 58, P = 0.045). While all 3 species had a similar number of seedlings that survived up to 2 months, only seedlings in E. camaldulensis plots survived for more than 2 months (Figure 2 4). Similar to these results, Emer and Fonseca (201 0) also found that Eucalyptus seedlings that were transplanted into various native and modified communities were short lived, surviving only 2.7 to 9.3 months. The significantly greater longevity of E. camaldulensis whenever differences were detected, in c ombination with a highest percent emergence observed in the seed addition studies, may indicate that E. camaldulensis does have a relatively greater ability to establish than the other species tested. These result s may support previous research that sugges t ed E. camaldulensis has high potential for invasiveness. Gordon et al. (2012) determined that E. camaldulensis has a Weed Risk Assessment (WRA) score of 18, representing a high invasion risk level and tying with E. globulus as the Eucalyptus species with the highest score out of the 38 species evaluated. Additionally, Rejmnek and Richardson (2011) reported that E. camaldulensis is the most widespread Eucalyptus species and has naturalized in 1 6 regions Conclusion The results of this study are valuable b ecause the potential for emergence and survival of Eucalyptus was tested in natural conditions that include the complex biotic and abiotic environmental components that can influence invasion success. The most potential for invasiveness was demonstrated fo r E. camaldulensis seedlings, in disturbed conditions and within Eucalyptus stands when compared to other levels of respec ti ve variables However, the probability of Eucalyptus seedling emergence in
46 seed addition subplots w as very low overall (0.0 to 0.32% ). While mean total emergence due to recruitment by the mature Eucalyptus was estimated to be as 28.3 thousand seedlings ha 1 (95% CI: 15.0 to 53.5 thousand seedlings ha 1 ) in the most favorable treatment combination for emergence no seedlings survived lo nger than 13 weeks. Similarly, the two site surveys found no Eucalyptus seedlings resulting from natural recruitment. Before making broad conclusions about the invasive potential of Eucalyptus, consideration needs to be given to the fact that seeding treat ments and site surveys were performed only once and may not necessarily reflect the continual propagule pressure of seed rain falling over entire seasons for multiple years. The scope of the seed addition studies in this study must also be considered, in t hat conclusions about invasiveness should be restricted to the studied conditions and Eucalyptus species tested. While it is not possible to guarantee that treatment combinations that demonstrated limited success in these studies will never result in highe r levels of seedling establishment, the trends observed in these studies are consistent with evidence from previous empirical research, which d id not support a particularly extensive invasive ability of Eucalyptus sp ecies.
47 Table 2 1. Expected germi nation was evaluated in a controlled environment growth chamber using eight 0.05 g samples of the seed lot for each species. The total number of normal germinated seeds in each sample was used to calculate the mean number of germinating seed per g for each seedlot The amount s of seed that were added into low and high seeding density subplots, which were expected to result in 500 and 1 000 germinating seeds m 2 w ere calculated. Eucalyptus species Mean number of germinating seeds g 1 seedlot Mass of seed s added in low seeding density subplots M ass of seed s added in high seeding density subplots E. grandis 655.1 0.048 0.095 E. amplifolia 92.3 0.339 0.678 E. camaldulensis 388.0 0.081 0.161 Table 2 2 Vegetation community characteristics at t he Gainesville, Florida study location. Characteristics Eucalyptus amplifolia seed orchard Abandoned forest road Native hardwood forest Soil series a Blitchton sand Lochloosa fine sand Blitchton sand Canopy cover b 74.89 1.86 % 64.03 4.27 % 98.25 0.18 % Predominant vegetation Albizia julibrissin Eucalyptus amplifolia Liquidambar styraciflua Quercus nigra Quercus virginiana Smilax auriculata Vitis sp. Dichondra carolinensis Quercus nigra Rubus sp. Smilax auriculata Parthenocissus quinquefolia Pin us taeda Quercus virginiana Smilax auriculata Toxicodendron radicans Ulmus americana a USDA 2012. Web Soil Survey. http://websoilsurvey.nrcs.usda.gov/app/HomePage.htm Accessed: January 1 0, 2013. b Means and standard errors for canopy cover measured over the center of each treatment plot. Table 2 3 Vegetation community characteristics at the Quincy, Florida study location. Characteristics Eucalyptus amplifolia progeny test Intensively prepared land for Eucalyptus planting Non grazed pasture Soil series a Dothan Furquay complex Dothan Furquay complex Dothan Furquay complex Canopy cover b 90.38 0.71 % 0.00 0.00% 0.00 0.00% Predominant vegetation Diospyros virginiana Eucalyptus a mplifolia Ligustrum sinense Populus deltoides Rubus sp. Eupatorium capillifolium Ipomoea coccinea Ipomoea cordatotriloba Oxalis stricta Passiflora incarnata Portulaca pilos a Senna obtusifolia L. Paspalum urvillei Paspalum laeve Oxalis stricta a USDA 2012. Web Soil Survey. http://websoilsurvey.nrcs.usda.gov/app/HomePage.htm Accessed: January 10, 2013. b Means and standard errors for canopy cover measured over the center of each treat ment plot.
48 Table 2 4 Mean percent Eucalyptus seedling emergence after seed s of three Eucalyptus species were added in to disturbance treatment plots within three communities at each of the two Florida study locations in June 2012. Seedling emergence was based on the total number of emerged seedlings that were counted in perio dic plot assessments over 25 weeks following seed addition. Mean percent emergence is presented for disturbance, species and vegetation community variables. There were no significant differences in the probability of emergence for these main effects at = 0.05. Variable Eucalyptus seedlings emerged % Disturbance Disturbed 0.000102 Non disturbed 0.000001 Eucalyptus species E. amplifolia 0.000000 E. camaldulensis 0.000228 E. grandis 0.000023 Vegetation community Gainesville Eucalyptus stand 0.000499 Forest road 0.015400 Native h ardwood forest 0.000000 Quincy Eucalyptus stand 0.000316 Non grazed pasture 0.000000 Intensive ly prep ared land 0.000000
49 Table 2 5 Mean percent Eucalyptus seedling emergence after seed s of three Eucalyptus species were added in to disturbance treatment plots within three communities at each of the two Florida study locations in June 2012. S eedling emergence was base d on the total number of emerged seedlings that were counted in periodic plot assessments over 25 weeks following seed addition. Mean percent emergence is presented for location, vegetation community and species treatment combinations There were no signif icant differences in the probability of emergence for these combinations of factor levels at = 0.05 Treatment combinations Eucalyptus seedlings emerged Location Vegetation community Eucalyptus species % Gainesville Eucalyptus stand E. amplifolia 0.000000 Gainesville Eucalyptus stand E. cam aldulensis 0.323500 Gainesville Eucalyptus stand E. grandis 0.126700 Gainesville Forest road E. amplifolia 0.000446 Gainesville Forest road E. camaldulensis 0.129000 Gainesville Forest road E. grandis 0.063400 Gainesville Native h ardwood forest E. amp lifolia 0.000000 Gainesville Native h ardwood forest E. camaldulensis 0.000000 Gainesville Native h ardwood forest E. grandis 0.000000 Quincy Eucalyptus stand E. amplifolia 0.000446 Quincy Eucalyptus stand E. camaldulensis 0.193700 Quincy Eucalyptus sta nd E. grandis 0.000000 Quincy Non grazed pasture E. amplifolia 0.000000 Quincy Non grazed pasture E. camaldulensis 0.000000 Quincy Non grazed pasture E. grandis 0.000000 Quincy Intensive ly prep ared land E. amplifolia 0.000000 Quincy Intensive ly prep a red land E. camaldulensis 0.000000 Quincy Intensive ly prep ared land E. grandis 0.000000
50 Table 2 6 Po pulation means for Eucalyptus seedling emergence across both seeded and non seeded areas of 21 disturbed and 21 non disturbed treatment pl ots within each of the various vegetati on communities at two Florida study l ocations. Since emergence was very close to zero in seed addition subplots, these means reflect recruitment by mature Eucalyptus trees. Po pulation means for s eedling emergence and 95% confidence intervals are also expressed on a per hectare basis for each treatment combination. Treatment combinations are presented in order from greatest emergence to least emergence. Treatment combinations a Eucalyptus seedlings emerged b c Seedlin g emergence ha 1 Location Vegetation community Disturbance Mean c Lower 95% CL Upper 95% CL Gainesville Eucalyptus stand Disturbed 45 28299 14976 53476 Gainesville Eucalyptus stand Non disturbed 7 4717 1100 20223 Quincy Eucalyptus stand Disturbed 6 3 810 738 19663 Gainesville Forest road Disturbed 3 2177 250 18927 Quincy Eucalyptus stand Non disturbed 1 635 78 5182 Gainesville Forest road Non disturbed 1 363 29 4546 a Not including non grazed pasture, intensively prepared land and native hardwood f orest communities that had zero emerged seedlings. b Total area of plots = 15.75 m 2 c Estimated population means from a generalized linear model.
51 Figure 2 1. T reatment plot dimensions and possible randomly assigned disturbance treatments within whole plots (1 x 1.5 m) is shown. Possible randomly assigned seeding density subplot treatments and non seeded areas used in the seed addition studies are also shown.
52 Figure 2 2. Seedling longevity for the Gainesville, Florida study location. No seedl ings survived more than 12 weeks. Figure 2 3. Seedling longevity for the Quincy, Florida study location. No seedlings survived more than 13 weeks.
53 Figure 2 4. Time of Eucalyptus seedling survival by species for all seedlings at both Quincy and Gaines ville study locations in Florida No seedlings survived more than 13 weeks.
