Citation

## Material Information

Title:
Monitoring and Assessing Longleaf Pine Ecosystem Restoration: A Case Study in North-Central Florida
Creator:
RASSER, MICHAEL K
2008

## Subjects

Subjects / Keywords:
Ecology ( jstor )
Ecosystems ( jstor )
Land management ( jstor )
Land use ( jstor )
Saplings ( jstor )
Seedlings ( jstor )
Shrubs ( jstor )
Species ( jstor )
Understory ( jstor )
Vegetation ( jstor )
Watermelon Pond ( local )

## Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Michael K Rasser. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
9/9/1999
Resource Identifier:
53345023 ( OCLC )

Full Text

MONITORING AND ASSESSING LONGLEAF PINE ECOSYSTEM
RESTORATION: A CASE STUDY IN NORTH-CENTRAL FLORIDA

By

MICHAEL K. RASSER

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

by

Michael K. Rasser

ACKNOWLEDGMENTS

There are many people I would like to thank for their contributions to this

manuscript. Loukas Arvanitis was a true mentor throughout my graduate student career

here at the University of Florida. My committee members, Wendell Cropper, Alan Long,

and George Tanner, always provided excellent advice and guidance in their particular

areas of expertise.

This project required extensive field data collection, in very hot and trying

conditions. Without David Zabriskie, Christine Housel, Nik Chourey, Willie Wood,

Brett Jestrow, and Louise Lundberg, this work would not have been possible. My co-

workers here in the Forest Information Systems Lab, Balaji Ramachandran, Steve Moore,

and Douglas Shipley, were essential in providing logistical support, especially when it

came to using GIS and GPS.

The Florida Division of Forestry provided the funding for this project. I would

especially like to thank Charlie Marcus, DOF State Lands Coordinator, for his generous

support and help throughout this project. Mpower3/Emerge donated high-resolution

aerial photography, which proved to be invaluable in the field. Robin Boughton, and the

Goethe State Forest Staff, were always very supportive.

Throughout my research there were many people who provided their personal

knowledge, especially Brian Olmert, President of the Loncala Corporation; Nancy Coile,

Division of Plant Industry, DOACS; Michael Drummond, Alachua County

Environmental Protection, and Dr. Doria Gordon, Department of Botany, UF/The Nature

Conservancy.

Of course, I have to thank my friends and family who provided endless moral

support. Soumya Mohan tirelessly reviewed this manuscript, and she and Arjun Mohan

tolerated my long physical and mental absences, for which I am greatly indebted.

page

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

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

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

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

CHAPTER

1 TH E PR OBLEM .................. ............................ .. .......... .............. .

Study O objectives ............................................. 1
O u tlin e ............................................................................ . 2

2 LITER A TU R E R EV IEW .............................................................. ....................... 3

History of the Longleaf Pine Ecosystem ............. ............................... ...............3
Sandhill E cology.................................................. 3
Restoration Ecology................ ...... ..................... .................. 4
Historical Inform ation in Restoration Ecology................................. .....................4
Sources of Historical Information ................... ............... ....................5
A erial Photography ............................................................... .. ................ .5
G general L and O office Survey............................................ ........... ............... 5
O their Sources of Inform ation............................................ ............... ............... 5
Challenges for Longleaf Pine Ecosystem Restoration ..............................................6
Ground Cover ............................. .......... .................. ..........
F ire .............................. .... ................................. . 7
M monitoring and A ssessing Restoration Success ........................................ .................8
What is Restoration Success................................................... 9
How is Restoration Success Monitored and Assessed .........................................9
Reference Sites ............. ............................... ............. .. ....... 10
V vegetation Sam pling .......................................... .. .. .... ....... .. .. .... 10
C om m unity M onitoring...................... .............. ....................... ............... 10
M easu ring V eg etation ........................................... ........................................ 10
Percent foliar cover ........................................... ...... .. .. .............. .. 11
D en sity ................................................................ ........ .... ...... 11
Frequency ......................................... .................. .... ........ 12

V

B asal area .............. .................................. ................. 12
Important Considerations for Sam pling................................... ....................... 12
Permanent or Temporary Sampling Locations .................................................12
N um ber of Sam ple U nits ......................................................................... ..... 13
S am ple D esign ...................................... .............................. ................. 14
Sim ple R andom Sam pling ................................................ .................. ...............14
Stratified R andom Sam pling ........................................ .......... ............... 15
Systematic Sampling ........... ................................................. ........ 15
Statistical Analysis of Plant Communities ............ ............................ ...........15
Indicators .......... ..... ......... ........................16
M easu res of Sim ilarity ........................................ ........................................ 16
Classification ............. .... ...... ..................................... ... ..... 17
O rd in atio n ........................ ..... ................................................. ............. 1 7
Statistical Analysis Based on Re-sampling .............. ................................ 17

3 STUDY AREA ................ ............. ............. .. ........ ........ ..... 19

L o catio n ......... ..... ............. ...................................... ..............................19
P h y sio g rap h y .................................................... ................ 19
L an d U se .................. .................................................... ............... 19
B biological C om m u nities ......................................................................................2 0
S an dhill U plan d L ak es ............................................................. .....................2 0
S a n d h ill ......................................................................................2 2
S c ru b ................. .................... ........................................................................ 2 3
H ardw ood H am m ock................................................. .............................. 23
P ra irie ...............................................................2 3
Ephemeral Pond..................... ..................................24
M anagem ent C considerations ........................................................... .....................24

4 M E T H O D O L O G Y ............................................................................ ................... 25

V egetation D ata C collection .............................................................. .....................25
G rou n d C ov er ...............................26.............................
Shrubs, Seedlings and Saplings........................................ ....................... 26
T re e s ............................................................2 7
Species Sampled .............. ................................. ....... .. ......... 27
Classification of Comparative Study Areas .................................. ...................27
N onreference A reas .............................................. ................ ...... 27
Site 4 .................................... ........................... .... ...... ........ 28
S ite 5 ............................................................................. 2 8
S ite 6 ............................................................................. 2 8
S ite 7 ............................................................................. 2 9
R reference Sites .....................................................................29
Field D ata Interpretation .................................................. ............................... 30
Im portance V values ............................................................... .. ......... ... .... 3 1
B asal A rea of Trees ............................................... .. ...... ................ 31
Sim ilarity A analysis ................................................ .... .... ... ....... .. ..31

Determination of Targets for Monitoring............................ .....................32
Proportional Abundance of Species ....................................... ............... 33

5 RESULTS AND DISCU SSION ........................................... .......................... 35

C o m p ariso n S ite s .................................................................................................. 3 5
U n d e rsto ry ..................................................................................................... 3 5
Shrub s ............................................................................ ....................35
Tree Seedlings and Saplings................................ ................... 36
Sum m ary of Sim ilarity Indices................................. ................. 36
Su gg ested T arg et Sp ecies ..................................................................................... 36
Understory .......................................... .......... 37
Wiregrass (Aristida beyrichiana) ........................................ ......37
Centipede grass (Eremochloe ophiurides) ....................................................38
Dog-fennel (Eupatorium capilifolium)...................................38
G reen briar (Sm ilax spp .) ....................................................... 38
T o tal fo rb co v er ...................................................................................... 3 9
Shrubs ............................................................39
Winged sumac (Rhus coppallinum) ........................................ ....39
Sand blackberry (Rubus cuneifolius)..................................... ......39
Florida rosemary (Ceratiola ericoides) ................................................ 40
Tree Seedlings and Saplings................................ ................... 40
T u rk ey o ak ...........................................................4 1
Black cherry (Prunus serotina) ................................. .. ................. 41
L au re l o a k ............................................................................................... 4 1
T re e s ..................................................................................................... 4 1
Longleaf pine .................. ................. .................... ............... 42
T u rk ey o ak ........................................................................................4 2

