<%BANNER%>

Flammability of native understory species in pine flatwood and hardwood hammock ecosystems

University of Florida Institutional Repository

PAGE 1

FLAMMABILITY OF NATIVE UNDERSTORY SPECIES IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS By ANNA LEE BEHM A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Anna Lee Behm

PAGE 3

This thesis is dedicated to my family Jim, Kathy, and Lisa Behm.

PAGE 4

ACKNOWLEDGMENTS This research was funded by the Southern Center for Wildland-Urban Interface Research and Information, Southern Research Station, USDA Forest Service under the National Fire Plan grant Assessing and Mitigating Wildfire Risk for Southern Wildland-Urban Interface Landowners. I am grateful to all those responsible for giving me this opportunity, especially Annie Hermansen, Ed Macie, Dr. Wayne Smith, and my advisor Dr. Mary Duryea. I am indebted to my committee, Dr. Mary Duryea, Dr. Alan Long, and Dr. Timothy Martin for their guidance. Additional guidance for this project came from the members of the Southern Wildland-Urban Interface Council (SWUIC). This study could not have been possible without the field and laboratory support from the USDA Forest Service, Florida Division of Forestry, Florida Suwannee River Water Management District, and Advanced Environmental Labs; and the Department of Animal Science, Department of Soil and Water Science, and Austin Cary Memorial Forest at the University of Florida. Gratitude is extended especially for the personal assistance of Cotton Randall, Sarah Bouchard, Dave Nolletti, and Al Boning. I am also grateful for all those who assisted me in sample and data collection in the field including Michele Nisi, Daniel Merced, Christian Quijano, Dr. Wayne Zipperer, Cotton Randall, Annie Hermansen, Dr. Mary Duryea, Eric Holzmueller, Brian Becker, and Dr. Jim Behm. I appreciate assistance with this research from the IFAS Department of Statistics and the University of Florida Herbarium. iv

PAGE 5

Sincere appreciation is extended to those who have gone out of their way to enrich my research and academic experiences over the years (especially Dr. Mark Gleason and Dr. Jan Thompson at Iowa State University, and Dr. Wayne Zipperer from the USDA Forest Service). I thank my colleagues and friends at the School of Forest Resources and Conservation for adding diversity of field and perspective to my education. I also appreciate the attention and patience offered by the faculty and staff at the School of Forest Resources and Conservation. Finally, I thank my dear friends Lisa Young, Brian Amidon, Angela Sokolowski, and Michel Masozera as well as my entire family for their support and encouragement felt from great distances. v

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES...........................................................................................................xi ABSTRACT......................................................................................................................xii CHAPTER 1 INTRODUCTION........................................................................................................1 Background Information...............................................................................................1 Problem..................................................................................................................6 Justification............................................................................................................7 Objectives.....................................................................................................................8 Associated Hypotheses..........................................................................................8 Assessment............................................................................................................9 Plant Flammability........................................................................................................9 Ecosystem Context.....................................................................................................10 Florida Ecosystems..............................................................................................10 Pine Flatwoods....................................................................................................11 Hardwood Hammock...........................................................................................11 Thesis Overview.........................................................................................................14 2 PLANT STRUCTURAL FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS...........................................................16 Introduction.................................................................................................................16 Materials and Methods...............................................................................................18 Selection of Sites and Plants................................................................................19 Litter Measurements............................................................................................20 Height Measurements..........................................................................................21 Biomass Measurements.......................................................................................21 Moisture Content.................................................................................................22 Statistical Analyses..............................................................................................23 Results.........................................................................................................................25 Site Characterization...........................................................................................25 Ecosystem Differences........................................................................................26 vi

PAGE 7

Litter measurements.....................................................................................26 Height measurements...................................................................................29 Biomass measurements................................................................................30 Species Differences.............................................................................................34 Litter measurements.....................................................................................35 Height measurements...................................................................................35 Biomass measurements................................................................................36 Discussion...................................................................................................................38 Ecosystem Differences........................................................................................39 Species Differences.............................................................................................42 Gaylussacia dumosa.....................................................................................43 Ilex glabra....................................................................................................44 Lyonia ferruginea.........................................................................................44 Vaccinium myrsinites...................................................................................44 Callicarpa americana..................................................................................44 Ilex opaca.....................................................................................................45 Quercus nigra...............................................................................................45 Vaccinium arboreum....................................................................................45 Myrica cerifera.............................................................................................45 Serenoa repens.............................................................................................45 Conclusions.................................................................................................................46 3 FOLIAR FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS....................................................................................48 Introduction.................................................................................................................48 Materials and Methods...............................................................................................50 Moisture Content.................................................................................................50 Leaf Area.............................................................................................................50 Volatile Solids.....................................................................................................51 Energy Content....................................................................................................51 Statistical Analyses..............................................................................................53 Results.........................................................................................................................54 Ecosystem Differences........................................................................................54 Leaf area.......................................................................................................54 Moisture content...........................................................................................55 Volatile solids...............................................................................................55 Energy content..............................................................................................55 Species Differences.............................................................................................56 Leaf area.......................................................................................................56 Moisture content...........................................................................................56 Volatile solids...............................................................................................57 Energy content..............................................................................................57 Serenoa repens.............................................................................................61 Discussion...................................................................................................................63 Ecosystem Differences........................................................................................63 Species Differences.............................................................................................65 vii

PAGE 8

Gaylussacia dumosa.....................................................................................66 Ilex glabra....................................................................................................66 Lyonia ferruginea.........................................................................................67 Vaccinium myrsinites...................................................................................67 Callicarpa americana..................................................................................68 Ilex opaca.....................................................................................................68 Quercus nigra...............................................................................................68 Vaccinium arboreum....................................................................................69 Myrica cerifera.............................................................................................69 Serenoa repens.............................................................................................70 Conclusions.................................................................................................................70 4 CONCLUSIONS AND RECOMMENDATIONS.....................................................72 Conclusions.................................................................................................................72 Recommendations.......................................................................................................74 Pine Flatwoods....................................................................................................74 Hardwood Hammocks.........................................................................................75 Future Research...................................................................................................75 APPENDIX A LOCATION OF PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS BY COUNTY IN FLORIDA...........................................................77 B LOCATION OF STUDY SITES................................................................................79 C ABSOLUTE AND RELATIVE DENSITIES OF UNDERSTORY AND MIDSTORY SPECIES AT EACH SITE...................................................................80 D ABSOLUTE AND RELATIVE DENSITY, RELATIVE DOMINANCE, RELATIVE FREQUENCY, AND IMPORTANCE VALUES FOR OVERSTORY SPECIES............................................................................................86 E MOISTURE CONTENT OF BIOMASS COMPONENTS.......................................90 F BIOMASS COMPONENTS FOR LOWEST 1-METER INTERVAL......................92 G EPA VOLATILE SOLID ANALYSIS METHODOLOGY......................................94 LIST OF REFERENCES...................................................................................................95 BIOGRAPHICAL SKETCH...........................................................................................100 viii

PAGE 9

LIST OF TABLES Table page 1-1 Common and scientific names of plants listed as appropriate for firewise landscaping according to two extension publications in Florida...............................4 1-2 Species common to hardwood hammock ecosystems across a moisture gradiant...12 2-1 Collection dates for study sites.................................................................................20 2-2 Percent of total understory stems characterized by the species studied...................25 2-3 Understory, midstory, and overstory stems per hectare at each site........................27 2-4 Basal area, height to lowest branch, and canopy closure at each site......................28 2-5 Litter depth and density for each species and ecosystem type.................................29 2-6 Dry weight of fine fuel biomass components for each species and ecosystem type.........................................................................................................31 3-1 Leaf area per leaf, specific leaf area, and leaf area per plant for species in flatwood and hardwood ecosystems.........................................................................58 3-2 Foliar moisture content and volatile solids for species within flatwood and hardwood ecosystems...............................................................................................59 3-3 Additional data for Serenoa repens included measurements of moisture content, volatile solid content, total energy content for dead foliage in addition to live foliage.......................................................................................................................62 4-1 Summary table comparing the results from flammability studies in Tasmania, South Africa, and Ethiopia.......................................................................................73 4-2 Proposed ranking mechanism of plant species by flammability..............................76 C-1 Absolute (stems per hectare) and relative densities (%) of understory species in 8, 12.56m2 plots at each study site.......................................................................81 C-2 Absolute (stems per hectare) and relative density (%) for midstory species (>3 m in height but <6.4 cm dbh) in 4, 400m2 plots at each study site....................84 ix

PAGE 10

D-1 Stems per hectare, relative density, relative dominance, relative frequency, and importance value for overstory species within each site...................................87 E-1 Moisture content (% dry weight) of fuel components..............................................91 F-1 Biomass separated into components for the lowest 1-m interval.............................93 x

PAGE 11

LIST OF FIGURES Figure page 1-1 Projected population pressures on forests through 2020............................................2 1-2 Lean, clean, and green defensible space zone surrounding a WUI structure in Florida after a wildfire................................................................................................4 1-3 Decision criteria for exclusion of garden or landscape vegetation as nonhazardous for vegetation clearance measurements in post-fire assessments.............8 1-4 Pine flatwood ecosystem with routine prescribed fire.............................................15 1-5 Typical hardwood hammock ecosystem..................................................................15 2-1 Total biomass by coarse fuel and fine fuel components for species within pine flatwoods and hardwood hammocks........................................................................32 2-2 Fuel loading for each species and ecosystem type...................................................33 2-3 Biomass per 1-meter height intervals by ecosystem type........................................34 2-4 Height and height to lowest branch of species within pine flatwoods and hardwood hammock ecosystems..............................................................................37 2-5 Biomass per 1-meter height intervals for species within pine flatwoods and hardwood hammocks................................................................................................39 2-6 Pine needles accumulated on a landscape plant.......................................................41 3-1 Foliar energy content in calories per gram, kilocalories per plant, and kilocalories per hectare for species within pine flatwood and hardwood hammock ecosystems...............................................................................................60 A-1 Distribution of pine flatwoods by county in Florida................................................78 A-2 Distribution of hardwood hammocks by county in Florida.....................................78 B-1 Location of pine flatwood and hardwood hammock study sites throughout North Central Florida...............................................................................................79 xi

PAGE 12

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FLAMMABILITY OF NATIVE UNDERSTORY SPECIES IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS By Anna Lee Behm August 2003 Chair: Mary L. Duryea Major Department: School of Forest Resources and Conservation The flammability of plants contributes to fire behavior and fire regimes in natural ecosystems. As urban development continues within and near fire-dependent ecosystems in the southern United States, the flammability of plants in those ecosystems may influence the survival of human-built structures during wildfire. To assess the importance of plant flammability in the wildland-urban interface (WUI), we compared flammability componentsignitability, sustainability, combustibility and consumabilityof native understory species commonly found in pine flatwood and hardwood hammock ecosystems in the southern United States. During the summer of 2002, six species from five pine flatwood sitesdwarf huckleberry (Gaylussacia dumosa [Andr.] A. Gray), gallbery (Ilex glabra [L.] A. Gray), rusty lyonia (Lyonia ferruginea [Walt.] Nutt.), evergreen blueberry (Vaccinium myrsinites Lam.), wax myrtle (Myrica cerifera L.), and saw palmetto (Serenoa repens [Bartr.] Small)and six species from five hardwood hammock sitesAmerican xii

PAGE 13

beautyberry (Callicarpa americana L.), American holly (Ilex opaca Ait. var opaca), water oak (Quercus nigra L.), sparkleberry (Vaccinium arboreum Marsh.), wax myrtle and saw palmettowere harvested for biomass analyses. Plant components were separated into live and dead foliage, accumulated litter on and under the plant, and small (<0.6 cm diameter) and large (>0.6 cm diameter) twigs, branches, and stems. Foliar biomass was further analyzed for leaf area, volatile solids, and energy content. Statistical analyses revealed differences between ecosystem types and among species. Understory plants in pine flatwoods have higher ignitability (lower moisture content), sustainability (higher fuel loading), and combustability (higher energy content), whereas the measurements for consumability was not different between ecosystem (fine fuel biomass and volatile solid content). Understory species in hardwood hammocks contain more total biomass because they contain more coarse fuel; yet coarse fuels do not affect the flammability of species as live coarse fuels are not typically consumed in a wildfire. Serenoa repens, a species common to both ecosystems, had relatively low foliar energy content per gram, but individual plants present a high hazard because they contain a great amount of biomass. Ilex glabra is also hazardous to WUI structures because it has high foliar energy content and a great amount of foliar biomass per hectare in flatwoods. Callicarpa americana plants present the lowest fire hazard to WUI structures compared to all other species studied. Results will assist the development of regionally specific guidelines for firewise landscaping and guide future plant flammability research. xiii

PAGE 14

CHAPTER 1 INTRODUCTION Background Information The wildland-urban interface (WUI) is a growing part of the landscape in the southern United States presenting new challenges to natural resource agencies, urban and rural fire agencies, city and county governments, and insurance companies. The wildland-urban interface can be defined from many perspectives: geographical, natural resource, sociopolitical, biophysical, and fire (Hermansen and Macie 2002). Geographically defined, the WUI is separated into three categories; classic, mixed, and isolated (Davis 1988). Classic interface areas occur where existing urban centers expand and invade wildlands. Mixed interface areas are formed as urban areas are built within and surrounded by wildland areas. Isolated interface areas are isolated islands of wildland vegetation are completely surrounded by urban areas which are sometimes used as city parks. From a natural resource perspective, the WUI is an area where increased human influence and land use conversion are changing natural resource goods, services, and management (Hermansen and Macie 2002). The WUI is not a new feature in the landscape of the southern United States. It has existed for as long as human habitation. Native Americans and early European settlers, living close to surrounding ecosystems, routinely cleared vegetation around homes through physical removal, routine prescribed fire, and livestock grazing (Pyne et al. 1996). However, the WUI is expanding at increasing rates in the southern US. During the 1980s, the rate of rural population growth was 3.1% in the South (Cordell and Macie 1

PAGE 15

2 2002). This rate increased to 7.5% in the South during the 1990s (Cordell and Macie 2002). Based on data from Cordell and Overdevest (2001), the population pressures on forests through year 2020 are projected in Figure 1-1. This shows areas in the South (by county) that are predicted to experience population growth in forested areas. Figure 1-1. Projected population pressures on forests through 2020. Darker colors indicate where population pressures on forests are expected to be the heaviest. From Figure 2.10 on page 27 in Cordell, H. K. and E. A. Macie. 2002. Population and demographic trends. Pages 11-35 in Human Influences on Forest Ecosystems: The Southern wildland-urban interface assessment. U. S. Department of Agriculture, Forest Service. GTR-SRS-55. 160 p. In the South, the forests and many other ecosystems associated with the WUI are dependent on fire for periodic renewal and to sustain species composition (Myers and Ewel 1990, Pyne et al. 1996, Walker and Oswald 1999). The warm and humid climate in the South supports rapid growth of understory plants; and frequent lightning storms provide a consistent source of ignition (Myers and Ewel 1990, Pyne et al. 1996, Walker

PAGE 16

3 and Oswald 1999). Wildfire, because of the direct effect on human health and safety, is one of the most recognized challenges within the WUI. Fire agencies define the WUI as an area where residential or commercial development is in or adjacent to areas prone to wildfire (Hermansen and Macie 2002 from Davis and Marker 1987, Tokle 1987). As population in the wildland-urban interface increases, the economic and social consequences of wildfire increase. In addition, even with advanced firefighting training and technology, firefighting agencies struggle to address complexities of fire in the wildland-urban interface (Swinford et al. 1987). The complexities are exacerbated when fire is unnaturally kept out of the interface, further increasing homeowners risk of wildfire damage. The scope of the WUI wildfire problem has prompted many fire and wildfire management agencies to initiate cooperative extension programs for communities and individual homeowners (Monroe 2002). In addition, governments have initiated policies to modify the behavior of developers, communities, and homeowners to alleviate the social burden of wildfire in the WUI (Monroe 2002). Many of these programs, often termed firewise, are focused on increasing the responsibility of the private sector and citizens in wildfire preparation. Recommendations include guidelines for entry and access, building materials, and landscaping. Firewise landscaping typically involves creating defensible space zones around homes at risk of wildfire. Defensible space is an area surrounding a home to allow easy access for firefighting equipment and personnel. Defensible space is also created to reduce the risk of wildfire damage if firefighting agencies are unable to defend each home. Typically, a defensible space zone of at least 30 feet (9 meters) surrounding a

PAGE 17

4 WUI structure should be established (Florida Firewise Communities 2000, Monroe and Long 2001). Within this zone, landowners are instructed to keep the area lean, clean, and green (Florida Firewise Communities 2000) (Figure 1-2). This includes pruning trees so that the lowest branches are at least 10 feet (3 meters) from the ground, removing vines and shrubs from under trees, clearing yard waste and firewood, and planting less flammable plant material in isolated landscape beds (Monroe and Long 2001). Figure 1-2. Lean, clean, and green defensible space zone surrounding a WUI structure in Florida after a wildfire. Photograph by Cotton Randall. Lists of plant species appropriate for firewise landscaping include native and ornamental species (Table 1-1). Table 1-1. Common and scientific names of plants listed as appropriate for firewise landscaping according to two extension publications in Florida. Common name Scientific name Monroe and Long 2001 MacCubbin and Mudge 2002 African iris Dietes iridioides [L.] N.E. Br. X Banana Musa spp. X Beautyberry Callicarpa americana L. X Bird of paradise Strelitzia reginae Banks X Black cherry Prunus serotina Ehrh. X Blue beech Carpinus caroliniana Walt. X

PAGE 18

5 Table 1-1. Continued Common name Scientific name Monroe and Long 2001 MacCubbin and Mudge 2002 Bottlebrush Callistemon spp. X Bugleweed Ajuga reptans L. X Century plant Agave decipiens Baker X Citrus Citrus spp. X Coontie Zamia pumila L. X Daylily Hemerocallis spp. X Dogwood Cornus spp. X Ferns multiple genus X Florida soapberry Sapindus marginatus Willd. X Fringetree Chionanthus virginicus L. X Hophornbeam Ostrya virginiana [Mill.] K. Koch X Indian hawthorn Rhaphiolepis indica [L.] Lindl. X Lantana Lantana camara L. X Ligustrum Ligustrum spp. X Lily of Nile Agapanthus praecox Willd. Subsp. Orientalis [F. M. Leight.] X Liriope Liriope muscari [Decne.] L.H. Bail. X Loquat Eriobotrya spp. X Magnolia Magnolia spp. X X Maples Acer spp. X Oaks Quercus spp. X Oleander Nerium oleander L. X Peach Prunus persica [L.] Batsch. X Periwinkle Vinca major L. X Persimmon Diospyros virginiana L. X Philodendron Philodendron spp. X Pineapple guava Acca sellowiana [O. Berg] Berret X Pittosporum Pittosporum tobira [Thunb.] Ait. X Plum Prunus spp. X Pomegranate Punica granatum L. X Pyrancantha Pyrancantha coccinea M.J. Roem X Red maple Acer rubrum L. X Redbud Cercis evolutes L. X X Sago Cycas evolute Thunb. X Society garlic Tulbaghia violacea L. X Sparkleberry Vaccinium arboreum Marsh. X Star jasmine Jasminum multiflorum [Burm.f.] Andr. X

PAGE 19

6 Table 1-1. Continued Common name Scientific name Monroe and Long 2001 MacCubbin and Mudge 2002 Sugarberry Celtis laevigata Willd. X Sweetgum Liquidambar styraciflua L. X Sycamore Platanus occidentalis L. X Viburnum Viburnum spp. X X Wild azalea Rhododendron canescens [Michx.] Sweet X X Wild olive Osmanthus americanus [L.] A. Gray X Wild plum Prunus umbellata Ell. X Winged elm Ulmus alata Michx. X In addition, highly flammable species such as saw palmetto, wax myrtle, yaupon holly (Ilex vomitoria Ait.), red cedar (Juniperus virginiana L.), gallberry, juniper (Juniperus spp.), pampas grass (Cortaderia selloana Schult. & Schult.f.), arborvitae (Platycladus orientalis L.), American holly, Italian cypress (Cupressus sempervirens L.), eucalyptus (Eucalyptus spp.), pine (Pinus spp.), Leyland cypress (Cupressocyparis leylandii [A.B. Jacks &Dallim.] Dallim.& AB. Jacks.), and fountain grass (Pennistetum spp.) are discouraged from being planted within firewise landscaping (Monroe and Long 2001, MacCubbin and Mudge 2002). Problem Landowners in the wildland-urban interface are instructed to reduce the number of flammable plants on their property and to plant species that are less flammable (Lippi and Kuypers 1998, Florida Firewise Communities 2000, Monroe and Long 2001). However, the species lists given to homeowners frequently have an unknown origin (Frommer and Weise 1995). In many cases, species lists are generated from other lists originating in different regions in the US (Lippi and Kuypers 1998). As firewise landscaping recommendations develop, there is a greater need for more regionally specific and

PAGE 20

7 scientifically founded species lists. Many characteristics are known to affect flammability making it difficult to generalize flammability based on a few characteristics. To complicate the situation further, no standard method of quantifying plant flammability has been fully accepted making comparing results from plant flammability studies difficult. Justification In post-fire assessments of structural survival, brush or shrub clearance >30 feet from a structure has shown to increase structural survival in wildfires in the southern United States (Abt et al. 1987, De Witt 2000). Appropriate flammability rankings for species within southern US ecosystems would be valuable to landowners, landscape architects, and nurseries in the wildland-urban interface. Given the definition of the WUI, the landscaping around interface homes includes many species native to the surrounding ecosystem. By comparing the flammability of understory species within two ecosystems, more information will be available to address landscaping hazards within these ecosystems. As firewise landscaping policies become more regionally specific, information relevant to specific ecosystem types will be desirable. In addition, knowledge of flammability characteristics of specific plants may support future post-fire research. Flammability ratings for species would be valuable to future post-fire analysis of structures. Foote et al. (1991) suggest a dichotomous approach to determining which vegetation should be measured to assess vegetation clearance for a structure threatened or damaged after a wildfire (Figure 1-3).

PAGE 21

8 Is vegetation YES Measure structure to vegetation clearance distance Ornamental shrubbery, ground cover, or an isolated tree? YES YES Able to readily transmit fire to structure? Flammable? NO NO NO Ignore Clearance Ignore Clearance Figure 1-3. Decision criteria for exclusion of garden or landscape vegetation as non-hazardous for vegetation clearance measurements in post-fire assessments. (Foote et al. 1991) Objectives The purpose of this study was to examine the characteristics that are known to influence flammability and to compare the flammability of understory species native to pine flatwood and hardwood hammock ecosystems. The specific objectives of the study were to determine the flammability in the context of structural and foliar characteristics. Structural components known to influence flammability at the plant level were quantified including fine fuel biomass and arrangement of biomass throughout the plant. Foliage from understory species was analyzed for moisture content, volatile solid content, and energy content. Associated Hypotheses This study had three major hypotheses: Hypothesis I: Understory plants in pine flatwoods are more flammable than understory plants in hardwood hammocks. Hypothesis II: Differences in the flammability of species will be significant. Hypothesis III: The flammability of wax myrtle and saw palmetto will be different between ecosystems.

PAGE 22

9 Assessment Beyond the above hypotheses, we interpreted the results of this study in the context of firewise landscaping recommendations. To do this, an assessment of the results centered around two questions: Assessment I: What are the implications for firewise landscaping in these ecosystems? Assessment II: In what way would individual plants contribute to structure survival or destruction in a wildfire? Plant Flammability Flammability was initially defined by Anderson in three components: ignitability, sustainability, and combustibility (1970). The ignitability component is the time until ignition once exposed to a heat source. Sustainability is the stability of burning rate, or the ability to sustain fire once ignited. Combustibility is defined as the rate of burn after ignition. The definition of flammability has since been expanded to include consumability, the proportion of mass or volume consumed by fire (Martin et al. 1994). Anderson (1970) related the flammability components of individual plants to fire characteristics at an ecosystem level. Ignitability of individual plants drives the chain of ignition in an ecosystem. Sustainability is related to the rate of fire spread and combustibility to fire intensity. The consumability of individual plants is analogous with the amount of fuel available for fire consumption on the ecosystem level (Martin et al. 1994). In this way, plant flammability is an important component of the natural fire regimes of ecosystems (Bond and van Wilgen 1996). In addition, plant flammability can influence wildfire behavior affecting the survivability of human-built structures.

PAGE 23

10 Ecosystem Context This study assesses flammability in the context of two ecosystems of Florida which are ecosystems in the coastal plain physiographic region of the southern United States. The natural role of fire in pine flatwoods and hardwood hammock ecosystems is discussed in this section. Florida Ecosystems Residential landscaping, like many characteristics of the interface, is a mix of both natural and synthetic landscapes. The natural ecosystem plays a large role in the design of landscapes around structures. Kuchler (1964) described eleven potential vegetation states for Florida in his categorization of vegetation types for the United States. The Florida Natural Areas Inventory (FNAI) in partnership with the Florida Department of Natural Resources and the Florida Nature Conservancy, have defined over sixty ecosystem types in Florida (1990). Myers and Ewel (1990) separate the upland, freshwater wetlands and aquatic, and coastal ecosystems into 13 Florida ecosystem types. Fire plays an important role in the function of almost every terrestrial ecosystem in Florida, and all terrestrial ecosystems will burn under the right conditions. Pine flatwoods and hardwood hammocks are named according to Myers and Ewel (1990), but the corresponding classification for FNAI (1990) and Kuchler (1964) are also given in the descriptions below. These two ecosystems, along with all others, are being affected by urbanization and are important in terms of the wildland-urban interface areas of Florida. See Appendix A for maps by county of each ecosystem type in Florida, adapted from Myers and Ewel (1990).

PAGE 24

11 Pine Flatwoods The Florida pine flatwoods (Abrahamson and Hartnett 1990) correspond to mesic flatwoods, scrubby flatwoods, wet flatwoods (FNAI 1990), and southern mixed forest (Kuchler 1964). Pine flatwoods are dominated by longleaf pine (Pinus palustris Mill.), slash pine (P. elliottii Engelm.), loblolly pine (P. taeda L.), south Florida slash pine (P. elliottii var. densa Little & Dorman), and pond pine (P. serotina Michx.) (Figure 1-4). In central and north Florida, live oak (Quercus virginiana Mill.), water oak, sweetgum, red maple, and ash (Fraxinus spp.) may add to the overstory composition. Understory composition consists of saw palmetto, gallberry, staggerbush or fetterbush (Lyonia spp.), dwarf huckleberry, wax myrtle, and tar flower (Befaria racemosa Vent.). Wiregrass (Astrida spp., Sporobolus spp.) is the dominant herbaceous ground cover in longleaf pine flatwoods. (Abrahamson and Hartnett 1990) The pine flatwood ecosystems are highly dependent on fire. Because they have a relatively open canopy, understory vegetation biomass can accumulate very rapidly. A fire frequency of every one to eight years is typical (FNAI 1990). When these areas burn, large areas of understory burn completely and because of the natural shedding of lower limbs by natural pine species, crowning of fires does not typically occur (Abrahamson and Hartnett 1990). Hardwood Hammock The temperate hardwood forests in Florida are referred to as hammocks (Platt and Schwartz 1990) (Figure 1-5). The temperate hardwoods inhabit xeric, mesic, and hydric moisture regimes with differing vegetation across this moisture gradient (Table 1-2). Temperate hardwood forests in xeric and mesic environments fit into the Kuchler (1964) classification of southern mixed forests. Those in hydric environments are classified as

PAGE 25

12 southern floodplain forest (Kuchler 1964). FNAI (1990) classifies temperate hardwood forests by moisture regime including xeric hammock, upland glade, upland hardwood forest, upland mixed forest, slope forest, and hydric hammock. Table 1-2. Species common to hardwood hammock ecosystems across a moisture gradiant (Platt and Schwartz 1990, FNAI 1990). Hardwood Hammock Common Name Scientific Name American holly Ilex opaca Ait. var opaca black cherry Prunus serotina Ehrh. blackjack oak Q. incana Bartr. Chapman oak Q. chapmanii Sarg. laurel oak Q. hemisphaerica Bartr. ex Willd. live oak Quercus virginiana Mill. persimmon Diospyros virginiana L. pignut hickory Carya glabra [Mill.] Sweet red bay Persea borbonia [L.] Spreng. sand live oak Q. geminata Small sand post oak Q. margaretta Ashe saw palmetto Serenoa repens [Bartr.] Small southern magnolia Magnolia grandiflora L. Southern red oak Q. falcata Michx. sparkleberry Vaccinium arboreum Marsh. staggerbush Lyonia spp. turkey oak Q. laevis Walt. wild olive Osmanthus americanus [L.] Benth. & Hook. F. ex A. Gray Xeric Hammock yaupon Ilex vomitoria Ait. American holly Ilex opaca Ait. var opaca Carolina holly Ilex ambigua [Michx.] Torr. var. ambigua devils walking stick Aralia spinosa L. eastern hophornbeam Ostrya virginiana [Mill.] K. Koch Florida maple Acer saccharum Marsh. flowering dogwood Cornus florida L. live oak Q. virginiana Mill. loblolly pine Pinus taeda L. pignut hickory Carya glabra (Mill.) Sweet redbud Cercis canadensis L. southern magnolia Magnolia grandiflora L. spruce pine Pinus glabra Walt. swamp chestnut oak Q. michauxii Nutt. Mesic Hammock sweetgum Liquidambar styraciflua L.

PAGE 26

13 Table 1-2. Continued Hardwood Hammock Common Name Scientific Name bluestem palmetto Sabal minor [Jacq.] Pers. cabbage palm Sabal palmetto [Walt.] Lodd. ex J.S. Schult. & J.H. Schult. dahoon holly Ilex cassine L. diamond-leaf oak Q. laurifolia Michx. hackberry Celtis laevigata Willd. loblolly pine Pinus taeda L. myrsine Myrsine floridana A. DC. needle palm Rhapidophyllum hystrix [Pursh] Wendle. & Drude pepper vine Ampelopsis arborea [L.] Koehne poison ivy Toxicodendron radicans [L.] Kuntze. rattanvine Berchemia scandens [Hill] K. Koch red cedar Juniperus virginiana L. red maple Acer rubrum L. royal fern Osmunda regalis L. saw palmetto Serenoa repens [Bartr.] Small southern magnolia Magnolia grandiflora L. swamp bay Persea palustris [Raf.] Sarg. swamp chestnut oak Q. michauxii Nutt. sweetbay Magnolia virginiana L. sweetgum Liquidambar styraciflua L. Virginia creeper Parthenocissus quinquefolia [L.] Planch. Walter viburnum Viburnus obovatum Walt. water oak Q. nigra L. wax myrtle Myrica cerifera L. Hydric Hammock yellow jessamine Gelsemium rankinii Small Temperate hardwood forests are not fire dependent and fires usually originate from surrounding ecosystems and rarely cover large areas (FNAI 1990). Fire frequencies are more frequent for xeric hardwood forests (> 30 years) and less frequent for hydric hardwood forests (very rarely) (FNAI 1990). Depending on the composition and structure of fuels, fires in temperate hardwood forests have the potential to become intense.

PAGE 27

14 Thesis Overview Chapter 2 explores the differences in structural flammability of understory species native to pine flatwood and hardwood hammock ecosystems. Plants were separated into live and dead foliage, accumulated litter on and under the plant, and small (<0.6 cm diameter) and large (>0.6 cm diameter) twigs, branches, and stems. Results were analyzed to determine the ecosystem and species differences in components relative to each other and in the context of potential hazards to WUI structures. Chapter 3 is a further analysis of foliar biomass characteristics. Leaf area, foliar moisture content, foliar volatile solids, and foliar energy content were quantified for each species within each ecosystem type. Results were analyzed, in conjunction with results from Chapter 2, to describe the flammability of understory species within each ecosystem type. Chapter 4 provides an overview of the results from this study. In addition, recommendations for firewise landscaping techniques within each ecosystem are described with recommendations for further research.

PAGE 28

15 Figure 1-4. Pine flatwood ecosystem with routine prescribed fire. Photograph by Larry Korhnak. Figure 1-5. Typical hardwood hammock ecosystem. Photograph by Larry Korhnak.

PAGE 29

CHAPTER 2 PLANT STRUCTURAL FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS Introduction Vegetation and litter provide the primary fuel source for wildfires which threaten WUI homes in the forests of the southern US. Periodic wildfire occurs in most southern forests removing accumulated biomass and litter, releasing nutrients, and facilitating regeneration (Myers and Ewel 1990, Pyne et al. 1996, Walker and Oswald 1999). As WUI areas expand within southern forests, challenges increase between the use of fire for natural resource management, range management, or silvicultural practices (Monroe 2002). In an attempt to reduce the burden of wildfire protection, policies and recommendations include the development of defensible space through firewise landscaping (Monroe 2002). As the plants within home landscapes are potential fuel for wildfires which may threaten human life and property, it has become important to evaluate the hazard of landscape plants. Quantifying the flammability of plant species is a way of evaluating the hazard of different plants which may be incorporated into landscaping. Flammability has been defined as having four components: ignitability, sustainability, combustibility, and consumability (Anderson 1970, Martin et al. 1994). Characteristics shown to influence flammability of plant material include moisture content (Gill et al. 1978), percent cellulose, hemicellulose, and lignin (Philpot 1970, Rundel 1981, Susott 1982a), volatile compounds (Shafizadeh et al. 1977, Susott 1982a, 16

PAGE 30

17 Wang and Huffman 1982, van Wilgen et al. 1990, Owens et al. 1998), silica-free mineral content (Mutch and Philpot 1970), leaf thickness (Montgomery and Cheo 1971), surface area-to-volume ratio (Rundel 1981, Papi and Trabaud 1990), and particle density (Brown 1970, Papi and Trabaud 1990). However, these characteristics have been studied to different extents by different methods, are not equally important to plant flammability, and are not all independent of one another (Shafizadeh et al. 1977, Etlinger 2000, Francis 2000). Etlinger (2000) found that the relative abundance of biomass components significantly affected the flammability of an entire plant, measured as heat release rate (sustainability and combustability), more than all other characteristics measured. The amount of dry mass consumed determined the total heat release while foliar biomass and foliar moisture content determined the peak heat release rate (Etlinger 2000). Etlinger (2000) also concluded that fine fuel biomass (foliage and small stems <0.6 cm) and fine fuel moisture content contribute more to the peak heat release rate of plants than many other characteristics. The height of understory plants and the arrangement of fuel within the plant can contribute to the intensity and height of surface fires (Pyne et al. 1996) and structural survivability in the WUI (Wilson and Ferguson 1986). Sources for plant ignition may also be different based on the arrangement of fuels. Fuel closer to the ground may be more susceptible to ignition from ground and surface fires. Fuel further from the ground may be more susceptible to ignition from firebrands. The density of fuel, or fuel loading, can affect the sustainability of fire within the plant; more dense fuels sustaining a more consistent fire (Rundel 1981). This is true until

PAGE 31

18 the fuel becomes so dense that oxygen becomes limiting (Rundel 1981). For this reason, litter with lower density may be more flammable than litter with higher density. The objectives of this study were to determine the biomass components and biomass arrangement of understory species within pine flatwood and hardwood hammock ecosystems. Based on these results, we attempt to determine the differences in potential structural flammability of understory species between the two ecosystem types. We also attempt to determine the relative hazards of species to WUI homes. Materials and Methods Species for this study were chosen based on their use in landscape plantings or their abundance in the understory of the two ecosystems. Species studied within pine flatwood ecosystems were dwarf huckleberry (Gaylussacia dumosa [Andr.] A. Gray), gallbery (Ilex glabra [L.] A. Gray), rusty lyonia (Lyonia ferruginea [Walt.] Nutt.), and evergreen blueberry (Vaccinium myrsinites Lam.). American beautyberry (Callicarpa americana L.), American holly (Ilex opaca Ait. var opaca), water oak (Quercus nigra L.), and sparkleberry (Vaccinium arboreum Marsh.) were studied within hardwood ecosystems. Wax myrtle (Myrica cerifera L.) and saw palmetto (Serenoa repens [Bartr.] Small) were studied in both ecosystems. For each plant collected, litter, height, and biomass measurements were made to determine the structural flammability of species within each ecosystem type. Litter underneath the plant was collected to determine litter depth and density. Next, plant height, height to lowest branch, and width measurements were taken to determine overall size of the plant. Plants were then cut at the base and separated into biomass components to characterize the potential fuel for each species.