54 CHAPTER 3 COMPARISON OF AMINOCYCLOPYRACHLOR TO STANDARD HERBICIDES FOR CONTROL OF EUCALYPTUS Introduction The Invasive Species Advisory Committee, a group of non federal experts and stakeholders which was established to provide advice on invasive species related issues to the interagency National Invasive Species Council (Federal Register 1999), made nine recommendations for federal biofuels programs to mini miz e the risk of bioenergy crop escape into the surrounding environment. Their recommendations included the need to establish protocols for rapid removal of bioenergy crops, should they disperse into surrounding areas or become abandoned and unwanted popul ations (NISC 2009). Herbicides are a relatively effective and inexpensive tool that can be used to manage Eucalyptus seedlings occurring from natural recruitment or for the removal of abandoned stands of mature trees To prepare for potential management of invasive Eucalyptus in the s outheastern US, effective chemical control methods need to be established. Herbicides for Eucalyptus Control The existing recommendations regarding methods for chemical control of Eucalyptus are inexact. They generally entail cut stump, basal frill (also known as cut stem treatment), or basal bark applications using concentrated herbicide solutions or emulsions containing the active ingredients triclopyr, imazapyr, or glyphosate. Individual stem treatments using these broad spe ctrum herbicides are preferred for their ease of use and targeted application. Triclopyr and glyphosate may be applied to the stems of target vegetation with minimal impact to nearby vegetation, because these herbicides are not readily absorbed from the so il by roots (Senseman 2007). However, imazapyr is
55 a soil active herbicide, thus injury to non target vegetation is more likely (Little and Shaner 1991). In Eucalyptus removal projects in Marin County, California, triclopyr and imazapyr had the best results and cut stump applications of 80.0% Garlon 4, or 100.0% Garlon 3A, Stalker, or Roundup were recommended (Bossard et al. 2000). Moore (2008) suggested that a lower rate (50.0%) of glyphosate or triclopyr is sufficient for basal bark, basal frill, or c ut stump applications. One study reported that very low rates of picloram (formulated as Tordon 22K) in partial and complete basal frills may provide satisfactory control (Bachelard et al.1965), h owever, control using these relatively low concentrations o f picloram was not consistent in all study areas or at all times. Additional trials with 12.5% Tordon 101 mixture (containing 240 g ae L 1 2,4 D amine plus 65 g ae L 1 picloram) as a basal frill treatment to different Eucalyptus species resulted in 90 100% flashback through natural root grafts) was a serious issue, as 30 m tall non treated trees were strongly affected. In general, complete control of Eucalyptus is rarely achieved through the use of single herbi cide treatments and reapplication to control sprouting is often necessary ( Bachelard et al. 1965; Morze 1971; Bossard et al. 2000; Little 2003; Little and van den Berg 2006), potentially taking up to three herbicide treatments to completely prevent resprou ting (Bossard et al. 2000). Furthermore, variable levels of tolerance among different species and the varying effectiveness of different application timings for commonly used herbicides make it difficult to make recommendations (Bachelard et al. 1965; Morz e 1971). Few recommendation sources consider tree size and vigor, which are also important for the success of herbicide treatments (Morze
56 1971). Research was needed to develop herbicide prescriptions for controlling Eucalyptus using basal frill and basal b ark approaches to refine the dose response for different tree diameters and to determine the impact to non target vegetation. Aminocyclopyrachlor for Control of Woody Plants Aminocyclopyrachlor (AMCP) is a new herbicide being developed by DuPont for use in non crop areas such as rights of way, turf, range and natural areas ( E.I. du Pont de Nemours and Company 2009). A pyrimidine carboxylic acid herbicide, it is structurally similar to pyridine carboxylic acid herbicides such as aminopyralid, picloram and triclopyr. It is believed that synthetic auxin mode of action is employed, interfering with normal plant growth (USDA 2012b). AMCP has good selectivity to established cool and warm season perennial grasses (Westra et al. 2009) and shows other positive ste wardship attributes including low volatility, the absence of bioaccumulation in animals, and low toxicity to terrestrial and aquatic organisms ( E.I. du Pont de Nemours and Company 2009). It has shown excellent control at rates as low as 140 g ae ha 1 for m any species including herbaceous weeds resistant to ALS inhibitors, triazines and glyphosate (Turner et al. 2009). AMCP has also demonstrated control at low rates for many woody plants in recent trials (Appendix B ), indicating its potential as an alternati ve to currently recommended herbicide treatments for control of Eucalyptus trees. Eucalyptus benthamii (Maiden et Cambage) is among the more promising species for wide scale planting in the southeast ern US because of its cold hardiness, success in out pla ntings under a variety of conditions and fast growth rates (Zalesny et al. 2011). In particular, the ability of this species to withstand cold weather could contribute to its potential invasiveness at a geographic scale. In September 2011, three studies we re
57 initiated in 29 month old E benthamii plantations to evaluate the control provided by the new herbicide AMCP The specific objectives of this study were as follows. Compare the effectiveness of four rates of AMCP to operational standard treatments, ima zapyr and triclopyr, for control of various diameter classes of Eucalyptus trees using basal bark and basal frill stem applications Describe the rate response of AMCP for control of various diameter classes of Eucalyptus using basal frill and basal bark a pplications. Determine t he effects of herbicide treatments to nearby non target Eucalyptus trees. Materials and Methods Study Areas Three experiments were installed on adjacent 0.2 ha E. benthamii plantations established in 2009 at the University of Flori da, North Florida Research and Education Center, south 3 m elevation. This location has a temperate climate with highest temperatures in July (mean 27 C), lowest temperatures in January (mean 10 C) a nd 143 cm average annual precipitation (NOAA 2002). The prevalent soil series in both plantations is Orangeburg fine sandy loam, but soils at the North plantation site were highly eroded, to the extent that the typic sandy surface horizon was absent (herea fter the two sites are referred to as eroded and non eroded ). Across the two sites, 960 seedlings had little winter dieback or mortality prior to the establishment of this study. Trees were generally smaller on the eroded site, but there was also high vari ability in size within both sites. As a part of a separate study (Osiecka and Minogue 2011), trees received different levels of competition control during the establishment year. This resulted in an array of tree
58 sizes, ranging from less than 1 to 10 m in height and from 4 to 20 cm in basal stem diameter (BSD) at groundline, when measured prior to treatment in late October 2011. Basal Bark Treatments and Experimental Design Identical experiments were conducted in each of the sites using a randomized comple te block design. One hundred and forty randomly selected healthy trees were ranked according to basal diameter and sequentially divided into 20 groups of seven trees each. The seven similarly sized trees of each group were randomly assigned one of seven he rbicide treatments (Table 3 1), resulting in 20 replicates of each herbicide treatment across the entire range of tree diameters. Non treated buffer trees were included in the design so that no treated tree was next to another treated tree. AMCP (120 g ae L 1 in the form of DPX MAT28 159, DuPont, Wilmington, DE) was tested as a basal bark treatment at four rates (5, 10, 20 and 40% v/v formulated material). Since AMCP is a new herbicide for which few guidelines exist, the range of treatment rates was select ed based on evidence from testing across other woody plants (Edwards and Beck 2011; Wilson et al. 2011 b ; Yeiser et al. 2011; J. Ferrell, University of Florida and M. Link, DuPont, personal communications, May, 2011). Other basal bark treatments included im azapyr (isopropylamine salt of imazapyr, 240 g ae L 1 in the form of Stalker, BASF, Research Triangle Park, NC), triclopyr (butoxyethyl ester, 480 g ae L 1 in the form of Garlon 4 Ultra, Dow AgroSciences, Indianapolis, IN) and a non herbicide treated c heck (applied seed oil carrier only). The formulated imazapyr product was applied at the mid range of prescribed label rates for thinline basal and stem applications (28.1% v/v) ( BASF Corporation 2008 ), as previous experience indicated this rate was more t han adequate Formulated triclopyr ester product was applied at the highest prescribed label rate for thinline basal bark treatment
59 (75% v/v) ( Dow AgroSciences LLC 2008 ). All herbicides were thoroughly mixed with 100% methylated soybean oil, alkylphenol et hoxylate, as the carrier (M.O.C., Helena, Collierville, TN). Five ml of herbicide/oil mixture per 2.5 cm basal stem diameter (BSD) were applied to the base of trees from 30 cm height to the groundline using a syringe. Preliminary testing indicated this vol ume generally sufficient to wet the stem completely from 30 cm height to the ground line for the various diameter classes studied. Herbicide treatments were applied on November 4 and 5, 2011 and no rainfall occurred within 48 hr following application. Bas al Frill Treatments and Experimental Design Basal frill application is recommended for trees that are greater than 5 cm in diameter at breast height (Miller 2010). Because f ew large trees were present in the eroded site, this experiment was performed only at the non eroded site. As in the basal bark studies, healthy trees with suitable diameters were ranked by BSD and sequentially divided into eight groups of seven trees with like diameters. The seven treatments (Table 3 2) were assigned in a randomized com plete block design with eight replications. A hatchet was used to make cup like downward incisions at 30 cm above the groundline. One cut was made per 2.5 cm BSD. Cuts were evenly spaced around the stem circumference. A syringe was used to apply one ml of a water and herbicide mixture to each cut. Trees in the non herbicide treated check were cut with a hatchet in the same manner as other treatments, but no herbicides were applied. AMCP was applied at four rates (12.5, 25, 50 and 100% formulated material). Herbicide concentrations tested were higher than in the basal bark treatments to accommodate the small volume of mixture that must be contained within each cut, but the actual amount of active ingredient applied to each tree was comparable to the
60 various r ates tested in basal bark treatments (Table 3 1). Standard treatments for Eucalyptus control included imazapyr (Stalker) and triclopyr (triclopyr triethylamine salt, 360 g ae L 1 in the form of Garlon 3A, Dow AgroSciences) applied at the respective reco mmended label rates of 7.81% and 50.0% v/v ( BASF Corporation 2008 Dow AgroSciences LLC 2003). Tree Assessments For all trees in both sites, pre treatment stem diameter was measured to the nearest millimeter at ground level (BSD) and at breast height, 13 7 cm from groundline (DBH), in October 2011. Pre treatment stem height was measured to the nearest centimeter using a height pole. In order to quantify stem dieback or growth, the live height was measured again 12 months after treatment ( MAT ). Additionally at 2, 6 and 12 MAT 12 MAT each tree was assessed for percent crown reduction. As defined by Miller and Glover (1991), this standard variable is an ocular estimate of reduction relative to the pre treatment condition, and includes stem dieback, leaf necro sis and defoliation. Estimates were made to the nearest five percent for values between 0 and 10%, and 90 and 100%; and to the nearest 10 percent for values between 10 and 90%. Additionally, phytotoxicity symptoms (mortality, foliar necrosis, defoliation, red foliage, chlorosis, stem sap flow, epinasty, basal sprouting and adventitious buds on stems) were noted as present or not present for each tree at 2, 6 and 12 MAT Statistical Analysis Stem live height and crown reduction For each experiment, SAS Proc GLM ( SAS v. 9.3, SAS Institute, Cary, NC) determine d differences between treatment groups for change in live stem height at 12 MAT = 0.05. Crown reduction
61 responses were primarily near to the fixed limits o f 0 and 100%. Homogeneity of variances and normality could not be achieved by standard data transformations (arcsine, arcsine square root and log), so traditional parametric and non parametric tests could not be used to determine if there were differences between treatment using SAS PROC GLM to determine if there were differences between treatments in each of the three experiments. The Ryan Einot Gabriel Welsch multiple range test (REGWQ) was performed to compare treatment means at = 0.05. These procedures are recommended for analysis of data that is simultaneously heteroscedastic and non normal, and they have been shown to provide good power and acceptable control for type I error rates (Cribbie et al. 2007). Phytotoxicity symptoms the significance of herbicide treatment for binary present/absent phytotoxicity responses (red foliage, chlorosis, e pinasty, basal sprouting and adventitious buds on stems ) that had not already been evaluated by percentage data. Red foliage and chlorosis were only evaluated for trees that did not show 100% crown reduction because of the mutual exclusivity of these condi tions. Aminocyclopyrachlor rate response Data from all three experiments were combined to model the rate response of AMCP relative to tree size using SAS Proc Logistic. Crown reduction was evaluated as the binomial response of mortality where trees were considered dead only when there was 100% crown reduction Logistic regression models examined the relationship between the probability of mortality and basal stem diameter (BSD) or diameter at