6 SUMMARY AND CONCLUSIONS ............................................ .................43

Suggestions for Future R research ....................................................... ............... 43
Further Exploratory Analysis of Field Data ..................... .................43
Determination of Target Abundances................. ........ ...............44
Effect of Centipede Grass on Longleaf Pine Restoration........................................44

ABSTRACT

A SPECIES SIM ILARITY ......... ......................... .....................45

B PROPORTIONAL DISTRIBUTION OF SPECIES AMONG REFERENCE AND
NONREFEREN CE SITES ............. ..................... ..................................... 48

C SPECIES RAN K IN G ............. ..................... ..........................................51

D IM PORTAN CE V ALUE S.......................................... ............... ............... 55

E TR EE B A SA L A R E A ............................................................... ............... ..... 58

L IST O F R E F E R E N C E S ........................................................................ .. ....................59

B IO G R A PH IC A L SK E T C H ...................................................................... ..................63

LIST OF TABLES

Table page

A-i Understory similarity among comparison sites based on Jaccard's Index ..............45

A-2 Understory similarity among comparison sites based on Sorensen's Index............45

A-3 Understory similarity among comparison sites based on Curtis Bray Index...........45

A-4 Shrub similarity among comparison sites based on Jaccard's Index .....................46

A-5 Shrub similarity among comparison sites based on Sorensen's Index...................46

A-6 Shrub similarity among comparison sites based on Curtis Bray Index....................46

A-7 Tree seedling and saplings similarity among comparison sites based on Jaccard's
coefficient............................................................................................. .47

A-8 Tree seedling and saplings similarity among comparison sites based on
Sorensen's coefficient ....................... .................. ................... .. ......47

A-9 Tree seedling and saplings similarity among comparison sites based on Curtis
B ray coefficient. ........................................................................47

B-1 Distribution of understory species among reference and nonreference points. .......49

B-2 Distribution of shrub species among reference and nonreference points.................50

B-3 Distribution of tree seedling and sapling species among reference and
nonreference points. ............................. .............. ......................... 50

C-1 U nderstory species ranking. ............................................. ............................ 52

C -2 Shrub species ranking ...................................................................... ...................53

C-3 Tree seedling and sapling species ranking. ................................... ............... 54

D-l Understory species importance value.............................................. ...............56

D-2 Tree seedling and sapling importance value. ................................ ..................57

D-3 Shrub species importance value. ........................................ ......................... 57

E-1 Overstory tree basal area (ft2/acre)................................ ......................... ....... 58

x

LIST OF FIGURES

Figure pge

3-1 Study area location ................................... ......... ............ .............. ..21

4-1 Sam pling architecture...................................................................... ...................26

4-2 Com parative study areas. ............................................... .............................. 30

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

MONITORING AND ASSESSING LONGLEAF PINE ECOSYSTEM
RESTORATION: A CASE STUDY IN NORTH-CENTRAL FLORIDA

By

Michael K. Rasser

August 2003

Chair: Loukas G. Arvanitis
Major Department: School of Forest Resources and Conservation

In a cooperative project with the Florida Division of Forestry, a methodology was

developed to monitor longleaf pine restoration. This project may serve as a model for

future management of the sandhill longleaf pine community at Goethe State Forest's

Watermelon Pond Unit (WPU).

A permanent network of sample points was established throughout the 4,778 acres

(1934 ha) of the WPU. Baseline data were collected on vegetative structure and

composition at the understory, midstory, and tree levels. Data collected included cover,

density, and frequency, of over 80 species. Sample points were permanently marked in

the field so that they could be relocated in the future for monitoring changes over time.

A series of comparison areas were identified throughout the property, that were

representative of the various ecological conditions of sandhill longleaf pine community,

based on anthropogenic influences. These sites include areas that had been severely

disturbed in the past, as well as reference sites representing desirable future conditions.

Site selection was based on available information regarding the presence of past

disturbances, such as timber production, farming, and grazing.

A statistical analysis was conducted to compare all study areas. The analysis

included determining importance values of the different species as well as measuring

similarity among sites. Species were ranked to determine which ones were responsible

for the greatest difference among the restored and reference sites. From this list, a series

of indicator species was chosen to allow land managers who are concerned about future

monitoring to measure the level of success of their restoration efforts.

CHAPTER 1
THE PROBLEM

Florida legislation, such as Preservation 2000 and Florida Forever, has placed a

large amount of land into public ownership for ecological restoration. Land managers

who are challenged with the task of restoring dynamic ecosystems must have a thorough

understanding of the current state of the vegetation and the desired future conditions in

order to establish realistic management goals. However, vegetation monitoring programs

are often hampered by the high cost of collecting, analyzing and interpreting field data.

Restoration of the longleaf pine (Pinuspalustris) sandhill community at the

Watermelon Pond Unit (WPU) of Goethe State Forest provides an example of such a

project. Once part of an extensive longleaf pine ecosystem, it has since been degraded

through anthropogenic influences such as clear cuts, mining, farming and ranching.

Current land managers are charged with the difficult task of returning this land to its

earlier, more natural state.

Study Objectives

This pilot study aims to address some of the challenges associated with monitoring

and measuring restoration success of the longleaf pine community at the WPU. The

specific objectives were to

* Establish a protocol for a statistical sampling design that allows adaptive
management of restoration success.

* Collect baseline field data.

* Conduct a comparative statistical analysis of the vegetative condition of the
longleaf pine community based on past land use.

* Develop a list of indicator species for future monitoring.

Outline

Chapter 2 reviews the literature on the science of restoring the longleaf pine

ecosystem. It focuses on the ecology of the longleaf pine ecosystem, obstacles to

restoration, and important aspects of sampling and monitoring such plant communities.

Chapter 3 describes the study site, the Watermelon Pond Unit of Goethe State Forest in

Florida. Chapter 4 outlines the methodologies used in our study and the rationale behind

their use. Chapter 5 examines the results obtained from using these methods. Chapter 6

summarizes the results with a discussion of the various indicators chosen for monitoring

restoration success. Suggestions are also made for future research and analysis related to

this area.

CHAPTER 2
LITERATURE REVIEW

History of the Longleaf Pine Ecosystem

Longleaf pine was once the dominant tree species throughout 24.3 million ha of

forest land in the southeastern United States (Outcalt 2000). A small fraction, perhaps as

little as 1.2 million ha, remains today (Hedman et al. 2000). The vast majority of this

forest is fragmented secondary growth that regenerated after the railroads expedited the

logging and clearing of these forests (Croker 1987). The largest remaining tract of old

growth longleaf is only about 200 acres (81 ha) in size (Johnson and Gjerstad 1998).