PAGE 32

19 Selection of Sites and Plants Five sites of each ecosystem type (pine flatwood and hardwood hammock) were located throughout North Central Florida (Appendix B). Criteria for site selection were presence of desired understory species in the desired ecosystem type. Fire had been absent in all sites for at least five years. Study sites included property owned by the USDA Forest Service (Osceola National Forest), Florida Division of Forestry (Jennings State Forest, Twin Rivers State Forest, Withlacoochee State Forest, and Welaka State Forest), Suwannee River Water Management District (Little River Springs and Steinhatchee), and the University of Florida (Austin Cary Memorial Forest). At each site, the vegetation was characterized by randomly selecting and measuring four circular tree plots (400 m2) and eight circular shrub plots (12.56 m2). Within the tree plots, diameter at breast height (dbh) and height to lowest branch was measured using a hypsometer (Haglf, Vertex III) was measured for tree species (>3 m in height (Foote et al. 1991) and >6.4 cm dbh). Species of midstory trees (>3 m in height but <6.4 cm dbh) were recorded. Canopy closure was measured from the center of each tree plot by averaging four readings from a concave spherical densiometer (Model-C, Forestry Supply, Inc.). Within the understory plots, the total number of stems by species was recorded and the height in 0.2 m increments from 0.4 m to 3 m for each individual. At each study site, three plants of each species were randomly selected and sampled between May and July 2002 (Table 2-1). Three transects were initiated at least 5 m away from any road or trail, and one plant of each species was randomly selected. There was a minimum of 5 m between flagged individuals of the same species to avoid microclimate influence. Plants were considered appropriate for this study between 1 to 3 meters in height, an interval identified by Foote et al. (1991). However, Gaylussacia dumosa and

PAGE 33

20 Vaccinium myrsinites were accepted into the random sample if over the height of 0.6 m and 0.4 m, respectively. These two species are common within pine flatwoods but do not typically grow very tall. A total of fifteen plants (three individuals at five sites) were sampled of Gaylussacia dumosa, Ilex glabra, Lyonia ferruginea, Vaccinium myrsinites, Callicarpa americana, Ilex opaca, Quercus nigra, and Vaccinium arboreum. A total of thirty plants (three individuals at five sites in both ecosystems) were sampled of Myrica cerifera and Serenoa repens. In total, 180 plants were harvested from the sites (Eq. 2-1). Biomass was measured on all individuals studied. Additional biomass data were taken on one randomly selected individual plant at each site; 60 plants were harvested at 1-m height intervals. 6 species 3 plants 5 sites 2 ecosystem types = 180 plants (2-1) Table 2-1. Collection dates for study sites. Ecosystem Site Plant collection dates, 2002 Flatwood Austin Cary M. F. May 23, 24 Osceola N. F. July 1, 2, 3 Jennings S. F. July 19, 22 Welaka S. F. June 11, 14, 19, 20 Withlacoochee S. F. June 3, 4 Hardwood Twin Rivers S. F. May 27, 28, 29 Osceola N. F. June 25, 26, 27 Jennings S. F. July 25, 29 Steinhatchee July 15, 16, 18 Little River Springs June 6, 7 Litter Measurements To determine the potential effects of species on the litter layer beneath understory plants, litter was measured with a quadrat. A 25 cm by 25 cm (internal dimensions) square quadrat made of PVC pipe and PVC joints was placed against the south side of the stem for consistancy. Within the quadrat, three readings of litter depth were made.

PAGE 34

21 Then the litter was removed and placed into a paper bag. Any living plants were removed from the litter sample before being weighed on an electronic balance (Ohaus, Scout II) with a maximum of 600 g and accuracy to 0.1 g. Litter samples were dried at 70oC for 72 h and weighed. Litter density is given in gcm3 (Eq. 2-2). Litter density (gcm3)= litter biomass / (average litter depth 25 cm 25 cm) (2-2) Height Measurements Before harvesting, total height and height to lowest branch were measured for each plant. The plant was not disturbed nor physically extended to take these measurements. The lowest branch measurement was made from the bottom of the litter layer to the point of the lowest vegetation on the branch, whether it was at the stem junction or at the terminal end of a branch. If multiple stems from the same individual emerged from beneath the litter layer, then the height to lowest branch was recorded as zero. Two measurements of crown width were taken at the widest point in perpendicular directions. Fuel density, or fuel loading (Rundel 1981), was calculated by dividing the total dry biomass by the plant volume (Eq. 2-3). The plant was then harvested at the soil line for biomass measurements. Fuel loading (mg/cm3)= total biomass/(height width 1 width 2) (2-3) Biomass Measurements Each plant was separated into components: live foliage, dead foliage, litter accumulated on plant (referred to as debris), small stems (<0.6 cm diameter), and coarse fuel (> or = 0.6 cm diameter) for biomass analyses. Dead foliage and debris were added to the fine fuel categorization of live foliage and small stems (<0.6 cm diameter) used by Etlinger (2000). If the amount of dead foliage was 0.1 g, it was included in the measurement of debris. In the field, fresh weights were recorded with the same

PAGE 35

22 electronic scale used for litter measurements. For foliar biomass samples, a subsample was removed for leaf area and volatile solid analyses (Chapter 3) and the sample was reweighed. One plant of each species at each site was sampled at 1-m intervals measured from the bottom of the litter layer. The plant was sampled by cutting off the highest interval first. The separation of height interval was only used in the biomass measurements as samples were pooled from all intervals for foliar analyses (Chapter 3). Foliar, small stem, and debris samples were dried at 70oC for 72 h and large stems dried until their weight was stable for at least 48 h. Live and dead petioles (course fuel) of Serenoa repens were weighed after drying for 144 h. The time of drying was determined based on results from a preliminary study. Samples were then weighed with the same electronic balance used in the field. The dry weight of the paper bag, depending on the size, was subtracted from the measured dry weight. The average dry weight of each bag size was determined by drying ten randomly selected bags of each bag size for 72 h at 70oC. Total dry foliar biomass was calculated based on the moisture content of the sub-sample dried in the oven. In this way, the sub-sample dry weight was multiplied by the ratio of the fresh weight of the whole sample divided by the fresh weight of the subsample. Moisture Content Because of the methodology used to determine the biomass components of each plant, moisture content of each component could also be calculated. Moisture content was calculated based on dry weight (Eq. 2-4). Moisture content (%) = [(fresh weight dry weight) / dry weight] 100 (2-4)

PAGE 36

23 Statistical Analyses Site characterization data, in stems per hectare, tree basal area, height to lowest branch, and canopy closure, were analyzed for overall ecosystem and site differences. The general linear model (glm) procedure in Statistical Analysis Software (SAS) was used. All pairwise comparisons of fixed means were performed using Tukeys test, rejecting the null hypothesis that there was no difference between treatments when p<0.05. Sites were considered fixed and were nested within ecosystem (Eq. 2-5). yijl = + i + j(i) + eijl (2-5) = true mean i = effect of level i of A (ecosystem) (df=1) j(i) = effect of level j of B (site) nested within level i of A (df=8) random eijl = experimental error [df=70 (understory) and df=30 (midstory and overstory)] Species data were analyzed for ecosystem, species, and site effects using the general linear model (glm) procedure in Statistical Analysis Software (SAS) (Eq. 2-6). Species and site effects were nested within ecosystem type. Site effects were considered random. yijkl = + i + Bj(i) + k(i) + Bjk(i) + eijkl (2-6) = true mean i = effect of level i of A (ecosystem) (df=1) random Bj(i) = effect of level j of B (site) nested within level i of A (df=8) k(i) = effect of level k of C (species) nested within level j of B within level i of A (df=10) random Bjk(i) = effect of level j of B (site) by the effect of level k of C (species) nested within level i of A (df=40) random eijkl = experimental error (df=120) When the site species interaction (Bjk(i) ) was not significant (p>0.1), then it was dropped from the model. When the site species interaction (Bjk(i) ) was significant (p<0.1), species were run individually to determine the random effect of site on

PAGE 37

24 individual species using the model, y= Bj(i). The forementioned model was used to analyze data for all species studied, treating Myrica cerifera in flatwood ecosystems as a different species than Myrica cerifera in hardwood ecosystems, and the same for Serenoa repens. A different statistical analysis was used for 1-m height interval data (Eq. 2-7). yikl = + i + k(i) + eikl (2-7) = true mean i = effect of level i of A (ecosystem) (df=1) k(i) = effect of level k of C (species) nested within level i of A (df=10) random eikl = experimental error (df=48) To determine the direct ecosystem effect on individual species, data from Myrica cerifera and Serenoa repens were analyzed separately (Eq. 2-8). yijkl = + i + j + ()ij + Bk(i) + lk(ij) + eijkl (2-8) = true mean i = effect of level i of A (ecosystem) (df=1) k = effect of level k of C (species) (df=1) ()ik = effect of level i of A (ecosystem) by level k of C (species) (df=1) random Bj(i) = effect of level j of B (site) nested within level i of A (ecosystem) (df=8) random lj(ik) = effect of level j of B (site) nested within A (ecosystem) by C (species) interaction (df=8) random eijkl = experimental error (df=39) When interaction variables were not significant (p<0.1), they were dropped from the model. All other tests were performed at =0.05. All pairwise comparisons of fixed means were performed using Tukeys test, rejecting the null hypothesis that there was no difference between treatments when p<0.05.

PAGE 38

25 Results Site Characterization The absolute and relative densities of understory species (Appendix C-1) were used to calculate the percentage of the understory that was characterized by this study (Table 2-2). Absolute and relative densities were also calculated for midstory species at each site (Appendix C-2). Absolute density, relative density, relative dominance, relative frequency, and importance values were calculated for overstory species at each site (Appendix D-1). Collectively, the flatwood sites contained higher understory density, lower midstory density, and lower overstory density than hardwood sites (Table 2-3). Further analyses of the overstory reveal that the flatwood sites contained less basal area per hectare than hardwood sites. The trees in hardwood sites had lower height to lowest branch and higher percentage of canopy closure than trees in flatwood sites (Table 2-4). Table 2-2. Percent of total understory stems characterized by the species studied at each site. Ecosystem Site % of total stems Flatwood Austin Cary M. F. 78.8 Osceola N. F. 97.2 Jennings S. F. 85.2 Welaka S. F. 84.7 Withlacoochee S. F. 77.4 Hardwood Twin Rivers S. F. 63.6 Osceola N. F. 62.1 Jennings S. F. 59.0 Steinhatchee 25.6 Little River Springs 59.0 Although there were collective differences between ecosystems, there were also similarities between some flatwood and hardwood sites. The understory density in the flatwood sites in Jennings State Forest and Withlacoochee State Forest were not different

PAGE 39

26 from the understory density in any of the hardwood sites (Table 2-3). There were no statistical differences between sites in midstory density (p=0.1140). A small difference existed in the overstory density between sites. The flatwood sites at Austin Cary Memorial Forest and Osceola National Forest contained fewer trees per hectare than the flatwood site at Jennings State Forest and the hardwood site at Little River Springs (Table 2-3). There was no difference in basal area between sites (p<0.0518). The flatwood site at Jennings State Forest contained trees with a lower height to lowest branch than the flatwood sites at Austin Cary Memorial Forest and Osceola National Forest (Table 2-4). The canopy closure at the flatwood site in Withlacoochee State Forest was not statistically different from all the hardwood sites except Twin Rivers State Forest (Table 2-4). Ecosystem Differences Ecosystems were significantly different in litter depth, but not litter density. Ecosystems were also significantly different for height and biomass measurements when all species were analyzed together. However, height and biomass measurements were not different for Myrica cerifera and Serenoa repens between ecosystems. Litter measurements Litter depth in the flatwood ecosystems was almost twice that in hardwood ecosystems (p=0.0025) (Table 2-5). This was also true for Myrica cerifera and Serenoa repens within flatwood and hardwood ecosystems (p=0.0030). However, litter density was the same in flatwoods and hardwoods (p=0.8541) and for Myrica cerifera and Serenoa repens within flatwood and hardwood ecosystems (p=0.2700). Site (nested within ecosystem) was significant for all litter measurements for all species studied.

PAGE 40

Table 2-3. Understory, midstory, and overstory stems per hectare for the study sites. Standard error is given in parentheses (n=8 for understory and n=4 for midstory and overstory site means; n=40 for understory and n=20 for midstory and overstory ecosystem means). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison between sites. 27 Understory Midstory Overstory Ecosystem Site Stems/ha Stems/ha Stems/ha Flatwood Austin Cary M. F. 136,000 (32,000) ab 0 (0) 150.0 (38.2) c Osceola N. F. 135,000(13,400) ab 0 (0) 206.3 (29.5) bc Jennings S. F. 79,000 (37,600) bcd 118 (64) 650.0 (110.4) a Welaka S. F. 177,000 (15,700) a 68 (32) 275.0 (25.0) abc Withlacoochee S. F. 85,400 (16,200) bc 175 (69) 562.5 (156.3) abc Flatwood Mean 123,000 (12,100) 72 (23) 380.3 (59.6) Hardwood Twin Rivers S. F. 2,190 (390) d 881 (160) 618.8 (64.9) ab Osceola N. F. 5,770 (792) cd 500 (245) 365.3 (90.9) abc Jennings S. F. 17,100 (1,750) cd 1137 (241) 606.3 (104.3) abc Steinhatchee 13,900 (2,300) cd 1000 (193) 600.0 (77.7) abc Little River Springs 7,960 (1,840) cd 1381 (344) 681.3 (74.6) a Hardwood Mean 9,390 (1,100) 980 (118) 572.5 (42.0) * indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems

PAGE 41

Table 2-4. Basal area, height to lowest branch, and canopy closure at each site. Standard error is given in parentheses (n=4 for site means and n=20 for ecosystem means). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison between sites. Ecosystem Site Basal area (cm2 per ha) Height to lowest branch (m) Canopy closure (%) ** Flatwood Austin Cary M. F. 14,200 (3400) 17.2 (1.4) a 52 (5) c Osceola N. F. 18,000 (2880) 16.4 (1.0) a 60 (1) c Jennings S. F. 29,600 (4190) 7.2 (0.0) bc 62 (6) c Welaka S. F. 20,000 (2380) 12.1 (2.0) ab 59 (6) c Withlacoochee S. F. 30,600 (4070) 12.5 (1.9) ab 84 (4) b Flatwood Mean 22,900 (2040) 12.8 (1.0) 63 (3) Hardwood Twin Rivers S. F. 32,900 (4660) 4.9 (0.4) c 98 (1) a Osceola N. F. 26,800 (5820) 10.3 (1.6) bc 85 (5) b Jennings S. F. 30.000 (5070) 5.9 (0.7) c 87 (2) b Steinhatchee 42,800 (5380) 7.1 (0.2) bc 88 (3) b Little River Springs 37,000 (4780) 6.0 (0.3) c 94 (1) ab Hardwood Mean 34,000 (2420) 6.8 (0.5) 90 (2) 28 indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems ** although percentage means are presented in the table statistical analysis on canopy closure was performed using an arcsin transformation.

PAGE 42

29 Site (nested within ecosystem) was also significant (p<0.0001) for litter moisture content, although there was no difference between ecosystems (Appendix E). Height measurements Based on all understory species studied, there was no difference between the average height to lowest branch between flatwood and hardwood sites (p=0.3968). In addition, ecosystem type did not affect height to lowest branch of Myrica cerifera and Serenoa repens (p=0.5293). On average, understory species were taller in hardwood ecosystems (147.8 cm) than flatwood ecosystems (108.5 cm), p<0.0001. However, ecosystem type did not affect the total height of Myrica cerifera and Serenoa repens (p=0.9495). Table 2-5. Litter depth and density for each species and ecosystem type. Standard error is given in parentheses (n=15 for species and n=90 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison. Ecosystem Species litter depth (cm) litter density (mg/cm3) Flatwood Gaylussacia dumosa 5.39 (0.59) bc 17.61 (2.44) ab Ilex glabra 4.87 (0.39) bc 20.54 (1.41) a Lyonia ferruginea 6.28 (0.46) b 13.42 (1.23) bc Vaccinium myrsinites 4.53 (0.47) bc 18.97 (3.46) ab Myrica cerifera 6.17 (0.99) b 16.95 (1.61) ab Serenoa repens 10.12 (0.92) a 9.49 (1.62) c Flatwood mean 6.23 (0.33) 16.16 (0.92) Hardwood Callicarpa americana 3.63 (0.36) c 16.02 (1.45) abc Ilex opaca 3.71 (0.44) c 16.80 (2.07) ab Quercus nigra 3.22 (0.14) c 19.47 (2.93) ab Vaccinium arboreum 3.77 (0.27) c 14.54 (1.73) abc Myrica cerifera 3.11 (0.23) c 19.39 (2.47) ab Serenoa repens 4.61 (0.50) bc 15.04 (1.49) abc Hardwood mean 3.67 (0.15) 16.88 (0.86) indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems

PAGE 43

30 Biomass measurements Total (per plant) fine fuel biomass and the fine fuel components live foliage, dead foliage, and debris were not different between pine flatwood and hardwood hammock ecosystems. However, small stem biomass was greater in hardwood ecosystems than flatwood ecosystems (p=0.0027) (Table 2-6). Total fine fuel and all individual fine fuel components of Myrica cerifera and Serenoa repens were not different between ecosystems. There was more coarse fuel in hardwood ecosystems than in flatwood ecosystems (p=0.0035) (Figure 2-1). However, coarse fuel biomass of Myrica cerifera and Serenoa repens were the same between ecosystems (p=0.2802). The results of total biomass analyses were similar to the results from the coarse fuel analyses. Total biomass per individual plant was greater in hardwood ecosystems than in flatwood ecosystems (p=0.0214) but total biomass per plant was the same between ecosystems for Myrica cerifera and Serenoa repens (p=0.4826). Fuel loading was higher for understory species in flatwood ecosystems than in hardwood ecosystems (p=0.0376) (Figure 2-2) although there was no difference in fuel loading of Myrica cerifera and Serenoa repens between ecosystems (p=0.9561). Live foliage collected in hardwood hammocks had greater foliar moisture content than pine flatwoods (Chapter 3) although there was no significant difference between moisture content for any other biomass component (Appendix E). There was significantly more live foliage and small stems per individual in the lowest 1-m interval in hardwood ecosystems than flatwood ecosystems (p=0.0001 and p=0.0016, respectively) (Figure 2-3). There was no significant difference between the dead foliage, debris, or large stem components of the lowest 1-meter interval between ecosystems (Appendix F). There was also no statistical difference between the total

PAGE 44

Table 2-6. Dry weight of fine fuel biomass components for each species and ecosystem types. Standard error is given in parentheses (n=15 for species and n=90 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison. Fine Fuels (total dry g) Ecosystem Species live foliage dead foliage debris small stems Flatwood Gaylussacia dumosa 3.9 (0.5) d 0.0 (0.0) b 1.0 (0.2) b 6.5 (1.0) e Ilex glabra 10.1 (2.2) cd 0.0 (0.0) b 1.8 (0.5) b 20.0 (3.4) cde Lyonia ferruginea 15.5 (2.5) cd 0.0 (0.0) b 3.0 (0.6) b 22.7 (3.0) cde Vaccinium myrsinites 3.6 (0.7) d 0.0 (0.0) b 2.7 (0.9) b 9.9 (2.4) e Myrica cerifera 40.3 (6.8) cd 0.0 (0.0) b 4.8 (1.0) b 45.4 (7.9) abc Serenoa repens 242.2 (45.3) b 192.1 (40.5) a 38.4 (8.9) a 0.0 (0.0) e Flatwood mean 52.6 (11.7) 32.0 (10.0) 8.7 (2.1) 17.4 (2.2) Hardwood Callicarpa americana 4.9 (0.9) cd 0.0 (0.0) b 0.3 (0.2) b 15.5 (2.6) de Ilex opaca 84.6 (15.9) c 0.1 (0.0) b 8.1 (4.2) b 69.0 (12.3) a Quercus nigra 14.8 (4.9) cd 0.1 (0.1) b 0.8 (0.4) b 20.1 (3.6) cde Vaccinium arboreum 34.0 (8.4) cd 0.0 (0.0) b 6.3 (1.9) b 60.6 (11.6) ab Myrica cerifera 28.3 (9.4) cd 0.0 (0.0) b 1.8 (0.6) b 38.1 (10.9) bcd Serenoa repens 324.5 (48.1) a 192.4 (44.0) a 40.4 (9.6) a 0.0 (0.0) e Hardwood mean 81.9 (14.5) 32.1 (10.4) 9.6 (2.3) 33.9 (4.2) 31 indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems

PAGE 45

32 Flatwood species Flatwood meanSerenoa repensMyrica ceriferaVaccinium myrsinitesLyonia ferrugineaIlex glabraGaylussacia dumosa coarse fuel fine fuel biomass (dry g) 020040060080010001200 Hardwood species Hardwood meanSerenoa repensMyrica ceriferaVaccinium arboreumQuercus nigraIlex opacaCallicarpa americana aabbbbbbbbbb*** Figure 2-1. Total biomass by coarse fuel and fine fuel components for species within pine flatwoods and hardwood hammocks (n=15 for species, n=90 for ecosystem). = significant difference in coarse fuel between ecosystems, ** = significant difference between total fuel between ecosystem type. Letters indicate significant (p<0.05) difference in Tukeys pairwise comparison for total biomass between species. In addition the same letters describe the differences between species for fine fuel and coarse fuel analyzed separately.

PAGE 46

33 Flatwood species Serenoa repensMyrica ceriferaVaccinium myrsinitesLyonia ferrugineaIlex glabraGaylussacia dumosa fuel loading Fuel loading (mg/cm3) 0.00.10.20.30.40.50.6 Hardwood species Serenoa repensMyrica ceriferaVaccinium arboreumQuercus nigraIlex opacaCallicarpa americana Figure 2-2. Fuel loading for each species and ecosystem type. Standard error is represented in error bars (n=15 for species, n=90 for ecosystem means). Star represents significant (p<0.05) difference between ecosystem type. There was no significant difference for fuel loading between species (p>0.05).

PAGE 47

34 biomass in the 1st (p=0.0717) or 2nd (p=0.1672) meter intervals between ecosystems. However, individual understory plants in hardwood hammocks had more total biomass in the 3rd meter interval (p<0.0001) than pine flatwood understory plants, although the actual amount of biomass in the 3rd meter interval was a small proportion of the total biomass (Figure 2-3). Analyses of biomass components within the 2nd and 3rd 1-m height intervals were not possible due to the low number of samples within these two intervals. dry weight (g) 050100150200250300 Hardwoods M1Hardwoods M2Hardwoods M3Flatwoods M1Flatwoods M2Flatwoods M3 large stems small stems live foliage dead foliage debris **** Figure 2-3. Biomass per 1-meter height intervals by ecosystem type (n=30). An * indicates significant (p<0.05) difference in meter 1 components between ecosystem types. An ** indicates significant (p<0.05) difference in total biomass per 1-meter height interval between ecosystem types. Species Differences Species effects on litter measurements were not pronounced. Overall height was different among species and the height to lowest branch was minimally different between

PAGE 48

35 species studied. Species differences in biomass were largely due to the greater overall biomass of Serenoa repens compared to all other species studied. Litter measurements Litter sampled beneath plants was deepest for Serenoa repens in flatwood ecosystems followed by Myrica cerifera and Lyonia ferruginea in flatwood ecosystems (Table 2-5). Serenoa repens, Myrica cerifera, and Lyonia ferruginea (all in flatwood sites) had greater litter depth beneath them than Callicarpa americana, Ilex opaca, Quercus nigra, Vaccinium myrsinites, and Myrica cerifera in hardwood ecosystems. Moisture content of litter was not different between species (Appendix E). Litter density sampled beneath Ilex glabra was greater than litter density sampled under Lyonia ferruginea and Serenoa repens in flatwood sites (Table 2-5). Serenoa repens in flatwood sites had lower litter density beneath it than Gaylucassia dumosa, Vaccinium myrsinites (both in flatwood ecosystems), and Ilex opaca, Quercus nigra, and Myrica cerifera (in hardwood ecosystems). There was a significant random site by species (nested within ecosystems) interaction for litter density (p=0.0091). Upon further analyses, litter density under Lyonia ferruginea, Callicarpa americana, Ilex opaca, Quercus nigra, and Vaccinium arboreum was affected by the random effect of site nested within ecosystem (p<0.05). Height measurements The height to lowest branch was significantly shorter for Vaccinium myrsinites and Serenoa repens (in either ecosystem) than Ilex glabra and Myrica cerifera (in flatwoods) and Ilex opaca, Quercus nigra, Vaccinium arboreum, and Myrica cerifera (in hardwoods) (Figure 2-4).

PAGE 49

36 Vaccinium arboreum, at 182.9 cm mean height, was taller than all flatwood species and all hardwood species except Ilex opaca and Quercus nigra. Vaccinium myrsinites and Gaylussacia dumosa were significantly shorter than all species studied averaging 51.9 cm and 71.0 cm in height, respectively. Myrica cerifera had the greatest variation in height in both flatwood and hardwood ecosystems. The total height of Lyonia ferruginea, Vaccinium myrsinites, and Myrica cerifera was significantly different between sites nested within ecosystems (p<0.05). This resulted in a significant site by species (nested within ecosystem) interaction (p=0.0415). Biomass measurements Total fine fuel biomass per individual was greatest for Serenoa repens in hardwood and flatwood ecosystems (Figure 2-1). Live foliage biomass was greatest for Serenoa repens in hardwood ecosystems, followed by Serenoa repens in flatwood ecosystems (Table 2-6). Dead foliage biomass was greatest for Serenoa repens in both flatwood and hardwood ecosystems. All other species had negligible dead foliage biomass. In addition, Serenoa repens in either ecosystem had more accumulated debris than all other species. Ilex opaca had greater small stem biomass than all other species except Vaccinium arboreum and Myrica cerifera (in flatwood ecosystems). Serenoa repens, in either ecosystem type, had higher fine fuel biomass and coarse fuel biomass per individual than all other species studied (Figure 2-1). There was no significant difference in fuel loading between species (p=0.5024) (Figure 2-2). Species by site interaction (nested within ecosystem), a random effect, was significant (p<0.1) for live foliage biomass, small stem biomass, large stem biomass, total biomass, and fuel loading. Further analyses showed that site (nested within ecosystem) was significant (p<0.05) for Ilex glabra (fuel loading), Vaccinium myrsinites (small stem

PAGE 50

Flatwood Species Gaylussacia dumosaIlex glabraLyonia ferrugineaVaccinium myrsinitesMyrica ceriferaSerenoa repens Height and height to lowest branch (cm) 050100150200250 total height (cm) height to lowest branch (cm) Hardwood Species Callicarpa americanaIlex opacaQuercus nigraVaccinium arboreumMyrica ceriferaSerenoa repens BC D BCD BCabaab b a b b BCCABABCA BC BCabaaaa b 37 Figure 2-4. Height and height to lowest branch of species within pine flatwoods and hardwood hammock ecosystems. Standard error is shown in error bars (n=15 for species and n=90 for ecosystems). Upper-case and lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison.

PAGE 51

38 biomass and fuel loading), Myrica cerifera in flatwoods (fuel loading), Serenoa repens in flatwoods (live foliage biomass and total biomass), and Myrica cerifera in hardwoods (large stem biomass and total biomass). Although there were differences between the foliar moisture content of species (Chapter 3), there was no difference between the moisture content of small stems between ecosystems (Appendix E). Species were significantly different in all components within the 1st 1-meter interval (p<0.05) (Appendix F, Figure 2-5). Serenoa repens contained more live foliage, dead foliage, and debris than most species studied. Serenoa repens, Ilex opaca, and Vaccinium arboreum contained more large stems than most species studied. Although the total biomass in the 1st 1-meter interval was different between species (p<0.0001) (Appendix F), there was no difference between species in the total biomass in the 2nd or 3rd 1-meter intervals. Discussion More variation in ecosystem structure existed among flatwood sites than hardwood sites. This is likely caused by the differences in stand age and that some flatwood sites were a combination of naturally regenerated and planted pines. The flatwood sites at Austin Cary Memorial Forest and Osceola National Forest had fewer, but larger trees per hectare. In addition, the trees at these two sites had greater height to lowest branch than other sites. The three flatwood sites with the fewest trees per hectare (Austin Cary Memorial Forest, Osceola National Forest, and Welaka State Forest) also had the highest understory density. It is likely that variation between the flatwood sites accounts for the variability between sites for litter and biomass measurements. Even with differences between sample sites, some general differences in plant structural components can be drawn from the ecosystem type as a whole.

PAGE 52

39 Hardwood species 0200400600800 Serenoa repensMyrica ceriferaVaccinium arboreumQuercus nigraIlex opacaCallicarpa americana large stems small stems live foliage dead foliage debris Flatwood species 0200400600800Meter 3 Serenoa repensMyrica ceriferaVaccinium myrsinitesLyonia ferrugineaIlex glabraGaylussacia dumosa Meter 2 Serenoa repensMyrica ceriferaVaccinium myrsinitesLyonia ferrugineaIlex glabraGaylussacia dumosa Serenoa repensMyrica ceriferaVaccinium arboreumQuercus nigraIlex opacaCallicarpa americana Biomass (dry g) 0200400600800 Meter 1 Serenoa repensMyrica ceriferaXia ferrugineaIlex glabrainium myrsinit Biomass (dry g) 0200400600800 Serenoa repensMyrica ceriferaVaccinium arboreumQuercus nigraIlex opacaCallicarpa americana Figure 2-5. Biomass per 1-meter height intervals for species within pine flatwoods and hardwood hammocks (n=5). Ecosystem Differences The amount of total fuel available for wildfire within the litter layer is greater in flatwood ecosystems and in addition, a surface fire would have flames that reached higher into the understory vegetation in the flatwoods than hardwoods. Acacia woodlands, a fire-prone ecosystem in southern Ethiopia, had twice the litter biomass

PAGE 53

40 compared to Afromontane forests, a less fire-prone ecosystem in the same region (Eriksson et al. 2003). The chemical and structural characteristics of the litter components will greatly influence the flammability of the litter layer, and the flammability of surrounding plants. For example, pine needles are known to be highly flammable because of high surface area to volume ratio and presence of volatile compounds (Fonda 2001, Schroeder et al. 2001). Although there was no difference between ecosystems in the height to lowest branch, the increased litter depth makes the lowest branch in flatwood ecosystems more susceptible to ignition from a surface fire. Once ignited, however, an individual understory plant common to hardwood ecosystems has the potential to carry fuel to a greater height than an individual understory plant common to flatwoods ecosystems. In addition, because of the increased total height and size, individual understory plants in hardwood ecosystems may be more susceptible to ignition from firebrands. The difference in height was largely based on the different species sampled within each ecosystem as there was no direct effect of ecosystem on the total height of Serenoa repens and Myrica cerifera. It also must be noted that the average height difference between pine flatwood and hardwood hammock may be biased due to the selection of specimens based on species-specific minimum and maximum height requirements. The ecosystem effect on the biomass components largely reflects the biomass difference between the species studied specific to eachecosystemas there were no direct ecosystem effect on the biomass components of Serenoa repens and Myrica cerifera. This is likely because these two species are adapted to a wide range of growing

PAGE 54

41 conditions and their requirements for growth are met within the microclimates in each ecosystem type. There was no difference between mass of debris between ecosystems. This was an unexpected result as pine needles appeared to accumulate more readily on understory plants than hardwood leaves given the grouped nature of pine needles (Figure 2-6). It is possible that the presence of even a few pine trees, as existed in most hardwood study sites (Appendix D), generate enough needles and debris to minimize the difference between ecosystem types. In addition, understory plants in hardwood hammock ecosystems may have intercepted more debris because of their greater overall size. Figure 2-6. Pine needles accumulated on a landscape plant. Even though there is more total biomass for individuals in hardwood ecosystems, the amount of fine fuels was not different between ecosystems. In a comparison of fire-prone fynbos ecosystems to fire-resistant forest patches in South Africa, understory species were found to contain more total biomass in fire-resistant forest patches (van Wilgen et al. 1990). Understory species in less fire-prone Afromontane forest also contained more total biomass than more fire-prone Acacia woodlands in southern

PAGE 55

42 Ethiopia (Eriksson et al. 2003). However, the forest patches and Afromontane forest also contained more fine fuel biomass (van Wilgen et al. 1990, Eriksson et al. 2003). In this study, the greater biomass in hardwood understory plants is due to increased biomass of large stems. Although there is more potential fuel for a fire per understory plant in hardwood ecosystems, the amount of that fuel that is likely to be consumed and contribute to the overall flammability of the plant is the same between ecosystems (Etlinger 2000). In addition, the fuel within individual understory plants in the flatwood ecosystems is more densely packed. Species within the flatwoods ecosystem may sustain a fire for a greater amount of time, therefore increasing the amount of fuel consumed and overall heat released per plant (Rundel 1981). Fire-prone fynbos ecosystems also contained a more dense fuel bed than forest patches (van Wilgen et al. 1990). We conclude that the potential structural flammability is higher for understory species within pine flatwood ecosystems than hardwood hammock ecosystems. Individual understory plants in pine flatwoods contain the same amount of fine fuel as understory plants in hardwood hammocks. However, the fuel is more densely packed within each understory plant and the litter depth underneath each plant is greater. Species Differences Serenoa repens, in either ecosystem, is highly structurally flammable with high ignitability (litter depth, litter density, height to lowest branch, dead foliage and debris) and sustainability (fine fuel biomass). Specific discussions are made of each species studied relative to each other. It must be noted that the lack of clear differences between species other than Serenoa repens may be due to the constraints of the statistical analysis

PAGE 56

43 in determining differences in species as biomass data for Serenoa repens were so much higher than all other species studied. The difference between the flammability of species within the same genus is an important consideration. Rankings of plant flammability typically are built from existing lists originating in different biographical regions. Many times, species in the same genera are considered to be of similar flammability. The results of biomass comparisons clearly show that species within the same genera (Ilex glabra versus Ilex opaca and Vaccinium myrsinites versus Vaccinium arboreum) do not always have similar flammability characteristics and represent different threats to WUI structures even within the same biographical region. Gaylussacia dumosa Gaylussacia dumosa along with Vaccinium myrsinites had the lowest average height than other species studied. Even though the average height was low compared to other species, the height to lowest stem was not significantly shorter than any other species. This means that most of the fine fuel components, including the foliage and debris, are not any more vulnerable to ignition from a surface fire than any other species studied. However, the consumability of Gaylussacia dumosa plants is high because of the high proportion of fine fuels in the total biomass. Therefore, once ignited, it is likely that the whole plant will be consumed, contributing to the spread of wildfire. Gaylussacia dumosa plants had relatively low debris biomass which may decrease the ignitability in wildfires. Based on the results of litter, height, and biomass measurements, an individual Gaylussacia dumosa plant does not cause a serious threat to WUI structures as it contains relatively little overall biomass, which would likely be entirely consumed if ignited.

PAGE 57

44 Ilex glabra There is no litter or biomass characteristic that would indicate that an individual Ilex glabra plant is highly flammable. Specifically, an individual plant may be slightly less flammable because the average litter density was slightly greater under Ilex glabra plants than other species studied. Lyonia ferruginea There is no litter or biomass characteristic that would indicate that an individual Lyonia ferruginea plant is highly flammable. However, the litter under Lyonia ferruginea (deeper depth and lower density) could increase the ignitability of Lyonia ferruginea plants compared to many of the other species studied. Vaccinium myrsinites Vaccinium myrsinites plants are highly flammable because of their low height and fine fuel composition making their ignitability and consumability quite high. However, the total potential fuel is very low, making the sustainability quite low and likely reducing the combustability of an individual plant. Therefore, individual plants of Vaccinium myrsinites are not a high hazard to WUI structures. Callicarpa americana Based on the results of litter, height, and biomass measurements, an individual Callicarpa americana plant does not cause a serious threat to WUI structures. The consumability is not high due to the low proportion of fine fuels. In addition this species has low total fuel content. Callicarpa americana had a relatively low height to lowest branch measurement compared to other species, which may increase the ignitability of the plant slightly.