62 breast height (DBH), AMCP concentration application method and the interactions between these explanatory variables. Logistic regression used logit transformation of probability, as the response va riable where The models fitted had the general form 3 1 where was a dummy variable for application method was diameter ( DBH or BSD ) was the AMCP concentration and values were the specific parameters of the model. Similar to the meth ods of Bergerud (1988), a final model was chosen using backward conditional selection starting with a full model and dropping insignificant explanatory variables and interactions one at a time. Herbicide impacts on non target trees As a measure of the po tential for flashback at various distances from treated trees symptoms of phytotoxicity were compared for non treated and buffer trees. Buffer trees were those directly adjacent to treated trees (1.5 m away) whereas non treated check trees were randomly assigned with the herbicide treatments to trees that were not adjacent to other treated trees (> 1.5 m away) Non treated trees in the basal frill study were cut but were grouped with non treated trees from the basal bark experiment for this analysis. For each assessment date, buffer trees were compared to non treated Crown reduction and stem height were compared for all trees in non treated and buffer groups using the methods described previousl y for each response variable. It was not possible to determine the impacts to herbaceous vegetation and grasses due to the highly variable vegetation
63 patterns throughout the sites and lack of data for a reference state of the vegetation when the herbicide treatments were applied. Results and Discussion Eucalyptus Control using Basal Bark Treatments Crown reduction highly significant (P < 0.0001) effect of basal bark treatments on crown reduction at 2 and 6 MAT for b oth sites, and at 12 MAT on the non eroded site. Differences in crown reduction due to treatments were not significant at 12 MAT on the eroded site, although all treatments had significantly greater crown reduction than the non treated check (Table 3 3). T he three highest AMCP rates resulted in 100% crown reduction in all assessment s at both site s The lowest rate of AMCP only differed from the higher rates initially, when at 2 MAT it resulted in less crown reduction (82%) on the non eroded site, which had larger trees. However, by 6 MAT 100% crown reduction was observed for nearly all of the trees that were treated with AMCP, regardless of rate, tree size, or site. This complete control was achieved using AMCP at 0.03 to 0.24 g ae per 2.5 cm BSD. No other studies were found regarding the use of basal bark herbicide treatments for the control of Eucalyptus but these results indicate that this method of application is effective for Eucalyptus control using AMCP Standard herbicide treatments always resulte d in significantly less crown reduction than AMCP treatments, except at 6 and 12 MAT on the eroded site (Table 3 3). Here, the population was comprised of small, less vigorous trees, which were perhaps more susceptible to herbicides. Mean crown reduction f or imazapyr treated trees increased from 62% to 100% on the eroded site and from 42% to 91% on the non
64 eroded site between 2 and 12 MAT, indicating that symptoms of imazapyr injury are slow to develop, as is commonly reported for herbicides with an amino a cid synthesis inhibit ing mode of action (Gunsolus and Curran 1999). Likewise, symptoms were also slow to appear for triclopyr ester with mean crown reduction increasing from 26% to 97% on the eroded site and 4% to 86% on the non eroded site between 2 and 12 MAT Triclopyr ester was consistently the least effective herbicide in both basal bark experiments. Stem live height The overall effect of basal bark treatment on live stem height at 12 MAT was highly significant at both sites (P < 0.0001). At the ero ded site, all herbicide treatments resulted in stem reduction whereas trees in the non treated check grew 335 cm on average (Figure 3 1). All herbicide treatments on the eroded site were significantly different from the non treated check, but there were no significant differences among herbicide treatments. At the non eroded site, all rates of AMCP and the imazapyr treatment resulted in significantly greater stem reduction compared to triclopyr ester or the non treated check, but reduction did not differ am ong AMCP and imazapyr treatments (Figure 3 1 ). Phytotoxicity symptoms For trees that did not have 100% crown reduction (143 trees at 2 MAT 66 trees at 6 MAT and 60 trees at 12 MAT ) significant relationships between phytotoxicity symptoms and basal bark herbicide treatments were only detected for the study on the non eroded site. Phytotoxic symptoms were most common for the standard imazapyr and triclopyr ester treatments, because AMCP treatments resulted in a high incidence of complete crown reduction, a nd this symptom also occurred more quickly than with the
65 standard treatments. Red foliage was more common (P = 0.014) at 2 MAT for triclopyr ester (44%) and imazapyr (17%) treatments than for the AMCP treatments, which did not result in red foliage symptom s. Treatment also had a significant effect (P = 0.002) on the occurrence of stem sap flow at 2 MAT At this assessment, stem sap flow was observed for 20% of the trees treated with triclopyr ester and for 20 % and 5% of the trees treated with the two lowes t rates of AMCP, respectively. Treatment had a highly significant effect (P = 0.002) on adventitious budding on stems, a known symptom of imazapyr in ang i osperms (Osiecka and Minogue 2012) at 12 MAT when it was observed at nodes on the trunks of 20% of im azapyr treated Eucalyptus trees. The incidence s of epinasty and basal sprouting were rare and did not differ by herbicide treatment at any time in either study. Basal sprouting was only observed on one triclopyr ester treated tree in each study at 12 MAT The effectiveness of AMCP treatments observed in these studies are consistent with those of Wilson et al. (2011b) who reported effective control by AMCP basal bark treatments for other fast growing woody plants such as Russian olive ( Elaeagnus angustifolia L.) and salt cedar ( Tamarix sp.). Edwards and Beck (2011) also reported similar results for Russian olive, although their trials were notably different because they applied approximately 30 ml per 2.5 cm of stem diameter, while 5 ml per 2.5 cm BSD was app lied in our study. Although a higher herbicide rate may be required for control of Russian olive as compared to Eucalyptus it appears that other studies have generally not investigated the lowest rates necessary to achieve control of some woody plants. Fu ture studies of woody plant control with AMCP should examine lower rates than used in our study.
66 Eucalyptus Control using Basal Frill Treatments Crown reduction The overall effect of basal frill herbicide treatment on crown reduction was highly significa nt (P < 0.0001). However, there was not a significant difference in crown reduction among the AMCP rates tested at any assessment, and by 6 MAT all 30 trees treated with AMCP had 100% crown reduction (Table 3 4). While AMCP treatments were significantly mo re effective at reducing live crown than triclopyr amine (49%) and imazapyr (19%) treatments at 2 MAT by 6 MAT there was no difference between triclopyr amine (80% at 6 MAT and 81% at 12 MAT ) and any of the AMCP rates (all 100%) Whereas imazapyr treatmen ts resulted in significantly greater crown reduction than the non treated check, imazapyr remained the least effective herbicide throughout the study, resulting in 58% mean crown reduction by 12 MAT In comparing the two application methods studied for th e assessment 12 MAT the mean crown reduction resulting from basal frill applications of imazapyr (58%) and triclopyr amine (81%) were qualitatively less than basal bark applications of imazapyr (91 100%) and triclopyr ester (97 100%). This trend contrasts the conclusion of Bossard et al. (2000) that application of triclopyr and imazapyr herbicides to the foliage or stems of sprouts is less effective than application directly to the vascular cambium immediately following cutting. This contrasting trend may have arisen because Bossard et al. (2000) applied herbicides at 80 100% v/v for both application methods, whereas relatively lower rates were applied for basal frill applications in of these herbicides our study
67 Stem live height The overall effect of ba sal frill herbicide treatment on stem height at 12 MAT was highly significant (P < 0.0001) and all herbicide treatments resulted in height reduction from pre treatment values due to stem dieback, whereas the non treated check grew 425 cm(Figure 3 2 ). There were no differences among the AMCP rates, but all AMCP treatments resulted in significantly greater stem dieback than either of the standard treatments. Phytotoxicity symptoms The overall effect of treatment was significant (P = 0.0196) for the presence of red foliage symptoms only at 2 MAT (results not shown). However, at this assessment it was commonly observed in all herbicide treatments, with the exception of the three highest AMCP rates, which had already resulted in 100% crown reduction. Treatment h ad a significant effect (P = 0.014) on stem sap flow, and was most common for imazapyr (44%) and triclopyr amine (71%) treatments. Similar to the basal bark experiment, treatment had a significant effect (P = 0.0149) on the presence of adventitious buds on stems at 12 MAT whereby 33.3% of trees treated with imazapyr had this symptom. Adventitious buds were not observed for trees in any other treatment. Although treatment effects were not significant, chlorosis was most common at 6 MAT when it was observed for 50% of the frilled non herbicide treated check trees, 11% of trees treated with imazapyr, and 14% of trees treated with triclopyr amine Epinasty or basal sprouting were not significantly a ffected by treatment nor were they common symptoms at any asses sment. Basal sprouting was observed for only one triclopyr amine treated tree and one imazapyr treated tree at 12 months following basal frill treatment.
68 Whereas no other literature reporting results for AMCP basal frill application for the control of woo dy plants was found, these new results demonstrate that it is an effective alternative to more labor intensive cut stump applications, which have been the focus of several studies (Edwards and Beck 2011; Wilson et al. 2011b; Yeiser et al. 2011). With respe ct to triclopyr amine basal frill treatments, control of E. benthamii in this study differed from results described by Little and van den Berg ( 2006) who applied the same formulation of triclopyr amine at 3% v/v in basal frill treatments to E. macarthurii trees and reported mortality of 83 90% of trees when evaluated nine months after treatment. O ur study found that only 57% of triclopyr amine treated trees had complete mortality by 12 MAT even though a much higher herbicide rate (50% v/v) was used. This d ifference in control may indicate differential susceptibility for these two Eucalyptus species, or possibly the need for continued assessments in our study to determine if greater levels of mortality will eventually occur. Aminocyclopyrachlor Rate Respons e Relative to Tree Diameter at Breast Height A significant model (Wald 2 = 29.82, 4 df, P < 0.0001) was developed that accurately predicts the actual mortality response observed for 161 out of 189 trees in this study (Table 3 5). The following l ogistic regression equation was used to predict the minimum concentratio n of AMC P (120 g ae L 1 AMCP formulation) required for a desired probability of mortality at 2 MAT of E benthamii of a given DBH, using basal bark or basal frill application s : 3 2
69 where appl ication method is dummy coded so that basal bark = 1 and basal frill = 0. For example, t he concentration of AMCP required to achieve 90% control of E. benthamii given by for basal bark applications 3 3 for basal frill applications 3 4 Even though application method was not a significant predictor, it was included in the model because it i ncreased the percent concordant. A logistic regressi on model was not fitted for mortality at 6 and 12 MAT because only one out of 199 trees treated with AMCP did not have 100% crown reduction at that point in time. This tree was one of the largest in the study (13.2 cm BSD) and received the treatment with t he lowest AMCP rate as a basal bark application using the lowest AMCP rate tested. Although complete mortality did not occur for the tree within the study period, at 6 MAT 45% of the crow n was reduced by 45% and by 12 MAT crown reduction increased to 75%. The model developed in this study is the first that considers tree size for the control of Eucalyptus These results demonstrate that tree diameter and application method have an important relationship with rate response that should be considered when pres cribing herbicide treatments. When applying AMCP to Eucalyptus Figures 3 4 and 3 5 developed from the logistic regression model can be used as a guide to select the lowest dosage for which death is likely to occur (predicted probability of death >50%) as early as two months after treatment, for a given stem diameter.