The degradation of this ecosystem may be attributed to a variety of factors including land

conversion, fire suppression, logging, and preference for other pine species in

silvicultural practices (Croker 1987, Gilliam and Platt 1999).

Sandhill Ecology

Large differences exist in the vegetative structure and composition of the longleaf

pine ecosystem. These may be attributed to fire regimes, physical soil variability,

anthropogenic soil disturbances, and geographic location (Myers and White 1987,

Rodgers and Provencher 1999). Myers (1990) lists clayhills, turkey oak (Quercus

laevis) barrens, longleaf pine/turkey oak, and upland pine forests among the many plant

communities that have been described for the longleaf pine ecosystem.

Sandhill pine, also referred to as High Pine, is a typical plant community on the

sandhill ridges of peninsular Florida. This community typically contains a sparse canopy

dominated by longleaf pine, low-stature midstory composed of xerophytic oaks and

clonal shrubs, and a herbaceous ground cover dominated by wiregrass (Aristida

beyrichiana) (Brockway et al. 1998, Myers 1990, Rodgers and Provencher 1999).

Long periods of fire suppression increase the importance of hardwood species and

decrease the forbs and graminoids component of the ground cover (Rodgers and

Provencher 1999). Sandhill diversity may decrease after fire suppression. However,

high levels of regional variation have made it very difficult to understand the community

dynamics of these systems.

Restoration Ecology

Ecological restoration is an emerging science, tracing its roots to the arboretum at

the University of Wisconsin, Madison (Jordan et al. 1987). It was there that a civilian

conservation group undertook a painstaking restoration of a prairie on several acres of

farmed land. The project was a landmark attempt to reverse damages that had been

caused by anthropogenic disturbances, and to return the land to an earlier, ecologically

more natural state.

Although the desired objective of ecological restoration may vary, the idea of

returning the land to historically natural conditions is ubiquitous in all restoration

projects. For example, in the case of reclaimed mine lands, the desired result may simply

be to vegetate an area in order to prevent erosion. However, more emphasis is now being

placed on overall ecosystem restoration, which involves a comprehensive process of

returning an ecological system to its naturally functioning state.

Historical Information in Restoration Ecology

One difficulty in restoring natural systems is the lack of knowledge about past

conditions. This is particularly evident in plant communities where few representative

examples still exist. To establish realistic management objectives, managers need

Sources of Historical Information

Aerial Photography

Aerial photography is a valuable tool for assessing past conditions of land by

providing a record of anthropogenic and natural disturbances at any given point in time.

The United States National Archives maintain a large collection of aerial photographs

taken before 1950. Local libraries, tax assessment offices, county offices and agricultural

units often make aerial photographs available to interested individuals, students, and land

managers.

General Land Office Survey

The United States General Land Office was responsible for creating land township

plat maps for the entire country. This survey was primarily completed in the mid 19th

century. In addition to installing survey markers at property corners, contractors usually

recorded the condition of the property (Wolf and Brinker 1994). These notes are often

readily available. For example, the Florida Department of Environmental Protection has

electronic versions of the original survey notes available online (FDEP 2003).

Other Sources of Information

Landowners and others intimately associated with an area may be valuable sources

of information regarding past land use and conditions of the land. For example, Brian

Olmert, President of the Loncala Corporation (High Springs, Florida), provided extensive

details on the company's prior land uses such as extent of cattle grazing, location and

details of timber operations, as well as personal observations regarding past ecological

conditions of the land. These details are extremely useful when assessing current land

conditions and establishing future restoration goals.

Challenges for Longleaf Pine Ecosystem Restoration

Restoration of the longleaf pine ecosystem is a challenging task for a variety of

reasons. The two major ones are returning the natural fire regime to the land and

recovering adversely affected ground cover (Glitzenstein et al. 1995). Fairly extensive

research has been conducted in recent years to surmount these obstacles.

Ground Cover

With the exception of tropical rainforests, longleaf pine forests are among the most

diverse ecosystems, sometimes containing over 40 species of plants in a square meter

(Hedman et al. 2000). However, this diversity poses additional problems in restoring this

unique ecosystem. Although much is known about planting trees, sufficient evidence is

not available about restoring the diverse understory (Horton 1995).

The presence of wiregrass is probably the most important component of the

understory in these forests (Brockway et al. 1998). In fact, longleaf pine and wiregrass

may be considered keystone species in the longleaf pine forest (Brockway and Outcalt

1998). The reason is that wiregrass both facilitates, and is tolerant of fire. An abundance

of wiregrass creates a continuous layer of fine fuels in the understory that encourages and

sustains frequent low intensity fires that were historically part of the fire regime.

However, there is a lack of knowledge about the historical extent of wiregrass, which is

very sensitive to soil disturbance (Rodgers and Provencher 1999).

During restoration, the actual planting of longleaf trees is relatively inexpensive.

Due to the economic value of the species, a fair amount of information is known about

how to plant and raise the seedlings to maturity. The difficulty lies in reconstructing the

diverse understory to include wiregrass and other important species. The Nature

Conservancy is currently conducting an ambitious restoration of a longleaf pine

ecosystem on a sandhill in northwest Florida (Seamon 1998). The project seeks to

restore both longleaf and the wiregrass dominated understory. Seeds are collected from

nearby areas that contain wiregrass, and are distributed in the area being restored.

Initially, manual methods were used to collect the seeds of wiregrass. This method was