PAGE 58

45 Ilex opaca With the exception of Serenoa repens, Ilex opaca plants contained slightly, but not always statistically significant more live foliage biomass, debris, fine fuel biomass, and total fuel biomass compared to most species studied. In addition, the great amount of small stem biomass makes the total fuel for fires very high for this species, giving it a moderately high flammability potential. Quercus nigra Quercus nigra plants retained slightly more debris, possibly increasing the ignitability of this species. However, no other characteristics indicate that individuals of this species in the understory are hazardous to WUI structures. Vaccinium arboreum Vaccinium arboreum plants in the understory are very similar in biomass components compared to Ilex opaca. Without knowing the chemical composition of the biomass, it is difficult to determine the differences in flammability between these two species. Myrica cerifera Myrica cerifera plants are more flammable in flatwood ecosystems than hardwood ecosystems because of deeper litter depth and increased small stem biomass. Generally, Myrica cerifera is comparable to both Ilex opaca and Vaccinium arboreum in biomass components. If ignited, there is a great amount of all fuel componentsfine fuel, coarse fuel, and total fuel. Serenoa repens Serenoa repens in either a pine flatwood or hardwood hammock ecosystem has many structural characteristics making it highly flammable. Serenoa repens along with

PAGE 59

46 Vaccinium myrsinites have low height to lowest branch making Serenoa repens highly vulnerable to ignition from surface fires. The large amount of debris and dead foliage biomass makes Serenoa repens highly flammable in either ecosystem. In addition, Serenoa repens within either ecosystem had the highest amount of total fine fuel, coarse fuel, and total biomass than any other species studied. Ignition is more facilitated within flatwood ecosystems, due to the decreased density and increased depth of litter. There was significantly more live foliage biomass per individual Serenoa repens plant in hardwood ecosystems than flatwoods ecosystem causing the Serenoa repens plant in hardwood ecosystems to be slightly more flammable (based on biomass) than a plant of the same species in flatwood ecosystems. This is especially true given the noted flammability of live Serenoa repens foliage (Hough and Albini 1978). Conclusions One result to consider from this study is that individual plants of the species studied may not pose a great threat to WUI structures, except for Serenoa repens. Based on the arrangement of biomass and great amount of total biomass, even individual plants of Serenoa repens can be a threat to WUI structures. Proximity of one landscape plant to another and the construction materials of the WUI home will also influence what species is or is not appropriate for firewise landscaping. The regular maintenance of plants can lower their flammability and therefore lower the wildfire hazard to WUI structures. Current recommendations for firewise landscaping guide landowners to remove accumulated pine needles within 30 feet (9 meters) surrounding a structure (Florida Firewise Communities 2000, Monroe and Long 2001). In addition, if debris and dead material accumulated on plants are removed, the flammability of those plants will be reduced. As Serenoa repens is notoriously difficult

PAGE 60

47 to remove once established, landowners may be able to greatly reduce the flammability of this species by removing dead fronds and debris, thinning, and pruning off lower fronds. However, great caution must be taken to keep up maintenance of Serenoa repens throughout the year and to make sure that individual plants are separated by adequate distance from each other and WUI structures. A more thorough examination of the arrangement of biomass may be a very important component to flammability of plants in the field. The location of different forms of fuel will influence the ignitability. Fuel close to the ground will be more susceptible to ignition from surface fires and fuel higher in the canopy may be more susceptible to ignition from firebrands. To make clear distinctions between species, a greater sample size and sampling at shorter height intervals may be necessary. In addition, knowing how landscape plants are typically ignited in wildfires, from post-fire studies, would improve future plant flammability studies. Chemical and structural composition of the individual biomass components between ecosystems or species may be different, greatly affecting plant flammability. In Chapter 3, we will examine those differences for foliar biomass.

PAGE 61

CHAPTER 3 FOLIAR FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS Introduction In Chapter 2, the distribution of available fuel into different structural components was presented for Gaylussacia dumosa, Ilex glabra, Lyonia ferruginea, and Vaccinium myrsinites in pine flatwoods; Callicarpa americana, Ilex opaca, Quercus nigra, and Vaccinium arboreum in hardwood hammocks; and Myrica cerifera and Serenoa repens in both ecosystems. However, foliar properties greatly affect the flammability of plants and it has been noted that species common in Southern pine ecosystems, especially Ilex glabra and Serenoa repens, have foliar characteristics which make them extremely flammable (Hough and Albini 1978). In addition, a study by Etlinger (2000) found that foliar biomass and foliar moisture content were the most important variables in determining the peak heat release rate of shrub species. Many acceptable methods to measure plant flammability exist. Measurements of plant tissue flammability have been made by thermal evolution analysis (Shafizadeh et al. 1977), oxygen combustion calorimetry (Dickinson and Kirkpatrick 1985, van Wilgen et al. 1990, Rodrguez-An et al. 1995, Nex-Regueira et al. 2001, Williamson and Agee 2002), thermogravimetric analysis (Mutch and Philpot 1970, Philpot 1970, Shafizadeh et al. 1977, Gill et al. 1978, Rogers et al. 1986, Dimitrakopoulos 2001), thermocouple analysis (Owens et al. 1998), evolved gas analysis (Susott 1982a), and differential scanning calorimetry (Susott 1982b). Muffle furnace tests (Cheo and Montgomery 1970, 48

PAGE 62

49 Montgomery and Cheo 1971), cone calorimetry (White et al. 1996), and the limiting oxygen index method (Mak 1988) have been used to test the flammability of entire leaves, stems, or branches of plants. Etlinger (2000) used an intermediate scale biomass calorimeter with a line burner to measure the flammability of entire plants. It is also possible to measure flammability in the field by timing the combustion of plants after being ignited (Ching and Stewart 1962). Isoperibol oxygen combustion calorimetry (formerly referred to as bomb calorimetry) is a technique which measures the total energy released (caldry g-1 or Jdry g-1) by complete combustion of plant tissue in an O2 gas enriched, sealed vessel. Oxygen combustion calorimetry has been used to determine the amount of energy in biomass fuel for energy production (Rodrguez-An et al. 1995, Nez-Regueira et al. 2001). In addition, oxygen combustion calorimetry has also been used to assess the impact of vegetation on wildfire behavior between different ecosystem types (Dickinson and Kirkpatrick 1985, van Wilgen et al. 1990). In this study, we used isoperibol oxygen combustion calorimetry to determine the combustability and muffle furnace tests to determine the comsumability of foliage between different species within different ecosystem types. The objective of this study was to compare the foliar flammability between species within pine flatwood and hardwood hammock ecosystems. To accomplish this, foliar leaf area, moisture content, volatile solids, and energy content were quantified. Foliar leaf area was measured to characterize the leaves. Foliar energy content measures the cumulative influence of flammability characteristics such as lignin, cellulose, and hemicellulose content, volatile compounds, and mineral content. However, foliar

PAGE 63

50 moisture content and low temperature volatile extractives are not incorporated into a measure of energy content. Therefore, foliar moisture content was measured separately. Based on the results of this study, in conjunction with the results from Chapter 2, differences between ecosystems and species were explored. Materials and Methods Live and dead foliage samples were collected through biomass separation as described in Chapter 2. Collected samples were weighed with an electronic balance with a maximum of 600 g and accuracy to 0.1 g both in the field and after being dried. Foliar samples were dried at 70oC for 72 h. The total of 180 plants used in these analyses. However, only one random individual of each species at each site was analyzed for leaf area (60 plants total). Moisture Content Moisture content of each sample was calculated based on dry weight (Eq. 3-1). Moisture content (%) = [(fresh weight dry weight) / dry weight] 100 (3-1) Leaf Area Leaf samples collected from one individual per species per site were placed into plastic bags and into a cooler in the field to reduce loss of leaf structure from water loss. Leaf samples were placed through a LI-COR 3100 Area Meter on the same day of sampling. Twenty-five (25) leaves from most species were used to make leaf area measurements. Only one leaf (frond without petiole) was measured for Serenoa repens and 5 leaves measured for Ilex opaca, Quercus nigra, and Callicarpa americana because of their large size. Fifty (50) leaves were used for Vaccinium myrsinites because of the small size of the leaves. Three measurements of the same leaves were made and the average was multiplied by two to calculate total leaf surface area for both the abaxial and

PAGE 64

51 adaxial leaf surfaces. Leaves were dried in a drying oven for 72 h at 70oC and then immediately weighed using a balance, accuracy 0.001 g. Leaf area per leaf was calculated by dividing the measured leaf area by the number of leaves used in the measurement. Specific leaf area (leaf area per gram of dry weight) was determined for each species. Also, total leaf area per individual was calculated based on the total dry weight of foliage. Volatile Solids Live foliar samples were collected from each plant to be tested for volatile solids content. In addition, dead foliar samples were collected from Serenoa repens plants for volatile solid analysis. Samples were collected and immediately placed in a cooler to prevent loss due to volatilization or decomposition. Samples were processed within 48 hours by Advanced Environmental Labs located in Tampa, FL by EPA standard 160.4 (Appendix G). Data is reported in mg volatile solids per kg dry weight. Energy Content Dried foliar samples were used to determine the energy content for each individual using standard isoperibol oxygen combustion calorimetry. The isoperibol oxygen combustion calorimeter (Parr Model 1261 Calorimeter, Parr Instrument Company) tests were performed in the Department of Animal Science at the University of Florida. Live foliar samples (also dead foliar samples for Serenoa repens) were tested. After drying, the samples were stored in an air-conditioned laboratory. Dried foliar samples were ground in an electric coffee grinder for sample homogenization in order to minimize sample loss in larger grinding apparatus. Ground samples were then stored in paper bags in the same laboratory until the calorimetry analysis was performed.

PAGE 65

52 Crucibles were placed in a drying oven for 24 h at 70oC. Dried crucibles were placed in a dessicator until cool and then weighed. Samples were processed in a random order. In this way, each sample from each site had equal opportunity to be processed on a given day. Approximately 30 mg of sample was measured into each crucible. Crucibles containing samples were placed back into the drying oven for 24 h at 70oC. Crucibles and samples were removed from the drying oven and placed into a dessicator until cool. The dried crucible plus sample dry weight was then measured and the sample dry weight was calculated. Samples were returned to the dessicator and taken out as needed to be analyzed with the calorimeter. Each ground foliar sample was processed in two runs. The two sample runs were processed on separate days. If the replicate run was greater than 2.5% (Dickinson and Kirkpatrick 1985) different from the first run, the sample was rejected and re-run twice at a later time. The two runs were averaged for statistical analyses. The calorimeter was calibrated using benzoic acid, ten ignitions per vessel. A fixed acid correction of 15 and a fixed acid correction of 10 (25 calories total) were automatically subtracted from the total energy released in combustion. This accounts for energy released from the production of nitric acid from atmospheric N2 gas in the vessel and the combustion of the fuse wire itself. NiChrome fuse wire (alloy of nickel and chromium) was used. The vessel was purged with 30 atm O2 gas and submerged in 2 L of deionized water. The calorimeter was left on throughout the study, but the water heating/cooling system was turned off over night for safety reasons. The energy content was calculated based on sample dry weight and expressed in calg-1.

PAGE 66

53 Foliar energy content per plant was calculated by multiplying the energy content (in kcal) per gram by the total foliar energy content per plant. In addition, foliar energy content per hectare per species was also calculated for each plant harvested (Eq. 3-2). kcalha-1= kcal per plant stems per ha of individual speices at individual site (3-2) Statistical Analyses Data were analyzed for ecosystem, species, and site effects using the general linear model (glm) procedure in Statistical Analysis Software (SAS). Species and site effects were nested within ecosystem type. Site effects were considered random. A different model was used for the leaf area data because only one measurement of each species was taken at each site (Eq. 3-3). yijkl = + i + k(i) + eikl (3-3) = true mean i = effect of level i of A (ecosystem) (df=1) k(i) = effect of level k of C (species) nested within level i of A (df=10) random eikl = experimental error (df=48) Moisture content, volatile solids, and energy content data were analyzed with the model in Equation 3-4. yijkl = + i + Bj(i) + k(i) + Bjk(i) + eijkl (3-4) = true mean i = effect of level i of A (ecosystem) (df=1) random Bj(i) = effect of level j of B (site) nested within level i of A (df=8) k(i) = effect of level k of C (species) nested within level j of B within level i of A (df=10) random Bjk(i) = effect of level j of B (site) by the effect of level k of C (species) nested within level i of A (df=40) random eijkl = experimental error (df=120) When the interaction Bjk(i) was not significant (p>0.1), then it was dropped from the model. When the interaction Bjk(i) was significant (p<0.1), species were run

PAGE 67

54 individually to determine the random effect of site on individual species using the model, y= Bj(i). To determine the ecosystem effect on individual species, data from Serenoa repens and Myrica cerifera were used (Eq. 3-5 and Eq. 3-6). For Other Measurements: yijkl = + i + j + ()ij + Bk(i) + lk(ij) + eijkl (3-5) For Leaf Area (LA): yijkl = + i + j + ()ij + eijkl (3-6) = true mean i = effect of level i of A (ecosystem) (LA df= 1), (df=1) k = effect of level k of C (species) (LA df=1), (df=1) ()ik = effect of level i of A (ecosystem) by level k of C (species) (LA df=1), (df=1) random Bj(i) = effect of level j of B (site) nested within level i of A (ecosystem) (df=8) random lj(ik) = effect of level j of B (site) nested within A (ecosystem) by C (species) interaction (df=8) random eijkl = experimental error (LA df=16), (df=39) Additional data collected on Serenoa repens was also analyzed (Eq. 3-7). yijl= + i + Bj(i) + eijl (3-7) = true mean i = effect of level i of A (ecosystem) (df=1) random Bj(i) = effect of level j of B (site) nested within level i of A (ecosystem) (df=8) random eijl = experimental error (df=20) When interaction variables were not significant (p<0.1), they were dropped from the model. All other tests were performed at =0.05. All pairwise comparisons of fixed means were performed using Tukeys test, rejecting the null hypothesis that there was no difference between treatments when p<0.05. Results Ecosystem Differences Leaf area Based on the understory species studied, hardwood ecosystems had higher (p<0.05) leaf area per leaf, specific leaf area, and leaf area per plant than flatwood ecosystems

PAGE 68

55 (Table 3-1). Leaf area per plant was greater for Myrica cerifera and Serenoa repens in hardwood ecosystems than flatwood ecosystems (p=0.0368) and there was no difference in leaf area per gram between ecosystems. There was a significant interaction between ecosystem and species on the leaf area per leaf (p=0.0298) for Myrica cerifera and Serenoa repens. Upon further analysis, leaf area per leaf was greater in hardwoods than flatwoods for Serenoa repens while Myrica cerifera was not affected by ecosystem type. Moisture content Hardwood ecosystems had significantly higher foliar moisture content during the sampling period, than flatwood ecosystems (p=0.0003) (Table 3-2). The foliar moisture content in understory species within flatwood sites was less variable than understory species within hardwood sites (Table 3-2). However, the moisture content of Myrica cerifera and Serenoa repens were the same between ecosystems (p=0.0771). Volatile solids Foliar volatile solids were not significantly different between ecosystems (p=0.5913). However, species sampled within flatwood sites had much higher variability in volatile solids than species sampled within hardwood sites (Table 3-2). There was also no difference between the foliar volatile solids of Myrica cerifera and Serenoa repens between ecosystems (p=0.2660). Energy content Total energy content per gram was significantly higher in flatwood ecosystems than in hardwood ecosystems (p<0.0001) (4,944.59 calories per gram and 4,776.98 calories per gram, respectively). However, total foliar energy content per plant was not significantly (p=0.0569) different between ecosystems. There was no difference between ecosystem type for total energy content per gram and total foliar energy content per plant

PAGE 69

56 measurements of Myrica cerifera and Serenoa repens. Live foliage of Myrica cerifera and Serenoa repens account for a mean 4,800,000 kilocalories per hectare in flatwood ecosystems and 500,000 kilocalories per hectare in hardwood ecosystems, although the direct ecosystem effect on the mean kilocalories per hectare released from the live foliage of Myrica cerifera and Serenoa repens was not significant (p=0.0961). Species Differences Leaf area Serenoa repens in hardwood ecosystems has significantly more leaf area per leaf than Serenoa repens in flatwood ecosystems (p<0.05). Serenoa repens, in either ecosystem type, had more leaf area per leaf than all other species examined in both ecosystems (Table 3-1). Vaccinium myrsinites had a very low leaf area at 0.34 cm2, but it was not statistically different from any species except Serenoa repens in either ecosystem type. Callicarpa americana had significantly higher specific leaf area (p<0.05) than Myrica cerifera (in flatwoods), Vaccinium arboreum, Vaccinium myrsinites, Ilex glabra, Lyonia ferruginea, Ilex opaca, and Serenoa repens (in either ecosystem type) (Table 3-1). Serenoa repens in hardwood ecosystems also had significantly (p<0.05) higher leaf area per plant than any other species (Table 3-1). Serenoa repens in flatwood ecosystems had significantly (p<0.05) greater leaf area per plant than both Gaylussacia dumosa and Vaccinium mysinities, which were significantly (p<0.05) lower than all other species. Moisture content Callicarpa americana had significantly higher foliar moisture content during the sampling period than any other species studied (p<0.05). Vaccinium arboreum had

PAGE 70

57 significantly higher foliar moisture content (p<0.05) than both Lyonia ferruginea and Serenoa repens (in either ecosystem type) (Table 3-2). Volatile solids Vaccinium myrsinites had significantly higher foliar volatile solid content (p<0.05) than Myrica cerifera (in flatwood ecosystems) and Serenoa repens (in either ecosystem type) (Table 3-2). Serenoa repens (in flatwood ecosystems) had significantly (p<0.05) lower foliar volatile solid content than Ilex opaca, Gaylussacia dumosa, Callicarpa americana, Quercus nigra, and Myrica cerifera (in hardwood ecosystems). Energy content Both energy content measurements (calories per gram and kilocalories per plant) had significant random interaction between site and species (p<0.10). By analyzing the data individually for each species, the random effect of site (nested within ecosystem) was significant (p<0.05) for Myrica cerifera and Serenoa repens in calories per gram and Serenoa repens in kilocalories per plant. The overall site effect was not significant (p>0.05) for either calories per gram or kilocalories per plant. Species was highly significant (p<0.0001) for both measurements. Total energy content per gram was distinctly different when comparing species (Figure 3-1). Ilex glabra and Lyonia ferruginea had significantly (p<0.05) higher total energy content per gram than all other species, followed by Gaylussacia dumosa, Vaccinium myrsinites, and Ilex opaca. Myrica cerifera (in either ecosystem type) and Vaccinium arboreum had significantly greater total energy content per gram (p<0.05) than Serenoa repens (in either ecosystem type), Quercus nigra, and Callicarpa americana.

PAGE 71

Table 3-1. Leaf area per leaf, specific leaf area, and leaf area per plant for species in flatwood and hardwood ecosystems. Standard error is given in parentheses (n=5 for species and n=30 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison between species. 58 Ecosystem Species cm2/leaf cm2/gram cm2/plant Flatwood Gaylussacia dumosa 6.46 (1.24) c 307.7 (23.9) ab 1,443.3 (269.4) c Ilex glabra 5.70 (0.27) c 169.4 (14.2) b 2,096.2 (831.7) bc Lyonia ferruginea 7.93 (0.92) c 155.7 (6.4) b 2,131.6 (323.9) bc Vaccinium myrsinites 0.34 (0.09) c 182.4 (17.6) b 823.2 (452.3) d Myrica cerifera 10.13 (2.69) c 237.1 (29.1) b 8,126.7 (1279.1) bc Serenoa repens 2597.14 (368.82) b 108.3 (9.2) b 19,959.4 (3129.6) b Flatwood mean 437.95 (187.83) 193.4 (13.7) 5,763.4 (1370.5) Hardwood Callicarpa americana 59.96 (7.76) c 440.9 (95.1) a 2,744.9 (917.0) bc Ilex opaca 41.71 (3.49) c 134.9 (22.4) b 12,334.9 (3887.4) bc Quercus nigra 56.38 (19.08) c 268.1 (42.4) ab 3,829.7 (2037.9) bc Vaccinium arboreum 7.75 (0.98) c 212.5 (47.4) b 4,847.7 (1635.0) bc Myrica cerifera 11.64 (0.44) c 283.3 (56.9) ab 13,995.1 (7788.2) bc Serenoa repens 4191.90 (557.02) a 135.4 (7.2) b 41,460.8 (8228.2) a Hardwood mean 728.22 (299.82) 245.9 (27.7) 13,202.2 (3099.4) * indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems.

PAGE 72

Table 3-2. Foliar moisture content and volatile solids for species within flatwood and hardwood ecosystems. Standard error is given in parentheses (n=15 for species and n=60 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison. Ecos y ste m S p ecies Moisture Content ( % ) Volatile Solids ( m g /k g) Flatwood Gaylussacia dumosa 124 (6) bcd 916,667 (9,644) ab Ilex glabra 139 (10) bcd 826,000 (54,465) abcd Lyonia ferruginea 119 (12) cd 841,333 (54,632) abcd Vaccinium myrsinites inites 133 (10) bcd 933,333 (17,284) a Myrica cerifera 141 (16) bcd 812,000 (11,960) bcd Serenoa repens 101 (2) d 732,667 (15,352) d Flatwoodmean 126 ( 4 ) 843 667 ( 15 060 ) Hardwoo d Callicar p a americana 460 ( 60 ) a 900 667 ( 7 836 ) ab Ilex opaca 144 (12) bcd 918,667 (5,845) ab Quercus nigra 187 (25) bcd 874,000 (16,784) ab Vaccinium arboreum 223 (14) b 837,333 (13,433) abcd Myrica cerifera 216 (24) bc 856,667 (26,071) abc Serenoa repens 113 (5) d 738,000 (9,962) cd Hardwoodmean 224 ( 17 ) 854 222 ( 8 569 ) 59 indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems.

PAGE 73

60 cal/gram 40004200440046004800500052005400 kcal/plant 050010001500 Flatwood species Gaylussacia dumosaIlex glabraLyonia ferrugineaVaccinium myrsinitesMyrica ceriferaSerenoa repens kcal/ha 0.02.0e+64.0e+66.0e+68.0e+61.0e+71.2e+7 40004200440046004800500052005400 050010001500 Hardwood species Callicarpa americanaIlex opacaQuercus nigraVaccinium arboreumMyrica ceriferaSerenoa repens 0.02.0e+64.0e+66.0e+68.0e+61.0e+71.2e+7 ababcddbdccdabcddcddcdcdcdcdcdabcccccccccc Calories per gram foliar b iomass (gram) per plant 0.001 kcal per cal = kcal per plant Kcal per plant stems per ha of individual species at individual site = foliar energy content (kcal) per ha Figure 3-1. Foliar energy content in calories per gram, kilocalories per plant, and kilocalories per hectare for species within pine flatwood and hardwood hammock ecosystems. Standard error is shown in error bars (n=15). Lower-case letter indicate significant (p<0.05) difference in Tukeys pairwise comparison

PAGE 74

61 Serenoa repens in hardwood ecosystems had significantly higher foliar energy per plant (p<0.05) than any other species, followed by Serenoa repens in flatwood ecosystems (Figure 3-1). Ilex opaca had significantly greater foliar energy per plant (p<0.05) than Callicarpa americana, Gaylussacia dumosa, and Vaccinium myrsinites. When foliar energy content of species is examined on a per hectare basis, Serenoa repens (in flatwoods), followed by Ilex glabra had higher kilocalories per hectare than all other species (Figure 3-1). The species by site interaction was significant in the model (p<0.0001) for energy content per hectare due to a significant random effect of site on Gaylucassia dumosa, Vaccinium myrsinites, Myrica cerifera, and Serenoa repens (in flatwood ecosystems) and Ilex opaca and Serenoa repens (in hardwood ecosystems). Serenoa repens Dead Serenoa repens foliage from hardwood ecosystems had higher moisture content than dead foliage from flatwood ecosystems (p=0.0397) (Table 3-3). There was no significant difference (p>0.05) between ecosystems for additional foliar measurements of Serenoa repens. There was a significant random effect (p<0.05) of site nested within ecosystem for live volatile solids, live and dead energy content, and kilocalories per plant. Energy content of dead foliage was generally lower than the energy content of live foliage (Table 3-3). Foliar energy content per plant was was nearly doubled by incorporating the dead foliar energy content into the calculations. In addition, potential foliar energy content per hectare was nearly doubled.

PAGE 75

Table 3-3. Additional data for Serenoa repens included measurements of moisture content, volatile solid content, total energy content for dead foliage in addition to live foliage. Standard error is given in parentheses (n=15). Total energy released by both live and dead foliar components are incorporated into the measurement of kilocalories per plant. Ecosystem effects were not significant (p>0.05) for all measurements except deaf foliar moisture content. Volatile Solids Energy Content Serenoa repens dead foliar moisture content (%) live foliage dead foliage live foliage dead foliage kilocalories/plant kilocalories/hectare Flatwood 13 (5) 732,666(15,352) 734,000 (11,944) 4,638.92 (18.62) 4,538.42 (21.82) 2,001.4 (391.7) 16,307,700 (5,341,485) Hardwood 30 (5) 738,000 (9,962) 722,000 (9,522) 4,702.24 (20.84) 4,612.24 (37.04) 2,222.8 (400.1) 1,223,420 (370,829) indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems. 62

PAGE 76

63 Discussion Ecosystem Differences Most of the ecosystem effects on flammability presented in this paper, except leaf area, are largely due to the differences in the species studied within each ecosystem type, as the foliar flammability of Myrica cerifera and Serenoa repens was not affected by ecosystem type. Generally, the species that represent the flatwood ecosystem in this study are more flammable than the species representing the hardwood ecosystem. Results of leaf area were expected because of the typical physiological response of leaves to shade. The hardwood sites had less variable and greater canopy closure; therefore the understory plants were in a more shaded environment than the flatwood sites (Table 2-2). It is largely recognized that leaves which grow in shady sites have high specific leaf area (leaf area per gram) (Lambers et al. 1998). This was reflected in our specific leaf area measurements as well in the leaf area per leaf and per plant. Certain species may be more responsive to the influence of shade as indicated in the leaf area per leaf difference of Serenoa repens between ecosystem types while Myrica cerifera did not change between ecosystems. The higher leaf area in hardwood ecosystems could increase the flammability of the understory species by increasing their ignitability. Ignitability may be higher based on the increased surface area for moisture to be evaporated from in a drought or more importantly, a nearby wildfire (Rundel 1981). It would be valuable to know the leaf area to volume for the species studied to relate the leaf structure more directly with flammability (Montgomery and Cheo 1971, Rundel 1981). Moisture content of foliage within hardwood hammocks may be greater compared to pine flatwoods. However, as this study indicates, this difference is likely due to the

PAGE 77

64 different understory species between the two ecosystems as there was no direct effect of ecosystem type on the foliar moisture content of Myrica cerifera and Serenoa repens. The more fire-prone Eucalyptus-Casuarina dry sclerophyll forests had species with lower foliar moisture content than species in less fire-prone woodlands in Tasmania (Dickinson and Kirkpatrick 1985). Foliar moisture content of species in fire-prone fynbos ecosystems was lower than species in forest patches in South Africa (van Wilgen et al. 1990). In addition, Eriksson et al. (2003) found higher fuel moisture content in less fire-prone Afromontane forest compared to the more fire-prone Acacia woodland. Moisture content strongly influences the flammability of plant tissue, likely more than any other characteristic (Gill et al. 1978, Rundel 1981, Etlinger 2001). The consumability, measured as volatile solids, of leaf tissue between ecosystems is more variable in pine flatwoods than hardwood hammocks, although there was not an overall difference between ecosystems. This may be related to the higher variability in ecosystem structure within sample sites (Chapter 2). The percent calorimeter ash (percent of initial sample left after calorimetry combustion) is another measure of the consumability of plant tissue. Rodrguez An et al. (1995) presented calorimeter ash values of 0.16 to 2.86% (Eq. 3-8) and Nez-Regueira et al. (2001) presented calorimeter ash values of 0.22 to 4.62%. These values are much lower (higher comsumability) than the values reported through the volatile solid analyses in this study. For example, the highest consumability was reported as 933,333 mg/kg, or 6.67%, for Vaccinium myrsinites. It is possible that through isoperibol oxygen consumption calorimetry, a more complete combustion is accomplished than accomplished with the method used in this study. However, the methodology used in this study may more closely represent wildfire

PAGE 78

65 conditions as plant tissue is combusted in an oxygen-rich environment within the calorimeter. calorimeter ash % = 100 (weight of final sample/weight of initial sample) (3-8) Combustability on a gram basis was higher for foliage from understory plants in flatwoods than hardwoods. These results are similar to the results of Dickinson and Kirkpatrick (1985) and van Wilgen (1990) who reported higher energy content per gram in foliage from more fire prone ecosystems (Eucalyptus-Casuarina dry sclerophyll and fynbos, respectively) than less fire prone ecosystems (woodlands and forest patches, respectively). This has been suggested as an adaptation by species in fire-prone ecosystems to improve species survivability (Bond and Midgley 1995 and Possingham et al. 1995). Species Differences Based on foliar flammability, clear distinctions can be made between the flammability of species. Callicarpa americana is the least flammable species studied, and Serenoa repens, Ilex glabra, Lyonia ferruginea, Myrica cerifera, and Ilex opaca are the most flammable species. However, these highly flammable species are flammable for different reasons. As mentioned before, the species found in flatwood ecosystems are generally more flammable than those found in hardwood ecosystems. It must be stated the energy content values expressed are solely a way to compare species to one another, not an absolute measure of energy content. Calorimetry, as a measure of total heat release, also does not distinguish between effective heat content (flaming heat content) of the tissue and heat released from the residue (smoldering heat content). In addition, oxygen consumption calorimetry is a complete combustion process that likely overestimates the amount of energy that would be released from the same

PAGE 79

66 sample in field conditions. Fuel and environmental conditions greatly affect the amount of potential fuel that will be available fuel to a wildfire and how the fuel will combust (Pyne et al. 1996). These conditions include: fuel moisture content, spatial arrangement, time of day, wind speed, weather, climatic conditions, and other microclimate influences of fire (Pyne et al. 1996). Those species that have high foliar energy content per hectare may be target species for fuel reduction strategies beyond defensible space in the WUI. These species include Ilex glabra and Myrica cerifera in flatwood ecosystems, and Serenoa repens in either ecosystem type. These species could be singled out to reduce the environmental impact of fuel reduction (mechanical, chemical, or biological) while still reducing the overall wildfire hazard of an area. However, by selectively removing individual species, the cost of fuel reduction would likely be higher. Gaylussacia dumosa Gaylussacia dumosa had a high specific leaf area, potentially exposing more of the foliar biomass to environmental extremes. However, this characteristic does not likely change the overall flammability of this species. The species does contain relatively high energy content per gram but because of the low foliar biomass, has low foliar energy content per plant. Therefore an individual Gaylussacia dumosa plant is likely to be highly flammable, but will not release much total heat. Ilex glabra The foliar energy content of Ilex glabra is similar to those reported by Dickinson and Kirkpatrick (1985) for foliar samples from Eucalyptus and Acacia species (Ilex glabra had a mean energy content converted to 21,482 Jg-1). The foliar energy content is high, increasing the amount of heat generated per gram of tissue consumed compared to

PAGE 80

67 other species studied. Although the energy content per individual plant is comparatively low, Ilex glabra grew so dense in the sites that Ilex glabra is a serious wildfire hazard to structures. In addition, Ilex glabra foliage is known to be highly flammable because of volatile extractives (Burgan and Susott 1991). Ilex glabra foliage contains 44.6% (based on dry weight) total ether and benzene-ethanol extractives which contribute greatly to the heat release of foliage (Shafizadeh et al. 1977). The substance present in Ilex glabra foliage contributing to the production of volatiles is cutin (Rogers et al. 1986). It is also possible that many volatile carbon compounds were incorporated into the total energy content measured through calorimetry. Burgan and Susott identified low-temperature volatiles as compounds becoming volatile up to 300oC (1991), whereas foliage in this study was dried at 70oC. Lyonia ferruginea Lyonia ferruginea foliage contains as much total energy per gram as Ilex glabra. In addition, the foliar energy content per plant is comparatively high. The foliar moisture content of this species is low, further increasing the flammability of this species. Lyonia ferruginea plants in pine flatwoods within the WUI can be a potential wildfire hazard if they are found in dense patches. Maintenance of Lyonia ferruginea in a landscape near a home should be to prune excess foliage and to not allow this species to be incorporated into dense plantings. Vaccinium myrsinites Although the foliar volatile content and energy content of Vaccinium myrsinites is relatively high, the low foliar biomass per individual reduces the hazard. Again, as with other plants, dense plantings of this species would present a hazard.

PAGE 81

68 Callicarpa americana Callicarpa americana foliage contains high specific leaf area, similar to Gaylussacia dumosa. However, the average foliar moisture content of this species was far greater than any other species studied. In addition, Callicarpa americana foliar energy content per gram, per plant, and per hectare was comparatively lower than all other species studied. For these reasons, Callicarpa americana is the least flammable species studied based on our results. This species has been cited as having low flammability on extension lists (Florida Firewise Communities 2000, Monroe and Long 2001). Ilex opaca For those species studied only in the hardwood ecosystemCallicarpa americana, Ilex opaca, Quercus nigra, and Vaccinium arboreumIlex opaca is the most flammable based on this study. The foliage had moderate moisture content, relatively high energy content per gram, and high energy content per plant and per hectare. This species has been cited as having high flammability by MacCubbin and Mudge (2002). Quercus nigra Foliage of Quercus nigra plants had a relatively high specific leaf area. However, the foliar energy content was one of the lowest values in this study. Foliar energy content per plant was also low. However, because the abundance of Quercus nigra in the understory of the study sites can be high, this species may contribute a large amount of energy content (heat) per hectare. Individual plants of Quercus nigra do not have characteristics of highly flammable plants, but could pose a hazard to WUI structures within dense plantings of small plants.

PAGE 82

69 Vaccinium arboreum As reported in Chapter 2, Ilex opaca and Vaccinium arboreum had similar biomass results. However, in the analyses of foliar biomass, differences between the species are more apparent. Vaccinium arboreum foliage contains less energy per gram which translates to less potential energy released per plant. In addition, Vaccinium arboreum does not contribute highly to foliar energy content per hectare. This species was listed as having low-flammability by Monroe and Long (2001). Myrica cerifera Foliage of Myrica cerifera did not have any characteristics exhibiting high flammability. The foliar moisture content, energy content per gram, and energy content per plant were moderate. However, given the frequency of Myrica cerifera listed as a highly flammable plant (Lippi and Kuypers 1998, Monroe and Long 2001, MacCubbin and Mudge 2002), there may be other characteristics of Myrica cerifera which were not detected with this methodology. As the common name wax myrtle suggests, the waxy substance on the leaves may be highly volatile. But a study by Burgan and Susott (1991) found that Myrica cerifera foliage contained significantly less low-temperature volatile compounds than either Ilex glabra or Serenoa repens. The morphology of Myrica cerifera plants growing naturally in wildland areas has been observed to be very different to the morphology of Myrica cerifera cultivated as a landscape specimen. These morphological differences may contribute to the high flammability of Myrica cerifera in a cultivated setting whereas Myrica cerifera in natural areas does not have high flammability. More study of this plant species, and how flammability changes with horticultural practices, is necessary to remove it from the lists as a highly flammable plant.