70 Impacts to Non target Vegetation There was no significant difference in crown reduction when buffer trees (those directly adjacent to treated trees) and non treated check trees were compared While non treated trees grew 377 cm and buffer trees only grew 337 cm, this difference in growth was not statistically significant. The most frequently observed phytotoxicity symptoms for both buffer and non treated check trees were foliar necrosis, defo liation and chlorosis (Table 3 presence of foliar necrosis was greater for buffer trees than for non treated check trees at 6 MAT (P = 0.0236). However, no mortality was observed in either gr oup at 2 MAT and by 6 MAT mortality occurred in only one percent of buffer trees, but not in any non treated check tree s At 12 MAT two of the five buffer trees that appeared to have 100% crown reduction at 6 MAT had sprouted new basal shoots. The large or medium buffer trees (5 15 cm BSD) that died were all neighbored on at least three sides by AMCP treated trees, and small trees (<5 cm BSD ) were bordered by one or two AMCP treated trees This suggest s that absorption of this herbicide from the soil by r oots, or perhaps herbicide uptake through root grafts, may be of concern when making applications in very close proximity to desirable trees. Other symptoms observed included epinasty, a symptom characteristic of the growth regulator herbicides triclopyr a nd AMCP. Various symptoms in buffer trees and non treated checks began to appear by 6 MAT but remained uncommon and differences were insignificant between buffer and non treated trees throughout the study period. While it was not possible to quantify imp acts to herbaceous vegetation in this study, no obvious symptoms of injury such as necrosis or chlorosis were observed for herbaceous vegetation near treated stems. Moreover, at 6 MAT several newly
71 established pine and oak seedlings were discovered throug hout the Eucalyptus understory. While there have been concerns about the effects of AMCP on non target trees and vegetation (USDA 2012b), these results suggest that risk injury to non target vegetation may be reduced when low rate, low volume directed appl ications are used. Conclusion Using a model for the predicted probabilities of mortality responses for a range of AMCP doses as a function of stem diameter and basal bark and basal frill herbicide treatment methods (Figures 3 4 and 3 5, Table 3 5), we can determine the appropriate dosage for any desired level of mortality at 2 MAT At 6 and 12 MAT the lowest AMCP rates tested provided complete or nearly complete control of E. benthamii and often provided significantly greater control than the standard tric lopyr and imazapyr treatments. At 12 MAT no basal sprouting was observed for AMCP treated trees, indicating long term control. Mortality was observed for less than 1% of non treated check trees that were within 1.5 m of herbicide treated trees in this stu dy. These results suggest that AMCP should be considered for new labeling for control of undesirable Eucalyptus trees. Based on this research, 0.03 g ae AMCP per 2.5 cm basal stem diameter (BSD) is effective as a basal stem treatment and 0.015 g ae per 2.5 cm BSD is effective for basal frill treatment to control Eucalyptus trees. Since these treatments gave complete control with basal frill treatment and near complete control with basal bark treatment when evaluated at 12 MAT future research should examine the efficacy of lower herbicide rates for Eucalyptus control. Although these studies tested AMCP rate response and the effects of tree diameter in comparison to the efficacy of standard herbicides, the potentially important effects of application timing a nd Eucalyptus species susceptibility were not determined.
72 This would require experiments across multiple seasons for those Eucalyptus species most commonly grown, such as E. camaldulensis, E. grandis, E. u rograndis, E. amplifolia, and others. Future resear ch regarding species susceptibility could look to the work of Morze (1971) which identified susceptible and resistant Eucalyptus species for picloram. Despite all factors that could affect herbicide treatment success not being evaluated, the strong control of E. benthamii in this study, with minimal injury to non target trees, represents significant progress in the development of effective herbicide prescriptions for Eucalyptus management and indicates that future research with different species and timings would be worthwhile.
73 Table 3 1. Treatments tested in the basal bark studies for Eucalyptus benthamii control in Florida. All herbicide treatments were applied in methylated soybean oil carrier. Five ml of the herbicide in oil mixtures were applied per 2.5 cm basal stem diameter (BSD), from 30 cm stem height to groundline. Herbicide Formulation tested Formulation concentration Applied formulation concentration in oil carrier Basal stem diameter specific dose g ae L 1 % v/v g ae per 2.5 cm BSD AMCP a DPX MAT28 159 120 5.0 0.030 AMCP DPX MAT28 159 120 10.0 0.060 AMCP DPX MAT28 159 120 20.0 0.120 AMCP DPX MAT28 159 120 40.0 0.240 Imazapyr Stalker 240 28.1 0.337 Triclopyr ester Garlon 4 Ultra 480 75.0 1.800 Non treated check a a AMCP is aminocyclopyrachlor. b Methylated soybean oil alone was applied to the non herbicide treated check. Table 3 2. Treatments tested in the basal frill studies for Eucalyptus benthamii control in Florida. One cut was made per 2.5 cm basal stem diameter ( BSD ) Cuts were evenly spaced around the stem circumference at 30 cm above the groundline. All herbicide treatments were diluted in water. One ml of herbicide mixtures was applied to each cut. Herbicide Formulation tested Formulation concentrati on Ap plied formulation concentration i n water Basal stem diameter specific dose g ae L 1 % v/v g ae per 2.5 cm BSD A MCP DPX MAT28 159 120 12.5 0.015 AMCP DPX MAT28 159 120 25.0 0.030 AMCP DPX MAT28 159 120 50.0 0.060 AMCP DPX MAT28 159 120 100.0 0 .120 Imazapyr Stalker 240 7.8 0.019 Triclopyr amine Garlon 3A 360 50.0 0.180 Non treated check a a AMCP is aminocyclopyrachlor. b The non treated check was cut without herbicide treatment.
74 Table 3 3. Crown reduction of Eucalyptus ben thamii at 2, 6 and 12 months after treatment (MAT) with diameter specific basal bark herbicide applications at the eroded and non eroded study sites Crown reduction was determined as an ocular estimate of reduction relative to the pre treatment condition, and included stem dieback, leaf necrosis and defoliation. Crown reduction a Herbicide Applied formulation conc. in oil carrier Basal stem diameter specific dose 2 MAT 6 MAT 12 MAT Eroded Non eroded Eroded Non eroded Eroded Non eroded (% v/ v) g ae per 2.5 cm BSD (%) AMCP b 5 .0 0.030 86 5 a 82 7 a 100 0 a 97 3 a 100 0 a 99 1 a AMCP 10 .0 0.060 98 2 a 98 2 b 100 0 a 100 0 a 100 0 a 100 0 a AMCP 20 .0 0.120 100 0 a 99 1 b 100 0 a 100 0 a 100 0 a 100 0 a AMCP 40 .0 0.240 100 0 a 99 1 b 100 0 a 100 0 a 100 0 a 100 0 a Imazapyr 28 .1 0.337 62 8 b 42 9 c 98 1 a 83 6 b 100 0 a 91 5 b Triclopyr ester 75 .0 1.800 26 8 c 4 2 d 92 5 a 73 9 b 97 2 a 86 6 b Non treated check 1 0 d 3 2 d 1 1 b 3 1 c 0 0 b 3 1 c a Treatment means (M SE) within a column followed by the same letter are not significantly different according to the Ryan Einot Gabriel Welsch multiple range test (REGWQ) on ranked data at = 0.05. b AMCP is aminocyclopyrachlor.
75 Table 3 4. Crown reduction of Eucalyptus benthamii at 2, 6 and 12 mon ths after treatment (MAT) with diameter specific basal frill herbicide applications at the non eroded study site. Crown reduction was determined as an ocular estimate of reduction relative to the pre treatment condition, and included stem dieback, leaf nec rosis and defoliation. Herbicide Applied formulation conc. in water Basal stem diameter specific dose Crown reduction a 2 MAT 6 MAT 12 MAT (% v/v) g ae per 2.5 cm BSD (%) AMCP b 1 2.5 0.015 91 9 a 100 0 a 100 0 a AMCP 25 .0 0.030 100 0 a 100 0 a 100 0 a AMCP 50 .0 0.060 99 2 a 100 0 a 100 0 a AMCP 100 .0 0.120 100 0 a 100 0 a 100 0 a Imazapyr 7.8 0.019 19 12 bc 49 11 b 58 11 b Triclopyr amine 50 .0 0.180 49 16 b 80 14 a 81 10 a Non treated check 3 2 c 0 0 c 2 1 c a Treatment means (M SE) within a column followed by the same letter are not significantly different according to the Ryan Einot Gabriel Welsch multiple range test (REGW Q) on ranks at = 0.05. b AMCP is aminocyclopyrachlor. Table 3 5. Logistic regression model variables for aminocyclopyrachlor (AMCP) rate response in Eucalyptus benthamii mortality at two months after treatment as a function of stem diameter at 137 cm h eight (diameter breast height, DBH). Predictor SE Wald df Intercept 2.3902 1.1183 4.5678 1 0.0326* DBH (cm) 0.3946 0.1376 8.2263 1 0.0041* AMCP conc entration a 0.1908 0.0612 9.7122 1 0.0018* Application method b 1.1556 0.7264 2.5308 1 0.1116 DBH AMCP concentration 0.0159 0.0075 4.5769 1 0.0324* a The formulation used (DPX MAT28 159) contained 120 g ae L 1 AMCP. b Basal bark is the reference level (i.e. application method = 1 for basal bark, application method = 0 for basal frill) Significant predictor at = 0.05 Table 3 6. P ercentage of buffer (n = 411) and non treated (n = 48) Eucalyptus benthamii trees that displayed symptoms of herbicide injury at 2, 6 and 12 months after treatment (MAT) in a study using aminocyclopyrachlor, imazapyr and tri clopyr herbicides. Buffer trees were adjacent to treated trees
76 (1.5 m away). Non treated trees were not adjacent to treated trees (> 1.5 m away). 2 MAT 6 MAT 1 2 MAT Symptom Buffer Non treated Buffer Non treated Buffer Non treated % Mortality 0 0 1 0 1 0 Necrosis 8 10 28* 13* 17 19 Defoliation 5 10 13 8 16 10 Red foliage 4 2 9 13 3 6 Chlorosis 7 10 15 25 4 0 Stem sap flow 0 0 1 0 0 0 Epinasty 0 0 2 2 7 6 Basal sprouting 0 0 1 2 1 0 Stem adventitious buds 0 0 0 0 0 0 Significant differences at = .0.05 between buffer and non Figure 3 1. Change in Eucalyptus benthamii live stem height from pre treatment values to 12 months after diameter specific basal bark treatments at the eroded (lighter bars) and non eroded (darker bars) study sites. Treatments
77 included various concentrations of 120 g ae L 1 aminocyclopyrachlor (AMCP), 240 g ae L 1 imazapyr (imaz) and 480 g ae L 1 triclopyr ester (triclo) in methylated soybean oil carrier, as compared to a non herbicide treated check receiving oi l carrier alone. Error bars show t he standard error of the mean. Treatments within each site that are noted with the same letter = 0.05. Figure 3 2 Change in Eucalyptus benthamii live stem height from pre treatment value s to 12 months after diameter specific basal frill treatments at the non eroded study site. Treatments included various concentrations of 120 g ae L 1 aminocyclopyrachlor (AMCP), 240 g ae L 1 imazapyr (imaz) and 360 g ae L 1 triclopyr amine (triclo) in wat er carrier, as compared to a basal frilled non herbicide treated check. Error bars show from the standard error of the mean. Treatments with the same letter are not significantly different = 0.05.