very expensive (about $3000 /acre). As a result, more cost-effective mechanical methods are being developed. The seeds are now distributed with a commercial hay blower. More research is needed to improve and refine these methods so that they may be applicable to a larger scale. Fire The most important, yet missing, disturbance in many current or potential sites of longleaf restoration, is fire, which is necessary to maintain forest diversity and structure. Fire suppression leads to an increase in hardwood trees and shrubs and eventually reduces plant diversity (Glitzenstein et al. 1995). To address this issue prescribed fire has become an invaluable tool for returning natural fire regimes (Long 2002, Wade 1988). However, due to smoke, local residents are often concerned about the use of prescribed fires (Monroe et al. 1999). In addition, there are dangers of bums escaping control and causing health problems and damage to nearby lands. As more people move into rural areas, the smoke shed, which is the area affected by smoke during a fire, impacts more people. Also, a return to a natural fire regime after decades of fire suppression, can be ecologically damaging (Baker 1992). Build up of fuel loads up can cause fires to bum more intensively than they would have when there were frequent low intensity fires. This unnatural fuel loading can create intense fires that may cause mortality in plants and animals that would have otherwise survived. An alternative approach to removing hardwood trees that are encroaching in areas where longleaf pine is under restoration is the use of herbicides. Recent study have used the herbicide hexazinone to eliminate the hardwoods (Brockway et al. 1998, Provencher et al. 2000). Broadly applied, the herbicide killed a diverse amount of plants and was not very effective in restoring the natural vegetation present in the understory. However, when applied more selectively in a spot application, particularly at levels of 2.2kg/ha, it was effective in eliminating hardwoods and allowing wiregrass to regain dominance. Monitoring and Assessing Restoration Success Restoration ecology is an experimental science. Much is to be learned about restoring longleaf pine communities. Long term monitoring allows restoration efforts to take place within the framework of a scientific experiment. The success or failure of management methods can be tested and assessed. This allows for adaptive management and may lead to a better understanding of the ecological dynamics of the system. Monitoring is conducted for two primary purposes: a) to allow land managers to monitor the relative success or failures of their efforts, and b) for adaptive management purposes. In order to accomplish these objectives, monitoring efforts must establish criteria for determining restoration success. For example, land managers may wish to reduce hardwood tree dominance in a longleaf community. Monitoring would allow land managers to examine the effects of prescribed bums on reducing hardwoods. If it were determined that a change in prescribed fire frequency or seasonality was more effective, management could be adapted accordingly. Whether or not a restoration plan fulfills all its desired objectives, valuable knowledge can be attained. Ecological restoration projects allow for basic research on how various components of ecosystems and plant communities function and interact. This is accomplished through a mechanistic approach to the restoration process. According to Jordan et al. (1987) ". .. one of the most valuable and powerful ways of studying something is to attempt to reassemble it, to repair it, and to adjust it so it works properly." Monitoring efforts allow us to measure the effects of our tinkering with the ecological communities. What is Restoration Success Restoration success is a relative term that largely depends upon the stated objectives. Restoration objectives can be clearly stated prior to the start of the project, such as to increase the population of a particular species, or group of species. For example, a wetland restoration objective may be to increase the number of wading birds. Sometimes goals are vague, such as 'the restoration of historic ecosystem conditions'. Ill defined projects run the risk of not being successful due to the inherent complexity of natural systems. Therefore, a restoration project must have clearly defined goals in order to establish measures of restoration success. How is Restoration Success Monitored and Assessed Monitoring vegetative community structure is one ways to assess restoration success. Restoration success is usually measured in relation to one or more representative reference sites that represent the ideal future condition of the system being restored. A variety of univariate, multivariate, and qualitative methods are used to compare and monitor areas being restored with reference sites (Elzinga et al. 2001). Reference Sites Choosing the correct reference site or sites is an important part of any monitoring program for restoration efforts. Reference sites should represent the ideal future conditions of the site that needs to be restored. In the case of ecosystems that have been reduced in extent, finding an ideal representative site is difficult. In addition, ecological communities can vary widely primarily, due to edaphic factors. Therefore, it is essential that reference sites be located as near as possible to the potential restoration site. Once the reference site, or sites have been chosen, monitoring restoration progress is accomplished by comparing the reference site(s) with those that are being restored. Monitoring efforts are usually administered at the ecological community level. As such, they focus upon the assemblage of plants and or animals at a particular site. Vegetation Sampling Community Monitoring Monitoring plant communities can be accomplished through statistical sampling. Vegetation data are collected from a representative sample of plants. Characteristics that may be measured include: foliar cover, density, frequency, and basal area. Following vegetation measurements, statistical analyses can be used to characterize and describe the plant community. In addition, there are a variety of indices that may be used to assess the similarities among reference sites and the sites to be restored. Measuring Vegetation There are a multitude of methods for measuring vegetation (Bonham 1989, Phillips 1959). The variables measured during the course of our study and discussed below include percent cover, density, and basal area. Each of these variables, including some of the associated sampling errors, are discussed below. Percent foliar cover Percent cover is one of the most commonly used measurements. Its popularity is probably attributed to the ease with which it can be ocularly estimated. One of the most common method for conducting this type of sampling is to use a fixed area quadrat within which species are identified along with their percent cover. However, there are many errors that must be taken into account when using ocularly estimated measurements of vegetative cover. Kennedy and Addison (1987) conducted a study that quantified some of the biases in such measurements. The authors found that species morphology, distribution of species, and species identification, were the most common sources of error. Measurements of plants with compound leaves, proportionally large amounts of woody material, and variable leaf size were more variable than others. Rare species or inconspicuous species are often missed because they are widely dispersed. In addition, misidentification of species accounted for a large amount of errors. Another method for assessing vegetative cover is the line-intercept method, which compares the length of each species along a line with total line length (Phillips 1959). This methods was not used in our study because it was found to be less efficient to permanently locate a line, as compared to a quadrat, due to an abundance of shrubs at the study site. These non-sampling errors can be reduced in several ways. Having more than one observer examine each quadrat may reduce the error due to species being identified incorrectly or from simply missing a plant. In addition, if the same work team is used throughout the sampling process, biases among individuals can be reduced. Density Density is the number of individuals per unit area. In vegetation sampling, density is usually determined by utilizing fixed area plots. Plots of all sizes and shapes are used. However, the ability to detect plants, and boundary errors, are important considerations for selecting appropriate size, shape, and number of plots (sample size). Quadrats that are too large may prevent the observer from locating all species. A significant source of error associated with using quadrats for measuring density occurs at the edges, as plants may falsely be identified as being in or out of the quadrat. Frequency Plant frequency may be defined as the percentage of total plots within which a species of interest can be found (Phillips 1959). Frequency is solely determined by the presence of the species, without regard to abundance. Obviously, larger quadrats may have a greater chance of including plants, and as such, frequency is strongly affected by plot size. Basal area Basal area is the cover of a plant's stem expressed as square feet per acre or square meters per hectare. The basal area of trees can be estimated in several ways, one of the most common being a properly calibrated prism. As opposed to fixed area plots, prisms deal with probabilities proportional to