PAGE 83

70 Serenoa repens Although the specific leaf area was not high compared to other species, Serenoa repens leaves and plants have high leaf area, especially when located in a hardwood ecosystem. The moisture content of Serenoa repens is comparatively low. The abundance of dead foliage with very low moisture content (13% in flatwoods and 30% in hardwoods) also makes Serenoa repens a hazard to WUI structures. The low heat release per gram of dry weight of Serenoa repens foliage has been documented by Shafizadeh et al. (1977). In addition, Serenoa repens contains only 13.1% ether and benzene-ethanol extractives that did not change the heat release of foliage after being removed (Shafizadeh et al. 1977). Although the species has relatively low foliar energy content, the large amount of biomass makes the total potential foliar energy release extremely high. This is especially true when incorporating the dead foliar energy content into the total foliar energy content per plant. Energy content per gram for dead foliage was lower than the energy content for live foliage which was expected as dead foliage contains less volatile compounds than live foliage (Mak 1988, Burgan and Susott 1991). It can be concluded that Serenoa repens can be a threat to WUI structures in either ecosystem type which has been recognized in many publications (Lippi and Kuypers 1998, Monroe and Long 2001, MacCubbin and Mudge 2002). Conclusions Many of the chemical characteristics of fuel have shown seasonal variation: moisture content (Pyne et al. 1996, Agee et al. 2002) volatile compounds (Bunting et al. 1983 and Owens et al. 1998), and energy content of forest wastes (Rodrguez An 1995 and Nez-Regueira et al. 2001). In addition, disturbance patterns may affect the

PAGE 84

71 chemical composition of fuels, such as time since last fire (Thackston et al. 1982, Mallik 1994, Rieske et al. 2002). Monitoring how the chemical components of these species change throughout the season or with different weather patterns may provide valuable information. We confidently conclude that plant species within flatwood ecosystems are likely to be more flammable than species within hardwood ecosystems. We can also make some general guidelines for species recommendations. For WUI structures in flatwood ecosystem types, Lyonia ferruginea, Myrica cerifera, and especially Ilex glabra and Serenoa repens should be discouraged from existing within defensible space zones. In hardwood ecosystems, Ilex opaca and Serenoa repens should be discouraged from being used in structural landscaping. Callicarpa americana is the species that, according to this study, has low flammability and can be incorporated into firewise landscaping.

PAGE 85

CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS Conclusions It first must be said that all plants are flammable. However, some plants are less flammable than others and these species are appropriate for firewise landscaping. Less flammable plants can be used with caution in arrangements that are aesthetically pleasing with low wildfire hazard. Plant selection for landscaping, even within fire-prone ecosystems, should be a balance of perceived benefits. These benefits include energy conservation, water conservation, wildlife habitat, aesthetics, privacy, noise reduction, and wildfire hazard reduction. It is also important to consider that choosing the right plant for the right site will provide healthier, more vibrant landscape plants with less water stress and less dead material. Maintenance is also very important as biomass can be kept low and moisture content kept high with proper maintenance. Based on the results from this study, we embrace Hypothesis I and II and reject Hypothesis III. Hypothesis I: Understory plants in pine flatwoods are more flammable than understory plants in hardwood hammocks. Understory plants in pine flatwoods are more flammable than understory plants in hardwood hammocks. We conclude that the ignitability (less distance between litter and lowest branch, lower foliar moisture content), sustainability (higher fuel loading), and combustability (total energy content) are higher for understory species within pine flatwood ecosystems than hardwood hammock ecosystems. However, the consumability 72

PAGE 86

73 (fine fuel biomass and volatile solids) was not different between pine flatwoods and hardwood hammocks. These results are similar to other comparisons of plant flammability between more fire prone ecosystems and less fire prone ecosystems (Table 4-1). Table 4-1. Summary table comparing the results from flammability studies in Tasmania (Dickinson and Kirkpatrick 1985), South Africa (van Wilgen et al. 1990), and Ethiopia (Eriksson et al. 2003). YES= results agree with statement in left column, NO= results do not agree with statement in left column, and na= information not available. The more fire-prone ecosystem contains Tasmania South Africa Ethiopia SE United States more litter biomass na na YES YES less total biomass na YES YES YES less fine fuels na YES YES NO more dense fuel bed na YES na YES lower foliar moisture content YES YES YES YES higher energy content YES YES na YES Hypothesis II: Differences in the flammability of species are significant. Flammability differences among species were the result of differences in foliar biomass, fine fuel biomass, foliar moisture content, and foliar energy content. Hypothesis III: The flammability of wax myrtle (Myrica cerifera) and saw palmetto (Serenoa repens) are different between ecosystems. By analyzing the data from Myrica cerifera and Serenoa repens separately, direct effects of ecosystems on the flammability of individual species were explored. The flammability characteristics were not different between ecosystems for these two species, except for litter measurements. Therefore we have to conclude that flammability of the same species in different ecosystem contexts is not different. This result is based only on species adapted to each ecosystem type. The same results are not predicted when comparing the flammability of a species adapted to a hardwood hammock ecosystem placed around a home in a pine flatwoods ecosystem.

PAGE 87

74 Also, the concept of wildfire risk must be addressed as plants of the same species within pine flatwood ecosystems are more likely to be exposed to wildfire in the first place. Therefore the flammability of a plant must also be evaluated within its context. Pine flatwoods contain more litter biomass which accumulates to a deeper depth than hardwood hammocks, therefore increasing the flammability of both Myrica cerifera and Serenoa repens. In addition, the density per hectare of Myrica cerifera and Serenoa repens is greater in pine flatwoods increasing the hazard of these species within an ecosystem associated with the WUI. Recommendations This section is devoted to recommendations based on the context of firewise landscaping recommendations through the following assessments: Assessment I: What are the implications for firewise landscaping in these ecosystems? Assessment II: In what way would individual plants contribute to structure survival or destruction in a wildfire? Pine Flatwoods It is widely recognized that WUI areas within southern pine ecosystems have higher wildfire hazard compared to WUI areas within hardwood ecosystems. This study shows that, in addition to the natural fire regime, the individual understory plants within the pine ecosystems are more flammable than those in hardwood ecosystems. Ilex glabra, Lyonia ferruginea, and Serenoa repens should not be planted or maintained within the defensible space zones of WUI structures. Ilex glabra and Lyonia ferruginea are intrinsically highly flammable as they produce a great amount of heat once ignited. Serenoa repens, however, is flammable based on the great amount of dead foliage and biomass in general. In addition, fire specialists do not recommend Myrica cerifera within

PAGE 88

75 WUI landscaping. Gaylussacia dumosa and Vaccinium myrsinites are acceptable within WUI landscaping if not densely or continuously planted, and may provide food for wildlife. Hardwood Hammocks Firewise landscaping techniques are not as necessary in hardwood hammock ecosystems. In these ecosystems, WUI landscaping can be designed more to meet other landowner needs than wildfire protection. However, it must be noted that within the WUI, the likelihood of ignition may be higher because of increased arson or accidents. Also, hardwood hammocks which are disturbed (fewer overstory trees, more dense understory, and more dead fuel) may be more prone to wildfire, especially in an extended drought. In addition, the presence of pine trees around a WUI structure, even within a hardwood hammock can cause a hazard because of the fallen pine needles. Firewise landscaping within hardwood hammocks, if necessary, does not need to be as stringent as within flatwood ecosystems. This study suggests that Serenoa repens and Ilex opaca should not be planted around homes within hardwood hammock ecosystems. However, if properly maintained on a routine basis, individuals of these species may be used. Proper maintenance includes removal of debris, thinning of foliage, and pruning of dead and lower branches. Landscaping arrangement should include vertical and horizontal separation of plants or landscape islands surrounding a WUI structure. Future Research Continued research on the social and economic dimensions of the WUI will improve our understanding of the WUI. In addition, it would be valuable to know how WUI residents decide what kind of landscaping and plants they maintain around their homes. In this study we focused on native understory species as these plants would likely

PAGE 89

76 exist within WUI in the South. However, people have a variety of choices for landscape plants. To make comprehensive lists of plant flammability, the author recommends a formal survey of fire professionals within the South. This survey would attempt to identify the most-flammable and least-flammable plants within specific biographical regions based on the accumulated experience of people with knowledge of local flora and wildfire. This experience would incorporate many more plant species within a variety of environmental conditions. Field and laboratory studies could then address discrepancies in the results. Another method of determining flammability of plant species may be to standardize a ranking process which would be easy to use throughout the South. For example, species could be examined either through landscaping publications, species descriptions, or through direct study to determine the potential level of hazard (Table 4-2). In this case, a complete list of characteristics would be used in a standardized method. Based on literature review and the results of this study, the author suggests that important characteristics for evaluation include foliar moisture content, foliar biomass, and fine fuel biomass. Additional tests could include either foliar volatile extractive analyses or foliar energy content. Table 4-2. Proposed ranking mechanism of plant species by flammability. Plant Flammability Ranking Criteria Low Flammability No hazardous characteristics Moderate Flammability 1-2 hazardous characteristics that can be mitigated through horticulture High Flammability 3 or more hazardous characteristics; characteristic of extreme risk; one or more characteristic that cannot be mitigated through horticulture

PAGE 90

APPENDIX A LOCATION OF PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS BY COUNTY IN FLORIDA

PAGE 91

Figure A-1. Distribution of pine flatwoods by county in Florida adapted from Myers and Ewel (1990). Dark gray areas represent counties where pine flatwoods exist. Figure A-2. Distribution of hardwood hammocks by county in Florida adapted from Myers and Ewel (1990). Dark gray areas represent counties where hardwood hammocks exist. 78

PAGE 92

APPENDIX B LOCATION OF STUDY SITES 6-7 2 9-10 5 1 4 Collection Sites: = hardwood site = flatwood site = both sites 8 3 1) ACMF (UF) 2) Twin Rivers S. F. (Ellaville Tract) (DOF) 3) Withlacooche S. F. (Richloam Tract) (DOF) 4) Little River Springs (SRWMD) 5) Welaka S. F. (DOF) 6) Osceola N. F. (USDA FS) 7) Osceola N. F. (USDA FS) 8) Steinhatchee (SRWMD) 9) Jennings S. F. (DOF) 10) Jennings S. F. (DOF) Figure B-1. Location of pine flatwood and hardwood hammock study sites throughout North Central Florida. Sites used were owned by the University of Florida (UF), Florida Division of Forestry (DOF), Suwannee River Water Management District (SRWMD), and USDA Forest Service (USDA FS). 79

PAGE 93

APPENDIX C ABSOLUTE AND RELATIVE DENSITIES OF UNDERSTORY AND MIDSTORY SPECIES AT EACH SITE

PAGE 94

Table C-1. Absolute (stems per hectare) and relative densities (%) of understory species in 8, 12.56m2 plots at each study site. Ecosystem Site Species Stems/ha Rel. density dead 3980.89 2.9 Gaylussacia dumosa 6070.86 4.5 Ilex glabra 94446.66 69.6 Lyonia lucida 23785.83 17.5 Myrica cerifera 199.04 0.1 Quercus nigra 99.52 0.1 Quercus pumila 895.70 0.7 Serenoa repens 4876.59 3.6 Austin Cary M. F. Vaccinium myrsinites 1293.79 1.0 dead 2786.62 2.1 Gaylussacia dumosa 2289.01 1.7 Hypericum brachyphyllum 99.52 0.1 Hypericum crux-andrae 298.57 0.2 Ilex glabra 111564.49 82.5 Licania michauxii 199.04 0.1 Lyonia ferruginea 99.52 0.1 Lyonia lucida 199.04 0.1 Magnolia virginiana 199.04 0.1 Myrica cerifera 4279.46 3.2 Serenoa repens 8459.39 6.3 Un-identified 99.52 0.1 Osceola N. F. Vaccinium myrsinites 4578.03 3.4 Aster walteri 497.61 0.6 Befaria racemosa 1393.31 1.8 dead 8957.01 11.3 Hypericum crux-andrea 99.52 0.1 Hypericum reductum 99.52 0.1 Ilex glabra 56926.75 71.9 Lyonia ferruginea 5573.25 7.0 Magnolia virginiana 99.52 0.1 Myrica cerifera 1393.31 1.8 Quercus pumila 398.09 0.5 Serenoa repens 3582.80 4.5 Jennings S. F. Un-identified 99.52 0.1 Clethra alnifolia 298.57 0.2 dead 8957.01 5.1 Gaylussacia dumosa 2886.15 1.6 Gordonia lasianthus 1492.83 0.8 Ilex glabra 130871.82 73.8 Flatwood Welaka S. F. Lyonia ferruginea 3980.89 2.2 81

PAGE 95

82 Table C-1. Continued Ecosystem Site Species Stems/ha Rel. density Lyonia lucida 16023.09 9.0 Serenoa repens 11942.68 6.7 Welaka S. F., cont. Vaccinium myrsinites 796.18 0.4 Hypericum prolificum 99.52 0.1 Ilex glabra 48865.45 58.5 Lyonia ferruginea 2089.97 2.5 Lyonia lucida 15824.04 19.0 Myrica cerifera 7364.65 8.8 Quercus hemisphaerica 696.66 0.8 Quercus nigra 1990.45 2.4 Sassafras albidum 99.52 0.1 Serenoa repens 4578.03 5.5 Un-identified 99.52 0.1 Flatwood, cont. Withlacoochee S. F. Vaccinium myrsinites 1791.40 2.1 Carya glabra 99.52 4.5 dead 99.52 4.5 Ilex opaca 597.13 27.3 Quercus nigra 398.09 18.2 Serenoa repens 99.52 4.5 Vaccinium arboreum 298.57 13.6 Twin Rivers S. F. Vaccinium corymbosum 597.13 27.3 dead 99.52 1.7 Ilex glabra 597.13 10.3 Liquidambar styraciflua 199.04 3.4 Myrica cerifera 796.18 13.8 Quercus hemisphaerica 199.04 3.4 Quercus nigra 2488.06 43.1 Serenoa repens 298.57 5.2 Vaccinium arboreum 298.57 5.2 Vaccinium corymbosum 199.04 3.4 Osceola N. F. Vaccinium stamineum 597.13 10.3 Acer rubrum 99.52 0.6 dead 696.66 4.3 Gaylussacia dumosa 99.52 0.6 Ilex glabra 398.09 2.5 Ilex opaca 497.61 3.1 Lyonia ligustrina var. foliosiflora 199.04 1.2 Magnolia virginiana 3184.71 19.6 Myrica cerifera 1393.31 8.6 Quercus hemisphaerica 1691.88 10.4 Hardwood Jennings S. F. Quercus nigra 5672.77 35.0

PAGE 96

83 Table C-1. Continued Ecosystem Site Species Stems/ha Rel. density Quercus stellata 99.52 0.6 Serenoa repens 398.09 2.5 Vaccinium arboreum 1592.36 9.8 Jennings S. F., cont. Vaccinium myrsinites 199.04 1.2 Acer rubrum 99.52 0.7 Crataegus marshallii 199.04 1.4 Cyrilla racemiflora 1990.45 14.3 dead 796.18 5.7 Diospyros virginiana 99.52 0.7 Ilex opaca 99.52 0.7 Lyonia lucida 4279.46 30.7 Myrica cerifera 1691.88 12.1 Quercus hemisphaerica 99.52 0.7 Quercus nigra 895.70 6.4 Ribus betulifolius 99.52 0.7 Serenoa repens 2.1 Un-identified 1492.83 10.7 Vaccinium arboreum 597.13 4.3 Vaccinium stamineum 497.61 3.6 Viburnum obovatum 398.09 2.9 Steinhatchee, SRWMD Viburnum rufidulum 298.57 2.1 Bafaria racemosa 99.52 1.3 dead 497.61 6.4 Fabaceae 99.52 1.3 Ilex opaca 398.09 5.1 Myrica cerifera 199.04 2.6 Quercus hemisphaerica 99.52 1.3 Quercus nigra 1592.36 20.5 Serenoa repens 1293.79 16.7 Vaccinium arboreum 1094.75 14.1 Vaccinium myrsinites 298.57 3.8 Vaccinium stamineum 99.52 1.3 Hardwood, cont. Little River Springs, SRWMD Viburnum obovatum 1990.45 25.6 298.57

PAGE 97

84 Table C-2. Absolute (stems per hectare) and relative density (%) for midstory species (>3 m in height but <6.4 cm dbh) in 4, 400m2 plots at each study site. Ecosystem Site Species Stems/ha Rel. density Pinus elliottii 6.25 5.3 Quercus hemisphaerica 87.50 73.7 Quercus nigra 18.75 15.8 Jennings S. F. Quercus stellata 6.25 5.3 Magnolia grandiflora 12.50 18.2 Magnolia virginiana 43.75 63.6 Welaka S. F. Pinus elliottii 12.50 18.2 Pinus elliottii 125.00 71.4 Pinus palustris 31.25 17.9 Quercus hemisphaerica 6.25 3.6 Flatwood Withlacoochee S. F. Quercus nigra 12.50 7.1 Carya glabra 68.75 7.8 dead 6.25 0.7 Ilex opaca 156.25 17.7 Juniperus virginiana 6.25 0.7 Liquidambar styraciflua 18.75 2.1 Magnolia virginiana 37.50 4.3 Nyssa biflora 43.75 5.0 Quercus hemisphaerica 6.25 0.7 Quercus nigra 25.00 2.8 Vaccinium arboreum 368.75 41.8 Vaccinium corymbosum 118.75 13.5 Twin Rivers S. F. Viburnum obovatum 25.00 2.8 Acer rubrum 6.25 1.3 Liquidambar styraciflua 31.25 6.3 Magnolia virginiana 6.25 1.3 Myrica cerifera 93.75 18.8 Nyssa biflora 18.75 3.8 Pinus elliottii 12.50 2.5 Quercus hemisphaerica 50.00 10.0 Quercus laurifolia 6.25 1.3 Quercus nigra 256.25 51.3 Osceola N. F. Vaccinium arboreum 18.75 3.8 Acer rubrum 43.75 3.8 Cornus florida 25.00 2.2 dead 18.75 1.6 Ilex opaca 18.75 1.6 Liquidambar styraciflua 12.50 1.1 Lyonia ferruginea 87.50 7.7 Lyonia ligustrina var. foliosiflora 93.75 8.2 Hardwood Jennings S. F. Magnolia virginiana 56.25 4.9

PAGE 98

85 Table C-2. Continued Ecosystem Site Species Stems/ha Rel. density Myrica cerifera 25.00 2.2 Nyssa biflora 25.00 2.2 Pinus elliottii 6.25 0.5 Quercus hemisphaerica 25.00 2.2 Quercus nigra 312.50 27.5 Quercus stellata 56.25 4.9 Vaccinium arboreum 306.25 26.9 Jennings S. F., cont. Viburnum obovatum 25.00 2.2 Acer rubrum 25.00 2.5 Crataegus marshallii 68.75 6.9 Cyrilla racemiflora 168.75 16.9 Ilex opaca 118.75 11.9 Liquidambar styraciflua 12.50 1.3 Myrica cerifera 125.00 12.5 Nyssa biflora 25.00 2.5 Quercus hemisphaerica 56.25 5.6 Quercus nigra 181.25 18.1 Ulmus americana 56.25 5.6 Vaccinium arboreum 100.00 10.0 Vaccinium corymbosum 6.25 0.6 Viburnum rufidulum 12.50 1.3 Steinhatchee, SRWMD Viburnum obovatum 43.75 4.4 Acer rubrum 168.75 12.2 Crataegus marshallii 62.50 4.5 dead 6.25 0.5 Diospyros virginiana 6.25 0.5 Ilex opaca 18.75 1.4 Liquidambar styraciflua 37.50 2.7 Nyssa biflora 6.25 0.5 Nyssa sylvatica 25.00 1.8 Quercus larifolia 6.25 0.5 Quercus nigra 100.00 7.2 Quercus sp. 18.75 1.4 Ulmus americana 6.25 0.5 Vaccinium arboreum 643.75 46.6 Vaccinium corymbosum 18.75 1.4 Vaccinium stamineum 87.50 6.4 Viburnum obovatum 162.50 11.7 Hardwood, cont. Little River Springs, SRWMD Viburnum rufidulum 6.25 0.5

PAGE 99

APPENDIX D ABSOLUTE AND RELATIVE DENSITY, RELATIVE DOMINANCE, RELATIVE FREQUENCY, AND IMPORTANCE VALUES FOR OVERSTORY SPECIES

PAGE 100

Table D-1. Stems per hectare, relative density, relative dominance, relative frequency, and importance value for overstory species within each site. Ecosystem Site Species Stems/ ha Rel. den. Rel. dom. Rel. freq. I. V. Pinus elliottii 87.50 51.9 55.9 40.0 147.8 Pinus palustris 75.00 44.4 41.0 40.0 125.4 Austin Cary M. F. Quercus nigra 6.25 3.7 3.1 20.0 26.8 Pinus elliottii 125.00 60.6 60.6 44.4 165.7 Pinus palustris 75.00 36.4 36.4 44.4 117.2 Osceola N. F. Pinus taeda 6.25 3.0 3.0 11.1 17.2 Pinus elliottii 593.75 91.3 93.3 66.7 251.3 Jennings S. F. Pinus palustris 56.25 8.7 6.7 33.3 48.7 Magnolia grandiflora 6.25 2.3 0.2 10.0 12.5 Magnolia virginiana 18.75 6.8 6.8 20.0 33.6 Persea palustris 6.25 2.3 1.0 10.0 13.3 Pinus elliottii 162.50 59.1 68.9 40.0 168.0 Welaka S. F. Pinus palustris 81.25 29.5 23.1 20.0 72.7 Pinus elliottii 306.25 79.0 83.1 50.0 212.1 Pinus palustris 50.00 12.9 14.9 12.5 40.3 Quercus hemisphaerica 6.25 1.6 0.2 12.5 14.3 Flatwood Withla-coochee S. F. Quercus nigra 25.00 6.5 1.8 25.0 33.3 Carya glabra 12.50 2.0 0.8 3.4 6.3 Ilex opaca 125.00 20.2 8.0 13.8 42.0 Juniperus virginiana 25.00 4.0 1.3 6.9 12.2 Liquidambar styraciflua 62.50 10.1 10.2 10.3 30.6 Nyssa aquatica 43.75 7.1 2.3 13.8 23.2 Nyssa biflora 12.50 2.0 1.1 3.4 6.5 Persea palustris 31.25 5.1 1.9 6.9 13.8 Pinus elliottii 6.25 1.0 2.3 3.4 6.8 Pinus glabra 6.25 1.0 1.0 3.4 5.4 Quercus nigra 218.75 35.4 26.1 13.8 75.2 Quercus virginiana 37.50 6.1 44.1 10.3 60.5 Un-identified 12.50 2.0 0.5 3.4 5.9 Twin Rivers S. F. Vaccinium arboreum 25.00 4.0 0.5 6.9 11.5 Hardwood Osceola N. F. Ilex opaca 18.75 5.3 0.6 9.1 15.0 87

PAGE 101

88 Table D-1. Continued Ecosystem Site Species Stems/ ha Rel. den. Rel. dom. Rel. freq. I. V. Liquidambar styraciflua 25.00 7.0 0.7 9.1 16.8 Nyssa biflora 18.75 5.3 2.5 9.1 16.8 Pinus elliottii 43.75 12.3 21.8 13.6 47.7 Pinus palustris 12.50 3.5 8.5 4.5 16.5 Quercus hemisphaerica 106.25 29.9 39.1 18.1 87.1 Quercus laurifolia 56.25 15.8 14.3 18.2 48.3 Quercus nigra 62.50 17.5 11.5 13.6 42.7 Osceola N. F., cont. Ulmus americana 12.50 3.5 0.9 4.5 9.0 Acer rubrum 25.00 4.1 0.9 6.3 11.3 Ilex opaca 6.25 1.0 1.1 3.1 5.3 Lyonia ferruginea 12.50 2.1 0.4 3.1 5.6 Lyonia ligustrina var. foliosiflora 68.75 11.3 3.8 9.4 24.5 Magnolia virginiana 25.00 4.1 0.9 6.3 11.2 Nyssa biflora 43.75 7.2 2.5 6.3 15.9 Persea hamulis 18.75 3.1 3.0 6.3 12.4 Pinus elliottii 50.00 8.2 45.9 9.4 63.5 Quercus hemisphaerica 37.50 6.2 5.1 12.5 23.8 Quercus incana 18.75 3.1 1.0 3.1 7.2 Quercus laurifolia 6.25 1.0 0.3 3.1 4.5 Quercus nigra 206.25 34.0 29.9 12.5 76.4 Quercus stellata 62.50 10.3 4.5 9.4 24.2 Jennings S. F. Vaccinium arboreum 25.00 4.1 0.7 9.4 14.2 Acer rubrum 50.00 8.3 3.3 7.1 18.8 Ilex opaca 25.00 4.2 0.8 7.1 12.1 Liquidambar styraciflua 37.50 6.3 1.4 7.1 14.8 Nyssa biflora 6.25 1.0 0.3 3.6 4.9 Hardwood, cont. Stein-hatchee, SRWMD Pinus elliottii 125.00 20.8 39.8 14.3 74.9

PAGE 102

89 Table D-1. Continued Quercus hemisphaerica 131.25 21.9 16.7 14.3 52.9 Quercus nigra 93.75 15.6 13.7 14.3 43.6 Quercus virginiana 43.75 7.3 17.9 10.7 35.9 Taxodium distichum 43.75 7.3 4.2 7.1 18.7 Ulmus americana 37.50 6.3 1.8 10.7 18.7 Stein-hatchee, SRWMD, cont. Viburnum rufidulum 6.25 1.0 0.1 3.6 4.7 Acer rubrum 275.00 40.4 14.4 12.1 66.9 Crataegus marshallii 6.25 0.9 0.1 3.0 4.1 Liquidambar styraciflua 68.75 10.1 4.4 12.1 26.6 Nyssa biflora 31.25 4.6 1.5 9.1 15.2 Pinus elliottii 18.75 2.8 15.1 6.1 23.9 Pinus palustris 6.25 0.9 2.4 3.0 6.3 Quercus hemisphaerica 6.25 0.9 0.8 3.0 4.8 Quercus laurifolia 68.75 10.1 10.5 12.1 32.8 Quercus nigra 87.50 12.8 13.7 12.1 38.7 Quercus virginiana 75.00 11.0 36.1 12.1 59.2 Vaccinium arboreum 18.75 2.8 0.4 6.1 9.2 Vaccinium stamineum 12.50 1.8 0.3 6.1 8.2 Hardwood, cont. Little River Springs, SRWMD Viburnum rufidulum 6.25 0.9 0.1 3.0 4.1

PAGE 103

APPENDIX E MOISTURE CONTENT OF BIOMASS COMPONENTS

PAGE 104

Table E-1. Moisture content (% dry weight) of fuel components. Standard error is given in parentheses (n=15 for species and n=90 for ecosystem for litter and live foliage; n=15 for species and n=75 for ecosystem for small stems; n for dead foliage, debris, and large stems varied because these samples were not available on all individuals). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison (for litter, live foliage, and small stems). Litter Fine Fuel Coarse Fuel Ecosystem Species litter live foliage dead foliage ** debris ** small stems large stems ** Flatwood Gaylussacia dumosa 53 (14) 124 (6) bcd na 21 (11) 171 (74) 136 (36) Ilex glabra 53 (15) 139 (10) bcd na 54 (31) 89 (9) 82 (2) Lyonia ferruginea 41 (12) 119 (12) cd na 47 (21) 93 (7) 88 (4) Vaccinium myrsinites 49 (11) 133 (10) bcd na 2 (11) 112 (27) na Myrica cerifera 57 (16) 141 (16) bcd 25 (na) 57 (37) 102 (12) 94 (6) Serenoa repens 60 (13) 101 (2) d 13 (5) 45 (10) na 179 (50) Flatwood mean 52 (5) 126 (4) 14 (4) 39 (9) 113 (16) 113 (13) Hardwood Callicarpa americana 71 (21) 460 (60) a na 84 (59) 193 (50) 189 (77) Ilex opaca 127 (22) 144 (12) bcd 35 (15) 72 (21) 109 (6) 88 (4) Quercus nigra 125 (25) 187 (25) bcd 5 (5) 162 (113) 102 (12) 89 (9) Vaccinium arboreum 132 (28) 223 (14) b na 58 (12) 88 (6) 91 (11) Myrica cerifera 96 (23) 216 (24) bc 100 (0) 76 (26) 145 (17) 119 (13) Serenoa repens 112 (26) 113 (5) d 30 (5) 141 (66) na 151 (16) Hardwood mean 111 (10) 224 (17) 33 (6) 98 (22) 127 (12) 119 (12) 91 indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems (for litter, live foliage, and small stems). ** Statistical analyses between species or ecosystem type were not possible for the moisture content of dead foliage, debris, or large stems.

PAGE 105

APPENDIX F BIOMASS COMPONENTS FOR LOWEST 1-METER INTERVAL

PAGE 106

Table F-1. Biomass separated into components for the lowest 1-m interval. Standard error is given in parentheses (n=5 for species and n=30 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukeys pairwise comparison. Meter 1 (lowest) Ecosystem Species live foliage dead foliage debris small stems large stems total Flatwood Gaylussacia dumosa 5.0 (1.3) d 0.0 (0.0) b 0.8 (0.4) c 8.6 (2.4) bc 0.4 (0.4) bc 14.8 (4.1) c Ilex glabra 3.6 (1.0) d 0.0 (0.0) b 1.5 (0.8) c 15.2 (2.8) bc 17.8 (4.3) bc 38.2 (5.9) c Lyonia ferruginea 7.7 (3.3) d 0.0 (0.0) b 1.7 (1.2) c 12.8 (3.1) bc 23.9 (11.8) bc 46.0 (10.4) c Vaccinium myrsinites 3.1 (0.5) d 0.0 (0.0) b 1.9 (0.7) c 10.1 (3.0) bc 0.0 (0.0) c 15.1 (3.7) c Myrica cerifera 19.4 (14.0) cd 0.0 (0.0) b 3.4 (1.2) c 33.9 (18.8) abc 50.3 (16.8) bc 107.1 (41.0) c Serenoa repens 132.6 (23.5) b 220.4 (65.0) a 29.9 (7.8) a 0.0 (0.0) c 359.6 (119.3) a 742.5 (170.6) a Flatwood mean 29.4 (10.0) 38.0 (18.8) 6.7 (2.4) 13.6 (3.7) 77.9 (31.0) 165.6 (57.2) Hardwood Callicarpa americana 4.0 (1.0) d 0.0 (0.0) b 0.1 (0.1) c 9.2 (1.1) bc 2.1 (1.6) c 15.4 (3.1) c Ilex opaca 93.7 (33.9) bc 0.0 (0.0) b 5.3 (2.4) bc 87.2 (23.4) a 182.4 (63.6) abc 368.6 (121.0) bc Quercus nigra 10.9 (2.4) d 0.04 (0.04) b 1.5 (1.2) c 22.5 (3.9) bc 75.9 (30.3) bc 109.8 (36.5) c Vaccinium arboreum 42.4 (15.3) cd 0.0 (0.0) b 3.3 (1.5) c 63.4 (22.4) ab 181.8 (62.2) abc 290.9 (94.9) bc Myrica cerifera 21.6 (8.4) cd 0.0 (0.0) b 1.5 (1.0) c 36.4 (12.3) abc 49.4 (21.7) bc 108.9 (42.3) c Serenoa repens 240.8 (29.5)a 90.1 (21.2) b 19.8 (7.0) ab 0.0 (0.0) c 243.0 (64.0) ab 593.8 (106.6) ab Hardwoodmean 68.9 (17.0) 15.0 (7.0) 5.3 (1.7) 36.4 (7.8) 122.3 (23.7) 247.9 (46.8) 93 indicates significant (p<0.05) difference in Tukeys pairwise comparison between ecosystems

PAGE 107

APPENDIX G EPA VOLATILE SOLID ANALYSIS METHODOLOGY METHOD #: 160.4 Approved for NPDES (Issued 1971) TITLE: Residue, Volatile (Gravimetric, Ignition at 550C) ANALYTE: Residue ,Volatile INSTRUMENTATION: Muffle Furnace STORET No. Total 00505 Non-Filterable 00535 Filterable 00520 1.0 Scope and Application 1.1 This method determines the weight of solid material combustible at 550C. 1.2 The test is useful in obtaining a rough approximation of the amount of organic matter present in the solid fraction of sewage, activated sludge, industrial wastes, or bottom sediments. 2.0 Summary of Method 2.1 The residue obtained from the determination of total, filterable or non-filterable residue is ignited at 550C in a muffle furnace. The loss of weight on ignition is reported as mg/L volatile residue. 3.0 Comments 3.1 The test is subject to many errors due to loss of water of crystallization, loss of volatile organic matter prior to combustion, incomplete oxidation of certain complex organics, and decomposition of mineral salts during combustion. 3.2 The results should not be considered an accurate measure of organic carbon in the sample, but may be useful in the control of plant operations. 3.3 The principal source of error in the determination is failure to obtain a representative sample. 4.0 Sample Handling and Preservation 4.1 Preservation of the sample is not practical; analysis should begin as soon as possible. Refrigeration or icing to 4C, to minimize microbiological decomposition of solids is recommended. 5.0 Precision and Accuracy 5.1 A collaborative study involving three laboratories examining four samples by means of ten replicates showed a standard deviation of 11 mg/L at 170 mg/L volatile residue concentration. 6.0 Reference 6.1 The procedure to be used for this determination is found in: Standard Methods for the Examination of Water and Wastewater, 14th Edition, p 95, Method 208E, (1975). 94

PAGE 108

LIST OF REFERENCES Abrahamson, W. G., and Hartnett, D. C. 1990. Pine Flatwoods and Dry Prairies. Pages 103-149 in: R. L. Myers and J. J. Ewel, eds. Ecosystems of Florida. University of Central Florida Press, Orlando. Abt, R., D. Kelly, and M. Kuypers. 1987. The Florida Palm Coast Fire: an analysis of fire incidence and residence characteristics. Fire Tech. 23:230-252. Agee, J. K., C. S. Wright, N. Williamson, and M. H. Huff. 2002. Foliar moisture content of Pacific Northwest vegetation and its relation to wildland fire behavior. For. Ecol. and Manage. 167:57-66. Anderson, H. E. 1970. Forest fuel ignitibility. Fire Tech. 6:312-319. Bond, W. J. and J. J. Midgley. 1995. Kill thy neighbour: an individualistic argument for the evolution of flammability. Oikos 73:79-85. Bond, W. J. and B. W. van Wilgen. 1996. Fire and plants. Chapman & Hall, London. 263 pp. Brown, J. K. 1970. Ratios of surface area to volume for common fine fuels. For. Sci. 16:101-105. Bunting, S. C., H. A. Wright, and W. H. Wallace. 1983. Seasonal variation in the ignition time of redberry juniper in West Texas. J. Range Manag. 36:169-171. Burgan, R. E. and R. A. Susott. 1991. Influence of sample processing techniques and seasonal variation on quantities of volatile compounds of gallberry, saw-palmetto, and wax myrtle. Int. J. Wildland Fire 101:57-62. Cheo, P. C. and K. R. Montgomery. 1970. The study of fire-retardance in plants. Lasca Leaves Vol:52-56. Ching, F. F. T. and W. S. Stewart. 1962. Research with slow burning plants. J. For. Vol:796-798. Cordell, H. K. and E. A. Macie. 2002. Population and demographic trends. Pages 11-35 in Human Influences on Forest Ecosystems: The Southern wildland-urban interface assessment. U. S. Department of Agriculture, Forest Service. GTR-SRS-55. 160 p. 95

PAGE 109

96 Cordell, H. K. and C. Overdevest. 2001. Footprints on the land: implications of population and economic growth for this countrys natural lands. Pages 229-284 in Footprints on the land: as assessment of demographic trends and the future of natural resources in the United States. Sagamore Publishing, Champaign, IL. Davis, J. and J. Marker. 1987. The wildland/urban fire problem. Fire Command. 54:26-27. Davis, J. B. 1988. The wildland-urban interface: what it is, where it is and its fire management problems. Pages 160-165 in Proceedings symposium and workshop: protecting homes from wildfire in the interior West. US Dept. Agric. For. Serv. GTR.INT-251. De Witt, J. L. 2000. Analysis of the utility of wildfire home protection strategies in Central Florida: a final report submitted to the Interagency Fire Science Team. Dickinson, K. J. M. and J. B. Kirkpatrick. 1985. The flammability and energy content of some important plant species and fuel components in the forests of southeastern Tasmania. J. Biogeography 12:121-134. Dimitrakopoulos, A. P. 2001. Thermogravimetric analysis of Mediterranean plant species. J. Analytical and Appl. Pyrolysis 60:123-130. Eriksson, I., D. Teketay, and A. Granstrm. 2003. Response of plant communities to fire in an Acacia woodland and a dry Afromontane forest, southern Ethiopia. For. Ecol. and Manage. 177:39-50. Etlinger, M. G. 2000. Fire performance of landscape plants. MS Thesis, University of California, Berkeley, 117 pp. Florida Firewise Communities. 2000. Are you firewise Florida? FL Div. of Forestry, FL Dept. of Ag. and Consumer Services; Div. of Emergency Manage., FL Dept. of Community Affairs. 2 pp. Florida Natural Areas Inventory (FNAI). 1990. Guide to the Natural Communities of Florida. FL Natl. Areas Invent. and FL Dept. of Natl. Reso. Tallahassee, FL. Fonda, R. W. 2001. Burning characteristics of needles from eight pine species. For. Sci. 47:390-396. Foote, E. I. D., Martin, R. E., and Gilless, J. K. 1991. The defensible space factor study: A survey instrument for post-fire structure loss analysis. Pages 66-73 in Proceedings of the 11th Conference on Fire and Forest Meteorology. Missoula, Montana. Francis, J. K. 2000. Comparison of hurricane damage to several species of urban trees in San Juan, Puerto Rico. J. of Arboriculture. 26:189-197.