78 Figure 3 3 Relationship between Eucalyptus benthamii stem diameter at breast height (DBH, measured at 137 cm height) and applied concentration of 120 g ae L 1 aminocyclopyrachlor (AMCP) in methylated soybean oil carrier for the predicted likelihood of mortality at two months after diameter specific basal bark treatment, as determined using a logistic regression model. The AMCP concentrations that were tested ranged from 5 % to 40%. Figure 3 4 Relationship between Eucalyptus benthamii s tem diameter at breast height (DBH, measured at 137 cm height) and applied concentration of 120 g ae L 1 aminocyclopyrachlor (AMCP) diluted in water for the predicted likelihood of mortality at two months after diameter specific basal frill treatment, as d etermined using a logistic regression model. The AMCP concentrations that were tested ranged from 12.5 % to 100%.
79 CHAPTER 4 SUMMARY AND IMPLICATIONS FOR MANAGEMENT AND FUTURE RESEARCH As hardy, fast growing grasses, trees and oil crops are being evaluated for an emerging bioenergy market, understanding the environmental issues associated with large scale exotic bioenergy plantings has become a priority. In particular, the potential for invasiveness has emerged as a critically important issue. This thesis in cluded two research approaches to examine the likelihood of important Eucalyptus species becom ing invasive. Another component of the thesis research identified a new herbicide that is highly effective in controlling unwanted or invasive Eucalyptus. In a s tudy of the potential invasiveness of Eucalyptus in the southeastern US the occurrence of Eucalyptus seedling recruitment within two geographically separate seed bearing E. amplifolia stands, and also in native and modified plant communities proximate to the stands, no Eucalyptus recruitment was found. T he other study w as the first experiment to examine invasive potential of Eucalyptus in the southeastern US Using an approach similar to the approach used in Brazil by da Silva et al. (2011) this study e valuated the invasive potential of three commercially important Eucalyptus species within the seed bearing E. amplifolia stands and in their proximate native and modified vegetation community types. The combined research findings demonstrated that Eucalypt us establishment and survival were generally low at the northern Florida locations, although greater success was sometimes observed for E. camaldulensis seedlings, in disturbed conditions and within Eucalyptus stands. These results support conclusions by G ordon et al. (2011, 2012) and suggest that caution is warranted regarding the cultivation of E. camaldulensis and concerning practices that might increase disturbance near Eucalyptus stands. The implications of this research are of
80 particular relevance for the common practice of establishing a bare soil or vegetative buffer zone around Eucalyptus plantings to mitigate their spread. The demonstrated role of disturbance in facilitating Eucalyptus seedling recruitment suggests that bare soil buffer zones shoul d be discontinued, and that a stable perennial plant community should be established instead. Considering that greater success was observed for Eucalyptus in Gainesville, the study location with higher rainfall and lower latitude, further experiments are a lso recommended for exploring the possibility for invasion in southern Florida and across sites with the widely varying rainfall patterns. In the second component of this thesis, the potential for effective Eucalyptus control by a promising new herbicide chemistry was evaluated and compared to standard applications of imazapyr and triclopyr herbicides. Rate response models were developed to predict the lowest effective AMCP rate for managing trees of various diameter sizes by two mo after treatment. Basal bark applications of 120 g ae L 1 AMCP at 5% v/v in methylated soybean oil resulted in 97 99% crown reduction of Eucalyptus and generally provided greater control than the standard 240 g ae L 1 imazapyr or 480 g ae L 1 triclopyr ester at 6 and 12 mo after treatment for all diameter classes. Similarly, basal frill applications of 120 g ae L 1 AMCP at 12.5% v/v in water resulted in 100% crown reduction of Eucalyptus and greater control than the standard imazapyr or 360 g ae L 1 triclopyr amine treatments at 6 and 12 mo after treatment for all diameter classes. In addition to the positive implications of these results for Eucalyptus control in the southeastern US, this research may be useful to land managers in other regions, especially in California and in Sou th Africa where extensive Eucalyptus removal projects are ongoing.
81 The results of this research are valuable because implications regarding both invasion ecology and management are considered, thus contributing to a more comprehensive knowledge of the pot ential environmental challenges posed by Eucalyptus culture than research in either field would have allowed alone. Although the definitive outcome of widespread Eucalyptus introductions in the southeast ern US cannot be determined, together the two compone nts of this research suggest that the level of invasiveness demonstrated by E. camaldulensis E. amplifolia and E. grandis is not overwhelming for northern and central Florida, and invading seedlings or unwanted populations could be effectively controlled using AMCP herbicide treatments.
82 APPENDIX A POTENTIAL INVASION RISK MANAGEME NT PRACTICES FOR EUCALYPTUS Table A 1. Practices to evaluate and manage potential invasiveness of Eucalyptus in the Southeastern US with respect to four commonly recognized phas es of intervention for managing invasive species including prevention, containment, control and management of the impacted ecosystem Supporting r eferences for management practices and their advantages and limitations are indicated in parentheses Interven tion type Management practice Goal of Practice Advantages of practice Limitation s of practice Prevention Weed Risk Assessment (WRA) (1, 2) Identify low/ high risk taxa Known level of accuracy (3). Widely tested (3). Time and cost efficient (4). Cannot a ssess novel cultivars (5). Imprecision in context of biomass crops (5, 6). Assess or bias. conclusion. E xperimental introductions (5, 7 9) Identify low/ high risk taxa and con ditions favorable to growth, spread Ability to ass ess novel cultivars (5). Can assess s pecies for which evaluat (5, 9). A novel tool in invasive context (5). Time and cost of experiments (5). Inexperience translating performance metrics into meaningful measures of invasion risk (5 ). Containment Selection for sterility or decreased fertility (10) Reduce/ eliminat e probability of seed dispersal Replaces the need for complex ecological research to evaluate spread and predict impacts (11). Costly high risk research (12). Complex regul atory process (12). Intellectual property concerns (11). Effectiveness uncertain due to novel use in forestry (11). Only attempted in one cultivar (14). Avoid trait selection that complicate s invasive control Maintain ability to control escaped Eucalypt us Little up front cost. Unknown if cold tolerance confers vigor that increases invasiveness. Does not affect traditionally bred trees. Harvest before seed maturation (5, 10) Prevent seed dispersal P otentially elimination of seed dispersal. Feasible to enforce. Unpredi ctable seed maturation. Possible e conomic losses if harvest must occur before optimum economic rotation length (15). Buffer zones around plantings (5, 10, 16) Reduce seed establishment Feasible to enforce maintenance of barrier. Low cos t. Disagreement over barrier size (5, 16) Does not account for long distance dispersal due to hurricanes (17). Unknown level of compliance for monitoring. Clean harvesting equipment (18) Reduce seed dispersal. Reduces risk of spread from easily identifia ble transportation vector. Impractical enforce ment Decreasing efficiency over time (19). Does not affect natural dispersal.
83 Tab le A 1. Continued. Intervention type Management practice Goal of Practice Advantages of practice Limitations of practice Monoclonal planting blocks (5) Reduce hybridization. Decrease area vulnerable to seed dispersal. Simple management, predictable and potentially increased yield from monoclonal plantings (20, 22) Increased risk of severe losses from disease. Negative impac ts on soil and biodiversity (20, 22). Control Mechanical control (22 24) Removal of escaped seedlings May be less injurious to environment compared to chemical control (22) Labor intensive and costly (23). Unlikely elimination of regrowth (22 24). Che mical control (23 29 ) Removal of escaped seedlings Better control compared to mechanical methods. Reapplication often necessary (23 26). Imprecise herbicide prescriptions (25, 28). Biocontrol (seeds) (30) Reduce rate of spread Growers can protect crops from damage. Self perpetuating barrier to spread. Lengthy regulatory process (31), costly to identify biocontrol agents (32), inability to protect seed supply (33), possible introduction of pathogens (34) Biocontrol (plants) Reduce vigor and population size Self perpetuating barrier to spread Lengthy regulatory process (31), costly to identify biocontrol agents (32), inability to protect crops from damage (33), possible introduction of pathogens (34) Impact management Ground and surface water protection (5, 10) Reduce adverse ecological impacts Planting away from waterways also reduces seed dispersal. Limited water availability to Eucalyptus may negatively affect yield. Wildfire protection (5, 10, 35) Reduce adverse ecological impacts Bare soil fire break also reduces dispersal and establishment. Labor and cost to reduce accumulated fuel load (23) Mitigate other ecological impacts Reduce adverse ecological impacts Management for key affected species may provide benefits to non target species. Impact s may not be easily quantified (36). May be incompatible with profitable Eucalyptus cultivation. Supporting references: (1) Gordon et al. 2011; (2) Gordon et al. 2012; (3) Gordon et al. 2008; (4) Daehler et al. 2004; (5) Flory et al. 2012; (6) Barney and DiTomaso 2008; (7) da Silva et al. 2011; (8) Emer and Fonseca 2010; (9) Davis et al. 2010; (10) Booth 2012; (11) FAO 2010a; ( 12) Wang and Brummer 2012; (13) Strauss and Viswanath 2011; (14) Hinchee 2011; (15) Langholtz et al. 2005; (16) FAC 2008; (17) Br ow der and Schroeder 1981; (18) NISC 2009; (19) Leung et al. 2005; (20) DeBell and Harrington 1993;(21) Zalesny et al. 2011; (22) Bean a nd Russo 1989; (23) NPS 2006; (24) Little and van den Berg (2006); (25) Bossard et al. 2000; (26) Bachelard et al. 1965; (2 7) Little 2003; (28) Moore 2002; (29) Morze 1971; (30) Wilson et al. 2011b; (31) Montgomery 2011; (32) Hobbs and Humphries 1995; (33) van Wilgen et al. 2011; (34) Hoffmann et al. 2011; (35) Goodrick and Santurf 2012; (36) Simberloff et al. 2012.