diameter at breast height of trees. The procedure is known as point sampling. Larger trees have a greater probability of being included in point sampling than smaller trees. Shrubs and other obstructions can obscure the observer's view of the tree stem, reducing the ability to detect the critical distance to determine whether a tree is included in a plot (Reed and Mroz 1997). Important Considerations for Sampling Permanent or Temporary Sampling Locations One decision that must be made when establishing a sampling protocol is whether the sample units should be located permanently or temporarily. If plots are temporary, it will not be possible to assess changes over time as easily as with permanent plots, points, transects or strips. This is because permanent sampling units eliminates the confounding effects of spatial heterogeneity inherent in plant populations. Overall, the main advantage of permanent sampling units is that it is easier to statistically prove changes over time than with temporary samples (Elzinga et al. 2001). The primary reason for not always using permanent sample units is the higher cost compared to temporary samples. Devices for permanent sample units include metal wire, posts, and other materials, all of which contributes substantially to the cost of the project. The advent of the Global Positioning System (GPS) in the 1990's has led to an increase in the cost effectiveness of utilizing permanent sampling units. GPS was developed by the U.S. Department of Defense to allow accurate location of nuclear submarines (Hum 1989). It has since gained widespread use for civilian applications. The system is based on a network of 24 satellites orbiting the earth at an altitude of 12,600 miles (Lewis 1998). Using a method similar to tri-lateration, distance from the ground and the orbital location of three or more satellites is used to accurately locate positions on earth. During the course of our study GPS reliably placed sample units less than one-meter from their intended location. Number of Sample Units The number and size of sampling units is an important consideration for any sampling design. The most important consideration when determining sample size is the objective of the sampling (Elzinga etal. 2001). A larger sample size increases precision, however the gain in precision lessens as sample size increases. Collecting samples is generally very expensive; therefore the purpose should be to collect enough samples in a cost effective manner that will estimate the variable of interest with a predetermined level of confidence. For example, suppose the management objective is to detect change in basal cover of wiregrass as an indicator of understory conditions in a longleaf pine sandhill (Brockway et al. 1998). The goal of the land managers may be to have wiregrass cover in a restoration site with + 10% error of sampling. The role of the investigator is to determine the sample size and the appropriate method of allocation of samples that will yield the reliable and acceptable sample estimates. Non-Sampling Errors Errors that are made during the process of measuring a variable are considered non- sampling errors (Thompson 1992). One source of error is transcription and recording errors. Field data are often recorded in a notebook, later entered into a computer, and transformed and analyzed. During each step of this process human errors can occur such as errors in measuring, recording, and eliminating or failing to include all sample units. In sampling vegetation, these errors can be significant. Sample Design Simple Random Sampling Simple random sampling requires that each sample unit be independent of each other and have the same probability of being selected. A variety of methods are employed to create random points such as the simple random-coordinate method (Elzinga et al. 2001). Using this method, a Cartesian plane is transposed onto a map of the sample area, and random X and Y coordinates can be selected. With the advent of geographic information systems (GIS) technologies, there are computer programs available that can easily create a number of random points throughout the sampling frame. Stratified Random Sampling Stratified sampling is a type of sampling where the sampling frame is first divided into strata, or relatively homogenous regions (Reed and Mroz 1997). Samples can be allocated to each stratum in different numbers. For example, if one stratum contains more variation in the attribute of interest, more samples can be allocated to it. As such, when the sample area can easily be allocated to strata, it usually results in greater sampling efficiency. Systematic Sampling In systematic sampling the starting point for sampling is chosen randomly, and successive samples are chosen at intervals from the starting point (Thompson 1992). Since the first sample unit is the only one chosen at random, each of the sampling units is not independent of each other. Estimates of means, totals and percent are usually similar to simple random sampling. The same cannot be said for their respective sample variances. An example of systematic sampling is a transect in which the starting point is chosen at random and samples are selected at regular 100 m intervals thereafter. An advantage of systematic sampling is the ability to increase sampling efficiency by reducing travel time between sample units. In addition, systematic sampling can be good at distributing samples uniformly throughout the sample area (Reed and Mroz 1997). Statistical Analysis of Plant Communities There are several methods to analyze plant communities that do not involve quantitative measures (Elzinga 2001). One qualitative method for monitoring plant communities the use of photography. By taking a series of photographs, over time in the same location, land managers can keep track of changes. Other methods include the use of site assessments and species checklists where vegetative communities can be categorized based on an individual's observations. Although qualitative methods can be valuable one must recognize the limitations of such methods given individual biases. Indicators Sometimes a single species or variable can be used as an indicator for monitoring changes in community structure and composition. Indicators are surrogate measures for other variables that are more difficult or more expensive to measure. For example, cattails (Typha spp.), that form the dominant vegetative component in areas of the oligotrophic everglades are affected by eutrophication from agricultural runoff. Rather than measure phosphorous levels in water, perhaps land managers could use cattails as an indicator of eutrophication. This could perhaps be more economical, but the effectiveness of all indicators is based upon a strong correlation with the attribute of interest. Measures of Similarity Ecologists have developed many methods to measure how closely one vegetative community resembles another. One of the most common is the use of similarity indices. Some similarity measures rely on the presence or absence of species and do not account for the difference in abundance among species (species evenness) (Digby 1987, Ludwig 1988). Two examples of such indices are that of Jaccards and Sorensen's. Jaccard's index is simply the proportion of species common to two sites divided by the total number of species found in each site. Sorensen's index divides the total number of common species by the mean, rather than total number of species (Reed and Mroz 1997). Other methods, such as the Curtis-Bray, incorporate species abundance, accounting for species evenness. Each of these measures has been used in our study and the respective equations can be found in Chapter 4. Classification Classification is the process of grouping items based on similar properties. Early classification methods relied primarily on the subjectivity of expert ecologists to place communities into classes (Pielou 1982). Today, the use of computers allows more sophisticated and quantitatively rigorous classifications to be made. To classify plant communities, it is necessary to have sample data, usually in the form of species presence and some measure of abundance. Since plant communities do not often occur in discrete classes, interpretation can be problematic. Ordination Unlike classification, which clusters data into groups, ordination arranges samples along one or more coordinate axes in multi-dimensional space (Pielou 1984). Axes are usually based upon environmental variables. As plant communities are often continuous, ordination is a natural way to represent data over gradients (Gauch 1982). However, the multi-dimensional aspect of ordination can make interpretation of results more complex than classification. Statistical Analysis Based on Re-sampling Recently, a variety of methods based on re-sampling data have become popular in analyzing ecological data. The value of these methods is that one does not have to adhere to the assumptions of normality required for parametric statistics. The major utility in these methods is the ability to construct confidence intervals and conduct significance tests that might not be possible with other statistical methods (Elzinga et al. 2001). Two popular methods are the bootstrap and the jackknife. The bootstrap technique involves re-sampling the samples as if it was an entire population. All the original data are pooled and recombined. Bootstrap sampling is conducted by taking a sample of these data and calculating the parameter of interest for each sample. After repeated sampling, often in the several thousands, one can calculate the