PAGE 110

97 Frommer, S. L. and D. R. Weise. 1995. The quest for all-purpose plants. USDA For. Serv. Gen. Tech. Rep. PSW-GTR-158. Gill, A. M., W. S. W. Trollope, and D. A. MacArthur. 1978. Role of moisture in the flammability of natural fuels in the laboratory. Aust. For. Res. 8:199-208. Hermansen L. A. and E. A. Macie. 2002. Introduction. Pages 1-7 in Human Influences on Forest Ecosystems: The Southern wildland-urban interface assessment. U. S. Department of Agriculture, Forest Service. GTR-SRS-55. 160 p. Hough, W. A. and F. A. Albini. 1978. Predicting fire behavior in palmetto-gallberry fuel complexes. USDA For. Serv. Res. Pap. SE-184, Southeast. For. Exp. Stn. Kuchler, A. W. 1964. Potential natural vegetation of the conterminous United States. Amer. Geog. Soc., New York, NY. 116 pp. Lambers, H., F. S. Chapin III, T. L. Pons. 1998. Plant physiological ecology. Springer-Verlag New York, Inc. New York, NY. 540 pp. Lippi, C. and M. Kuypers. 1998. Flagler horticulture: making your landscape more resistant to wildfires. Flagler County Extension, FL Cooperative Extension Service, IFAS, University of Florida. 6 pp. MacCubbin, T. and D. Mudge. 2002. Fire wise landscaping: making sensible choices. Orange County Extension, FL Cooperative Extension Service, IFAS, University of Florida. 2 pp. Mak, E. H. T. 1988. Measuring foliar flammability with the limiting oxygen index method. For. Sci. 34:523-529. Mallik, A. U. 1994. Autecological response of Kalmia angustifolia to forest types and disturbance regimes. For. Ecol. and Manage. 65:231-249. Martin, R. E., D. A. Gordon, M. A. Gutierrez. 1994. Assessing the flammability of domestic and wildland vegetation. Page 796 in 12th Conference on Fire and Forest Meteorology. Soc. of Amer. For., Bethesda, MD. Monroe, M. C. 2002. Fire. Pages 133-150 in Human Influences on Forest Ecosystems: The Southern wildland-urban interface assessment. U. S. Department of Agriculture, Forest Service. GTR-SRS-55. 160 p. Monroe, M. C. and A. J. Long. 2001. Landscaping in Florida with fire in mind. FL Cooperative Extension Service, IFAS, University of Florida. FOR 71. 2 pp. Montgomery, K. R. and P. C. Cheo. 1971. Effect of leaf thickness on ignitibility. For. Sci. 17:475-478.

PAGE 111

98 Mutch, R. W. and C. W. Philpot. 1970. Relation of silica content to flammability in grasses. For. Sci. 16:64-65. Myers, R. L. and Ewel, J. J. (eds.). 1990. Ecosystems of Florida. University of Central Florida Press, Orlando. 765 pp. Nex-Regueira, L., J. Rodrguez-An, J. Proupn-Castieiras, and A. Romero-Garcia. 2001. Energetic evaluation of biomass originating from forest waste by bomb calorimetry. J. Thermal Analysis and Calorimetry 66:281-292. Owens, M. K., Lin Chii-dean, C. A. Taylor, Jr., and S. G. Whisenant. 1998. Seasonal patterns of plant flammability and monoterpenoid content in Juniperus ashei. J. of Chem. Ecol. 24:2115-2129. Papi, C. and L. Trabaud. 1990. Structural characteristics of fuel components of five Mediterranean shrubs. For. Ecol. and Manage. 35:249-259. Philpot, C. W. 1970. Influence of mineral content on the pyrolysis of plant materials. For. Sci. 16:461-471. Platt, W. J., and Schwartz, M. W. 1990. Temperate Hardwood Forests. Pages 194-229 in: R. L. Myers and J. J. Ewel, eds. Ecosystems of Florida. University of Central Florida Press, Orlando. Possingham, H. P, H. N. Comins, and I. R. Noble. 1995. The fire and flammability niches in plant communities. J. theor. Biol. 174:97-108. Pyne, S. J., P. L. Andrews, and R. D. Laven. 1996. Introduction to wildland fire, 2nd Edition. John Wiley & Sons, Inc., New York, NY. 769 pp. Rieske, L. K., H. H. Housman, M. A. Arthur. 2002. Effects of prescribed fire on canopy foliar chemistry and suitability for an insect herbivore. For. Ecol. and Manage. 160:177-187. Rodrguez-An, J. A., F. Fraga-Lpez, J. Proupn-Castieiras, J. Palacios-Ledo, and L. Nez-Regueira. 1995. Calorific values and flammability for forest wastes during the seasons of the year. Bioresource Tech. 52:269-274. Rogers, J. M., R. A. Susott, and R. G. Kelsey. 1986. Chemical composition of forest fuels affecting their thermal behavior. Can. J. For. Res. 16:721-726. Rundel, P. W. 1981. Structural and chemical components of flammability. US Dept. Agric. For. Serv. Gen. Tech. Rep. WO: 26:183-207. Schroeder, R. A., R. E. Martin, and K. Buteau. 2001. Ember ignitability of Pinus radiata and Sequoia sempervirens litter: methodology and results. Pages 119-124 in Proceedings of the Californias 2001 Wildfire Conference: 10 years after the 1991 East Bay Hills Fire. Oakland, CA.

PAGE 112

99 Shafizadeh, F., P. P. S. Chin, and W. F. DeGroot. 1977. Effective heat content of green forest fuels. For. Sci. 23:81-89. Susott, R. A. 1982a. Characterization of the thermal properties of forest fuels by combustible gas analysis. For. Sci. 28:404-420. Susott, R. A. 1982b. Differential scanning calorimetry of forest fuels. For. Sci. 28:839-851. Swinford, R. M., Tokle, G. O., Bethea, J., and Erb, R. 1987. Protecting your home from wildfire. Accompanist to broadcast by National Fire Protection, USFS, Dept. of Interior, Nat. Assoc. of State Forest., US Fire Administration. Thankston, R. E., P. E. Hale, A. S. Johnson, and M. J. Harris. 1982. Chemical composition of mountain-laurel Kalmia leaves from burned and unburned sites. J. Wildl. Manage. 46:492-496. Tokle, G. O. 1987. The wildland/urban interface 2025. Pages 49-52 in Symposium on wildland fire 2000. USDA Forest Service. Gen. Tech. Rep. PSW-101. van Wilgen, B. W., K. B. Higgins, and D. U. Bellstedt. 1990. The role of vegetation structure and fuel chemistry in excluding fire from forest patches in the fire-prone fynbos shrublands of South Africa. J. Ecology 78:210-222. Walker, L. C. and B. P. Oswald. 1999. The southern forest: Geography, ecology, and silviculture. CRC Press, Boca Raton, FL. 332 pp. Wang, S. and J. B. Huffman. 1982. Effect of extractives on heat content of melaleuca and eucalyptus. Wood Sci. 15:33-38. White, R. H., D. R. Weise, and S. Frommer. 1996. Preliminary evaluation of the flammability of native and ornamental plants with the cone calorimeter. Pages 256-265 in: Proceedings of the International Conference on Fire Safety. Vol. 21. Millbrae, California, January 8-12. Williamson, N. M. and J. K. Agee. 2002. Heat content variation of interior Pacific Northwest conifer foliage. Intl. J. Wildland Fire 11:91-94. Wilson, A. A. G. and I. S. Ferguson. 1986. Predicting the probability of house survival during bushfires. J. of Env. Manage. 23:259-270.

PAGE 113

BIOGRAPHICAL SKETCH Anna L. Behm was raised in the fields and prairies of the Midwest United States. During her formal education in the ecology of horticultural and agricultural systems, she participated in several urban forestry courses. These courses changed her perspective on ecology and sparked intense curiosity in the urban environment. Upon graduating from Iowa State University in 2001 with a Bachelor of Science degree in environmental horticulture, she began a Master of Science degree program at the University of Florida focusing on the issue of fire in the wildland-urban interface. She looks forward to further involvement with the study and management of natural resources for human use and ecological integrity in urban environments. 100


Permanent Link: http://ufdc.ufl.edu/UFE0001167/00001

Material Information

Title: Flammability of native understory species in pine flatwood and hardwood hammock ecosystems
Physical Description: Mixed Material
Creator: Behm, Anna Lee ( Author, Primary )
Publication Date: 2003
Copyright Date: 2003

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001167:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001167/00001

Material Information

Title: Flammability of native understory species in pine flatwood and hardwood hammock ecosystems
Physical Description: Mixed Material
Creator: Behm, Anna Lee ( Author, Primary )
Publication Date: 2003
Copyright Date: 2003

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001167:00001


This item has the following downloads:


Full Text












FLAMMABILITY OF NATIVE UNDERSTORY SPECIES IN PINE FLATWOOD
AND HARDWOOD HAMMOCK ECOSYSTEMS
















By

ANNA LEE BEHM


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

UNIVERSITY OF FLORIDA


2003

































Copyright 2003

by

Anna Lee Behm

































This thesis is dedicated to my family- Jim, Kathy, and Lisa Behm.















ACKNOWLEDGMENTS

This research was funded by the Southern Center for Wildland-Urban Interface

Research and Information, Southern Research Station, USDA Forest Service under the

National Fire Plan grant "Assessing and Mitigating Wildfire Risk for Southern Wildland-

Urban Interface Landowners." I am grateful to all those responsible for giving me this

opportunity, especially Annie Hermansen, Ed Macie, Dr. Wayne Smith, and my advisor

Dr. Mary Duryea.

I am indebted to my committee, Dr. Mary Duryea, Dr. Alan Long, and Dr. Timothy

Martin for their guidance. Additional guidance for this project came from the members

of the Southern Wildland-Urban Interface Council (SWUIC). This study could not have

been possible without the field and laboratory support from the USDA Forest Service,

Florida Division of Forestry, Florida Suwannee River Water Management District, and

Advanced Environmental Labs; and the Department of Animal Science, Department of

Soil and Water Science, and Austin Cary Memorial Forest at the University of Florida.

Gratitude is extended especially for the personal assistance of Cotton Randall, Sarah

Bouchard, Dave Nolletti, and Al Boning. I am also grateful for all those who assisted me

in sample and data collection in the field including Michele Nisi, Daniel Merced,

Christian Quijano, Dr. Wayne Zipperer, Cotton Randall, Annie Hermansen, Dr. Mary

Duryea, Eric Holzmueller, Brian Becker, and Dr. Jim Behm. I appreciate assistance with

this research from the IFAS Department of Statistics and the University of Florida

Herbarium.









Sincere appreciation is extended to those who have gone out of their way to enrich

my research and academic experiences over the years (especially Dr. Mark Gleason and

Dr. Jan Thompson at Iowa State University, and Dr. Wayne Zipperer from the USDA

Forest Service). I thank my colleagues and friends at the School of Forest Resources and

Conservation for adding diversity of field and perspective to my education. I also

appreciate the attention and patience offered by the faculty and staff at the School of

Forest Resources and Conservation. Finally, I thank my dear friends Lisa Young, Brian

Amidon, Angela Sokolowski, and Michel Masozera as well as my entire family for their

support and encouragement felt from great distances.
















TABLE OF CONTENTS
page

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

LIST OF TABLES .................. .. ............... ......... ... .. .... ....... ... ..... ..... ix

LIST OF FIGURES ......... ....... .................... .......... ....... ............ xi

ABSTRACT ........ .............. ............. ........ .......... .......... xii

CHAPTER

1 IN TR O D U C T IO N ............................................................. .. ......... ...... .....

B background Inform ation ...................................................... ............................... .
P ro b lem ................................................ 6
Ju stification ............................................................................................. ............ 7
O bjectiv e s ..................................................................................................... . 8
A associated H ypotheses .............................................................. ............... 8
A ssessm en t ....................................................... 9
Plant Flammability............................ .........9
E co sy stem C o n tex t ................................................................................................ 10
F lo rid a E co sy stem s ......................................................................................... 10
Pine Flatw oods ............................................... .............. ...... ...............11
H ardw ood H am m ock ...................................................................................... 11
T hesis O overview ................................. ........................... ....... 14

2 PLANT STRUCTURAL FLAMMABILITY IN PINE FLATWOOD AND
HARDWOOD HAMMOCK ECOSYSTEMS ........................................ ....16

In tro d u ctio n ........................................................................................16
M materials an d M eth o d s .......................................................................................... 18
Selection of Sites and Plants ................................. .................... .......19
Litter M easurem ents .............................................................................20
H eight M easu rem ents ..................................................................................... 2 1
B iom ass M easurem ents .................................................................................. 2 1
M oistu re C content ............................................................................. 2 2
Statistical A nalyses.................................................. 23
R e su lts ...........................................................................................2 5
S ite C h aracterization ...................................................................................... 2 5
Ecosystem D differences ............................................................. 26









Litter m easurem ents ............................................................................. 26
H eight m easurem ents ...........................................................................29
B iom ass m easurem ents ........................................... ........................... 30
Species D differences ................................................. .. .............. ........ ..34
L itter m easurem ents .......................... ...... .............. ............ .. ........ .... 35
H eight m easurem ents ...........................................................................35
B iom ass m easurem ents ........................................... ........................... 36
D isc u ssio n ...............................................................................................3 8
E cosystem D differences ............................................... ............................... 39
Species D differences ................................................. .. .............. ........ ..42
G aylussacia dum osa ................................................................. ... ............ 43
Ilex g la bra ..............................................................4 4
Lyoniaferruginea ............................................................. ................. 44
Vaccinium myrsinites ............................................................................44
Callicarpa am ericana ............................................................................ 44
Ilex opaca ........................................................................ ......... .................. 45
Q uercus nigra ........ .................................................. .. ............ .. .. 45
Vaccinium arboreum ......... ......... ............ ........................... 45
M yrica cerifera .............................................................. .. .. ........ .... 45
Serenoa rep ens ........................ .. ........................ .. ....... .... ............4 5
C o n c lu sio n s............................................................................................................ 4 6

3 FOLIAR FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD
HAMMOCK ECOSYSTEMS ............................................................................48

Introdu action ....................................................................................................4 8
M materials and M methods ........................................................................ ..................50
M moisture Content .......... .. .................... ........ .......... .............. 50
Leaf Area ................ ........ ....... ....... ...............50
V o latile S o lid s ....................................................................... 5 1
Energy Content ................ ... ......... ..............51
Statistical A n aly ses........... .......................................... ................ ... .... ... .. .. 53
Results .............................................. 54
E cosystem D differences ............................................... ............................... 54
Leaf area ........................................................................... ........................... 54
M oistu re content............ ....................................................... .. .... ...... .. 55
V olatile solids................................................... 55
Energy content................................................. 55
Species D differences ................................................. .. .............. .............56
Leaf area ........................................................................... ........................... 56
M oistu re content............ ....................................................... .. .... ...... .. 56
V olatile solid s ....................................................57
E n ergy content........... ........................................ .................. .. .... .. .... 57
S eren o a rep en s ........................................................................................ 6 1
D isc u ssio n .............................................................................................................. 6 3
E cosystem D differences ............................................... ............................... 63
Species D differences ................................................. .. .............. .............65









G aylussacia dum osa ............................................................... ............... 66
Ilex g la bra ..............................................................6 6
Lyonia ferruginea ........................................................ .... .. .... ........ 67
Vaccinium myrsinites ............................................................................67
Callicarpa am ericana ............................................................................ 68
Ilex opaca ........................................................................ ......... .................. 68
Q uercus nigra ............................. .. .. ................. ............... .. 68
Vaccinium arboreum ............................ ......... ....................... 69
M yrica cerifera .................. .......................... .... .... ................. 69
Serenoa repens .......................... ........... ........ ............ 70
C o n c lu sio n s........................................................................................................... 7 0

4 CONCLUSIONS AND RECOMMENDATIONS ............................................... 72

C o n c lu sio n s ........................................................................................................... 7 2
Recommendations.................. ......... .......................... 74
P in e F latw ood s ..............................................................74
H ardw ood H am m ocks ............................................................................ 75
Future R research ................................................... .. .......... .............. .. 75

APPENDIX

A LOCATION OF PINE FLATWOOD AND HARDWOOD HAMMOCK
ECOSYSTEMS BY COUNTY IN FLORIDA................................ ...............77

B LO CA TION OF STU D Y SITES.................................................................... ...... 79

C ABSOLUTE AND RELATIVE DENSITIES OF UNDERSTORY AND
MIDSTORY SPECIES AT EACH SITE ....................................... ............... 80

D ABSOLUTE AND RELATIVE DENSITY, RELATIVE DOMINANCE,
RELATIVE FREQUENCY, AND IMPORTANCE VALUES FOR
O V E R ST O R Y SP E C IE S ................................................................. .....................86

E MOISTURE CONTENT OF BIOMASS COMPONENTS ......................................90

F BIOMASS COMPONENTS FOR LOWEST 1-METER INTERVAL....................92

G EPA VOLATILE SOLID ANALYSIS METHODOLOGY ....................................94

L IST O F R E F E R E N C E S ......... ................. ...................................................................95

BIOGRAPHICAL SKETCH ...... ........ ................... ............................ 100
















LIST OF TABLES


Table p

1-1 Common and scientific names of plants listed as appropriate for firewise
landscaping according to two extension publications in Florida. ...........................4

1-2 Species common to hardwood hammock ecosystems across a moisture gradiant... 12

2-1 C collection dates for study sites.......................................... ........................... 20

2-2 Percent of total understory stems characterized by the species studied .................25

2-3 Understory, midstory, and overstory stems per hectare at each site........................27

2-4 Basal area, height to lowest branch, and canopy closure at each site ....................28

2-5 Litter depth and density for each species and ecosystem type................................29

2-6 Dry weight of fine fuel biomass components for each species and
ecosystem type. .......................................................................31

3-1 Leaf area per leaf, specific leaf area, and leaf area per plant for species in
flatwood and hardwood ecosystems............................................... ................... 58

3-2 Foliar moisture content and volatile solids for species within flatwood and
hardw ood ecosystem s......... .......................................................... ..... .... .....59

3-3 Additional data for Serenoa repens included measurements of moisture content,
volatile solid content, total energy content for dead foliage in addition to live
fo lia g e ...................................... .................................................... 6 2

4-1 Summary table comparing the results from flammability studies in Tasmania,
South A frica, and E thiopia. ........................................ ........................................73

4-2 Proposed ranking mechanism of plant species by flammability............................76

C-1 Absolute (stems per hectare) and relative densities (%) of understory species
in 8, 12.56m 2 plots at each study site. .......................................... ............... 81

C-2 Absolute (stems per hectare) and relative density (%) for midstory species
(>3 m in height but <6.4 cm dbh) in 4, 400m2 plots at each study site ..................84









D-l Stems per hectare, relative density, relative dominance, relative frequency,
and importance value for overstory species within each site................................87

E-l Moisture content (% dry weight) of fuel components............................................91

F-l Biomass separated into components for the lowest 1-m interval.............................93
















LIST OF FIGURES


Figure pge

1-1 Projected population pressures on forests through 2020........................................2

1-2 Lean, clean, and green defensible space zone surrounding a WUI structure in
Florida after a w ildfire............ .......................................................... ................ .4

1-3 Decision criteria for exclusion of garden or landscape vegetation as non-
hazardous for vegetation clearance measurements in post-fire assessments ............8

1-4 Pine flatwood ecosystem with routine prescribed fire. .......................................15

1-5 Typical hardwood hammock ecosystem. ..................................... ............... 15

2-1 Total biomass by coarse fuel and fine fuel components for species within pine
flatwoods and hardwood hammocks......................................................................32

2-2 Fuel loading for each species and ecosystem type ................................................33

2-3 Biomass per 1-meter height intervals by ecosystem type. ............. ..................34

2-4 Height and height to lowest branch of species within pine flatwoods and
hardw ood ham m ock ecosystem s ........................................ ......................... 37

2-5 Biomass per 1-meter height intervals for species within pine flatwoods and
hardw ood ham m ocks ..................................................................... .....................39

2-6 Pine needles accumulated on a landscape plant. ............................................. 41

3-1 Foliar energy content in calories per gram, kilocalories per plant, and
kilocalories per hectare for species within pine flatwood and hardwood
ham m ock ecosystem s. ...................................................................... ..................60

A-i Distribution of pine flatwoods by county in Florida............... ................ ..........78

A-2 Distribution of hardwood hammocks by county in Florida. ...................................78

B-l Location of pine flatwood and hardwood hammock study sites throughout
N north C central F lorida. ..................................................................... ...................79















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

FLAMMABILITY OF NATIVE UNDERSTORY SPECIES IN PINE FLATWOOD
AND HARDWOOD HAMMOCK ECOSYSTEMS

By

Anna Lee Behm

August 2003

Chair: Mary L. Duryea
Major Department: School of Forest Resources and Conservation

The flammability of plants contributes to fire behavior and fire regimes in natural

ecosystems. As urban development continues within and near fire-dependent ecosystems

in the southern United States, the flammability of plants in those ecosystems may

influence the survival of human-built structures during wildfire. To assess the

importance of plant flammability in the wildland-urban interface (WUI), we compared

flammability components-ignitability, sustainability, combustibility and

consumability-of native understory species commonly found in pine flatwood and

hardwood hammock ecosystems in the southern United States.

During the summer of 2002, six species from five pine flatwood sites-dwarf

huckleberry (Gaylussacia dumosa [Andr.] A. Gray), gallery (Ilex glabra [L.] A. Gray),

rusty lyonia (Lyoniaferruginea [Walt.] Nutt.), evergreen blueberry (Vaccinium

myrsinites Lam.), wax myrtle (Myrica cerifera L.), and saw palmetto (Serenoa repens

[Bartr.] Small)-and six species from five hardwood hammock sites-American









beautyberry (Callicarpa americana L.), American holly (Ilex opaca Ait. var opaca),

water oak (Quercus nigra L.), sparkleberry (Vaccinium arboreum Marsh.), wax myrtle

and saw palmetto-were harvested for biomass analyses. Plant components were

separated into live and dead foliage, accumulated litter on and under the plant, and small

(<0.6 cm diameter) and large (>0.6 cm diameter) twigs, branches, and stems. Foliar

biomass was further analyzed for leaf area, volatile solids, and energy content.

Statistical analyses revealed differences between ecosystem types and among

species. Understory plants in pine flatwoods have higher ignitability (lower moisture

content), sustainability (higher fuel loading), and combustability (higher energy content),

whereas the measurements for consumability was not different between ecosystem (fine

fuel biomass and volatile solid content). Understory species in hardwood hammocks

contain more total biomass because they contain more coarse fuel; yet coarse fuels do not

affect the flammability of species as live coarse fuels are not typically consumed in a

wildfire.

Serenoa repens, a species common to both ecosystems, had relatively low foliar

energy content per gram, but individual plants present a high hazard because they contain

a great amount of biomass. Ilex glabra is also hazardous to WUI structures because it

has high foliar energy content and a great amount of foliar biomass per hectare in

flatwoods. Callicarpa americana plants present the lowest fire hazard to WUI structures

compared to all other species studied. Results will assist the development of regionally

specific guidelines for firewise landscaping and guide future plant flammability research.














CHAPTER 1
INTRODUCTION

Background Information

The wildland-urban interface (WUI) is a growing part of the landscape in the

southern United States presenting new challenges to natural resource agencies, urban and

rural fire agencies, city and county governments, and insurance companies. The

wildland-urban interface can be defined from many perspectives: geographical, natural

resource, sociopolitical, biophysical, and fire (Hermansen and Macie 2002).

Geographically defined, the WUI is separated into three categories; classic, mixed, and

isolated (Davis 1988). Classic interface areas occur where existing urban centers expand

and invade wildlands. Mixed interface areas are formed as urban areas are built within

and surrounded by wildland areas. Isolated interface areas are isolated islands of

wildland vegetation are completely surrounded by urban areas which are sometimes used

as city parks. From a natural resource perspective, the WUI is an area where increased

human influence and land use conversion are changing natural resource goods, services,

and management (Hermansen and Macie 2002).

The WUI is not a new feature in the landscape of the southern United States. It has

existed for as long as human habitation. Native Americans and early European settlers,

living close to surrounding ecosystems, routinely cleared vegetation around homes

through physical removal, routine prescribed fire, and livestock grazing (Pyne et al.

1996). However, the WUI is expanding at increasing rates in the southern US. During

the 1980s, the rate of rural population growth was 3.1% in the South (Cordell and Macie










2002). This rate increased to 7.5% in the South during the 1990s (Cordell and Macie

2002). Based on data from Cordell and Overdevest (2001), the population pressures on

forests through year 2020 are projected in Figure 1-1. This shows areas in the South (by

county) that are predicted to experience population growth in forested areas.

















'' 1 -
N ligiibl.
Ligh[ t
Moderate
Moderately heavy
Heavy



Figure 1-1. Projected population pressures on forests through 2020. Darker colors
indicate where population pressures on forests are expected to be the
heaviest. From Figure 2.10 on page 27 in Cordell, H. K. and E. A. Macie.
2002. Population and demographic trends. Pages 11-35 in Human
Influences on Forest Ecosystems: The Southern wildland-urban interface
assessment. U. S. Department of Agriculture, Forest Service. GTR-SRS-55.
160 p.

In the South, the forests and many other ecosystems associated with the WUI are

dependent on fire for periodic renewal and to sustain species composition (Myers and

Ewel 1990, Pyne et al. 1996, Walker and Oswald 1999). The warm and humid climate in

the South supports rapid growth of understory plants; and frequent lightning storms

provide a consistent source of ignition (Myers and Ewel 1990, Pyne et al. 1996, Walker









and Oswald 1999). Wildfire, because of the direct effect on human health and safety, is

one of the most recognized challenges within the WUI. Fire agencies define the WUI as

an area where residential or commercial development is in or adjacent to areas prone to

wildfire (Hermansen and Macie 2002 from Davis and Marker 1987, Tokle 1987).

As population in the wildland-urban interface increases, the economic and social

consequences of wildfire increase. In addition, even with advanced firefighting training

and technology, firefighting agencies struggle to address complexities of fire in the

wildland-urban interface (Swinford et al. 1987). The complexities are exacerbated when

fire is unnaturally kept out of the interface, further increasing homeowners' risk of

wildfire damage.

The scope of the WUI wildfire problem has prompted many fire and wildfire

management agencies to initiate cooperative extension programs for communities and

individual homeowners (Monroe 2002). In addition, governments have initiated policies

to modify the behavior of developers, communities, and homeowners to alleviate the

social burden of wildfire in the WUI (Monroe 2002). Many of these programs, often

termed "firewise," are focused on increasing the responsibility of the private sector and

citizens in wildfire preparation. Recommendations include guidelines for entry and

access, building materials, and landscaping.

Firewise landscaping typically involves creating defensible space zones around

homes at risk of wildfire. Defensible space is an area surrounding a home to allow easy

access for firefighting equipment and personnel. Defensible space is also created to

reduce the risk of wildfire damage if firefighting agencies are unable to defend each

home. Typically, a defensible space zone of at least 30 feet (9 meters) surrounding a









WUI structure should be established (Florida Firewise Communities 2000, Monroe and

Long 2001). Within this zone, landowners are instructed to keep the area "lean, clean,

and green" (Florida Firewise Communities 2000) (Figure 1-2). This includes pruning

trees so that the lowest branches are at least 10 feet (3 meters) from the ground, removing

vines and shrubs from under trees, clearing yard waste and firewood, and planting less

flammable plant material in isolated landscape beds (Monroe and Long 2001).




















Figure 1-2. Lean, clean, and green defensible space zone surrounding a WUI structure in
Florida after a wildfire. Photograph by Cotton Randall.

Lists of plant species appropriate for firewise landscaping include native and

ornamental species (Table 1-1).

Table 1-1. Common and scientific names of plants listed as appropriate for firewise
landscaping according to two extension publications in Florida.
Common name Scientific name Monroe and MacCubbin
Long 2001 and Mudge
2002
African iris Dietes iridioides [L.] N.E. Br. X
Banana Musa spp. X
Beautyberry Callicarpa americana L. X
Bird of paradise Strelitzia reginae Banks X
Black cherry Prunus serotina Ehrh. X
Blue beech Carpinus caroliniana Walt. X









Table 1-1. Continued
Common name


Bottlebrush
Bugleweed
Century plant
Citrus
Coontie
Daylily
Dogwood
Ferns
Florida soapberry
Fringetree
Hophornbeam

Indian hawthorn
Lantana
Ligustrum
Lily of Nile


Liriope

Loquat
Magnolia
Maples
Oaks
Oleander
Peach
Periwinkle
Persimmon
Philodendron
Pineapple guava
Pittosporum
Plum
Pomegranate
Pyrancantha

Red maple
Redbud
Sago
Society garlic
Sparkleberry
Star jasmine


Scientific name


Monroe and
Long 2001


Callistemon spp.
Ajuga reptans L.
Agave decipiens Baker
Citrus spp.
Zamia pumila L.
Hemerocallis spp.
Cornus spp.
multiple genus
Sapindus marginatus Willd.
Chionanthus virginicus L.
Ostrya virginiana [Mill.] K.
Koch
Rhaphiolepis indica [L.] Lindl.
Lantana camera L.
Ligustrum spp.
Agapanthus praecox Willd.
Subsp. Orientalis [F. M.
Leight.]
Liriope muscari [Decne.] L.H.
Bail.
Eriobotrya spp.
Magnolia spp.
Acer spp.
Quercus spp.
Nerium oleander L.
Prunuspersica [L.] Batsch.
Vinca major L.
Diospyros virginiana L.
Philodendron spp.
Acca sellowiana [0. Berg] Berret
Pittosporum tobira [Thunb.] Ait.
Prunus spp.
Punica granatum L.
Pyrancantha coccinea M.J.
Roem
Acer rubrum L.
Cercis evolute Os L.
Cycas 0evolute Thunb.
Tulbaghia violacea L.
Vaccinium arboreum Marsh.
Jasminum multiflorum [Burm.f.]
Andr.


MacCubbin
and Mudge
2002
X
x
X
X
X









Table 1-1. Continued
Common name Scientific name Monroe and MacCubbin
Long 2001 and Mudge
2002
Sugarberry Celtis laevigata Willd. X
Sweetgum Liquidambar styraciflua L. X
Sycamore Platanus occidentalis L. X
Viburnum Viburnum spp. X X
Wild azalea Rhododendron canescens X X
[Michx.] Sweet
Wild olive Osmanthus americanus [L.] A. X
Gray
Wild plum Prunus umbellata Ell. X
Winged elm Ulmus alata Michx. X

In addition, highly flammable species such as saw palmetto, wax myrtle, yaupon

holly (Ilex vomitoria Ait.), red cedar (Juniperus virginiana L.), gallberry, juniper

(Juniperus spp.), pampas grass (Cortaderia selloana Schult. & Schult.f.), arborvitae

(Platycladus orientalis L.), American holly, Italian cypress (Cupressus sempervirens L.),

eucalyptus (Eucalyptus spp.), pine (Pinus spp.), Leyland cypress (Cupressocyparis

leylandii [A.B. Jacks &Dallim.] Dallim.& AB. Jacks.), and fountain grass (Pennistetum

spp.) are discouraged from being planted within firewise landscaping (Monroe and Long

2001, MacCubbin and Mudge 2002).

Problem

Landowners in the wildland-urban interface are instructed to reduce the number of

flammable plants on their property and to plant species that are less flammable (Lippi and

Kuypers 1998, Florida Firewise Communities 2000, Monroe and Long 2001). However,

the species lists given to homeowners frequently have an unknown origin (Frommer and

Weise 1995). In many cases, species lists are generated from other lists originating in

different regions in the US (Lippi and Kuypers 1998). As firewise landscaping

recommendations develop, there is a greater need for more regionally specific and









scientifically founded species lists. Many characteristics are known to affect

flammability making it difficult to generalize flammability based on a few characteristics.

To complicate the situation further, no standard method of quantifying plant flammability

has been fully accepted making comparing results from plant flammability studies

difficult.

Justification

In post-fire assessments of structural survival, brush or shrub clearance >30 feet

from a structure has shown to increase structural survival in wildfires in the southern

United States (Abt et al. 1987, De Witt 2000). Appropriate flammability rankings for

species within southern US ecosystems would be valuable to landowners, landscape

architects, and nurseries in the wildland-urban interface. Given the definition of the

WUI, the landscaping around interface homes includes many species native to the

surrounding ecosystem. By comparing the flammability of understory species within two

ecosystems, more information will be available to address landscaping hazards within

these ecosystems. As firewise landscaping policies become more regionally specific,

information relevant to specific ecosystem types will be desirable.

In addition, knowledge of flammability characteristics of specific plants may

support future post-fire research. Flammability ratings for species would be valuable to

future post-fire analysis of structures. Foote et al. (1991) suggest a dichotomous

approach to determining which vegetation should be measured to assess vegetation

clearance for a structure threatened or damaged after a wildfire (Figure 1-3).









Is vegetation...


Figure 1-3. Decision criteria for exclusion of garden or landscape vegetation as non-
hazardous for vegetation clearance measurements in post-fire assessments.
(Foote et al. 1991)

Objectives

The purpose of this study was to examine the characteristics that are known to

influence flammability and to compare the flammability of understory species native to

pine flatwood and hardwood hammock ecosystems.

The specific objectives of the study were to determine the flammability in the

context of structural and foliar characteristics. Structural components known to influence

flammability at the plant level were quantified including fine fuel biomass and

arrangement of biomass throughout the plant. Foliage from understory species was

analyzed for moisture content, volatile solid content, and energy content.

Associated Hypotheses

This study had three major hypotheses:

* Hypothesis I: Understory plants in pine flatwoods are more flammable than
understory plants in hardwood hammocks.
* Hypothesis II: Differences in the flammability of species will be significant.
* Hypothesis III: The flammability of wax myrtle and saw palmetto will be different
between ecosystems.









Assessment

Beyond the above hypotheses, we interpreted the results of this study in the context

of firewise landscaping recommendations. To do this, an assessment of the results

centered around two questions:

* Assessment I: What are the implications for firewise landscaping in these
ecosystems?
* Assessment II: In what way would individual plants contribute to structure
survival or destruction in a wildfire?

Plant Flammability

Flammability was initially defined by Anderson in three components: ignitability,

sustainability, and combustibility (1970). The ignitability component is the time until

ignition once exposed to a heat source. Sustainability is the stability of burning rate, or

the ability to sustain fire once ignited. Combustibility is defined as the rate of burn after

ignition. The definition offlammability has since been expanded to include

consumability, the proportion of mass or volume consumed by fire (Martin et al. 1994).

Anderson (1970) related the flammability components of individual plants to fire

characteristics at an ecosystem level. Ignitability of individual plants drives the chain of

ignition in an ecosystem. Sustainability is related to the rate of fire spread and

combustibility to fire intensity. The consumability of individual plants is analogous with

the amount of fuel available for fire consumption on the ecosystem level (Martin et al.

1994). In this way, plant flammability is an important component of the natural fire

regimes of ecosystems (Bond and van Wilgen 1996). In addition, plant flammability can

influence wildfire behavior affecting the survivability of human-built structures.









Ecosystem Context

This study assesses flammability in the context of two ecosystems of Florida which

are ecosystems in the coastal plain physiographic region of the southern United States.

The natural role of fire in pine flatwoods and hardwood hammock ecosystems is

discussed in this section.