84 APPENDIX B AMINOCYCLOPYRACHLOR FOR CONTROL OF WOODY PLANTS : LITERATURE REVIEW Table B 1. A list of reported research for the control of woody plant species using aminocyclopyrachlor (AMCP) herbicides is given below The rate of AMCP tested other herbicides includ ed, method s of herbicide application and the outcome for control of each taxa are described as they were reported in the original reference s Tested taxa AMCP r ate a Ot her herbicides in tank mix b Application m ethod Reported o utcome Reference Acacia sp. SU F oliar E xcellent control Rick et al. 2009 B roadcast C ontrol Alford et al. 2012 Acer negundo SU F oliar E xcellent control Rick et al. 2009 Acer rubrum SU F oliar E xcellent control Rick et al. 2009 F oliar V ariable control Ezell et al. 2012 Imaz F oliar E xcellent control Ezell et al. 2012 Acropitlon repens 140 g ai ha 1 P r e emergence 26 37% control Sebastian et al. 2011 Albizia julibrissin 8.75 70 g ha 1 F oliar 53 100 % control 1 MAT Ko e pke Hill et al. 2012 Artes ima spp. 180 g ai ha 1 B roadcast G Hergert et al. 2011 Baccharis halimifolia 280 g ai ha 1 F oliar 95 100% control Ezell et al. 2012 280 g ai ha 1 M etsulfuron F oliar 95 100% control Ezell et al. 2012 Carya sp. 280 350 g pro duct ha 1 F oliar 70 80% control Ezell et al. 2012 Celtis laevigata SU F oliar E xcellent control Rick et al. 2009 Celtis occidentalis SU F oliar E xcellent control Rick et al. 2009 Diospyros virginiana 280 350 g product ha 1 > 90% control E zell et al. 2012 Elaeagnus angustifolia 140 280 g ai ha 1 F oliar 99% control 1 YAT Wilson et al. 2011 b 280 g ai ha 1 M etsulfuron F oliar 99% control 1 YAT Wilson et al. 2011 b 5% v/v B asal bark 99% control 1 YAT Wilson et al. 2011 b 5 25% v/v C ut stump S ig. diff. from check Edwards and Beck 2011 15% v/v B asal bark E ffective Edwards and Beck 2011 Fraxinus sp. Fo liar C ontrol Turner et al. 2009 P oor control Ezell et al. 2012
85 Table B 1. Continued. Tested taxa AMC P rate a Other herbicides in tank mix b Application method Reported outcome Reference Ilex vomitoria 2.5 15% v/v Cut stump 97 99% control 1.5 YAT Yeiser et al. 2011 10% v/v Triclo Cut stum p 100% control 540 DAT Yeiser et al. 2011 Kochia scoparia 140 315 g ai ha 1 SU Foliar Excellent control 1 12 MAT Turner et al. 2009 Lantana camara 200 g ai ha 1 Broadcast 98 100% 1 YAT Ferrell et al. 2012 Liquidambar styraciflua 2.5 15% v/v Cut stump 73 81% control 1.5 YAT Yeiser et al. 2011 10% v/v Triclo Cut stump 83% control 1.5 YAT Yeiser et al. 2011 Foliar Poor control Ezell et al. 2012 Imaz Fo liar Good/excellent control Ezell et al. 2012 Morella cerifera Triclo Marginal control Ezell and Yeiser 2010 Pinus sp. Excellent control Ezell and Yeiser 2010 Pro sopis gradnulosia Broadcast Control Alford et al. 2012 Prosopis sp. SU Foliar Excellent control Rick et al. 2009 Quercus alba 280 350 g product ha 1 Foliar Very good/excellent control Ezell et al. 2012 280 350 g product ha 1 Imaz or Gly F oliar Very good/excellent control Ezell et al. 2012 Quercus rubra Foliar Variable control Ezell et al. 2012 Imaz or Gly Foliar > 75% control Ezell et al. 2012 Rhamnus cathartica Foliar Control Turner et al. 2009 Rhus sp. Foliar Control Ezell et al. 2012 Robinia sp. Foliar Control Turner et al. 2009 Rosa multiflora Foliar Control Turner et al. 2009 Sarcobatus 329 g ai ha 1 Broadcast LaFantasie et al. 2012 Symphoricarpos occidentalis 126 g ai ha 1 Metsulfuron Foliar 99% control 1 YAT Wilson et al. 2011b Tamarix spp. 5% v/v Basal bark 99% control 1 YAT Wilson et al. 2011b 280 g ai ha 1 Foliar 33% control 1 YAT Wilson et al. 2011b Triadica sebifera 2.5 15% v/v Cut stump 77 100% control 1.5 YAT Yeiser et al. 2011 10% v/v Triclo Cut stump 100% control 1.5 YAT Yeiser et al. 2011 a A dash indicates that application rate was not described in the study report. b Product formulations of glyphosate (gly), imazapyr (imaz), triclopyr (triclo), metsulfuron methyl (metsulfuron), or sulfonylu rea herbicides (SU). A dash indicates no other herbicides were used in the tank mixture
86 LITERATURE CITED Adam s, P., C. Beadle, N. Mendham, and P. Smethurst. 2003. The impact of timing and duration of grass control on growth of a young Eucalyptus globulus Labill. plantation New Forest. 26:147 165.Alford, C. M., J. H. Meredith, E. P. Castner, and C. Medlin. 2012. Aminocyclopyrachlor: A New Active for Brush Control in Range and Pasture. Proc. West. Soc. Weed Sci. 65: 61 62. [Abstract] Anderson, M. C., H. Adams, B. Hope, and M. Powell. 2004. Risk Assessment for Invasive Species. Risk Anal. 24:787 793. Bachelard, E. P., A. Sarfaty, and P.M. Attiwill. 1965. Chemical Control of Eucalypt Vegetation. Aust. For. 29:181 191. Barney, J. N., and J. M. DiTomaso. 2008. Nonnative Species and Bioenergy: Are W e Cultivating the Next Invader? BioScience 58:1 7. BASF Corporation. 2008 Stalker herbicide product label. R esearch Triangle Park, NC. 9 p. Bean, C., and M. J. Russo. 1989. Environmental Stewardship Abstract for Eucalyptus globulus Arlington, VA: The Na ture Conservancy. 21 p. Becerra, P. I., and R. O. Bustamente. 2008. The effect of herbivory on seedling survival of the invasive exotic species Pinus radiata and Eucalyptus globulus in a Mediterranean ecosystem of Central Chile. Forest. Ecol. Manag. 256:15 73 1578. Bergerud, W.A. 1988. Dose Herbicide Injection System. Victoria, BC, Canada: Ministry of Forests and Lands Research Note 102. 28 p. Blazier, M. A., J. Johnson, E. L. Taylor, and B. Osbon. 2012. Herbicide site preparation and release options for Eucalyptus plantation establishment in the western gulf. Pages 19 23 in J. R. Butnor, ed. Proceedings of the 16th Biennial Southern Silvicultural Research Conference. Asheville, NC: USDA, Forest Service, Southern Research Station e Gen. Tech. Rep SRS 156. Boland, D. J. 1986. Testing and Storage of Eucalyptus and Acacia Seed. Pages 75 94 in R. D. Ayling, B.R.T. Seward, eds., Proceedings of a Workshop on Seed Handling and Eucalypt Taxonomy. Harare, Zimbabwe Booth, T. H. 2013. Eucalypts and Their Potential for Invasiveness Particularly in Frost Prone Regions. Int. J. For. Res. doi:10.1155/2012/837165. Bossard, C. C., J. M. Randall, and M. C. Hoshovsky. 2000. Invasive Plants of ey, CA: University of California Press.www.cal ipc.org. Accessed: January 22, 2013.
87 Bradley, B. A., D. M. Blumenthal, D. S. Wilcove, and L. H. Ziska. 2010. Predicting plant invasions in an era of global change. Trends Ecol. Evol. 25:310 18. Browder, J. A. and P. B. Schroeder. 1981. Melaleuca seed dispersal and perspective on control. Pages 17 21 in R. K. Geiger, ed., Proceedings of the Melaleuca Symposium Ft. Myers, Florida : Florida Dept. of Agriculture and Consumer Services, Division of Forestry, Tallah assee. Buddenhagen, C. E., C. Chimera, and P. Clifford. 2009. Assessing Biofuel Crop Invasiveness: A Case Study. PLoS ONE 4:1 6. Callaham Jr., M. A., J. A. Stanturf, W. J. Hammond, D. L. Rockwood, E. S. Wenk, and cape of Eucalyptus spp. Seedlings from Plantations in Southeastern USA. Int. J. For. Res. doi:10.1155/2013/946374. Carillo Gaviln, A., J. M. Espelta, and M. Vil. 2012. Establishment constraints of an alien and a native conifer in different habitats. Biol Invasions 13:1279 1289. Casler, M. D., K. P. Vogel, C. Taliaferro and R. L. Wynia. 2004. Latitudinal adaptation of switchgrass populations. Crop. Sci. 44:293 303. Chimera, C. G ., C. E. Buddenhagen, and P. M. Clifford. 2010. Biofuels: the risks and dange rs of introducing invasive species. Biofuels 1:785 596. Cremer K. W.1977. Distance of seed dispersal in Eucalypts estimated from seed weights. Aust. For. Restor. 7:225 228. Cribbie, R. A., R. R. Wilcox, C. Bewell, and H. J. Keselman. 2007. Tests for Treatm ent Group Equality When Data are Nonnormal and Heteroscedastic. J. Mod. Appl. Statistical Method. 6:117 132. da Silva, P.H.M., F. Poggiani, A. M. Sebbenn, and E. S. Mori. 2011. Can Eucalyptus invade native forest fragments close to commercial stands? Fores t Ecol. Manag. 26:2075 2080. Daehler, C. C., J. S. Denslow, S. Ansari, and H. C. Kuo. 2004. A Risk Assessment System for Screening out Invasive Pest Plants from Hawaii and Other Pacific Islands. Conserv. Biol. 18:360 368. Davis, A. S., R. D. Cousens, J. Hi ll, R. N. Mack, D. Simberloff, and S. Raghu. 2010. Screening bioenergy feedstock crops to mitigate invasion risk. Frontiers Ecol. Environ. 8:533 539. Davis, P. B., F. D. Menalled, R.K.D. Peterson, and B. D. Maxwell. 2011. Refinement of weed risk assessment s for biofuels using Camelina sativa as a model species. J. Appl. Ecol. 48:989 997.