mean and standard error from the bootstrap estimates (Krebs 1998). From these intervals confidence values and significance values can be assigned. The jackknife is similar to the bootstrap, except that re-sampling is done without replacement. The data are recombined by removing one random sample each time. Every time the samples are recombined, the mean and standard error are calculated. The jacknife, due to the limited amount of combinations inherent with sampling without replacement does not require intensive calculations (Krebs 1998). CHAPTER 3 STUDY AREA Location Our study was conducted at Goethe State Forest, in Florida's Levy and Alachua Counties. The 4,778-acre (1911 ha) Watermelon Pond Unit (WPU) is a disjoint collection of parcels primarily to the south and west of Watermelon Pond (latitude 290 32' 42", longitude 82 36' 20"), on the Brooksville Ridge in north-central Florida (Fig. 3- 1). Physiography The WPU is on the northern part of the Brooksville Range, an area characterized by deep marine deposited soils of the Candler-Apopka-Astatula associations (Dickinson et al. 1982). The elevation is quite variable for this region of Florida, ranging from 45 to 125 feet (14 to 38 m) above sea level. The areas of lowest elevation may contain sandhill upland lakes or depressional marshes. These lakes are usually acidic and fairly oligotrophic (Griffith 1998) Land Use During the past century, a variety of land uses have impacted the natural communities of the WPU. These include phosphate mining, timber removal, and cattle grazing. General land office survey records indicate that much of the upland areas of the WPU were pine forest in 1840s (FDEP 2003). Logging operations cleared most of the longleaf pine trees by the 1930's and they were apparently left to regenerate (Olmert 2001). Very few mature longleaf trees are present on the property today. More than 1000 acres (405 ha) of the property were planted with slash pine (Pinus elliotti) between 1969-1971 and cleared before purchase of the land by the state of Florida in 1995. Previously planted areas are primarily located on the Watermelon Pond North tracts (Fig. 3-1). The trees were planted "in the rough" with site preparation consisting of girdling oak trees. The Florida Division of Forestry has now planted these areas with longleaf pine with the intention of restoring the sandhill community. Phosphate mining was a profitable business at the turn of the century the profits from exporting phosphates to foreign markets created an economic boon in some of the small towns near WPU. There is evidence of three such mines on the property. Phosphate was removed from the open mines without the benefit of machinery, using mules and forced labor (Olmert 2001). Biological Communities According to descriptions of the Florida Natural Areas Inventory (1990) there are eight basic community types present in the WPU. These community types exist in mosaics throughout the landscape, often grading into one another. Presently there are no accurate maps available to describe the extent and location of these communities. However, there is a complete set of high resolution (0.3m ground sampling distance) aerial photographs taken in the 2002 by Emerge (Andover, Massachusetts). Sandhill Upland Lakes Watermelon Pond is a complex system of interconnected bodies of water known as sandhill upland lakes (The Florida Natural Areas Inventory 1990). Seasonal fluctuations of water levels can be quite extreme. Watermelon pond contains 551 acres (223 ha) of wetlands and surface water (Dickinson et al. 1982). However, at the time of our study there was no water present in any of these lakes due to an extended drought. Watermelon .. / .i ', *- T "---"i l' ~i , IT "2 _: i-- I -K Legend i i I i i Arterial Roads -Si/yAr I Figure 3-1 Study area location Pond has been used recreationally for boating and fishing. The County of Alachua operates a boat ramp on the northeastern portion of the pond. Prior to drought of recent years, it was well known as an excellent venue for large-mouth bass (Micropterus salmoides) fishing. Sandhill Sandhill is by far the most widespread vegetative community in the WPU. Fire suppression, soil disturbances, and other anthropogenic impacts, have created large differences in the structure and composition of this community. Typically a sandhill community is described as containing a sparse canopy of longleaf pine with a midstory of small xerophytic oaks, such as turkey oak (Quercus laevis), bluejack oak (Q. incana) and sand post oak (Q. stelleta). The understory is dominated by wiregrass (Aristida beyrichiana), pineywoods-dropseed (Sporobolusjunceus), and a diverse number of herbaceous species and several species of small shrubs. A very small part of the WPU sandhill contains all of the characteristics described above. The primary missing component is a mature longleaf canopy, apparently as a result of historic logging. There is also evidence of prolonged fire suppression, which has resulted in an increase in hardwood species. This is also evidenced by the high density and large sizes ofxeropyhtic oaks throughout much of the property. In some cases "Turkey Oak Barrens" have formed where large turkey oaks dominate the canopy. Soil disturbance has also had a profound impact upon the ground cover in some areas. In particular, wiregrass, a key component of the understory, does not regenerate after severe soil disturbances, such as plowing. There is a 150-acre previously farmed section that contains virtually no wiregrass. Scrub At higher elevations and in well-drained sandy soils, rosemary/oak scrub can be found. Dominant species in this community include the Florida rosemary (Ceratiola ericoides) and sand live oak (Q. geminata). The ground area typically contains patches of bare soil and wiregrass, as well as lichen (primarily Cladonia and Cladinia spp.). There is an ongoing debate as to whether many of these areas are naturally occurring scrub community or have been formed as the result of sandhill being degraded by poor logging practices, followed by decades of fire. Some of the scrub dominated by Florida rosemary still contains longleaf pine. However, longleaf pine is conspicuously absent in some areas of higher elevation. This suggests that perhaps a more typical scrub community may have always existed in areas of higher elevation and deep sands. Hardwood Hammock At lower elevations and on less permeable soils, hardwood hammocks frequently occur. In particular, the eastern most section of WPU east, which lies on an Otela- Candler-Taveras soil association, supports a n extensive area of hardwood hammock. Much of the hardwood hammocks within the study area are dominated by laurel oak (Quercus hemisphaerica) with relatively low tree diversity. Historical aerial photography indicates that these hammocks of low diversity are the result of recent succession on cleared land. In more diverse areas, commonly occurring tree species include red bay (Persea borbonia), pignut hickory (Carya glabra), and eastern redcedar (Juniperus virginiana). Prairie The land surrounding Watermelon Pond is seasonally inundated by water, forming a prairie community, which is dominated by maidencane (Panicum hemitomum), sand cord grass (Spartina bakeri) and other species of grass. There are scattered trees and shrubs located on the prairie such as slash pine, gallberry (Ilex glabra), wax myrtle (Myrica cerifera) and St. John's wort (Hypericum spp.). Ephemeral Pond Depressional areas within the prairie that have long hydroperiods are called ephemeral ponds. They contain a different assemblage of plants, including spatterdock (Nuphar luteum), redroot (Ii /lc l, u/he / caroliniana), and pennywort (Hydrocotyle spp.). Ephemeral ponds are unable to sustain populations of fish, however they are particularly important breeding habitat for amphibians (Babitt and Tanner 1997). Management Considerations By far the greatest challenge to management is the fact that the property is a disjoint collection of parcels, with little connectivity among them (see Fig. 3-1). Problems associated with small nature reserves include genetic isolation, and detrimental edge effects, such as those caused by invasive species, and anthropogenic influences. Numerous ranchettes and other residential structures border many of the property boundaries. These may complicate the use of prescribed fire, which is necessary to maintain the fire adapted plant communities present in the WPU. Other negative impacts caused by humans are illegal dumping, all-terrain vehicle use, and feral dogs. The main office at Goethe State Forest, which maintains regulatory and administrative control, is more than 20 miles (8 kilometers) away from the main tract of the forest. It is therefore difficult, if not impossible, to maintain a continuous presence in the WPU. Despite these challenges, protecting the WPU is essential, since it is one of the last remaining examples of the diverse upland and lacustrine communities of the Northern Brooksville Ridge. CHAPTER 4 METHODOLOGY Vegetation Data Collection The initial vegetation data collection consisted of 321 sample points located systematically throughout the property that were marked with metal stakes, identified with aluminum tags and geo-referenced using a Trimble (Sunnyvale, California) Pro- XRS GPS with real time differential correction (Fig. 4-1). The distance between sample points in each transect is 100 meters. Systematic sampling was chosen for this part of the study as it allows a large number of samples to be distributed evenly throughout the entire property. Each permanent sample point