Florida Ecosystems

Residential landscaping, like many characteristics of the interface, is a mix of both

natural and synthetic landscapes. The natural ecosystem plays a large role in the design

of landscapes around structures. Kuchler (1964) described eleven potential vegetation

states for Florida in his categorization of vegetation types for the United States. The

Florida Natural Areas Inventory (FNAI) in partnership with the Florida Department of

Natural Resources and the Florida Nature Conservancy, have defined over sixty

ecosystem types in Florida (1990). Myers and Ewel (1990) separate the upland,

freshwater wetlands and aquatic, and coastal ecosystems into 13 Florida ecosystem types.

Fire plays an important role in the function of almost every terrestrial ecosystem in

Florida, and all terrestrial ecosystems will burn under the right conditions. Pine

flatwoods and hardwood hammocks are named according to Myers and Ewel (1990), but

the corresponding classification for FNAI (1990) and Kuchler (1964) are also given in

the descriptions below. These two ecosystems, along with all others, are being affected

by urbanization and are important in terms of the wildland-urban interface areas of

Florida. See Appendix A for maps by county of each ecosystem type in Florida, adapted

from Myers and Ewel (1990).









Pine Flatwoods

The Florida pine flatwoods (Abrahamson and Hartnett 1990) correspond to mesic

flatwoods, scrubby flatwoods, wet flatwoods (FNAI 1990), and southern mixed forest

(Kuchler 1964). Pine flatwoods are dominated by longleaf pine (Pinuspalustris Mill.),

slash pine (P. elliottii Engelm.), loblolly pine (P. taeda L.), south Florida slash pine (P.

elliottii var. densa Little & Dorman), and pond pine (P. serotina Michx.) (Figure 1-4). In

central and north Florida, live oak (Quercus virginiana Mill.), water oak, sweetgum, red

maple, and ash (Fraxinus spp.) may add to the overstory composition. Understory

composition consists of saw palmetto, gallberry, staggerbush or fetterbush (Lyonia spp.),

dwarf huckleberry, wax myrtle, and tar flower (Befaria racemosa Vent.). Wiregrass

(Astrida spp., Sporobolus spp.) is the dominant herbaceous ground cover in longleaf pine

flatwoods. (Abrahamson and Hartnett 1990)

The pine flatwood ecosystems are highly dependent on fire. Because they have a

relatively open canopy, understory vegetation biomass can accumulate very rapidly. A

fire frequency of every one to eight years is typical (FNAI 1990). When these areas burn,

large areas of understory burn completely and because of the natural shedding of lower

limbs by natural pine species, crowning of fires does not typically occur (Abrahamson

and Hartnett 1990).

Hardwood Hammock

The temperate hardwood forests in Florida are referred to as hammocks (Platt and

Schwartz 1990) (Figure 1-5). The temperate hardwoods inhabit xeric, mesic, and hydric

moisture regimes with differing vegetation across this moisture gradient (Table 1-2).

Temperate hardwood forests in xeric and mesic environments fit into the Kuchler (1964)

classification of southern mixed forests. Those in hydric environments are classified as









southern floodplain forest (Kuchler 1964). FNAI (1990) classifies temperate hardwood

forests by moisture regime including xeric hammock, upland glade, upland hardwood

forest, upland mixed forest, slope forest, and hydric hammock.

Table 1-2. Species common to hardwood hammock ecosystems across a moisture
gradiant (Platt and Schwartz 1990, FNAI 1990).


Hardwood Hammock
Xeric Hammock





















Mesic Hammock


Common Name
American holly
black cherry
blackjack oak
Chapman oak
laurel oak
live oak
persimmon
pignut hickory
red bay
sand live oak
sand post oak
saw palmetto
southern magnolia
Southern red oak
sparkleberry
staggerbush
turkey oak
wild olive

yaupon
American holly
Carolina holly

devil's walking stick
eastern
hophornbeam
Florida maple
flowering dogwood
live oak
loblolly pine
pignut hickory
redbud
southern magnolia
spruce pine
swamp chestnut oak
sweetgum


Scientific Name
Ilex opaca Ait. var opaca
Prunus serotina Ehrh.
Q. incana Bartr.
Q. chapmanii Sarg.
Q. hemisphaerica Bartr. ex Willd.
Quercus virginiana Mill.
Diospyros virginiana L.
Carya glabra [Mill.] Sweet
Persea borbonia [L.] Spreng.
Q. geminata Small
Q. margaretta Ashe
Serenoa repens [Bartr.] Small
Magnolia grandiflora L.
Q. falcata Michx.
Vaccinium arboreum Marsh.
Lyonia spp.
Q. laevis Walt.
Osmanthus americanus [L.] Benth.
& Hook. F. ex A. Gray
Ilex vomitoria Ait.
Ilex opaca Ait. var opaca
Ilex ambigua [Michx.] Torr. var.
ambigua
Aralia spinosa L.
Ostrya virginiana [Mill.] K. Koch

Acer saccharum Marsh.
Cornusflorida L.
Q. virginiana Mill.
Pinus taeda L.
Carya glabra (Mill.) Sweet
Cercis canadensis L.
Magnolia grandiflora L.
Pinus glabra Walt.
Q. michauxii Nutt.
Liquidambar styraciflua L.









Table 1-2. Continued
Hardwood Hammock
Hydric Hammock


Common Name
bluestem palmetto
cabbage palm

dahoon holly
diamond-leaf oak
hackberry
loblolly pine
myrsine
needle palm

pepper vine
poison ivy

rattanvine
red cedar
red maple
royal fern
saw palmetto
southern magnolia
swamp bay
swamp chestnut oak
sweetbay
sweetgum
Virginia creeper

Walter viburnum
water oak
wax myrtle
fellow iessamine


Scientific Name
Sabal minor [Jacq.] Pers.
Sabalpalmetto [Walt.] Lodd. ex
J.S. Schult. & J.H. Schult.
Ilex cassine L.
Q. laurifolia Michx.
Celtis laevigata Willd.
Pinus taeda L.
Myrsinefloridana A. DC.
Rhapidophyllum hystrix [Pursh]
Wendle. & Drude
Ampelopsis arborea [L.] Koehne
Toxicodendron radicans [L.]
Kuntze.
Berchemia scandens [Hill] K. Koch
Juniperus virginiana L.
Acer rubrum L.
Osmunda regalis L.
Serenoa repens [Bartr.] Small
Magnolia grandiflora L.
Perseapalustris [Raf.] Sarg.
Q. michauxii Nutt.
Magnolia virginiana L.
Liquidambar styraciflua L.
Parthenocissus quinquefolia [L.]
Planch.
Viburnus obovatum Walt.
Q. nigra L.
Myrica cerifera L.
Gelsemium rankinii Small


Temperate hardwood forests are not fire dependent and fires usually originate

from surrounding ecosystems and rarely cover large areas (FNAI 1990). Fire frequencies

are more frequent for xeric hardwood forests (> 30 years) and less frequent for hydric

hardwood forests (very rarely) (FNAI 1990). Depending on the composition and

structure of fuels, fires in temperate hardwood forests have the potential to become


intense.









Thesis Overview

Chapter 2 explores the differences in structural flammability of understory species

native to pine flatwood and hardwood hammock ecosystems. Plants were separated into

live and dead foliage, accumulated litter on and under the plant, and small (<0.6 cm

diameter) and large (>0.6 cm diameter) twigs, branches, and stems. Results were

analyzed to determine the ecosystem and species differences in components relative to

each other and in the context of potential hazards to WUI structures.

Chapter 3 is a further analysis of foliar biomass characteristics. Leaf area, foliar

moisture content, foliar volatile solids, and foliar energy content were quantified for each

species within each ecosystem type. Results were analyzed, in conjunction with results

from Chapter 2, to describe the flammability of understory species within each ecosystem

type.

Chapter 4 provides an overview of the results from this study. In addition,

recommendations for firewise landscaping techniques within each ecosystem are

described with recommendations for further research.































Figure 1-4. Pine flatwood ecosystem with routine prescribed fire. Photograph by Larry
Korhnak.


Figure 1-5. Typical hardwood hammock ecosystem. Photograph by Larry Korhnak.














CHAPTER 2
PLANT STRUCTURAL FLAMMABILITY IN PINE FLATWOOD AND
HARDWOOD HAMMOCK ECOSYSTEMS

Introduction

Vegetation and litter provide the primary fuel source for wildfires which threaten

WUI homes in the forests of the southern US. Periodic wildfire occurs in most southern

forests removing accumulated biomass and litter, releasing nutrients, and facilitating

regeneration (Myers and Ewel 1990, Pyne et al. 1996, Walker and Oswald 1999). As

WUI areas expand within southern forests, challenges increase between the use of fire for

natural resource management, range management, or silvicultural practices (Monroe

2002). In an attempt to reduce the burden of wildfire protection, policies and

recommendations include the development of defensible space through "firewise"

landscaping (Monroe 2002). As the plants within home landscapes are potential fuel for

wildfires which may threaten human life and property, it has become important to

evaluate the hazard of landscape plants. Quantifying the flammability of plant species is

a way of evaluating the hazard of different plants which may be incorporated into

landscaping.

Flammability has been defined as having four components: ignitability,

sustainability, combustibility, and consumability (Anderson 1970, Martin et al. 1994).

Characteristics shown to influence flammability of plant material include moisture

content (Gill et al. 1978), percent cellulose, hemicellulose, and lignin (Philpot 1970,

Rundel 1981, Susott 1982a), volatile compounds (Shafizadeh et al. 1977, Susott 1982a,









Wang and Huffman 1982, van Wilgen et al. 1990, Owens et al. 1998), silica-free mineral

content (Mutch and Philpot 1970), leaf thickness (Montgomery and Cheo 1971), surface

area-to-volume ratio (Rundel 1981, Papi6 and Trabaud 1990), and particle density

(Brown 1970, Papi6 and Trabaud 1990). However, these characteristics have been

studied to different extents by different methods, are not equally important to plant

flammability, and are not all independent of one another (Shafizadeh et al. 1977, Etlinger

2000, Francis 2000).

Etlinger (2000) found that the relative abundance of biomass components

significantly affected the flammability of an entire plant, measured as heat release rate

(sustainability and combustability), more than all other characteristics measured. The

amount of dry mass consumed determined the total heat release while foliar biomass and

foliar moisture content determined the peak heat release rate (Etlinger 2000). Etlinger

(2000) also concluded that fine fuel biomass (foliage and small stems <0.6 cm) and fine

fuel moisture content contribute more to the peak heat release rate of plants than many

other characteristics.

The height of understory plants and the arrangement of fuel within the plant can

contribute to the intensity and height of surface fires (Pyne et al. 1996) and structural

survivability in the WUI (Wilson and Ferguson 1986). Sources for plant ignition may

also be different based on the arrangement of fuels. Fuel closer to the ground may be

more susceptible to ignition from ground and surface fires. Fuel further from the ground

may be more susceptible to ignition from firebrands.

The density of fuel, or fuel loading, can affect the sustainability of fire within the

plant; more dense fuels sustaining a more consistent fire (Rundel 1981). This is true until









the fuel becomes so dense that oxygen becomes limiting (Rundel 1981). For this reason,

litter with lower density may be more flammable than litter with higher density.

The objectives of this study were to determine the biomass components and

biomass arrangement of understory species within pine flatwood and hardwood hammock

ecosystems. Based on these results, we attempt to determine the differences in potential

structural flammability ofunderstory species between the two ecosystem types. We also

attempt to determine the relative hazards of species to WUI homes.

Materials and Methods

Species for this study were chosen based on their use in landscape plantings or their

abundance in the understory of the two ecosystems. Species studied within pine flatwood

ecosystems were dwarf huckleberry (Gaylussacia dumosa [Andr.] A. Gray), gallery

(Ilex glabra [L.] A. Gray), rusty lyonia (Lyoniaferruginea [Walt.] Nutt.), and evergreen

blueberry (Vaccinium myrsinites Lam.). American beautyberry (Callicarpa americana

L.), American holly (Ilex opaca Ait. var opaca), water oak (Quercus nigra L.), and

sparkleberry (Vaccinium arboreum Marsh.) were studied within hardwood ecosystems.

Wax myrtle (Myrica cerifera L.) and saw palmetto (Serenoa repens [Bartr.] Small) were

studied in both ecosystems.

For each plant collected, litter, height, and biomass measurements were made to

determine the structural flammability of species within each ecosystem type. Litter

underneath the plant was collected to determine litter depth and density. Next, plant

height, height to lowest branch, and width measurements were taken to determine overall

size of the plant. Plants were then cut at the base and separated into biomass components

to characterize the potential fuel for each species.









Selection of Sites and Plants

Five sites of each ecosystem type (pine flatwood and hardwood hammock) were

located throughout North Central Florida (Appendix B). Criteria for site selection were

presence of desired understory species in the desired ecosystem type. Fire had been

absent in all sites for at least five years. Study sites included property owned by the

USDA Forest Service (Osceola National Forest), Florida Division of Forestry (Jennings

State Forest, Twin Rivers State Forest, Withlacoochee State Forest, and Welaka State

Forest), Suwannee River Water Management District (Little River Springs and

Steinhatchee), and the University of Florida (Austin Cary Memorial Forest).

At each site, the vegetation was characterized by randomly selecting and measuring

four circular tree plots (400 m2) and eight circular shrub plots (12.56 m2). Within the tree

plots, diameter at breast height (dbh) and height to lowest branch was measured using a

hypsometer (Haglof, Vertex III) was measured for tree species (>3 m in height (Foote et

al. 1991) and >6.4 cm dbh). Species of midstory trees (>3 m in height but <6.4 cm dbh)

were recorded. Canopy closure was measured from the center of each tree plot by

averaging four readings from a concave spherical densiometer (Model-C, Forestry

Supply, Inc.). Within the understory plots, the total number of stems by species was

recorded and the height in 0.2 m increments from 0.4 m to 3 m for each individual.

At each study site, three plants of each species were randomly selected and sampled

between May and July 2002 (Table 2-1). Three transects were initiated at least 5 m away

from any road or trail, and one plant of each species was randomly selected. There was a

minimum of 5 m between flagged individuals of the same species to avoid microclimate

influence. Plants were considered appropriate for this study between 1 to 3 meters in

height, an interval identified by Foote et al. (1991). However, Gaylussacia dumosa and









Vaccinium myrsinites were accepted into the random sample if over the height of 0.6 m

and 0.4 m, respectively. These two species are common within pine flatwoods but do not

typically grow very tall.

A total of fifteen plants (three individuals at five sites) were sampled of

Gaylussacia dumosa, Ilex glabra, Lyoniaferruginea, Vaccinium myrsinites, Callicarpa

americana, Ilex opaca, Quercus nigra, and Vaccinium arboreum. A total of thirty plants

(three individuals at five sites in both ecosystems) were sampled ofMyrica cerifera and

Serenoa repens. In total, 180 plants were harvested from the sites (Eq. 2-1). Biomass

was measured on all individuals studied. Additional biomass data were taken on one

randomly selected individual plant at each site; 60 plants were harvested at 1-m height

intervals.

6 species 3 plants 5 sites 2 ecosystem types =180 plants (2-1)

Table 2-1. Collection dates for study sites.
Ecosystem Site Plant collection dates, 2002
Flatwood Austin Cary M. F. May 23, 24
Osceola N. F. July 1, 2, 3
Jennings S. F. July 19, 22
Welaka S.F. June 11, 14, 19, 20
Withlacoochee S. F. June 3, 4
Hardwood Twin Rivers S. F. May 27, 28, 29
Osceola N. F. June 25, 26, 27
Jennings S. F. July 25, 29
Steinhatchee July 15, 16, 18
Little River Springs June 6, 7

Litter Measurements

To determine the potential effects of species on the litter layer beneath understory

plants, litter was measured with a quadrat. A 25 cm by 25 cm (internal dimensions)

square quadrat made of 1/2" PVC pipe and PVC joints was placed against the south side of

the stem for consistency. Within the quadrat, three readings of litter depth were made.









Then the litter was removed and placed into a paper bag. Any living plants were

removed from the litter sample before being weighed on an electronic balance (Ohaus,

Scout II) with a maximum of 600 g and accuracy to 0.1 g. Litter samples were dried at

70C for 72 h and weighed. Litter density is given in g-cm3 (Eq. 2-2).

Litter density (g-cm3)= litter biomass / (average litter depth 25 cm 25 cm) (2-2)

Height Measurements

Before harvesting, total height and height to lowest branch were measured for each

plant. The plant was not disturbed nor physically extended to take these measurements.

The lowest branch measurement was made from the bottom of the litter layer to the point

of the lowest vegetation on the branch, whether it was at the stem junction or at the

terminal end of a branch. If multiple stems from the same individual emerged from

beneath the litter layer, then the height to lowest branch was recorded as zero. Two

measurements of crown width were taken at the widest point in perpendicular directions.

Fuel density, or fuel loading (Rundel 1981), was calculated by dividing the total dry

biomass by the plant volume (Eq. 2-3). The plant was then harvested at the soil line for

biomass measurements.

Fuel loading (mg/cm3)= total biomass/(height width 1 width 2) (2-3)

Biomass Measurements

Each plant was separated into components: live foliage, dead foliage, litter

accumulated on plant (referred to as debris), small stems (<0.6 cm diameter), and coarse

fuel (> or = 0.6 cm diameter) for biomass analyses. Dead foliage and debris were added

to the fine fuel categorization of live foliage and small stems (<0.6 cm diameter) used by

Etlinger (2000). If the amount of dead foliage was <0.1 g, it was included in the

measurement of debris. In the field, fresh weights were recorded with the same









electronic scale used for litter measurements. For foliar biomass samples, a subsample

was removed for leaf area and volatile solid analyses (Chapter 3) and the sample was

reweighed.

One plant of each species at each site was sampled at 1-m intervals measured from

the bottom of the litter layer. The plant was sampled by cutting off the highest interval

first. The separation of height interval was only used in the biomass measurements as

samples were pooled from all intervals for foliar analyses (Chapter 3).

Foliar, small stem, and debris samples were dried at 700C for 72 h and large stems

dried until their weight was stable for at least 48 h. Live and dead petioles (course fuel)

of Serenoa repens were weighed after drying for 144 h. The time of drying was

determined based on results from a preliminary study. Samples were then weighed with

the same electronic balance used in the field. The dry weight of the paper bag, depending

on the size, was subtracted from the measured dry weight. The average dry weight of

each bag size was determined by drying ten randomly selected bags of each bag size for

72 h at 70C. Total dry foliar biomass was calculated based on the moisture content of

the sub-sample dried in the oven. In this way, the sub-sample dry weight was multiplied

by the ratio of the fresh weight of the whole sample divided by the fresh weight of the

subsample.

Moisture Content

Because of the methodology used to determine the biomass components of each

plant, moisture content of each component could also be calculated. Moisture content

was calculated based on dry weight (Eq. 2-4).

Moisture content (%) = [(fresh weight dry weight) / dry weight] 100 (2-4)









Statistical Analyses

Site characterization data, in stems per hectare, tree basal area, height to lowest

branch, and canopy closure, were analyzed for overall ecosystem and site differences.

The general linear model (glm) procedure in Statistical Analysis Software (SAS) was

used. All pairwise comparisons of fixed means were performed using Tukey's test,

rejecting the null hypothesis that there was no difference between treatments when

p<0.05. Sites were considered fixed and were nested within ecosystem (Eq. 2-5).

y = tA + a, + j,) + e,, (2
It = true mean
a, = effect of level i of A (ecosystem) (df=l)
f3j0) = effect of level of B (site) nested within level i of
A (df=8)
random el = experimental error [df=70 (understory) and df=30


-5)


(midstory and overstory)]

Species data were analyzed for ecosystem, species, and site effects using the

general linear model (glm) procedure in Statistical Analysis Software (SAS) (Eq. 2-6).

Species and site effects were nested within ecosystem type. Site effects were considered

random.

yykl = [ + a, + Bj) + 6k() + B6jk() + ejkl (2-6)
It = true mean
a, = effect of level i of A (ecosystem) (df=l)
random B, = effect of level of B (site) nested within
level i of A (df=8)
k(i) = effect of level k of C (species) nested within
level of B within level i of A (df=10)
random B6jk( = effect of level of B (site) by the effect of level
k of C (species) nested within level i of A (df=40)
random eykl = experimental error (df=120)

When the site x species interaction (B6jk() was not significant (p>0.1), then it was

dropped from the model. When the site x species interaction (B6Jk() was significant

(p<0.1), species were run individually to determine the random effect of site on









individual species using the model, y= Bj,(. The aforementioned model was used to

analyze data for all species studied, treating Myrica cerifera in flatwood ecosystems as a

different species than Myrica cerifera in hardwood ecosystems, and the same for Serenoa

repens.

A different statistical analysis was used for 1-m height interval data (Eq. 2-7).

y&ki = + + + 6k() + ek (2-7)
I = true mean
a, = effect of level i of A (ecosystem) (df=l)
6k() = effect of level k of C (species) nested within level i of A
(df=10)
random ekl = experimental error (df=48)

To determine the direct ecosystem effect on individual species, data from Myrica

cerifera and Serenoa repens were analyzed separately (Eq. 2-8).

yki = [t + a, + 65 + (a6), + Bk() + lk) + ekl (2-8)
It = true mean
a, = effect of level i of A (ecosystem) (df=l)
6k = effect of level k of C (species) (df=l)
(a6),k = effect of level i of A (ecosystem) by level k of C
(species) (df=1)
random B,(i = effect of level of B (site) nested within level i of
A (ecosystem) (df=8)
random 1l0k) = effect of level of B (site) nested within A
(ecosystem) by C (species) interaction (df=8)
random ejk = experimental error (df=39)

When interaction variables were not significant (p<0.1), they were dropped from

the model. All other tests were performed at a=0.05. All pairwise comparisons of fixed

means were performed using Tukey's test, rejecting the null hypothesis that there was no

difference between treatments when p<0.05.









Results

Site Characterization

The absolute and relative densities of understory species (Appendix C-1) were used

to calculate the percentage of the understory that was characterized by this study (Table

2-2). Absolute and relative densities were also calculated for midstory species at each site

(Appendix C-2). Absolute density, relative density, relative dominance, relative

frequency, and importance values were calculated for overstory species at each site

(Appendix D-l).

Collectively, the flatwood sites contained higher understory density, lower

midstory density, and lower overstory density than hardwood sites (Table 2-3). Further

analyses of the overstory reveal that the flatwood sites contained less basal area per

hectare than hardwood sites. The trees in hardwood sites had lower height to lowest

branch and higher percentage of canopy closure than trees in flatwood sites (Table 2-4).

Table 2-2. Percent of total understory stems characterized by the species studied at each
site.
Ecosystem Site % of total stems
Flatwood Austin Cary M. F. 78.8
Osceola N. F. 97.2
Jennings S. F. 85.2
Welaka S. F. 84.7
Withlacoochee S. F. 77.4
Hardwood Twin Rivers S. F. 63.6
Osceola N. F. 62.1
Jennings S. F. 59.0
Steinhatchee 25.6
Little River Springs 59.0

Although there were collective differences between ecosystems, there were also

similarities between some flatwood and hardwood sites. The understory density in the

flatwood sites in Jennings State Forest and Withlacoochee State Forest were not different









from the understory density in any of the hardwood sites (Table 2-3). There were no

statistical differences between sites in midstory density (p=0.1140). A small difference

existed in the overstory density between sites. The flatwood sites at Austin Cary

Memorial Forest and Osceola National Forest contained fewer trees per hectare than the

flatwood site at Jennings State Forest and the hardwood site at Little River Springs (Table

2-3).

There was no difference in basal area between sites (p<0.0518). The flatwood site

at Jennings State Forest contained trees with a lower height to lowest branch than the

flatwood sites at Austin Cary Memorial Forest and Osceola National Forest (Table 2-4).

The canopy closure at the flatwood site in Withlacoochee State Forest was not

statistically different from all the hardwood sites except Twin Rivers State Forest (Table

2-4).

Ecosystem Differences

Ecosystems were significantly different in litter depth, but not litter density.

Ecosystems were also significantly different for height and biomass measurements when

all species were analyzed together. However, height and biomass measurements were not

different for Myrica cerifera and Serenoa repens between ecosystems.

Litter measurements

Litter depth in the flatwood ecosystems was almost twice that in hardwood

ecosystems (p=0.0025) (Table 2-5). This was also true for Myrica cerifera and Serenoa

repens within flatwood and hardwood ecosystems (p=0.0030). However, litter density

was the same in flatwoods and hardwoods (p=0.8541) and forMyrica cerifera and

Serenoa repens within flatwood and hardwood ecosystems (p=0.2700). Site (nested

within ecosystem) was significant for all litter measurements for all species studied.












Table 2-3. Understory, midstory, and overstory stems per hectare for the study sites. Standard error is given in parentheses (n=8 for
understory and n=4 for midstory and overstory site means; n=40 for understory and n=20 for midstory and overstory
ecosystem means). Lower-case letters indicate significant (p<0.05) difference in Tukey's pairwise comparison between
sites.


Understory


Midstory


Ecosystem Site Stems/ha Stems/ha
Flatwood Austin Cary M. F. 136,000 (32,000) ab 0 (0)
Osceola N.F. 135,000(13,400) ab 0(0)
Jennings S.F. 79,000 (37,600) bcd 118(64)
Welaka S. F. 177,000 (15,700) a 68 (32)
Withlacoochee S. F. 85,400 (16,200) bc 175 (69)
Flatwood Mean 123,000 (12,100) 72 (23)
Hardwood Twin Rivers S. F. 2,190 (390) d 881 (160)
Osceola N. F. 5,770 (792) cd 500 (245)
Jennings S.F. 17,100 (1,750) cd 1137(241)
Steinhatchee 13,900 (2,300) cd 1000 (193)
Little River Springs 7,960 (1,840) cd 1381 (344)
Hardwood Mean 9,390 (1,100) 980 (118) *
indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems


Overstory


Stems/ha
150.0 (38.2) c
206.3 (29.5) bc
650.0 (110.4) a
275.0(25.0) abc
562.5 (156.3) abc
380.3 (59.6)
618.8(64.9) ab
365.3 (90.9) abc
606.3 (104.3) abc
600.0 (77.7) abc
681.3 (74.6) a
572.5 (42.0) *













Table 2-4. Basal area, height to lowest branch, and canopy closure at each site. Standard error is given in parentheses (n=4 for site
means and n=20 for ecosystem means). Lower-case letters indicate significant (p<0.05) difference in Tukey's pairwise
comparison between sites.


Site
Austin Cary M. F.
Osceola N. F.
Jennings S. F.
Welaka S. F.
Withlacoochee S. F.
Flatwood Mean
Twin Rivers S. F.
Osceola N. F.
Jennings S. F.
Steinhatchee
Little River Springs


Basal area (cm2 per ha)
14,200 (3400)
18,000 (2880)
29,600 (4190)
20,000 (2380)
30,600 (4070)
22,900 (2040)
32,900 (4660)
26,800 (5820)
30.000 (5070)
42,800 (5380)
37,000 (4780)


Height to lowest branch (m)
17.2(1.4) a
16.4(1.0) a
7.2 (0.0) bc
12.1(2.0) ab
12.5(1.9) ab
12.8(1.0)
4.9 (0.4) c
10.3 (1.6) bc
5.9(0.7) c
7.1 (0.2) bc
6.0(0.3) c


Canopy
52(5)
60(1)
62 (6)
59(6)
84 (4)
63(3)
98(1)
85(5)
87(2)
88(3)
94(1)


closure (%) **
c
c
c
c
b


Hardwood Mean 34,000 (2420) 6.8 (0.5) 90 (2) *
* indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems
** although percentage means are presented in the table statistical analysis on canopy closure was performed using an arcsin
transformation.


Ecosystem
Flatwood


Hardwood









Site (nested within ecosystem) was also significant (p<0.0001) for litter moisture content,

although there was no difference between ecosystems (Appendix E).

Height measurements

Based on all understory species studied, there was no difference between the

average height to lowest branch between flatwood and hardwood sites (p=0.3968). In

addition, ecosystem type did not affect height to lowest branch of Myrica cerifera and

Serenoa repens (p=0.5293). On average, understory species were taller in hardwood

ecosystems (147.8 cm) than flatwood ecosystems (108.5 cm), p<0.0001. However,

ecosystem type did not affect the total height of Myrica cerifera and Serenoa repens

(p=0.9495).

Table 2-5. Litter depth and density for each species and ecosystem type. Standard error
is given in parentheses (n=15 for species and n=90 for ecosystem). Lower-
case letters indicate significant (p<0.05) difference in Tukey's pairwise
comparison.
Ecosystem Species litter depth (cm) litter density (mg/cm3)
Flatwood Gaylussacia dumosa 5.39 (0.59) bc 17.61 (2.44) ab
Ilexglabra 4.87 (0.39) bc 20.54 (1.41) a
Lyoniaferruginea 6.28(0.46) b 13.42 (1.23) bc
Vaccinium myrsinites 4.53 (0.47) bc 18.97 (3.46) ab
Myrica cerifera 6.17(0.99) b 16.95 (1.61) ab
Serenoa repens 10.12 (0.92) a 9.49 (1.62) c
Flatwood mean 6.23 (0.33) 16.16 (0.92)
Hardwood Callicarpa americana 3.63 (0.36) c 16.02 (1.45) abc
Ilex opaca 3.71 (0.44) c 16.80(2.07) ab
Quercusnigra 3.22 (0.14) c 19.47(2.93) ab
Vaccinium arboreum 3.77 (0.27) c 14.54(1.73) abc
Myrica cerifera 3.11 (0.23) c 19.39 (2.47) ab
Serenoa repens 4.61 (0.50) bc 15.04(1.49) abc
Hardwood mean 3.67 (0.15) 16.88 (0.86)
* indicates significant (p<0.05) difference in Tukey's pairwise comparison between
ecosystems









Biomass measurements

Total (per plant) fine fuel biomass and the fine fuel components- live foliage,

dead foliage, and debris- were not different between pine flatwood and hardwood

hammock ecosystems. However, small stem biomass was greater in hardwood

ecosystems than flatwood ecosystems (p=0.0027) (Table 2-6). Total fine fuel and all

individual fine fuel components ofMyrica cerifera and Serenoa repens were not different

between ecosystems.

There was more coarse fuel in hardwood ecosystems than in flatwood ecosystems

(p=0.0035) (Figure 2-1). However, coarse fuel biomass ofMyrica cerifera and Serenoa

repens were the same between ecosystems (p=0.2802). The results of total biomass

analyses were similar to the results from the coarse fuel analyses. Total biomass per

individual plant was greater in hardwood ecosystems than in flatwood ecosystems

(p=0.0214) but total biomass per plant was the same between ecosystems for Myrica

cerifera and Serenoa repens (p=0.4826). Fuel loading was higher for understory species

in flatwood ecosystems than in hardwood ecosystems (p=0.0376) (Figure 2-2) although

there was no difference in fuel loading ofMyrica cerifera and Serenoa repens between

ecosystems (p=0.9561). Live foliage collected in hardwood hammocks had greater foliar

moisture content than pine flatwoods (Chapter 3) although there was no significant

difference between moisture content for any other biomass component (Appendix E).

There was significantly more live foliage and small stems per individual in the

lowest 1-m interval in hardwood ecosystems than flatwood ecosystems (p=0.0001 and

p=0.0016, respectively) (Figure 2-3). There was no significant difference between the

dead foliage, debris, or large stem components of the lowest 1-meter interval between

ecosystems (Appendix F). There was also no statistical difference between the total












Table 2-6. Dry weight of fine fuel biomass components for each species and ecosystem types. Standard error is given in parentheses
(n=15 for species and n=90 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in Tukey's pairwise
comparison.

Fine Fuels (total dry g)
Ecosystem Species live foliage dead foliage debris small stems
Flatwood Gaylussacia dumosa 3.9(0.5) d 0.0(0.0) b 1.0(0.2) b 6.5(1.0) e
Ilexglabra 10.1 (2.2) cd 0.0(0.0) b 1.8(0.5) b 20.0(3.4) cde
Lyoniaferruginea 15.5 (2.5) cd 0.0(0.0) b 3.0(0.6) b 22.7 (3.0) cde
Vaccinium myrsinites 3.6(0.7) d 0.0 (0.0) b 2.7 (0.9) b 9.9 (2.4) e
Myrica cerifera 40.3(6.8) cd 0.0(0.0) b 4.8(1.0) b 45.4(7.9) abc
Serenoa repens 242.2 (45.3) b 192.1 (40.5) a 38.4 (8.9) a 0.0 (0.0) e
Flatwood mean 52.6(11.7) 32.0 (10.0) 8.7 (2.1) 17.4 (2.2)
Hardwood Callicarpa americana 4.9 (0.9) cd 0.0 (0.0) b 0.3 (0.2) b 15.5 (2.6) de
Ilex opaca 84.6 (15.9) c 0.1 (0.0) b 8.1 (4.2) b 69.0 (12.3) a
Quercusnigra 14.8(4.9) cd 0.1(0.1) b 0.8(0.4) b 20.1 (3.6) cde
Vaccinium arboreum 34.0(8.4) cd 0.0(0.0) b 6.3 (1.9) b 60.6 (11.6) ab
Myrica cerifera 28.3 (9.4) cd 0.0(0.0) b 1.8(0.6) b 38.1 (10.9) bcd
Serenoa repens 324.5 (48.1) a 192.4 (44.0) a 40.4 (9.6) a 0.0 (0.0) e
Hardwood mean 81.9 (14.5) 32.1 (10.4) 9.6 (2.3) 33.9 (4.2) *
* indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems











Gaylussacia dumosa -


Ilex glabra -I


Lyoniaferruginea -

Vaccinium myrsinites -

Myrica cerifera -


b
M coarse fuel
b fine fuel

b

b

b


Serenoa repens


Flatwood mean-


Callicarpa americana -


Quercus nigra -


Vaccinium arboreum


Myrica cerifera -

Serenoa repens

Hardwood mean -


U


a


*


0 200 400 600 800 1000
0 200 400 600 800 1000


1200


biomass (dry g)


Total biomass by coarse fuel and fine fuel components for species within
pine flatwoods and hardwood hammocks (n=15 for species, n=90 for
ecosystem). = significant difference in coarse fuel between ecosystems,
** = significant difference between total fuel between ecosystem type.
Letters indicate significant (p<0.05) difference in Tukey's pairwise
comparison for total biomass between species. In addition the same letters
describe the differences between species for fine fuel and coarse fuel
analyzed separately.


Figure 2-1.


II


Ilex opaca b














Gaylussacia dumosa


Ilex glabra


Lyoniaferruginea


Vaccinium myrsinites


Myrica cerifera


Serenoa repens






Callicarpa americana


Ilex opaca


Quercus nigra


Vaccinium arboreum


Myrica cerifera


Serenoa repens


Sfuel loading


0.0 0.1 0.2 0.3 0.4 0.5


Fuel loading (mg/cm3)


Fuel loading for each species and ecosystem type. Standard error is
represented in error bars (n=15 for species, n=90 for ecosystem means). Star
represents significant (p<0.05) difference between ecosystem type. There
was no significant difference for fuel loading between species (p>0.05).


Figure 2-2.


1


i


H










biomass in the 1st (p=0.0717) or 2nd (p=0.1672) meter intervals between ecosystems.

However, individual understory plants in hardwood hammocks had more total biomass in

the 3rd meter interval (p<0.0001) than pine flatwood understory plants, although the

actual amount of biomass in the 3rd meter interval was a small proportion of the total

biomass (Figure 2-3). Analyses of biomass components within the 2nd and 3rd 1-m height

intervals were not possible due to the low number of samples within these two intervals.



I I large stems
Flatwoods M3 i small stems
l live foliage
B E dead foliage
Flatwoods M2 debris


Flatwoods Ml


Hardwoods M3 -


Hardwoods M2


Hardwoods Ml



0 50 100 150 200 250 300

dry weight (g)

Figure 2-3. Biomass per 1-meter height intervals by ecosystem type (n=30). An '*'
indicates significant (p<0.05) difference in meter 1 components between
ecosystem types. An '**' indicates significant (p<0.05) difference in total
biomass per 1-meter height interval between ecosystem types.

Species Differences

Species effects on litter measurements were not pronounced. Overall height was

different among species and the height to lowest branch was minimally different between









species studied. Species differences in biomass were largely due to the greater overall

biomass of Serenoa repens compared to all other species studied.