88 DeBell, D. S., and C. A. Harrington. 1993. Deploying genotypes in short rotation plantations: Mixtures and pure cultures of clones and species. Forest. Chron. 69:705 713. D odet, M., and C. Collet. 2012. When should exotic forest plantation tree species be considered as an invasive threat and how should we treat them? Biol. Invasions 14:1765 1778. Dougherty, D., and J. Wright. 2012. Silviculture and economic evaluation of Euc alypt plantations in the southern US. BioResources 7:1994 2001. Dow AgroSciences LLC 2003. Garlon 3A herbicide product label. LOES No. 010 00084. Indianapolis, IN. 9 p. Dow AgroSciences LLC. 2008 Garlon 4 Ultra herbicide product label. LOES No. 010 02 127. Indianapolis, IN. 9 p. Edwards, R. J., and K. G. Beck. 2011. Control of Russian Olive Through Cut Stump and Basal Bark Herbicide Applications. Proc. West. Soc. Weed Sci. 64:70 71. [Abstract] E.I. du Pont de Nemours and Company. 2009. DuPont DPX MAT28 herbicide. Technical Bulletin No. K15023. Wilmington, DE. 7 p. Emer, C., and C. R. Fonseca. 2010. Araucaria Forest conservation: mechanisms providing resistance to invasion by exotic timber trees. Biol. Invasion 13: 189 202. Environmental Law Institute. 2 010. Status and Trends in State Invasive Species Policy: 2002 2009. Washington, D.C. http://100thmeridian.org/documents/ELI_Invasive_ Species_State_Po licy_Report_05_2010.pdf Accessed: December 6, 2012. Ewel, J. J., D. J. O'Dowd, J. Bergelson, C. C. Daehler, C. M. D'Antonio, L. D. G mez, D. R. Gordon, R. J. Hobbs, A. Holt, K. R. Hopper, C. E. Hughes, M. LaHart, R.R.B. Leakey, W. G. Lee, L. L. Loope, D H. Lorence, S. M. Louda, A. E. Lugo, P. B. McEvoy, D. M. Richardson, and P. M. Vitousek. Deliberate introductions of species: research needs. Bioscience 49:619 30. Ezell A. W., R. Turner, and J. L. Yeiser. 2012. Use of Aminocyclopyrachlor for forestry si te preparation in the southeastern U.S. Pages 120 121 in J. R. Butnor, ed., 2012. Proceedings of the 16th Southern Silvicultural Research Conference. Asheville, NC: USDA, Forest Service, Southern Research Station e Gen. Tech. Rep. SRS 156. Ezell, A. and J. L. Yeiser. 2010. Aminocyclopyrachlor (MAT28) for brush control. http://conference.ifas.ufl.edu/aw10/presentations/Wed/Session%20B/1340%20Ferrell.p df Acce ssed: January 23, 2012. [Presentation] [FAC] Florida Administrative Code. 2008. 5B 57.011 Biomass Plantings. Pages 6 7 in Introduction or Release of Plant Pests, Noxious Weeds, Arthropods, and Biological
89 Control Agents. Gainesville, FL: Florida Dept. of Agriculture and Consumer Services, Division of Plant Industry. [FAO] Food and Agriculture Organization of the United Nations. 2010a. Forests and Genetically Modified Trees. Rome, Italy: FAO. 240 p. [FAO] Food and Agriculture Organization of the United Nat ions. 2010b. Eucalyptus in East Africa socio economic and environmental issues. Planted Forests and Trees Working Papers. Rome, Italy: FAO. 30 p. Federal Register. 1999. Executive Order 13112 of February 3, 1999 Invasive Species. Federal Register, Feb 8, 1999. 64(25):6183 6186. Ferrell, J., B. Sellers, and E. Jennings. 2012. Herbicidal Control of Largeleaf Lantana ( Lantana camara ). Weed Technol. 26:554 558. Flory, S. L., K. A. Lorentz, D. R. Gordon, and L. E. Sollenberger. 2012. Experimental approaches fo r evaluating the invasion risk of biofuel crops. Environ. Res. Lett. doi:10.1088/1748 9326/7/4/045904 Weeds in Eucal yptus globulus subsp. maidenii ( F. Muell ) establishment: effects of competition on sapling growth and survivorship New Forest. 37:251 264. Gonzalez, R., T. Treasure, J. Wright, D. Saloni, R. Phillips, R. Abt, and H. Jameel. 2011. Exploring the potential of Eucalyptus for energy production in the Southern United States: Financial analysis of delivered biomass. Part I. Biomass Bioenerg. 35:755 766. Goodrick, S. L., and J. A. Stanturf. 2012. Evaluating Potential Changes in Fire Risk from Eucalyptus Plantings in the Southern United States. Int. J. For. Res. doi:10.1155/2012/680246. Gootee, R. S., E. P. Weber, K. Blatner, M. Carroll, and D. Baumgartner. 2012. Regulation, Knowledge Transfer, and Forestry Policy Implementation: Different Strokes for Different Fol ks? Sustain. Ag. Res. 155 65. Gordon, D. R., D. A. Onderdonk, A. M. Fox, and R. K. Stocker. 2008. Consistent accuracy of the Australian weed risk assessment system across varied geographies. Divers. Distrib. 14:234 242. Gordon, D. R., S. L. Flory, A. L. C ooper, and S. K. Morris. 2012. Assessing the Invasion Risk of Eucalyptus in the Unites States using the Australian Weed Risk Assessment. Int. J. For. Res. doi:10.1155/2012/203768. Gordon, D.R., K. J. Tancig, D. A. Onderdonk, and C. A. Gantz. 2011. Assessin g the invasive potential of biofuel species proposed for Florida and the United States using the Australian Weed Risk Assessment. Biomass Bioenerg. 35:74 79.
90 Gunsolus, J. L., and W. S. Curran. 1999. Herbicide Mode of Action and Injury Symptoms. North Centr al Regional Publication 377. Harcourt, R. L., J. Kyozuka R. B. Floyd, K. S. Bateman, H. Tanaka, V. D e croocq, D. J. Llewellyn, X. Zhu, W. J. Peacock, and E. S. Dennis. 2000. Insect and herbicide resistant transgenic eucalypts. Mol. Breeding 6:307 315. H arfouche, A., R. Meilan, and A. Altman. 2011. Tree genetic engineering and applications to sustainable forestry and biomass production. Trends. Biotechnol. 29:9 17. Hergert, H. J., B. Mealor, R. D. Mealor, and A. D. Kniss. 2011. Seedling Response of 27 Nat ive Species and 2 Exotic Weeds to Aminocyclopyrachlor. Proc. West. Soc. Weed Sci. 64:22. [Abstract] Hinchee, M., C. Zhang, S. Chang, M. Cunningham, W. Hammond, and N. Nehra. 2011. Biotech Eucalyptus ple of freeze tolerant Eucalyptus in the southeastern U.S. BMC Proc. 5 (Suppl 7):I24. Hobbs, R. J., and S. E. Humphries. 1995. An integrated approach to the ecology and management of invasive plants. Conserv. Biol. 9 :761 770. Hodges, A. W., T. J. Stevens, and M. Rahmani. 2010. Economic Impacts of Expanded Woody Biomass Utilization on the Bioenergy and Forest Products Industries in Florida. Sponsored project final report to FDACS Division of Forestry. Gainesville, FL: University of Florida, IFAS, Food and R esource Economics Department. 36 p Hodgson, L. M. 1976. Some aspects of flowering and reproductive behaviour in Eucalyptus grandis (Hill) Maiden at JDM Keet Forest Research Station formally Zomerkomst Forest Research Station. S. Afr. For. J. 97:18 28. Ho ffmann, J. H., V. C. Moran, and B. W. van Wilgen. 2011. Prospects for the biological control of invasive Pinus species (Pinaceae) in South Africa. Afr. Entomol. 19:393 401. Hulme, P. E. 2011. Weed risk assessment: a way forward or a waste of time? J. Appl. Ecol. 49: 10 19. Humber, J. M., and L. Hermanutz. 2011. Impacts of non native plant and animal invaders on gap regeneration in a protected boreal forest. Biol. Invasions 13:2361 2377. [IFAS] Institute of Food and Agricultural Sciences. 2012. Florida Aut omated Weather Network. fawn.ifas.ufl.edu/ Accessed: January 10, 2013. Jenkins, P. T. 2012. Invasive animals and wildlife pathogens in the United States: the economic case for more risk assessments and regulation. Biol. Invasions doi:10.1007/s10530 012 0296 8.
91 Koepke Hill, R., G. R. Armel, J. T. Brosnan, G. K. Breeden, J. J. Vargas, and T. C. Mueller. 2012. Control of Silk Tree ( Albizia julib rissin ) with Aminocyclopyrachlor and Other Herbicides. Weed Sci. 60:345 349. Kull, C. A., and J. Tassin. 2012. Australian acacias: useful and (sometimes) weedy. Biol. Invasions doi:10.1007/s10530 012 0244 7. LaFantasie, J. J., B. A. Mealor, and A. R. Kniss 2012. Black Greasewood Community Response to Aminocyclopyrachlor. Proc. West. Soc. Weed Sci. 65: 25 26. [Abstract] Langholtz, M., D. R. Carter, D. L. Rockwood, J.R.R. Alavalapati, and A. Green. 2005. Effect of dendroremediation incentives on the profitab ility of short rotation woody cropping of Eucalyptus grandis Forest Policy Econ. 7:806 817. Le Maitre, D. C., B. W. van Wilgen, C. M. Gelderblom, C. Bailey, R. A. Chapman, and J. A. Nel. 2002. Invasive alien trees and water resources in South Africa: case studies of the costs and benefits of management. For. Ecol. Manag. 160: 143 159. Ledgard, N. 2001. The spread of lodgepole pine ( Pinus contorta Dougl ) in New Zealand. For. Ecol. Manag.141:43 57 Lee, C. E. 2002. Evolutionary genetics of invasive species. Trends Ecol. Evol. 17:386 391. Leung, B., D. M. Lodge, D. Finnoff, J. F. Shogren, and D. Lodge. 2005. Managing invasive species: Rules of thumb for rapid assessment. Ecol. Econ. 55:24 36. Lewandowski, I., and A. Faaij. 2006. Steps towards the development of a certification system for sustainable bio energy trade. Biomass Bioenerg. 30:83 104. Li, J., J. A. Duggin, C. D. Grant, and W. A. Loneragan. 2003. Changes in the early survival of Eucalyptus blakelyi in grasslands of the New England Tablelands, NSW, Au stralia. Forest Ecol. Manag.173:319 334. Little, D. L. and D. L. Shaner. 1991. Absorption and translocation of the imidazolinone herbicides. Pages 53 69 in The Imidazolinone Herbicides. CRC Press, Boca Raton, FL. Litt le, K. M. 2003. Killing Eucalyptus grandis cut stumps after multiple coppice rotations in the KwaZulu Natal midlands, South Africa. S. Afr. For. J. 199:7 13. Little, K. M., and G. J. van den Berg. 2006. First rotation Eucalyptus macarthurii cut stump contr ol in KwaZulu Natal. S. Afr. For. J. 207:15 20. Lockwood, J., M. Hoopes, and M. Marchetti. 2007. Invasion Ecology. Malden, MA: Blackwell Publishing. p 219. Lockwood, J.L.L., D. Simberloff, M. L. McKinney, and B. Von Holle. 2001. How many, and which, plant s will invade natural areas? Biol. Invasions 3:1 8.
92 Low, T. 2012. Australian acacias: weeds or useful trees? Biol. Invasions 14:2217 2227. Lowery, R. F., and D. H. Gjerstad. 1991. Chemical and mechanical site preparation. Pages 251 262 in M. L. Duryea, and P. M. Dougherty, eds., Forest Regeneration Manual. Dordrecht, Netherlands: Kluwer Academic Publishers. McCormick, N., and G. Howard. 2013. Beating back biofuel crop invasions: Guidelines on managing the invasion risk of biofuel developments. Renew. Energ. 49:263 266. Miller, J. H., and G. R. Glover, eds. 1991. Standard Methods for Forest Herbicide Research. Champaign, IL: Southern Weed Science Society. 68 p. Miller, J. H., S. T. Manning, and S. F. Enloe. 2010. A Management Guide for Invasive Plants in Sout hern Forests. Asheville, NC: USDA, Forest Service, Southern Research Station e Gen. Tech. Rep. SRS 156. 120 p. Minton, M. S., and R. N. Mack. 2010. Naturalization of plant populations: the role of cultivation and population size and density. Oecologia 164: 399 409. Montgomery, M. E. 2011. Understanding Federal Regulations as Guidelines for Classical Biological Control Programs. Pages 25 42 in B. Onken, and R. Reardon, eds., Implementation and Status of Biological control of the Hemlock Wooly Adelgid. USDA, Forest Service Publication FHTET 2011 04. Moore, K. 2008. Eradicating Eucalyptus Acacia and Other Invasive Trees. Capitola, CA: Wildlands Restoration Team. 5 p. Morze, J. 1971. Chemical control of Eucalyptus species. For. S. Afr. 12:49 53. Myers, R. L. 1 983. Site susceptibility to invasion by the exotic tree Melaleuca quinquenervia in southern Florida. J. Appl. Ecol. 20: 645 648. Naval Meteorology and Oceanography Command. 2012. Complete Sun and Moon Data for One Day: U.S. Cities and Towns. Stennis Space Center, MS. http://aa.usno.navy.mil/data/docs/RS_OneDay.php Accessed : May 15, 2012. [NISC] National Invasive Species Council. 2009. Biofuels: Cultivating Energy, not Invasive Species. Washi ngton, DC: National Invasive Species Council Paper 11. 4 p. [NOAA] National Oceanic and Atmospheric Administration. 2002. Monthly Station Normals of Temperature, Precipitation, and Heating and Cooling Degree Days 1971 2000. In Climatography of the United States No. 81.08 Florida. Asheville, NC: NOAA National Climatic Data Center and National Environmental Satellite, Data, and Information Ser vice. http://cdo.ncdc.noaa.gov/ climatenormal s/clim81/FLnorm.pdf Accessed: June 10, 2011. [NPS] National Park Service 2006. Eucalyptus, a Transcontinental Legacy. Fire management, Resource Protection, and the Challenges of Tasmanian Blue Gum.