is the focal point from which all sampling is conducted. Measurements were taken on three categories of the vegetation: a) ground cover b) shrubs, seedlings and saplings, and c) trees. The methodology and observed variables varied among these three classes to ensure reliable estimates. Seven areas were chosen for a comparative analysis. These sites included three reference sites and four nonreference sites. These areas were delineated using aerial photographs in a geographic information system (GIS). To increase sampling intensity in areas selected for comparative analysis, random sampling was used within the seven selected sites. The methodology for data collection was the same as for the systematic sampling. Random points were generated in ArcView 3.2 GIS software (ESRI, Redlands, CA). These points were navigated to in the field using GPS. There were a total of 184 sample points located in the 7 test areas, 128 in nonreference sites and 56 in reference sites. Ground Cover Ground cover was ocularly assessed in two 1- m2 quadrats at each sample point. Estimates were recorded in 5% categories from 0 to 100. Location of quadrats was determined by a random azimuth between 0and 360 relative to the permanent point. Quadrat locations were marked with a metal stake and embossed aluminum tags so that they could be relocated in the future. Each random quadrat was positioned 2 m away from the permanent point (Fig. 4-1). Percent cover of each species was ocularly estimated and recorded. In addition, the number of stems of each species was recorded. 2m 7S 1 m LQadrat Location 2 Figure 4-1 Sampling architecture Shrubs, Seedlings and Saplings All shrubs, seedlings and saplings within a 2m radius of the permanent plot center were counted. Saplings were defined as trees less than 10 cm DBH but greater than 1.37 m in height. The seedlings were less than 1.37 m in height. Every shrub, sapling, and seedling within the plot was classified by species. Trees A 10-factor prism (variable radius plot or point sampling) was used to sample trees greater than 10cm DBH to estimate the basal area per acre in each sample point. The species and diameter at breast height (1.37 m) was recorded for each tree in the plot. Species Sampled Almost all woody plants that were encountered were identified to species in the sampling process. Woody species whose identification proved taxonomically difficult, such as those requiring fertile specimens when none were available, were identified to genus. Nomenclature in our study followed that of Godfrey (1988) for woody plants and Wunderlin (1998) for herbaceous species. Classification of Comparative Study Areas Several study sites were selected to provide additional information on the sandhill community, targeted for ecological restoration on this property. The purpose of these sites was to sample as much of the vegetative diversity in the sandhill longleaf pine community as possible. Both high quality sites (reference) and highly disturbed areas were chosen to represent the wide range of vegetation at the WPU. Nonreference Areas Sites 4, 5, 6, and 7 have a similar land-use history. The nonreference sites were all former plantation that were established between 1969-1972 using small tractors with a drag-type setter. The only site preparation at the time of planting was hand girdling of the oaks, which were primarily turkey oak and bluejack oak. The comparison regions (Fig. 4-2), classified for the purposes of our study, are as follows: Site 4 Before 1969 this land was leased for cattle. Centipede grass was planted as cattle forage, and in many areas this species dominates the ground cover in lieu of wiregrass. Site 4 was planted in slash pine in 1969-70. Samples for this comparative analysis were stratified so as to avoid inclusion of vegetation in and around a historic phosphate mine, that is a xeric hardwood community. All harvestable timber was extracted in 1995 by the Loncala Corporation before sale of the land to the Florida Division of Forestry. Although prescribed burning was carried out in the fall of 2000, it has not yet been planted with longleaf pine seedlings. Site 5 This site was apparently plowed for agriculture at one point due to the state of the existing ground cover. For example, wiregrass, a species highly susceptible to soil disturbance, was not found on this site. Site 5 was planted in slash pine 1968-1969, in the same manner as Section 4. This timber was clear cut prior to acquisition by DOF, and later planted twice with bare-root longleaf seedlings in 1996/1997 and 1997/1998. Site 6 Site 6 is a former slash pine plantation that was established between 1968 and 1970. After being clearcut in 1996 it was planted with longleaf pine seedlings in December 1999 at a density of 590 trees per acre (1458 per ha). However, only 30% of them survived. There are some scattered remnant mature longleaf pines in this area. The ground cover is in better condition than Section 5. Site 7 Site 7 is also a former slash pine plantation, established in 1969/1970, and cut in 1996. This section was planted in January 1999 with longleaf pine seedlings, utilizing scalping as a ground treatment. A prescribed burn took place in 2000. Reference Sites Three reference sites were chosen. These areas were smaller in size than the nonreference sites because there are few tracts of quality sandhill left. Criteria used for choosing these sites were the presence of mature longleaf pine trees and intact ground cover. Field reconnaissance was conducted on the ground to verify the condition of these sites prior to their selection as reference sites. In addition, assessments done in 1995 by the Florida Natural Areas Inventory (FNAI) have identified these areas as being in excellent condition. The general land office records from the mid-19th century (FDEP 2003) state that these areas sustained pineland, sometimes referring to them in their notes as "first rate pineland". It is likely that all reference sites were historically logged in the 1930's and low intensity cattle grazing was undertaken from the 1930's through the 1990's. Although ranching may have taken place on reference sites, the extent of centipede grass present is much less than in the nonreference sites. However, due to the extensive conversion of the northern Brooksville range to agriculture and housing development, it is unlikely that a more representative portion of longleaf sandhill community for the WPU exists today. Legend N Goethe State For t g mrr rir E.or.nrdr", Sam ple Pont Locations 0 1 ,2I 2 4,800 Meters I I ii I Figure 4-2 Comparative study areas. Sites 1, 2, and 3 are reference sites. Sites 4,5,6 and 7 are non reference sites. Field Data Interpretation The purpose of this data analysis was to determine a list of indicator species for monitoring restoration efforts. This process involved comparing values for species importance and tree basal area. It was done in a preliminary assessment that attempted to sites. In addition, ecological similarity was measured between sites. The second stage of the data analysis involved ranking those species that accounted for the greatest difference between the reference and nonreference sites. McCoy and Mushinsky ( 2002) used statistical analysis of binomial probability to effectively determine which fauna were responsible for differences in community structure of wildlife species in restored phosphate mine lands. A similar methodology was used in examining the vegetative composition in our study. Following ranking, a list of indicator species was developed based upon statistical significance. The methods used for the purpose of our study are described in detail below. Importance Values For understory species importance values are the sum of the relative frequency and relative percent cover of each of the species. For shrubs, seedlings and saplings importance value is the sum of the relative frequency and relative density. For example, the relative frequency of Florida rosemary (Ceratiola ericoides) at site 3 is 1.67% and the relative density is 0.09%, so the importance value is 1.76. Basal Area of Trees Basal area of trees was estimated with probability proportional to tree diameters at breast height using a 10- factor prism. A random sample of 24 plots was drawn from each of the areas of interest for this analysis. Similarity Analysis Similarity indices were used to compare the vegetative structure of each of the study areas. The three measures of similarity used in our study included Jaccard's coefficient of similarity, Sorensen's index, and the Curtis Bray index. To make comparison easier all similarity measures were expressed in percentages. The Jaccard's Index is a percentage of the species that are common between two sites. Sorensen's index is the number of common species between both sites. The Curtis Bray Index incorporates measures of abundance, in this case density and foliar cover, to account for differences in species evenness. The equations for these measures are as follows: Jaccard's Coefficient of Similarity is shown in Equation 4-1 s = Number of Common Species10 (4-1) Total Numberof Species Sorensen's Index is shown in Equation 4-2 Ss Numberof CommonSpecies 100 (4-2) (S + Sb )/2 Sa = The number of species in community A Sb = The number of species in community B Curtis Bray Index is shown in equation 4-3 2 min(X,, Y) Cs = -1 (4-3) 1=1 1=1 where (Xi,Yi) are the abundance measures for a given species in a population for community A and B. Determination of Targets for Monitoring. In order to determine indicator species for monitoring, a ranking system was employed. It relies on