Litter measurements

Litter sampled beneath plants was deepest for Serenoa repens in flatwood

ecosystems followed by Myrica cerifera and Lyoniaferruginea in flatwood ecosystems

(Table 2-5). Serenoa repens, Myrica cerifera, and Lyoniaferruginea (all in flatwood

sites) had greater litter depth beneath them than Callicarpa americana, Ilex opaca,

Quercus nigra, Vaccinium myrsinites, and Myrica cerifera in hardwood ecosystems.

Moisture content of litter was not different between species (Appendix E).

Litter density sampled beneath Ilex glabra was greater than litter density sampled

under Lyoniaferruginea and Serenoa repens in flatwood sites (Table 2-5). Serenoa

repens in flatwood sites had lower litter density beneath it than Gaylucassia dumosa,

Vaccinium myrsinites (both in flatwood ecosystems), and Ilex opaca, Quercus nigra, and

Myrica cerifera (in hardwood ecosystems). There was a significant random site by

species (nested within ecosystems) interaction for litter density (p=0.0091). Upon further

analyses, litter density under Lyoniaferruginea, Callicarpa americana, Ilex opaca,

Quercus nigra, and Vaccinium arboreum was affected by the random effect of site nested

within ecosystem (p<0.05).

Height measurements

The height to lowest branch was significantly shorter for Vaccinium myrsinites and

Serenoa repens (in either ecosystem) than Ilex glabra and Myrica cerifera (in flatwoods)

and Ilex opaca, Quercus nigra, Vaccinium arboreum, and Myrica cerifera (in hardwoods)

(Figure 2-4).









Vaccinium arboreum, at 182.9 cm mean height, was taller than all flatwood species

and all hardwood species except Ilex opaca and Quercus nigra. Vaccinium myrsinites

and Gaylussacia dumosa were significantly shorter than all species studied averaging

51.9 cm and 71.0 cm in height, respectively. Myrica cerifera had the greatest variation in

height in both flatwood and hardwood ecosystems. The total height ofLyonia

ferruginea, Vaccinium myrsinites, and Myrica cerifera was significantly different

between sites nested within ecosystems (p<0.05). This resulted in a significant site by

species (nested within ecosystem) interaction (p=0.0415).

Biomass measurements

Total fine fuel biomass per individual was greatest for Serenoa repens in hardwood

and flatwood ecosystems (Figure 2-1). Live foliage biomass was greatest for Serenoa

repens in hardwood ecosystems, followed by Serenoa repens in flatwood ecosystems

(Table 2-6). Dead foliage biomass was greatest for Serenoa repens in both flatwood and

hardwood ecosystems. All other species had negligible dead foliage biomass. In

addition, Serenoa repens in either ecosystem had more accumulated debris than all other

species. Ilex opaca had greater small stem biomass than all other species except

Vaccinium arboreum and Myrica cerifera (in flatwood ecosystems). Serenoa repens, in

either ecosystem type, had higher fine fuel biomass and coarse fuel biomass per

individual than all other species studied (Figure 2-1). There was no significant difference

in fuel loading between species (p=0.5024) (Figure 2-2).

Species by site interaction (nested within ecosystem), a random effect, was

significant (p<0.1) for live foliage biomass, small stem biomass, large stem biomass, total

biomass, and fuel loading. Further analyses showed that site (nested within ecosystem)

was significant (p<0.05) for Ilex glabra (fuel loading), Vaccinium myrsinites (small stem













200

150

100

50

0 -


Flatwood Species Hardwood Species


Figure 2-4. Height and height to lowest branch of species within pine flatwoods and hardwood hammock ecosystems. Standard error
is shown in error bars (n=15 for species and n=90 for ecosystems). Upper-case and lower-case letters indicate significant
(p<0.05) difference in Tukey's pairwise comparison.


total height (cm) A
II height to lowest branch (cm) A B
ABC
BC BC BC BC C BC BC



D D

a a a a a
b b b b
I L 1, 11111,


I I I I I I


I I I I









biomass and fuel loading), Myrica cerifera in flatwoods (fuel loading), Serenoa repens in

flatwoods (live foliage biomass and total biomass), and Myrica cerifera in hardwoods

(large stem biomass and total biomass). Although there were differences between the

foliar moisture content of species (Chapter 3), there was no difference between the

moisture content of small stems between ecosystems (Appendix E).

Species were significantly different in all components within the 1st 1-meter

interval (p<0.05) (Appendix F, Figure 2-5). Serenoa repens contained more live foliage,

dead foliage, and debris than most species studied. Serenoa repens, Ilex opaca, and

Vaccinium arboreum contained more large stems than most species studied. Although

the total biomass in the 1st 1-meter interval was different between species (p<0.0001)

(Appendix F), there was no difference between species in the total biomass in the 2nd or

3rd 1-meter intervals.

Discussion

More variation in ecosystem structure existed among flatwood sites than

hardwood sites. This is likely caused by the differences in stand age and that some

flatwood sites were a combination of naturally regenerated and planted pines. The

flatwood sites at Austin Cary Memorial Forest and Osceola National Forest had fewer,

but larger trees per hectare. In addition, the trees at these two sites had greater height to

lowest branch than other sites. The three flatwood sites with the fewest trees per hectare

(Austin Cary Memorial Forest, Osceola National Forest, and Welaka State Forest) also

had the highest understory density. It is likely that variation between the flatwood sites

accounts for the variability between sites for litter and biomass measurements. Even with

differences between sample sites, some general differences in plant structural

components can be drawn from the ecosystem type as a whole.













Flatwood species
0 200 400 600
1 1 1


Hardwood species
0 200 400 600


Gaylussacia dumosa -

Ilex glabra -

m
Lyoniaferruginea -

SVaccinium myrsinites -

Myrica cer-fera -

Serenoa repens -



Gaylussacia dumosa -

Ilex glabra -

- -
Lyona ferruginea -

SVaccinum myrsinites -

Myrica cernfera -

Serenoa repens -



Ya ainumn myrsinit -

Ilex glabra -

Xa la ferruginea



Myrica cernfera -

Serenoa repens -


Calhcarpa americana

Ilex opaca

Quercus nigra -

Vaccinmum arboreum

Myrica cernfera

Serenoa repens -



Calhcarpa americana

Ilex opaca

Quercus nigra

Vaccinmum arboreum

Myrica cernfera

Serenoa repens -



Calhcarpa americana -

Ilex opaca

Quercus nigra

Vacciium arboreum

Myrica cerifera

Serenoa repens


0 200 400 600 800 0 200 400 600 800

Biomass (dry g) Biomass (dry g)



Figure 2-5. Biomass per 1-meter height intervals for species within pine flatwoods and

hardwood hammocks (n=5).


Ecosystem Differences


The amount of total fuel available for wildfire within the litter layer is greater in


flatwood ecosystems and in addition, a surface fire would have flames that reached


higher into the understory vegetation in the flatwoods than hardwoods. Acacia


woodlands, a fire-prone ecosystem in southern Ethiopia, had twice the litter biomass


large stems
I I small stems
live foliage
I I dead foliage
I debris






















I








I


I





EU







Ii









compared to Afromontane forests, a less fire-prone ecosystem in the same region

(Eriksson et al. 2003). The chemical and structural characteristics of the litter

components will greatly influence the flammability of the litter layer, and the

flammability of surrounding plants. For example, pine needles are known to be highly

flammable because of high surface area to volume ratio and presence of volatile

compounds (Fonda 2001, Schroeder et al. 2001).

Although there was no difference between ecosystems in the height to lowest

branch, the increased litter depth makes the lowest branch in flatwood ecosystems more

susceptible to ignition from a surface fire. Once ignited, however, an individual

understory plant common to hardwood ecosystems has the potential to carry fuel to a

greater height than an individual understory plant common to flatwoods ecosystems. In

addition, because of the increased total height and size, individual understory plants in

hardwood ecosystems may be more susceptible to ignition from firebrands. The

difference in height was largely based on the different species sampled within each

ecosystem as there was no direct effect of ecosystem on the total height of Serenoa

repens and Myrica cerifera. It also must be noted that the average height difference

between pine flatwood and hardwood hammock may be biased due to the selection of

specimens based on species-specific minimum and maximum height requirements.

The ecosystem effect on the biomass components largely reflects the biomass

difference between the species studied specific to eachecosystemas there were no direct

ecosystem effect on the biomass components of Serenoa repens and Myrica cerifera.

This is likely because these two species are adapted to a wide range of growing









conditions and their requirements for growth are met within the microclimates in each

ecosystem type.

There was no difference between mass of debris between ecosystems. This was an

unexpected result as pine needles appeared to accumulate more readily on understory

plants than hardwood leaves given the grouped nature of pine needles (Figure 2-6). It is

possible that the presence of even a few pine trees, as existed in most hardwood study

sites (Appendix D), generate enough needles and debris to minimize the difference

between ecosystem types. In addition, understory plants in hardwood hammock

ecosystems may have intercepted more debris because of their greater overall size.

















Figure 2-6. Pine needles accumulated on a landscape plant.

Even though there is more total biomass for individuals in hardwood ecosystems,

the amount of fine fuels was not different between ecosystems. In a comparison of fire-

prone fynbos ecosystems to fire-resistant forest patches in South Africa, understory

species were found to contain more total biomass in fire-resistant forest patches (van

Wilgen et al. 1990). Understory species in less fire-prone Afromontane forest also

contained more total biomass than more fire-prone Acacia woodlands in southern









Ethiopia (Eriksson et al. 2003). However, the forest patches and Afromontane forest also

contained more fine fuel biomass (van Wilgen et al. 1990, Eriksson et al. 2003). In this

study, the greater biomass in hardwood understory plants is due to increased biomass of

large stems. Although there is more potential fuel for a fire per understory plant in

hardwood ecosystems, the amount of that fuel that is likely to be consumed and

contribute to the overall flammability of the plant is the same between ecosystems

(Etlinger 2000).

In addition, the fuel within individual understory plants in the flatwood ecosystems

is more densely packed. Species within the flatwoods ecosystem may sustain a fire for a

greater amount of time, therefore increasing the amount of fuel consumed and overall

heat released per plant (Rundel 1981). Fire-prone fynbos ecosystems also contained a

more dense fuel bed than forest patches (van Wilgen et al. 1990).

We conclude that the potential structural flammability is higher for understory

species within pine flatwood ecosystems than hardwood hammock ecosystems.

Individual understory plants in pine flatwoods contain the same amount of fine fuel as

understory plants in hardwood hammocks. However, the fuel is more densely packed

within each understory plant and the litter depth underneath each plant is greater.

Species Differences

Serenoa repens, in either ecosystem, is highly structurally flammable with high

ignitability (litter depth, litter density, height to lowest branch, dead foliage and debris)

and sustainability (fine fuel biomass). Specific discussions are made of each species

studied relative to each other. It must be noted that the lack of clear differences between

species other than Serenoa repens may be due to the constraints of the statistical analysis









in determining differences in species as biomass data for Serenoa repens were so much

higher than all other species studied.

The difference between the flammability of species within the same genus is an

important consideration. Rankings of plant flammability typically are built from existing

lists originating in different biographical regions. Many times, species in the same

genera are considered to be of similar flammability. The results of biomass comparisons

clearly show that species within the same genera (Ilex glabra versus Ilex opaca and

Vaccinium myrsinites versus Vaccinium arboreum) do not always have similar

flammability characteristics and represent different threats to WUI structures even within

the same biographical region.

Gaylussacia dumosa

Gaylussacia dumosa along with Vaccinium myrsinites had the lowest average

height than other species studied. Even though the average height was low compared to

other species, the height to lowest stem was not significantly shorter than any other

species. This means that most of the fine fuel components, including the foliage and

debris, are not any more vulnerable to ignition from a surface fire than any other species

studied. However, the consumability of Gaylussacia dumosa plants is high because of

the high proportion of fine fuels in the total biomass. Therefore, once ignited, it is likely

that the whole plant will be consumed, contributing to the spread of wildfire.

Gaylussacia dumosa plants had relatively low debris biomass which may decrease the

ignitability in wildfires. Based on the results of litter, height, and biomass measurements,

an individual Gaylussacia dumosa plant does not cause a serious threat to WUI structures

as it contains relatively little overall biomass, which would likely be entirely consumed if

ignited.









Ilex glabra

There is no litter or biomass characteristic that would indicate that an individual

Ilex glabra plant is highly flammable. Specifically, an individual plant may be slightly

less flammable because the average litter density was slightly greater under Ilex glabra

plants than other species studied.

Lyoniaferruginea

There is no litter or biomass characteristic that would indicate that an individual

Lyoniaferruginea plant is highly flammable. However, the litter under Lyonia

ferruginea (deeper depth and lower density) could increase the ignitability of Lyonia

ferruginea plants compared to many of the other species studied.

Vaccinium myrsinites

Vaccinium myrsinites plants are highly flammable because of their low height and

fine fuel composition making their ignitability and consumability quite high. However,

the total potential fuel is very low, making the sustainability quite low and likely

reducing the combustability of an individual plant. Therefore, individual plants of

Vaccinium myrsinites are not a high hazard to WUI structures.

Callicarpa americana

Based on the results of litter, height, and biomass measurements, an individual

Callicarpa americana plant does not cause a serious threat to WUI structures. The

consumability is not high due to the low proportion of fine fuels. In addition this species

has low total fuel content. Callicarpa americana had a relatively low height to lowest

branch measurement compared to other species, which may increase the ignitability of

the plant slightly.









Ilex opaca

With the exception of Serenoa repens, Iex opaca plants contained slightly, but not

always statistically significant more live foliage biomass, debris, fine fuel biomass, and

total fuel biomass compared to most species studied. In addition, the great amount of

small stem biomass makes the total fuel for fires very high for this species, giving it a

moderately high flammability potential.

Quercus nigra

Quercus nigra plants retained slightly more debris, possibly increasing the

ignitability of this species. However, no other characteristics indicate that individuals of

this species in the understory are hazardous to WUI structures.

Vaccinium arboreum

Vaccinium arboreum plants in the understory are very similar in biomass

components compared to Iex opaca. Without knowing the chemical composition of the

biomass, it is difficult to determine the differences in flammability between these two

species.

Myrica cerifera

Myrica cerifera plants are more flammable in flatwood ecosystems than hardwood

ecosystems because of deeper litter depth and increased small stem biomass. Generally,

Myrica cerifera is comparable to both Iex opaca and Vaccinium arboreum in biomass

components. If ignited, there is a great amount of all fuel components-fine fuel, coarse

fuel, and total fuel.

Serenoa repens

Serenoa repens in either a pine flatwood or hardwood hammock ecosystem has

many structural characteristics making it highly flammable. Serenoa repens along with









Vaccinium myrsinites have low height to lowest branch making Serenoa repens highly

vulnerable to ignition from surface fires. The large amount of debris and dead foliage

biomass makes Serenoa repens highly flammable in either ecosystem. In addition,

Serenoa repens within either ecosystem had the highest amount of total fine fuel, coarse

fuel, and total biomass than any other species studied. Ignition is more facilitated within

flatwood ecosystems, due to the decreased density and increased depth of litter. There

was significantly more live foliage biomass per individual Serenoa repens plant in

hardwood ecosystems than flatwoods ecosystem causing the Serenoa repens plant in

hardwood ecosystems to be slightly more flammable (based on biomass) than a plant of

the same species in flatwood ecosystems. This is especially true given the noted

flammability of live Serenoa repens foliage (Hough and Albini 1978).

Conclusions

One result to consider from this study is that individual plants of the species studied

may not pose a great threat to WUI structures, except for Serenoa repens. Based on the

arrangement of biomass and great amount of total biomass, even individual plants of

Serenoa repens can be a threat to WUI structures. Proximity of one landscape plant to

another and the construction materials of the WUI home will also influence what species

is or is not appropriate for firewise landscaping.

The regular maintenance of plants can lower their flammability and therefore lower

the wildfire hazard to WUI structures. Current recommendations for firewise

landscaping guide landowners to remove accumulated pine needles within 30 feet (9

meters) surrounding a structure (Florida Firewise Communities 2000, Monroe and Long

2001). In addition, if debris and dead material accumulated on plants are removed, the

flammability of those plants will be reduced. As Serenoa repens is notoriously difficult









to remove once established, landowners may be able to greatly reduce the flammability of

this species by removing dead fronds and debris, thinning, and pruning off lower fronds.

However, great caution must be taken to keep up maintenance of Serenoa repens

throughout the year and to make sure that individual plants are separated by adequate

distance from each other and WUI structures.

A more thorough examination of the arrangement of biomass may be a very

important component to flammability of plants in the field. The location of different

forms of fuel will influence the ignitability. Fuel close to the ground will be more

susceptible to ignition from surface fires and fuel higher in the canopy may be more

susceptible to ignition from firebrands. To make clear distinctions between species, a

greater sample size and sampling at shorter height intervals may be necessary. In

addition, knowing how landscape plants are typically ignited in wildfires, from post-fire

studies, would improve future plant flammability studies.

Chemical and structural composition of the individual biomass components

between ecosystems or species may be different, greatly affecting plant flammability. In

Chapter 3, we will examine those differences for foliar biomass.














CHAPTER 3
FOLIAR FLAMMABILITY IN PINE FLATWOOD AND HARDWOOD HAMMOCK
ECOSYSTEMS

Introduction

In Chapter 2, the distribution of available fuel into different structural components

was presented for Gaylussacia dumosa, Ilex glabra, Lyoniaferruginea, and Vaccinium

myrsinites in pine flatwoods; Callicarpa americana, Ilex opaca, Quercus nigra, and

Vaccinium arboreum in hardwood hammocks; and Myrica cerifera and Serenoa repens

in both ecosystems. However, foliar properties greatly affect the flammability of plants

and it has been noted that species common in Southern pine ecosystems, especially Ilex

glabra and Serenoa repens, have foliar characteristics which make them extremely

flammable (Hough and Albini 1978). In addition, a study by Etlinger (2000) found that

foliar biomass and foliar moisture content were the most important variables in

determining the peak heat release rate of shrub species.

Many acceptable methods to measure plant flammability exist. Measurements of

plant tissue flammability have been made by thermal evolution analysis (Shafizadeh et al.

1977), oxygen combustion calorimetry (Dickinson and Kirkpatrick 1985, van Wilgen et

al. 1990, Rodriguez-Afi6n et al. 1995, Nufiex-Regueira et al. 2001, Williamson and Agee

2002), thermogravimetric analysis (Mutch and Philpot 1970, Philpot 1970, Shafizadeh et

al. 1977, Gill et al. 1978, Rogers et al. 1986, Dimitrakopoulos 2001), thermocouple

analysis (Owens et al. 1998), evolved gas analysis (Susott 1982a), and differential

scanning calorimetry (Susott 1982b). Muffle furnace tests (Cheo and Montgomery 1970,









Montgomery and Cheo 1971), cone calorimetry (White et al. 1996), and the limiting

oxygen index method (Mak 1988) have been used to test the flammability of entire

leaves, stems, or branches of plants. Etlinger (2000) used an intermediate scale biomass

calorimeter with a line burner to measure the flammability of entire plants. It is also

possible to measure flammability in the field by timing the combustion of plants after

being ignited (Ching and Stewart 1962).

Isoperibol oxygen combustion calorimetry (formerly referred to as bomb

calorimetry) is a technique which measures the total energy released (cal-dry g-1 or J-dry

g-1) by complete combustion of plant tissue in an 02 gas enriched, sealed vessel. Oxygen

combustion calorimetry has been used to determine the amount of energy in biomass fuel

for energy production (Rodriguez-Afi6n et al. 1995, Nufiez-Regueira et al. 2001). In

addition, oxygen combustion calorimetry has also been used to assess the impact of

vegetation on wildfire behavior between different ecosystem types (Dickinson and

Kirkpatrick 1985, van Wilgen et al. 1990). In this study, we used isoperibol oxygen

combustion calorimetry to determine the combustability and muffle furnace tests to

determine the comsumability of foliage between different species within different

ecosystem types.

The objective of this study was to compare the foliar flammability between species

within pine flatwood and hardwood hammock ecosystems. To accomplish this, foliar

leaf area, moisture content, volatile solids, and energy content were quantified. Foliar

leaf area was measured to characterize the leaves. Foliar energy content measures the

cumulative influence of flammability characteristics such as lignin, cellulose, and

hemicellulose content, volatile compounds, and mineral content. However, foliar









moisture content and low temperature volatile extractives are not incorporated into a

measure of energy content. Therefore, foliar moisture content was measured separately.

Based on the results of this study, in conjunction with the results from Chapter 2,

differences between ecosystems and species were explored.

Materials and Methods

Live and dead foliage samples were collected through biomass separation as

described in Chapter 2. Collected samples were weighed with an electronic balance with

a maximum of 600 g and accuracy to 0.1 g both in the field and after being dried. Foliar

samples were dried at 700C for 72 h. The total of 180 plants used in these analyses.

However, only one random individual of each species at each site was analyzed for leaf

area (60 plants total).

Moisture Content

Moisture content of each sample was calculated based on dry weight (Eq. 3-1).

Moisture content (%) = [(fresh weight dry weight) / dry weight] 100 (3-1)

Leaf Area

Leaf samples collected from one individual per species per site were placed into

plastic bags and into a cooler in the field to reduce loss of leaf structure from water loss.

Leaf samples were placed through a LI-COR 3100 Area Meter on the same day of

sampling. Twenty-five (25) leaves from most species were used to make leaf area

measurements. Only one leaf (frond without petiole) was measured for Serenoa repens

and 5 leaves measured for Ilex opaca, Quercus nigra, and Callicarpa americana because

of their large size. Fifty (50) leaves were used for Vaccinium myrsinites because of the

small size of the leaves. Three measurements of the same leaves were made and the

average was multiplied by two to calculate total leaf surface area for both the abaxial and









adaxial leaf surfaces. Leaves were dried in a drying oven for 72 h at 700C and then

immediately weighed using a balance, accuracy 0.001 g. Leaf area per leaf was

calculated by dividing the measured leaf area by the number of leaves used in the

measurement. Specific leaf area (leaf area per gram of dry weight) was determined for

each species. Also, total leaf area per individual was calculated based on the total dry

weight of foliage.

Volatile Solids

Live foliar samples were collected from each plant to be tested for volatile solids

content. In addition, dead foliar samples were collected from Serenoa repens plants for

volatile solid analysis. Samples were collected and immediately placed in a cooler to

prevent loss due to volatilization or decomposition. Samples were processed within 48

hours by Advanced Environmental Labs located in Tampa, FL by EPA standard 160.4

(Appendix G). Data is reported in mg volatile solids per kg dry weight.

Energy Content

Dried foliar samples were used to determine the energy content for each individual

using standard isoperibol oxygen combustion calorimetry. The isoperibol oxygen

combustion calorimeter (Parr Model 1261 Calorimeter, Parr Instrument Company) tests

were performed in the Department of Animal Science at the University of Florida. Live

foliar samples (also dead foliar samples for Serenoa repens) were tested. After drying,

the samples were stored in an air-conditioned laboratory. Dried foliar samples were

ground in an electric coffee grinder for sample homogenization in order to minimize

sample loss in larger grinding apparatus. Ground samples were then stored in paper bags

in the same laboratory until the calorimetry analysis was performed.









Crucibles were placed in a drying oven for 24 h at 700C. Dried crucibles were

placed in a dessicator until cool and then weighed. Samples were processed in a random

order. In this way, each sample from each site had equal opportunity to be processed on

a given day. Approximately 30 mg of sample was measured into each crucible.

Crucibles containing samples were placed back into the drying oven for 24 h at 700C.

Crucibles and samples were removed from the drying oven and placed into a dessicator

until cool. The dried crucible plus sample dry weight was then measured and the sample

dry weight was calculated. Samples were returned to the dessicator and taken out as

needed to be analyzed with the calorimeter.

Each ground foliar sample was processed in two runs. The two sample runs were

processed on separate days. If the replicate run was greater than 2.5% (Dickinson and

Kirkpatrick 1985) different from the first run, the sample was rejected and re-run twice at

a later time. The two runs were averaged for statistical analyses.

The calorimeter was calibrated using benzoic acid, ten ignitions per vessel. A fixed

acid correction of 15 and a fixed acid correction of 10 (25 calories total) were

automatically subtracted from the total energy released in combustion. This accounts for

energy released from the production of nitric acid from atmospheric N2 gas in the vessel

and the combustion of the fuse wire itself. NiChrome fuse wire (alloy of nickel and

chromium) was used. The vessel was purged with 30 atm 02 gas and submerged in 2 L

of deionized water. The calorimeter was left on throughout the study, but the water

heating/cooling system was turned off over night for safety reasons. The energy content

was calculated based on sample dry weight and expressed in cal-g1.









Foliar energy content per plant was calculated by multiplying the energy content

(in kcal) per gram by the total foliar energy content per plant. In addition, foliar energy

content per hectare per species was also calculated for each plant harvested (Eq. 3-2).

kcal-ha-= kcal per plant stems per ha of individual speices at individual site (3-2)

Statistical Analyses

Data were analyzed for ecosystem, species, and site effects using the general linear

model (glm) procedure in Statistical Analysis Software (SAS). Species and site effects

were nested within ecosystem type. Site effects were considered random. A different

model was used for the leaf area data because only one measurement of each species was

taken at each site (Eq. 3-3).

Yjkl = [t + a, + 6k() + ekl (3-3)
I = true mean
a, = effect of level i of A (ecosystem) (df=l)
k(i) = effect of level k of C (species) nested within level
i of A (df=10)
random e,kl = experimental error (df=48)

Moisture content, volatile solids, and energy content data were analyzed with the

model in Equation 3-4.

,jkl = t + a, + Bj) + 6k(i) + B6k(i) + ejk (3-4)
It = true mean
a, = effect of level i of A (ecosystem) (df=l)
random BjI = effect of level of B (site) nested within level i of
A (df=8)
6k() = effect of level k of C (species) nested within level
j of B within level i of A (df=10)
random Bjk() = effect of level of B (site) by the effect of level k
of C (species) nested within level i of A (df=40)
random ejk = experimental error (df=120)

When the interaction Bjk() was not significant (p>0.1), then it was dropped from

the model. When the interaction B6jk( was significant (p<0.1), species were run









individually to determine the random effect of site on individual species using the model,

y= B,/).

To determine the ecosystem effect on individual species, data from Serenoa repens

and Myrica cerifera were used (Eq. 3-5 and Eq. 3-6).

For Other Measurements: ykl = p + a, + 6, + (a6), + Bk() + lk(i) + e;kl (3-5)
For Leaf Area (LA): ykl= + + a + 6 + (a6) + eyk (3-6)
It = true mean
a, = effect of level i of A (ecosystem) (LA df= 1), (df=l)
6k = effect of level k of C (species) (LA df=l), (df=1)
(a6),k = effect of level i of A (ecosystem) by level k of C (species)
(LA df=l), (df=l)
random B,( = effect of level of B (site) nested within level i of A
(ecosystem) (df=8)
random 1(Jk) = effect of level of B (site) nested within A (ecosystem)
by C (species) interaction (df=8)
random e,kl = experimental error (LA df=16), (df=39)

Additional data collected on Serenoa repens was also analyzed (Eq. 3-7).

yi= [i + a, + Bj, + ei (3-7)
I = true mean
a, = effect of level i of A (ecosystem) (df=l)
random BJ( = effect of level of B (site) nested within level i of A
(ecosystem) (df=8)
random e,, = experimental error (df=20)

When interaction variables were not significant (p<0.1), they were dropped from

the model. All other tests were performed at a=0.05. All pairwise comparisons of fixed

means were performed using Tukey's test, rejecting the null hypothesis that there was no

difference between treatments when p<0.05.

Results

Ecosystem Differences

Leaf area

Based on the understory species studied, hardwood ecosystems had higher (p<0.05)

leaf area per leaf, specific leaf area, and leaf area per plant than flatwood ecosystems









(Table 3-1). Leaf area per plant was greater for Myrica cerifera and Serenoa repens in

hardwood ecosystems than flatwood ecosystems (p=0.0368) and there was no difference

in leaf area per gram between ecosystems. There was a significant interaction between

ecosystem and species on the leaf area per leaf (p=0.0298) for Myrica cerifera and

Serenoa repens. Upon further analysis, leaf area per leaf was greater in hardwoods than

flatwoods for Serenoa repens while Myrica cerifera was not affected by ecosystem type.

Moisture content

Hardwood ecosystems had significantly higher foliar moisture content during the

sampling period, than flatwood ecosystems (p=0.0003) (Table 3-2). The foliar moisture

content in understory species within flatwood sites was less variable than understory

species within hardwood sites (Table 3-2). However, the moisture content of Myrica

cerifera and Serenoa repens were the same between ecosystems (p=0.0771).

Volatile solids

Foliar volatile solids were not significantly different between ecosystems

(p=0.5913). However, species sampled within flatwood sites had much higher variability

in volatile solids than species sampled within hardwood sites (Table 3-2). There was also

no difference between the foliar volatile solids of Myrica cerifera and Serenoa repens

between ecosystems (p=0.2660).

Energy content

Total energy content per gram was significantly higher in flatwood ecosystems than in

hardwood ecosystems (p<0.0001) (4,944.59 calories per gram and 4,776.98 calories per

gram, respectively). However, total foliar energy content per plant was not significantly

(p=0.0569) different between ecosystems. There was no difference between ecosystem

type for total energy content per gram and total foliar energy content per plant









measurements of Myrica cerifera and Serenoa repens. Live foliage of Myrica cerifera

and Serenoa repens account for a mean 4,800,000 kilocalories per hectare in flatwood

ecosystems and 500,000 kilocalories per hectare in hardwood ecosystems, although the

direct ecosystem effect on the mean kilocalories per hectare released from the live foliage

of Myrica cerifera and Serenoa repens was not significant (p=0.0961).

Species Differences

Leaf area

Serenoa repens in hardwood ecosystems has significantly more leaf area per leaf

than Serenoa repens in flatwood ecosystems (p<0.05). Serenoa repens, in either

ecosystem type, had more leaf area per leaf than all other species examined in both

ecosystems (Table 3-1). Vaccinium myrsinites had a very low leaf area at 0.34 cm2, but it

was not statistically different from any species except Serenoa repens in either ecosystem

type.

Callicarpa americana had significantly higher specific leaf area (p<0.05) than

Myrica cerifera (in flatwoods), Vaccinium arboreum, Vaccinium myrsinites, Ilex glabra,

Lyoniaferruginea, Ilex opaca, and Serenoa repens (in either ecosystem type) (Table 3-1).

Serenoa repens in hardwood ecosystems also had significantly (p<0.05) higher leaf area

per plant than any other species (Table 3-1). Serenoa repens in flatwood ecosystems had

significantly (p<0.05) greater leaf area per plant than both Gaylussacia dumosa and

Vaccinium mysinities, which were significantly (p<0.05) lower than all other species.

Moisture content

Callicarpa americana had significantly higher foliar moisture content during the

sampling period than any other species studied (p<0.05). Vaccinium arboreum had









significantly higher foliar moisture content (p<0.05) than both Lyoniaferruginea and

Serenoa repens (in either ecosystem type) (Table 3-2).

Volatile solids

Vaccinium myrsinites had significantly higher foliar volatile solid content (p<0.05)

than Myrica cerifera (in flatwood ecosystems) and Serenoa repens (in either ecosystem

type) (Table 3-2). Serenoa repens (in flatwood ecosystems) had significantly (p<0.05)

lower foliar volatile solid content than Ilex opaca, Gaylussacia dumosa, Callicarpa

americana, Quercus nigra, and Myrica cerifera (in hardwood ecosystems).

Energy content

Both energy content measurements (calories per gram and kilocalories per plant)

had significant random interaction between site and species (p<0.10). By analyzing the

data individually for each species, the random effect of site (nested within ecosystem)

was significant (p<0.05) for Myrica cerifera and Serenoa repens in calories per gram and

Serenoa repens in kilocalories per plant. The overall site effect was not significant

(p>0.05) for either calories per gram or kilocalories per plant. Species was highly

significant (p<0.0001) for both measurements.

Total energy content per gram was distinctly different when comparing species

(Figure 3-1). Ilex glabra and Lyoniaferruginea had significantly (p<0.05) higher total

energy content per gram than all other species, followed by Gaylussacia dumosa,

Vaccinium myrsinites, and Ilex opaca. Myrica cerifera (in either ecosystem type) and

Vaccinium arboreum had significantly greater total energy content per gram (p<0.05)

than Serenoa repens (in either ecosystem type), Quercus nigra, and Callicarpa

americana.













Table 3-1. Leaf area per leaf, specific leaf area, and leaf area per plant for species in flatwood and hardwood ecosystems. Standard


error is given in parentheses (n=5 for species and n=
difference in Tukey's pairwise comparison between


-30 for ecosystem). Lower-case letters indicate significant (p<0.05)
species.


Species
Gaylussacia dumosa
Ilex glabra
Lyoniaferruginea
Vaccinium myrsinites
Myrica cerifera
Serenoa repens
Flatwood mean


cm2/leaf
6.46(1.24)
5.70 (0.27)
7.93 (0.92)
0.34 (0.09)
10.13 (2.69)
2597.14 (368.82)
437.95 (187.83)


cm2/gram
307.7 (23.9)
169.4 (14.2)
155.7(6.4)
182.4 (17.6)
237.1 (29.1)
108.3 (9.2)
193.4 (13.7)


cm2/plant
1,443.3 (269.4)
2,096.2 (831.7)
2,131.6 (323.9)
823.2 (452.3)
8,126.7 (1279.1)
19,959.4 (3129.6)
5,763.4 (1370.5)


Hardwood Callicarpa americana 59.96 (7.76) c 440.9 (95.1) a 2,744.9 (917.0) bc
Ilex opaca 41.71 (3.49) c 134.9 (22.4) b 12,334.9 (3887.4) bc
Quercusnigra 56.38 (19.08) c 268.1 (42.4) ab 3,829.7 (2037.9) bc
Vaccinium arboreum 7.75 (0.98) c 212.5 (47.4) b 4,847.7 (1635.0) bc
Myrica cerifera 11.64 (0.44) c 283.3 (56.9) ab 13,995.1 (7788.2) bc
Serenoa repens 4191.90 (557.02) a 135.4 (7.2) b 41,460.8 (8228.2) a
Hardwood mean 728.22 (299.82) 245.9 (27.7) 13,202.2 (3099.4) *


* indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems.


Ecosystem
Flatwood


J


J











Table 3-2. Foliar moisture content and volatile solids for species within flatwood and hardwood ecosystems. Standard error is given
in parentheses (n=15 for species and n=60 for ecosystem). Lower-case letters indicate significant (p<0.05) difference in
Tukey's pairwise comparison.

Ecosystem SDecies Moisture Content (%) Volatile Solids (mg/ka)
Flatwood Gaylussacia dumosa 124 (6) bcd 916,667 (9,644) ab
Ilex glabra 139 (10) bcd 826,000 (54,465) abcd
Lyoniaferruginea 119(12) cd 841,333 (54,632) abcd
Vaccinium myrsinites inites 133 (10) bcd 933,333 (17,284) a
Myrica cerifera 141 (16) bcd 812,000 (11,960) bcd
Serenoa repens 101(2) d 732,667 (15,352) d
Flatwood mean 126 (4) 843.667 (15.060)
Hardwood Callicarva americana 460 (60) a 900.667 (7.836) ab
Ilex opaca 144 (12) bcd 918,667 (5,845) ab
Quercus nigra 187 (25) bcd 874,000 (16,784) ab
Vaccinium arboreum 223 (14) b 837,333 (13,433) abcd
Myrica cerifera 216 (24) bc 856,667 (26,071) abc
Serenoa repens 113 (5) d 738,000 (9,962) cd
Hardwood mean 224 (17) 854.222 (8.569)
indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems.














a a


5400 -

5200 -

5000 -

4800 -

4600 -

4400 -

4200 -

4000 -



1500 -




1000 -




500 -




0
1.2e+7 -


1.0e+7 -


8.0e+6 -


6.0e+6 -


4.0e+6 -


2.0e+6 -


I rF-


Kcal per plant stems
per ha of individual
species at individual site
= foliar energy content
(kcal) per ha

b


V
1'


S.\ ,


Flatwood species


Hardwood species


Foliar energy content in calories per gram, kilocalories per plant, and
kilocalories per hectare for species within pine flatwood and hardwood
hammock ecosystems. Standard error is shown in error bars (n=15). Lower-
case letter indicate significant (p<0.05) difference in Tukey's pairwise
comparison


Calories per gram foliar
biomass (gram) per plant
* 0.001 kcal per cal=
kcal per plant






cd cd cd


5400


5200

5000

4800

4600

4400

4200

4000



1500




1000




500




0
1.2e+7


1.0e+7


8.0e+6


6.0e+6


4.0e+6


2.0e+6


Figure 3-1.