93 Washington, DC: US Department of the Interior, NPS. http://biomass.forestguild.org/casestudies/1001/Eucalyptus.pdf Accessed : December 6, 2012. Osiecka, A. and P. J. Minogue. 2011. Preliminary results: Development of selective herbi cide treatments for establishment of Eucalyptus urograndis (FTE) and Eucalyptus benthamii plantations. Quincy, FL: Institute of Food and Agricultural Sciences, North Florida Research and Education Center Research Report 2011 01.14 p. Osiecka, A. and P. J. Minogue. 2012. Selective Herbicides for Bald Cypress Restoration and Cultivation. Weed Tech. 26:460 468. Pagni, P. J. 1993. Causes of the 20 October 1991 Oakland Hills Conflagration. Fire Safety J. 2:331 339. Paine, T. D., M. J. Steinbauer, and S. A. Law son. 2011. Native and Exotic Pests of Eucalyptus : A Worldwide Perspective. Annu. Rev. Entomol. 56:181 201. Panetta, F. D. 1993. A system for assessing proposed plant introductions for weed potential. Plant. Prot. Q. 8:10 14. Parker, I. M., D. Simberloff, W. M. Lonsdale, K. Goodell, M. Wonham, P. M. Kareiva, M. H. Williamson, B. Von Holle, P. B. Moyle, J. E. Byers, and L. Goldwasser. 1999. Impact : T oward a F ramework for U nderstanding the E cological E ffects of Invaders Biol. Invasions 1:3 19. Parker, I. M. and P. Kareiva. 1996. Assessing the risks of invasion for genetically engineered plants: acceptable evidence and reasonable doubt. Biol. Conserv. 78:196 208. Raghu, S., R. C. Anderson, C. C. Daehler, A. S. Davis, R. N. Wiedenmann, D. Simberloff, and R. N. Mack. 2006. Adding Biofuels to the Invasive Species Fire? Science 313:1742. Pheloung, P. C., P. A. Williams, and S. R. Halloy. 1999. A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. J. Environ. Manage. 57 : 239 25 1. Ramensteiner, E., and M. Simula 2002. Forest certification an instrument to promote sustainable forest management? J. Env. Manage.67:87 98. Rejmnek, M., and D. M. Richardson. 2011. Eucalypts. Pages 203 209 in D. Simberloff, and M. Rejmnek, eds., E ncyclopedia of Biological Invasions. Berkeley, CA: University of California Press. Richardson, D. M. 1998. Forestry trees as invasive aliens. Conserv. Biol. 12:18 26. Richardson, D. M., and M. Rejmnek. 2011. Trees and shrubs as invasive alien species a global review. Divers. Distrib. 17:788 809.
94 Richardson, D. M., and R. Blanchard. 2011. Learning from our mistakes: minimizing problems with invasive biofuel plants. Curr. Opin. Environ. Sustain. 3:36 42. Rick, S. K., R. G. Turner, J. R. Pitts, E. Hidalgo, and J. C. Claus. 2009. Aminocyclopyrachlor Blend Products for Brush and Weed Control on Utility and Roadside Rights of Way. North Central Weed Sci. Soc. Proc. 64:129. [Abstract] Rockwood, D., D. Carter, M. Langholtz, and J. Stricker. 2006. Eucalyptus and P opulus short rotation woody crops for phosphate mined lands in Florida USA. Biomass Bioenerg. 30:728 734. [RSB] Roundtable on Sustainable Biofuels 2010. RSB Principles and Criteria for Sustainable Biofuel Production. RSB STD 01 001 Version 2.0. http://rsb.org/pdfs/standards/11 03 08 RSB PCs Version 2.pdf Accessed : December 6, 2012. Sax, D. F. 2002. Equal diversity in disparate species assemblages: a comparison of native and exoti c woodlands in California. Global. Ecol. Biogeog. 11:49 57. Scarlat, N., and J.F. Dallemand. 2011 Recent developments of biofuels/bioenergy sustainability certification: A global review. Energ. Policy 39 :1630 1646. Sebastian, J. R., J. G. Beck, S. Nissen, D. Sebastian, and S. Rodgers. 2011. Native Species Establishment on Russian Knapweed Infested Rangeland Following Pre Plant Herbicides Applications. Proc. West. Soc. Weed Sci. 64:16. [Abstract] Senbeta, F., A. Seyoum, and T. Woldemariam. 2010. Is Eucalypt us farming a blessing Pages 160 170 in W. Tadesse, E. Tolosana, and R. Lpez, eds., Eucalyptus Species Management, History, Status and Trends in Ethiopia. Proceedings fr om the Congress held in Addis Ababa. Senseman, S. A., ed. 2007. Herbicide Handbook. Weed Science Society of America. Lawrence, KS. Pp. 244 245, 360. Simberloff, D., J. L. Martin, P. Genovesi, V. Maris, D. A. Wardle, J. Aronson, F. Courchamp, B. Galil, E. G arca h i, Trends Ecol. Evol. 1578:1 9. Strauss, S. H., and V. Viswanath. 2011. Field trials of GM trees in the USA: activi ty and regulatory developments. BMC Proc. 5 (Suppl 7):O61. Turner R. G., J. R. Pitts and D. D. Ganske. 2009. Aminocyclopyrachlor blend products for vegetation management on railroad and utility sites. Proc. North Central Weed Sci. Soc. 64:130. [Abstract]
95 [ USDA] United States Department of Agriculture, Agricultural Research Service. 2012a. Plant Hardiness Zone Map http://planthardiness.ars.usda.gov/PHZMWeb/ Accessed: January 10, 2013. [USDA] Unit ed States Department of Agriculture, Forest Service. 2008. Agriculture Handbook 727: The Woody Plant Seed Manual. Pp. 504 512. [USDA] United States Department of Agriculture, Forest Service. 2012b. Aminocyclopyrachlor Human Health and Ecological Risk Asses sment Final Report. Morgantown, WV: Forest Health Technology Enterprise Team SERA TR 056 01 03a. 212 p. van Wilgen, B. W., C. Dyer, J. H. Hoffmann, P. Ivey, D. C. Le Maitre, J. L. Moore, D. M. Richardson, M. Rouget, A. Wannenburgh, and J.R.U. Wilson. 2011. National scale strategic approaches for managing introduced plants: insights from Australian acacias in South Africa. Divers. Distrib. 17:1060 1075. van Wilgen, B. W., G. G. Forsyth, D. C. Le Maitre, A. Wannenburgh, J.D.F. Kotz, E. van den Berg, and L. H enderson. 2012. An assessment of the effectiveness of a large, national scale invasive alien plant control strategy in South Africa. Biol. Cons. 148:28 38. Virtue, J. G., and R. L. Melland. 2003. The Environmental Weed Risk of Revegetation and Forestry Pla nts. Adelaide, South Australia: Department of Water, Land and Biodiversity Conservation Report DWLBS2003/02. 184 p. Wang, Z., and C. E. Brummer. 2012. Is genetic engineering ever going to take off in forage, turf and bioenergy crop breeding? Ann. Bot. Lon don doi:10.1093/aob/mcs027. Westra, P., S. Nissen, D. Shaner, B. Lindenmayer, and G. Brunk. 2009. Invasive Weed Management with Aminocyclopyrachlor in the Central Great Plains. Proc. Weed Sci. Soc. Am. 49:407. [Abstract] Wevill, T., and J. Read. 2010. Fine scale patterns in the distribution of semi arid tree species at Wyperfeld National Park, southeastern Australia the potential roles of resource gradients vs. disturbance. J. Arid Environ. 74:482 490. Williamson, M. 1993. Invaders, weeds and the risk fro m genetically manipulated organisms. Experentia. 49:219 224. Wilson, J.R.U., C. Garifo, M. R. Gibson M. Arianoutsou, B. B. Bakar, S. Baret, L. Celesti Grapow, J. M. DiTomaso, J. Dufour Dror, C. Kueffer, C. A. Kull, J. H. Hoffman, F.A.C. Impson, L. L. Loop e, E. Marchante, H. Marchante, J. L. Moore, D. J. Murphy, J. Tassin A. Witt, R. D. Zenni, and D. M. Richardson. 2011 a Risk assessment, eradication, and biological control: global efforts to limit Australian acacia invasions. Divers. Distrib. 17:1030 1046
96 Wilson, R., G. Sbatella, G. and S. Young. 2011 b New Herbicides for Managing Invasive Plants in Range, Pasture, and Riparian Areas. Lincoln, NE: University of Nebraska Crop Production Clinics. 2 p. Witt, A.B.R. 2010. Biofuels and invasive species from an African perspective a review. GCB Bioenerg. 2:321 329. Wunderlin, R. P., and B. F. Hansen. 2008. Atlas of Florida Vascular Plants. Tampa, FL: Institute for Systematic Botany, University of South Florida. http://florida.plantatlas.usf.edu/ Accessed: December 6, 2012. Yeiser, J. L., M. Link and J. Grogan. 2011. Screening Cut Stump Control of Chinese Tallowtree, Sweetgum, and Yaupon with Aminocyclopyrachlor. Pages 389 393 in J. R. Butnor, ed., Proceedi ngs of the16th Biennial Southern Silvicultural Research Conference. Asheville, NC: USDA, Forest Service, Southern Research Station e Gen. Tech. Rep. SRS 156. Zalesn y, R. S., R. W. Cunningham, R. B. Hall, J. Mirck, and D. L. Rockwood, J. Santurf, and T. A. Volk. 2011. Woody Biomass from Short Rotation Energy Crops. Pages 27 63 in J. Zhu, ed., Sustainable Production of Fuels, Chemicals, and Fibers from Forest Biomass. Washington DC: American Chemical Society.
97 BIOGRAPHICAL SKETCH Kimberly Lore ntz was born and raised in Cleve land, Ohio. She received her Bachelor of Science degree in b iology from the University of Akron in 2011. As a n undergraduate research assistant there, she contributed to projects in the fields of genetics, biomechanics, anim al behavior and plant ecology. In the summer of 2010 she took an in intern ship at the Rocky Mountain Biological Laboratory in Colorado, where she completed research for her undergraduate honors thesis that compared nectar and pollen rewards for pollinator s across native and invasive confamilial plants. Her time at Rocky Mountain Biological Laboratory inspired her to enroll in the graduate program at the University of Florida to further pursue her interest in invasive plant ecology and management. Kimberly received her Master of Science in forest resources and conservation from the University of Florida in the spring of 2013.