O'C'


S.









Serenoa repens in hardwood ecosystems had significantly higher foliar energy per

plant (p<0.05) than any other species, followed by Serenoa repens in flatwood

ecosystems (Figure 3-1). Ilex opaca had significantly greater foliar energy per plant

(p<0.05) than Callicarpa americana, Gaylussacia dumosa, and Vaccinium myrsinites.

When foliar energy content of species is examined on a per hectare basis, Serenoa

repens (in flatwoods), followed by Ilex glabra had higher kilocalories per hectare than all

other species (Figure 3-1). The species by site interaction was significant in the model

(p<0.0001) for energy content per hectare due to a significant random effect of site on

Gaylucassia dumosa, Vaccinium myrsinites, Myrica cerifera, and Serenoa repens (in

flatwood ecosystems) and Ilex opaca and Serenoa repens (in hardwood ecosystems).

Serenoa repens

Dead Serenoa repens foliage from hardwood ecosystems had higher moisture

content than dead foliage from flatwood ecosystems (p=0.0397) (Table 3-3). There was

no significant difference (p>0.05) between ecosystems for additional foliar measurements

of Serenoa repens. There was a significant random effect (p<0.05) of site nested within

ecosystem for live volatile solids, live and dead energy content, and kilocalories per

plant.

Energy content of dead foliage was generally lower than the energy content of live

foliage (Table 3-3). Foliar energy content per plant was was nearly doubled by

incorporating the dead foliar energy content into the calculations. In addition, potential

foliar energy content per hectare was nearly doubled.












Table 3-3. Additional data for Serenoa repens included measurements of moisture content, volatile solid content, total energy content
for dead foliage in addition to live foliage. Standard error is given in parentheses (n=15). Total energy released by both
live and dead foliar components are incorporated into the measurement of kilocalories per plant. Ecosystem effects were
not significant (p>0.05) for all measurements except deaf foliar moisture content.


Volatile Solids


Energy Content


dead foliar
moisture
content (%)


live foliage


dead foliage live foliage


dead foliage


kilocalories/plant kilocalories/hectare


Flatwood 13 (5) 732,666 734,000 4,638.92 4,538.42 2,00
(15,352) (11,944) (18.62) (21.82)
Hardwood 30 (5)* 738,000 722,000 4,702.24 4,612.24 2,22
(9,962) (9,522) (20.84) (37.04)
* indicates significant (p<0.05) difference in Tukey's pairwise comparison between ecosystems.


1.4(391.7)

2.8(400.1)


16,307,700
(5,341,485)
1,223,420 (370,829)


Serenoa
repens









Discussion

Ecosystem Differences

Most of the ecosystem effects on flammability presented in this paper, except leaf

area, are largely due to the differences in the species studied within each ecosystem type,

as the foliar flammability ofMyrica cerifera and Serenoa repens was not affected by

ecosystem type. Generally, the species that represent the flatwood ecosystem in this

study are more flammable than the species representing the hardwood ecosystem.

Results of leaf area were expected because of the typical physiological response of

leaves to shade. The hardwood sites had less variable and greater canopy closure;

therefore the understory plants were in a more shaded environment than the flatwood

sites (Table 2-2). It is largely recognized that leaves which grow in shady sites have high

specific leaf area (leaf area per gram) (Lambers et al. 1998). This was reflected in our

specific leaf area measurements as well in the leaf area per leaf and per plant. Certain

species may be more responsive to the influence of shade as indicated in the leaf area per

leaf difference of Serenoa repens between ecosystem types while Myrica cerifera did not

change between ecosystems. The higher leaf area in hardwood ecosystems could

increase the flammability of the understory species by increasing their ignitability.

Ignitability may be higher based on the increased surface area for moisture to be

evaporated from in a drought or more importantly, a nearby wildfire (Rundel 1981). It

would be valuable to know the leaf area to volume for the species studied to relate the

leaf structure more directly with flammability (Montgomery and Cheo 1971, Rundel

1981).

Moisture content of foliage within hardwood hammocks may be greater compared

to pine flatwoods. However, as this study indicates, this difference is likely due to the









different understory species between the two ecosystems as there was no direct effect of

ecosystem type on the foliar moisture content ofMyrica cerifera and Serenoa repens.

The more fire-prone Eucalyptus-Casuarina dry sclerophyll forests had species with lower

foliar moisture content than species in less fire-prone woodlands in Tasmania (Dickinson

and Kirkpatrick 1985). Foliar moisture content of species in fire-prone fynbos

ecosystems was lower than species in forest patches in South Africa (van Wilgen et al.

1990). In addition, Eriksson et al. (2003) found higher fuel moisture content in less fire-

prone Afromontane forest compared to the more fire-prone Acacia woodland. Moisture

content strongly influences the flammability of plant tissue, likely more than any other

characteristic (Gill et al. 1978, Rundel 1981, Etlinger 2001).

The consumability, measured as volatile solids, of leaf tissue between ecosystems

is more variable in pine flatwoods than hardwood hammocks, although there was not an

overall difference between ecosystems. This may be related to the higher variability in

ecosystem structure within sample sites (Chapter 2). The percent calorimeter ash

(percent of initial sample left after calorimetry combustion) is another measure of the

consumability of plant tissue. Rodriguez Afi6n et al. (1995) presented calorimeter ash

values of 0.16 to 2.86% (Eq. 3-8) and Nufiez-Regueira et al. (2001) presented calorimeter

ash values of 0.22 to 4.62%. These values are much lower (higher comsumability) than

the values reported through the volatile solid analyses in this study. For example, the

highest consumability was reported as 933,333 mg/kg, or 6.67%, for Vaccinium

myrsinites. It is possible that through isoperibol oxygen consumption calorimetry, a more

complete combustion is accomplished than accomplished with the method used in this

study. However, the methodology used in this study may more closely represent wildfire









conditions as plant tissue is combusted in an oxygen-rich environment within the

calorimeter.

calorimeter ash % = 100 (weight of final sample/weight of initial sample) (3-8)

Combustability on a gram basis was higher for foliage from understory plants in

flatwoods than hardwoods. These results are similar to the results of Dickinson and

Kirkpatrick (1985) and van Wilgen (1990) who reported higher energy content per gram

in foliage from more fire prone ecosystems (Eucalyptus-Casuarina dry sclerophyll and

fynbos, respectively) than less fire prone ecosystems (woodlands and forest patches,

respectively). This has been suggested as an adaptation by species in fire-prone

ecosystems to improve species survivability (Bond and Midgley 1995 and Possingham et

al. 1995).

Species Differences

Based on foliar flammability, clear distinctions can be made between the

flammability of species. Callicarpa americana is the least flammable species studied,

and Serenoa repens, Ilex glabra, Lyoniaferruginea, Myrica cerifera, and Ilex opaca are

the most flammable species. However, these highly flammable species are flammable for

different reasons. As mentioned before, the species found in flatwood ecosystems are

generally more flammable than those found in hardwood ecosystems.

It must be stated the energy content values expressed are solely a way to compare

species to one another, not an absolute measure of energy content. Calorimetry, as a

measure of total heat release, also does not distinguish between effective heat content

(flaming heat content) of the tissue and heat released from the residue (smoldering heat

content). In addition, oxygen consumption calorimetry is a complete combustion process

that likely overestimates the amount of energy that would be released from the same









sample in field conditions. Fuel and environmental conditions greatly affect the amount

of potential fuel that will be available fuel to a wildfire and how the fuel will combust

(Pyne et al. 1996). These conditions include: fuel moisture content, spatial arrangement,

time of day, wind speed, weather, climatic conditions, and other microclimate influences

of fire (Pyne et al. 1996).

Those species that have high foliar energy content per hectare may be target species

for fuel reduction strategies beyond defensible space in the WUI. These species include

Ilex glabra and Myrica cerifera in flatwood ecosystems, and Serenoa repens in either

ecosystem type. These species could be singled out to reduce the environmental impact

of fuel reduction (mechanical, chemical, or biological) while still reducing the overall

wildfire hazard of an area. However, by selectively removing individual species, the cost

of fuel reduction would likely be higher.

Gaylussacia dumosa

Gaylussacia dumosa had a high specific leaf area, potentially exposing more of the

foliar biomass to environmental extremes. However, this characteristic does not likely

change the overall flammability of this species. The species does contain relatively high

energy content per gram but because of the low foliar biomass, has low foliar energy

content per plant. Therefore an individual Gaylussacia dumosa plant is likely to be

highly flammable, but will not release much total heat.

Ilex glabra

The foliar energy content ofllex glabra is similar to those reported by Dickinson

and Kirkpatrick (1985) for foliar samples from Eucalyptus and Acacia species (Ilex

glabra had a mean energy content converted to 21,482 J-g-1). The foliar energy content is

high, increasing the amount of heat generated per gram of tissue consumed compared to









other species studied. Although the energy content per individual plant is comparatively

low, Ilex glabra grew so dense in the sites that Ilex glabra is a serious wildfire hazard to

structures. In addition, Ilex glabra foliage is known to be highly flammable because of

volatile extractives (Burgan and Susott 1991). Ilex glabra foliage contains 44.6% (based

on dry weight) total ether and benzene-ethanol extractives which contribute greatly to the

heat release of foliage (Shafizadeh et al. 1977). The substance present in Ilex glabra

foliage contributing to the production of volatiles is cutin (Rogers et al. 1986). It is also

possible that many volatile carbon compounds were incorporated into the total energy

content measured through calorimetry. Burgan and Susott identified low-temperature

volatiles as compounds becoming volatile up to 300C (1991), whereas foliage in this

study was dried at 700C.

Lyoniaferruginea

Lyoniaferruginea foliage contains as much total energy per gram as Ilex glabra. In

addition, the foliar energy content per plant is comparatively high. The foliar moisture

content of this species is low, further increasing the flammability of this species. Lyonia

ferruginea plants in pine flatwoods within the WUI can be a potential wildfire hazard if

they are found in dense patches. Maintenance of Lyoniaferruginea in a landscape near a

home should be to prune excess foliage and to not allow this species to be incorporated

into dense plantings.

Vaccinium myrsinites

Although the foliar volatile content and energy content of Vaccinium myrsinites is

relatively high, the low foliar biomass per individual reduces the hazard. Again, as with

other plants, dense plantings of this species would present a hazard.









Callicarpa americana

Callicarpa americana foliage contains high specific leaf area, similar to

Gaylussacia dumosa. However, the average foliar moisture content of this species was

far greater than any other species studied. In addition, Callicarpa americana foliar

energy content per gram, per plant, and per hectare was comparatively lower than all

other species studied. For these reasons, Callicarpa americana is the least flammable

species studied based on our results. This species has been cited as having low

flammability on extension lists (Florida Firewise Communities 2000, Monroe and Long

2001).

Ilex opaca

For those species studied only in the hardwood ecosystem-Callicarpa

americana, Ilex opaca, Quercus nigra, and Vaccinium arboreum-lex opaca is the most

flammable based on this study. The foliage had moderate moisture content, relatively

high energy content per gram, and high energy content per plant and per hectare. This

species has been cited as having high flammability by MacCubbin and Mudge (2002).

Quercus nigra

Foliage of Quercus nigra plants had a relatively high specific leaf area. However,

the foliar energy content was one of the lowest values in this study. Foliar energy content

per plant was also low. However, because the abundance of Quercus nigra in the

understory of the study sites can be high, this species may contribute a large amount of

energy content (heat) per hectare. Individual plants of Quercus nigra do not have

characteristics of highly flammable plants, but could pose a hazard to WUI structures

within dense plantings of small plants.









Vaccinium arboreum

As reported in Chapter 2, Ilex opaca and Vaccinium arboreum had similar biomass

results. However, in the analyses of foliar biomass, differences between the species are

more apparent. Vaccinium arboreum foliage contains less energy per gram which

translates to less potential energy released per plant. In addition, Vaccinium arboreum

does not contribute highly to foliar energy content per hectare. This species was listed as

having low-flammability by Monroe and Long (2001).

Myrica cerifera

Foliage of Myrica cerifera did not have any characteristics exhibiting high

flammability. The foliar moisture content, energy content per gram, and energy content

per plant were moderate. However, given the frequency ofMyrica cerifera listed as a

highly flammable plant (Lippi and Kuypers 1998, Monroe and Long 2001, MacCubbin

and Mudge 2002), there may be other characteristics of Myrica cerifera which were not

detected with this methodology. As the common name wax myrtle suggests, the waxy

substance on the leaves may be highly volatile. But a study by Burgan and Susott (1991)

found that Myrica cerifera foliage contained significantly less low-temperature volatile

compounds than either Ilex glabra or Serenoa repens. The morphology of Myrica

cerifera plants growing naturally in wildland areas has been observed to be very different

to the morphology of Myrica cerifera cultivated as a landscape specimen. These

morphological differences may contribute to the high flammability of Myrica cerifera in

a cultivated setting whereas Myrica cerifera in natural areas does not have high

flammability. More study of this plant species, and how flammability changes with

horticultural practices, is necessary to remove it from the lists as a highly flammable

plant.









Serenoa repens

Although the specific leaf area was not high compared to other species, Serenoa

repens leaves and plants have high leaf area, especially when located in a hardwood

ecosystem. The moisture content of Serenoa repens is comparatively low. The

abundance of dead foliage with very low moisture content (13% in flatwoods and 30% in

hardwoods) also makes Serenoa repens a hazard to WUI structures. The low heat release

per gram of dry weight of Serenoa repens foliage has been documented by Shafizadeh et

al. (1977). In addition, Serenoa repens contains only 13.1% ether and benzene-ethanol

extractives that did not change the heat release of foliage after being removed

(Shafizadeh et al. 1977).

Although the species has relatively low foliar energy content, the large amount of

biomass makes the total potential foliar energy release extremely high. This is especially

true when incorporating the dead foliar energy content into the total foliar energy content

per plant. Energy content per gram for dead foliage was lower than the energy content

for live foliage which was expected as dead foliage contains less volatile compounds than

live foliage (Mak 1988, Burgan and Susott 1991). It can be concluded that Serenoa

repens can be a threat to WUI structures in either ecosystem type which has been

recognized in many publications (Lippi and Kuypers 1998, Monroe and Long 2001,

MacCubbin and Mudge 2002).

Conclusions

Many of the chemical characteristics of fuel have shown seasonal variation:

moisture content (Pyne et al. 1996, Agee et al. 2002) volatile compounds (Bunting et al.

1983 and Owens et al. 1998), and energy content of forest wastes (Rodriguez Afi6n 1995

and Nfifez-Regueira et al. 2001). In addition, disturbance patterns may affect the









chemical composition of fuels, such as time since last fire (Thackston et al. 1982, Mallik

1994, Rieske et al. 2002). Monitoring how the chemical components of these species

change throughout the season or with different weather patterns may provide valuable

information.

We confidently conclude that plant species within flatwood ecosystems are likely

to be more flammable than species within hardwood ecosystems. We can also make

some general guidelines for species recommendations. For WUI structures in flatwood

ecosystem types, Lyoniaferruginea, Myrica cerifera, and especially Ilex glabra and

Serenoa repens should be discouraged from existing within defensible space zones. In

hardwood ecosystems, Ilex opaca and Serenoa repens should be discouraged from being

used in structural landscaping. Callicarpa americana is the species that, according to this

study, has low flammability and can be incorporated into firewise landscaping.














CHAPTER 4
CONCLUSIONS AND RECOMMENDATIONS

Conclusions

It first must be said that all plants are flammable. However, some plants are less

flammable than others and these species are appropriate for firewise landscaping. Less

flammable plants can be used with caution in arrangements that are aesthetically pleasing

with low wildfire hazard. Plant selection for landscaping, even within fire-prone

ecosystems, should be a balance of perceived benefits. These benefits include energy

conservation, water conservation, wildlife habitat, aesthetics, privacy, noise reduction,

and wildfire hazard reduction. It is also important to consider that choosing the right

plant for the right site will provide healthier, more vibrant landscape plants with less

water stress and less dead material. Maintenance is also very important as biomass can

be kept low and moisture content kept high with proper maintenance.

Based on the results from this study, we embrace Hypothesis I and II and reject

Hypothesis III.

S Hypothesis I: Understory plants in pine flatwoods are more flammable than
understory plants in hardwood hammocks.

Understory plants in pine flatwoods are more flammable than understory plants in

hardwood hammocks. We conclude that the ignitability (less distance between litter and

lowest branch, lower foliar moisture content), sustainability (higher fuel loading), and

combustability (total energy content) are higher for understory species within pine

flatwood ecosystems than hardwood hammock ecosystems. However, the consumability









(fine fuel biomass and volatile solids) was not different between pine flatwoods and

hardwood hammocks. These results are similar to other comparisons of plant

flammability between more fire prone ecosystems and less fire prone ecosystems (Table

4-1).

Table 4-1. Summary table comparing the results from flammability studies in Tasmania
(Dickinson and Kirkpatrick 1985), South Africa (van Wilgen et al. 1990), and
Ethiopia (Eriksson et al. 2003). YES= results agree with statement in left
column, NO= results do not agree with statement in left column, and na=
information not available.
The more fire-prone ecosystem Tasmania South Ethiopia SE United
contains... Africa States
more litter biomass na na YES YES
less total biomass na YES YES YES
less fine fuels na YES YES NO
more dense fuel bed na YES na YES
lower foliar moisture content YES YES YES YES
higher energy content YES YES na YES

* Hypothesis II: Differences in the flammability of species are significant.

Flammability differences among species were the result of differences in foliar

biomass, fine fuel biomass, foliar moisture content, and foliar energy content.

* Hypothesis III: The flammability of wax myrtle (Myrica cerifera) and saw
palmetto (Serenoa repens) are different between ecosystems.

By analyzing the data from Myrica cerifera and Serenoa repens separately, direct

effects of ecosystems on the flammability of individual species were explored. The

flammability characteristics were not different between ecosystems for these two species,

except for litter measurements. Therefore we have to conclude that flammability of the

same species in different ecosystem contexts is not different.

This result is based only on species adapted to each ecosystem type. The same

results are not predicted when comparing the flammability of a species adapted to a

hardwood hammock ecosystem placed around a home in a pine flatwoods ecosystem.









Also, the concept of wildfire risk must be addressed as plants of the same species within

pine flatwood ecosystems are more likely to be exposed to wildfire in the first place.

Therefore the flammability of a plant must also be evaluated within its context. Pine

flatwoods contain more litter biomass which accumulates to a deeper depth than

hardwood hammocks, therefore increasing the flammability of both Myrica cerifera and

Serenoa repens. In addition, the density per hectare ofMyrica cerifera and Serenoa

repens is greater in pine flatwoods increasing the hazard of these species within an

ecosystem associated with the WUI.

Recommendations

This section is devoted to recommendations based on the context of firewise

landscaping recommendations through the following assessments:

* Assessment I: What are the implications for firewise landscaping in these
ecosystems?
* Assessment II: In what way would individual plants contribute to structure
survival or destruction in a wildfire?

Pine Flatwoods

It is widely recognized that WUI areas within southern pine ecosystems have

higher wildfire hazard compared to WUI areas within hardwood ecosystems. This study

shows that, in addition to the natural fire regime, the individual understory plants within

the pine ecosystems are more flammable than those in hardwood ecosystems. Ilex

glabra, Lyoniaferruginea, and Serenoa repens should not be planted or maintained

within the defensible space zones of WUI structures. Ilex glabra and Lyoniaferruginea

are intrinsically highly flammable as they produce a great amount of heat once ignited.

Serenoa repens, however, is flammable based on the great amount of dead foliage and

biomass in general. In addition, fire specialists do not recommend Myrica cerifera within









WUI landscaping. Gaylussacia dumosa and Vaccinium myrsinites are acceptable within

WUI landscaping if not densely or continuously planted, and may provide food for

wildlife.

Hardwood Hammocks

Firewise landscaping techniques are not as necessary in hardwood hammock

ecosystems. In these ecosystems, WUI landscaping can be designed more to meet other

landowner needs than wildfire protection. However, it must be noted that within the

WUI, the likelihood of ignition may be higher because of increased arson or accidents.

Also, hardwood hammocks which are disturbed (fewer overstory trees, more dense

understory, and more dead fuel) may be more prone to wildfire, especially in an extended

drought. In addition, the presence of pine trees around a WUI structure, even within a

hardwood hammock can cause a hazard because of the fallen pine needles. Firewise

landscaping within hardwood hammocks, if necessary, does not need to be as stringent as

within flatwood ecosystems. This study suggests that Serenoa repens and Ilex opaca

should not be planted around homes within hardwood hammock ecosystems. However,

if properly maintained on a routine basis, individuals of these species may be used.

Proper maintenance includes removal of debris, thinning of foliage, and pruning of dead

and lower branches. Landscaping arrangement should include vertical and horizontal

separation of plants or landscape islands surrounding a WUI structure.

Future Research

Continued research on the social and economic dimensions of the WUI will

improve our understanding of the WUI. In addition, it would be valuable to know how

WUI residents decide what kind of landscaping and plants they maintain around their

homes. In this study we focused on native understory species as these plants would likely









exist within WUI in the South. However, people have a variety of choices for landscape

plants.

To make comprehensive lists of plant flammability, the author recommends a

formal survey of fire professionals within the South. This survey would attempt to

identify the most-flammable and least-flammable plants within specific biographical

regions based on the accumulated experience of people with knowledge of local flora and

wildfire. This experience would incorporate many more plant species within a variety of

environmental conditions. Field and laboratory studies could then address discrepancies

in the results.

Another method of determining flammability of plant species may be to standardize

a ranking process which would be easy to use throughout the South. For example,

species could be examined either through landscaping publications, species descriptions,

or through direct study to determine the potential level of hazard (Table 4-2). In this

case, a complete list of characteristics would be used in a standardized method. Based on

literature review and the results of this study, the author suggests that important

characteristics for evaluation include foliar moisture content, foliar biomass, and fine fuel

biomass. Additional tests could include either foliar volatile extractive analyses or foliar

energy content.

Table 4-2. Proposed ranking mechanism of plant species by flammability.
Plant Flammability Ranking Criteria
Low Flammability No hazardous characteristics
Moderate Flammability 1-2 hazardous characteristics that can be mitigated through horticulture
High Flammability 3 or more hazardous characteristics; characteristic of extreme risk; one or
more characteristic that cannot be mitigated through horticulture















APPENDIX A
LOCATION OF PINE FLATWOOD AND HARDWOOD HAMMOCK ECOSYSTEMS
BY COUNTY IN FLORIDA





























Figure A-1. Distribution of pine flatwoods by county in Florida adapted from Myers and
Ewel (1990). Dark gray areas represent counties where pine flatwoods exist.


Figure A-2. Distribution of hardwood hammocks by county in Florida adapted from
Myers and Ewel (1990). Dark gray areas represent counties where
hardwood hammocks exist.


u-i














APPENDIX B
LOCATION OF STUDY SITES


2 6-7


Collection Sites:

= hardwood site

= flatwood site

=both sites


1) ACMF(UF)
2) Twin Rivers S. F. (Ellaville Tract) (DOF)
3) Withlacooche S. F. (Richloam Tract) (DOF)
4) Little River Springs (SRWMD)
5) Welaka S. F. (DOF)
6) Osceola N. F. (USDA FS)
7) Osceola N. F. (USDA FS)
8) Steinhatchee (SRWMD)
9) Jennings S. F. (DOF)
10) Jennings S. F. (DOF)


Figure B-1. Location of pine flatwood and hardwood hammock study sites throughout
North Central Florida. Sites used were owned by the University of Florida
(UF), Florida Division of Forestry (DOF), Suwannee River Water
Management District (SRWMD), and USDA Forest Service (USDA FS).















APPENDIX C
ABSOLUTE AND RELATIVE DENSITIES OF UNDERSTORY AND MIDSTORY
SPECIES AT EACH SITE









Table C-1. Absolute (stems per hectare) and relative densities (%) ofunderstory species
in 8, 12.56m2 plots at each study site.
Ecosystem Site Species Stems/ha Rel.
density


Austin Cary M. F.


Osceola N. F.


Jennings S. F.


dead


3980.89


Gaylussacia dumosa 6070.86 4.5
Ilex glabra 94446.66 69.6
Lyonia lucida 23785.83 17.5
Myrica cerifera 199.04 0.1
Quercus nigra 99.52 0.1
Quercus pumila 895.70 0.7
Serenoa repens 4876.59 3.6


Vaccinium myrsinites
dead


1293.79
2786.62


Gaylussacia dumosa 2289.01 1.7
Hypericum
brachyphyllum 99.52 0.1
Hypericum crux-andrae 298.57 0.2
Ilex glabra 111564.49 82.5
Licania michauxii 199.04 0.1
Lyoniaferruginea 99.52 0.1
Lyonia lucida 199.04 0.1
Magnolia virginiana 199.04 0.1
Myrica cerifera 4279.46 3.2
Serenoa repens 8459.39 6.3
Un-identified 99.52 0.1


Vaccinium myrsinites
Aster walteri


4578.03
497.61


Befaria racemosa 1393.31 1.8
dead 8957.01 11.3
Hypericum crux-andrea 99.52 0.1
Hypericum reductum 99.52 0.1
Ilex glabra 56926.75 71.9
Lyonia ferruginea 5573.25 7.0
Magnolia virginiana 99.52 0.1
Myricacerifera 1393.31 1.8
Quercus pumila 398.09 0.5
Serenoa repens 3582.80 4.5


Un-identified


99.52


-I -I 4


Welaka S. F.


Clethra alnifolia


298.57


dead 8957.01 5.1
Gaylussacia dumosa 2886.15 1.6
Gordonia lasianthus 1492.83 0.8
Ilex glabra 130871.82 73.8


Flatwood


Lyonia ferruginea


3980.89









Table C-1. Continued
Ecosystem Site Species Stems/ha Rel.
density
Flatwood, Lyonia lucida 16023.09 9.0
cont. Welaka S. F., cont. Serenoa repens 11942.68 6.7
Vaccinium myrsinites 796.18 0.4
Hypericum prolificum 99.52 0.1
Ilex glabra 48865.45 58.5
Lyonia ferruginea 2089.97 2.5
Lyonia lucida 15824.04 19.0
Myrica cerifera 7364.65 8.8
Withlacoochee S. F. Quercus hemisphaerica 696.66 0.8
Quercus nigra 1990.45 2.4
Sassafras albidum 99.52 0.1
Serenoa repens 4578.03 5.5
Un-identified 99.52 0.1
Vaccinium myrsinites 1791.40 2.1


Twin Rivers S. F.


Carya glabra


99.52


dead 99.52 4.5
Ilex opaca 597.13 27.3
Quercus nigra 398.09 18.2
Serenoa repens 99.52 4.5
Vaccinium arboreum 298.57 13.6


Vaccinium corvmbosum


597.13


27.3


dead 99.52 1.7
Ilex glabra 597.13 10.3
Liquidambar styraciflua 199.04 3.4
Myrica cerifera 796.18 13.8
Osceola N.F. Quercus hemisphaerica 199.04 3.4
Osceola N. F.
Quercus nigra 2488.06 43.1
Serenoa repens 298.57 5.2
Vaccinium arboreum 298.57 5.2
Vaccinium corymbosum 199.04 3.4
Vaccinium stamineum 597.13 10.3


Jennings S. F.


Acer rubrum


99.52


dead 696.66 4.3
Gaylussacia dumosa 99.52 0.6
Ilex glabra 398.09 2.5
Ilex opaca 497.61 3.1
Lyonia ligustrina var.
foliosiflora 199.04 1.2
Magnolia virginiana 3184.71 19.6
Myrica cerifera 1393.31 8.6
Quercus hemisphaerica 1691.88 10.4


ur 5


Hardwood


Ouercus nigra


5672.77


35.0









Table C-1. Continued
Ecosystem Site Species Stems/ha Rel.
density


Jennings S. F., cont.


Steinhatchee,
SRWMD


Little River Springs,
SRWMD


Ouercus stellata


99.52


Serenoa repens 398.09 2.5
Vaccinium arboreum 1592.36 9.8


Vaccinium myrsinites
Acer rubrum


199.04
99.52


Crataegus marshallii 199.04 1.4
Cyrilla racemiflora 1990.45 14.3
dead 796.18 5.7
Diospyros virginiana 99.52 0.7
Ilex opaca 99.52 0.7
Lyonia lucida 4279.46 30.7
Myrica cerifera 1691.88 12.1
Quercus hemisphaerica 99.52 0.7
Quercus nigra 895.70 6.4
Ribus betulifolius 99.52 0.7
Serenoa repens 298.57 2.1
Un-identified 1492.83 10.7
Vaccinium arboreum 597.13 4.3
Vaccinium stamineum 497.61 3.6
Viburnum obovatum 398.09 2.9


Viburnum rufidulum
Bafaria racemosa


298.57
99.52


dead 497.61 6.4
Fabaceae 99.52 1.3
Ilex opaca 398.09 5.1
Myrica cerifera 199.04 2.6
Quercus hemisphaerica 99.52 1.3
Quercus nigra 1592.36 20.5
Serenoa repens 1293.79 16.7
Vaccinium arboreum 1094.75 14.1
Vaccinium myrsinites 298.57 3.8
Vaccinium stamineum 99.52 1.3


Hardwood,
cont.


Viburnum obovatum


1990.45


25.6









Absolute (stems per hectare) and relative density (%) for midstory species
(>3 m in height but <6.4 cm dbh) in 4, 400m2 plots at each study site.


Ecosystem Site Species Stems/ha Rel. density
Flatwood Pinus elliottii 6.25 5.3
Jennings S.F. Quercus hemisphaerica 87.50 73.7
Jennings S. F.
Quercus nigra 18.75 15.8
Quercus stellata 6.25 5.3
Magnolia grandiflora 12.50 18.2
Welaka S. F. Magnolia virginiana 43.75 63.6
Pinus elliottii 12.50 18.2
Pinus elliottii 125.00 71.4
Withlacoochee Pinus palustris 31.25 17.9
S. F. Quercus hemisphaerica 6.25 3.6
Quercus nigra 12.50 7.1


Twin Rivers
S.F.


Carva glabra


68.75


dead 6.25 0.7
Ilex opaca 156.25 17.7
Juniperus virginiana 6.25 0.7
Liquidambar styraciflua 18.75 2.1
Magnolia virginiana 37.50 4.3
Nyssa biflora 43.75 5.0
Quercus hemisphaerica 6.25 0.7
Quercus nigra 25.00 2.8
Vaccinium arboreum 368.75 41.8
Vaccinium corymbosum 118.75 13.5


Viburnum obovatum


25.00


Acer rubrum 6.25 1.3
Liquidambar styraciflua 31.25 6.3
Magnolia virginiana 6.25 1.3
Myrica cerifera 93.75 18.8
OsceolaN.F. Nyssa biflora 18.75 3.8
Pinus elliottii 12.50 2.5
Quercus hemisphaerica 50.00 10.0
Quercus laurifolia 6.25 1.3
Quercus nigra 256.25 51.3
Vaccinium arboreum 18.75 3.8


Jennings S. F.


Acer rubrum


43.75


Cornus florida 25.00 2.2
dead 18.75 1.6
Ilex opaca 18.75 1.6
Liquidambar styraciflua 12.50 1.1
Lyonia ferruginea 87.50 7.7
Lyonia ligustrina var.
foliosiflora 93.75 8.2


Manli i,"naa 62


Table C-2.


Hardwood


Ma-nolia virginiana


56.25









Table C-2. Continued
Ecosystem Site Species Stems/ha Rel. density


Jennings S. F.,
cont.


Steinhatchee,
SRWMD


Little River
Springs,
SRWMD


Myrica cerifera


25.00


Nyssa biflora 25.00 2.2
Pinus elliottii 6.25 0.5
Quercus hemisphaerica 25.00 2.2
Quercus nigra 312.50 27.5
Quercus stellata 56.25 4.9
Vaccinium arboreum 306.25 26.9


Viburnum obovatum
Acer rubrum


25.00
25.00


Crataegus marshallii 68.75 6.9
Cyrilla racemiflora 168.75 16.9
Ilex opaca 118.75 11.9
Liquidambar styraciflua 12.50 1.3
Myrica cerifera 125.00 12.5
Nyssa biflora 25.00 2.5
Quercus hemisphaerica 56.25 5.6
Quercus nigra 181.25 18.1
Ulmus americana 56.25 5.6
Vaccinium arboreum 100.00 10.0
Vaccinium corymbosum 6.25 0.6
Viburnum rufidulum 12.50 1.3


Viburnum obovatum
Acer rubrum


43.75
168.75


Crataegus marshallii 62.50 4.5
dead 6.25 0.5
Diospyros virginiana 6.25 0.5
Ilex opaca 18.75 1.4
Liquidambar styraciflua 37.50 2.7
Nyssa biflora 6.25 0.5
Nyssa sylvatica 25.00 1.8
Quercus larifolia 6.25 0.5
Quercus nigra 100.00 7.2
Quercus sp. 18.75 1.4
Ulmus americana 6.25 0.5
Vaccinium arboreum 643.75 46.6
Vaccinium corymbosum 18.75 1.4
Vaccinium stamineum 87.50 6.4
Viburnum obovatum 162.50 11.7


Hardwood,
cont.


Viburnum rufidulum


6.25















APPENDIX D
ABSOLUTE AND RELATIVE DENSITY, RELATIVE DOMINANCE, RELATIVE
FREQUENCY, AND IMPORTANCE VALUES FOR OVERSTORY SPECIES









Table D-1.


Etems per hectare, relative density, relative dominance, relative frequency,
and importance value for overstoy spec e


Ecosystem Site Species Stems/ Rel. Rel. Rel. I. V.
ha den. dom. freq.
Flatwood Pinus elliottii 87.50 51.9 55.9 40.0 147.8
Austin Cary
Ms Pinus palustris 75.00 44.4 41.0 40.0 125.4
Quercus nigra 6.25 3.7 3.1 20.0 26.8
Pinus elliottii 125.00 60.6 60.6 44.4 165.7
Osceola
N.F. Pinus palustris 75.00 36.4 36.4 44.4 117.2
Pinus taeda 6.25 3.0 3.0 11.1 17.2
Jennings Pinus elliottii 593.75 91.3 93.3 66.7 251.3
S. F. Pinus palustris 56.25 8.7 6.7 33.3 48.7
Magnolia
grandiflora 6.25 2.3 0.2 10.0 12.5
Magnolia
Welaka virginiana 18.75 6.8 6.8 20.0 33.6
S. F. Persea
palustris 6.25 2.3 1.0 10.0 13.3
Pinus elliottii 162.50 59.1 68.9 40.0 168.0
Pinus palustris 81.25 29.5 23.1 20.0 72.7
Pinus elliottii 306.25 79.0 83.1 50.0 212.1
Withla- Pinus palustris 50.00 12.9 14.9 12.5 40.3
coochee Quercus
S.F. hemisphaerica 6.25 1.6 0.2 12.5 14.3
Quercus nigra 25.00 6.5 1.8 25.0 33.3
Hardwood Caryaglabra 12.50 2.0 0.8 3.4 6.3
Ilex opaca 125.00 20.2 8.0 13.8 42.0
Juniperus
virginiana 25.00 4.0 1.3 6.9 12.2
Liquidambar
styraciflua 62.50 10.1 10.2 10.3 30.6
Nyssa aquatica 43.75 7.1 2.3 13.8 23.2
Nyssa biflora 12.50 2.0 1.1 3.4 6.5
Twin Rivers Persea
S.F. palustris 31.25 5.1 1.9 6.9 13.8
Pinus elliottii 6.25 1.0 2.3 3.4 6.8
Pinus glabra 6.25 1.0 1.0 3.4 5.4
Quercus nigra 218.75 35.4 26.1 13.8 75.2
Quercus
virginiana 37.50 6.1 44.1 10.3 60.5
Un-identified 12.50 2.0 0.5 3.4 5.9
Vaccinium
arboreum 25.00 4.0 0.5 6.9 11.5
Osceola
N.F. Ilex opaca 18.75 5.3 0.6 9.1 15.0