comparing the binomial proportion of species occurrence between the reference and nonreference sites (McCoy and Mushinsky 2002). Species were ranked according to their statistical significance of occurrence between reference and nonreference sites. Proportional Abundance of Species The first step for ranking the species was to determine the proportion of sample points at which each species was present. This was done by aggregating all samples for the 56 plots in the reference sites, and all samples for the 128 nonreference sites. Appendix B lists the proportional abundance of each species. It was calculated based on the respective number of plots at which specific species was found, without regard to abundance (Appendix B). For example, sand blackberry (Rubus cuneifolius), was found in 10 of the 56 reference plots and in 83 of the 128 nonreference plots. Therefore, the proportional abundance of this species was 0.18 in the reference and 0.65 in the nonreference plots. The statistical differences between the two binomial proportions was determined by approximating the normal distribution. In this case, a two-tailed test was performed with the null hypothesis that the proportion between the two sites was equal. The formula used for the statistical test of comparison of two binomial proportions (Ott, 1993) was: z 7 / 2 1 1 12 where = ,r 2 and Y1 + Y2 nl + n2 Zr = Proportion of sample points at which a species is observed for the reference sites 72 = Proportion of sample points at which a species is observed for the non- reference sites. n = Proportion of sample points at which a species is observed for both the reference and nonreference sites. y, = Number of occurrences of a species at the nonreference sites. y = Number of occurrences of species at the reference sites. nl = Number of sample points in the reference sites. n2 = Number of sample points in the nonreference sites. Since this is a two-tailed test, the null hypothesis z1 = I2 is rejected for any given value of a if: Z > Z,/2 Next, the species are ranked according to this value (Appendix C). Those species with a statistically significance level of p=0.05 or greater were considered focal species. Those not found to possess a statistical significance were defined as non-focal species. In addition, there was a number of species that were not found in a sufficient number of the plots for the normal approximation to be valid in this test, that is N;r or N(1 -) were not greater than or equal to 5. Species included in this category were listed as un- common. CHAPTER 5 RESULTS AND DISCUSSION Comparison Sites Appendix A includes the results of the Jaccard's (J), Sorensen's (S), and Curtis Bray (CB) similarity indices, as calculated for the understory, shrubs, and tree seedlings and saplings, which compare ecological similarity among the sites. Importance values (Appendix D) are also incorporated into the following discussion. Understory Site 5 is least similar to the reference site for all three indices. The similarity was only 32.4% with the CB measure, and 34.6% and 51.4%, respectively, using the J and S measure of similarity. This is an obvious finding considering that Site 5, having been previously farmed, has sustained the most anthropogenic disturbance of all the sites. Tillage of the soil has eliminated wiregrass, which is present in all of the other sites. In addition, there is a greater abundance of forbs and grasses, compared to the other sites (Table D-l). Sites 6 (CB=57.8%, J=54.2%, S=70.3%), and 7 (CB=56.7%, J=54.5%, S=70.6%), were the most similar to the reference site. Site 4 and 5 were least similar according to the CB, likely due to the relatively greater importance of forbs and grasses at Site 5 (Table D-l). Shrubs For the shrub layer, the CB showed significantly less similarity than the J and S, among most sites. This can be attributed to differences in species evenness, relative to the number of species. For example, site 5 proved to be most similar to the reference site, using the J (55.6%) and S (63.2%), but least similar according to the CB (27.8%). J and S do not account for the abundance of species. They merely account for the presence or absence of common species. Site 5 and 7 were the most similar of all measures (CB= 85.2, J=71.4, S=83.3). This may be attributed to the fact that these sites have been disturbed the most, Site 5 being tilled historically and Site 7 prescribed burned in 2000. Tree Seedlings and Saplings Site 5 showed the least similarity with the reference site. This may be partially attributed to the absence of turkey oak in Site 5, which is very common in the understory of all the other sites, especially the reference areas. Sites 4 and 7, which have been prescribed burned in 2000 and 1998, respectively, are most similar to the reference sites. Summary of Similarity Indices The results of the similarity indices seem to suggest that although similar species composition may exist between reference and nonreference sites, there may be large differences in evenness. This is especially evident with mid-story shrubs and tree seedlings whose abundance and sizes may be a function of fire history. For example, a more recently burned site may contain a higher abundance of small oaks, such as turkey oak seedlings. This suggests that the Curtis Bray similarity index may be the most appropriate for measuring ecological similarity between reference and nonreference sites. Suggested Target Species The following is an outline of the suggested target species based upon species ranking, importance value, and basal area. The ranking methods employed in our study provide a clear indication of the significance of the proportion of sampling points that contained species without regard to abundance. Appendix C contains a list of the corresponding statistical significance. Appendix D includes a complete list of all of the importance values for each of these ranked species. These importance values serve as a useful tool to discuss the reasons why some species may have been found more often at the reference or the nonreference sites. In addition, importance values give relative measures of abundance in each of the comparison areas. It should be noted that there is some overlap in the sampling of species in the understory, shrub, and tree seedling and sapling layer. For example, the presence of sand- blackberry is noted in both the understory quadrats, and in the 2m radius fixed area plots intended to measure shrubs, saplings and tree seedlings. In such situations, the observations from the 2m radius plot are used, as the larger sampling area provides a better estimate. Understory Five targets for monitoring were determined for the understory. Among these, wiregrass was found more often in the reference than in the nonreference sites. The others, centipede grass, dog-fennel, green briar, and total forb cover, were more often detected in the nonreference areas. Wiregrass (Aristida beyrichiana) Wiregrass was identified in the early planning stages of this project as a potential target species. There is a significantly higher proportion of wiregrass in the reference sites than in the nonreference sites (P<0.01). However, Site 6, a nonreference site, had the highest importance value for wiregrass (31.24). The difference in proportional abundance between reference and nonreference sites (Appendix B) may be due to the fact that Site 5, a nonreference site, does not contain any wiregrass due to past tillage (Table D-l). This suggests that even though the nonreference sites were converted to plantation for about a 25-year period, the impact on wiregrass might be negligible. This corroborates research by Hedman et al. (2000) that past agriculture use has a greater impact on ground cover than many forest management activities. Centipede grass (Eremochloe ophiurides) Centipede grass is a rhizomotous perennial grass (Godfrey 1988) introduced in the WPU to improve grazing (Olmert 2001). Centipede grass was observed more often in the nonreference plots (P<0.01 ) than the reference plots and has become established as a dominant ground cover species throughout many areas of the nonreference sites. The effects of centipede grass on restoration of longleaf pine ecosystems have not been extensively studied. Two possible ways in which centipede grass could inhibit restoration efforts are to compete with native plants and to alter the ground cover fuel composition. Centipede grass should continue to be monitored over time so that its persistence in the restoration sites may be measured. For example, it may be that once a tree canopy is established, this species will be competitively excluded by trees or more shade tolerant understory species. Dog-fennel (Eupatorium capilifolium) Dog-fennel is a ruderal species that is often found in disturbed woods and fields (Radford et al 1968) and, as such, is very common in the nonreference sites. Expectedly, dog fennel was found significantly more often at the nonreference sites (P<0.01). This species is easy to identify, and should be a useful target species for monitoring. Green briar (Smilax spp.) Green briar is a woody vine that was most prevalent in the nonreference sites (P<0.01). The increase of woody vines with fire suppression has been documented (Rodgers and Provencher. 1999), and the importance values for this shrub seem to indicate a similar finding. Green briar was less important in Site 4 (imp. = 1.48), prescribed burned in 2000 and Site 7 (imp =4.29), than Sites 5 (imp=8.89) and 6 (imp=5.19), which have not been prescribed burned. In comparison, the reference sites contained relatively little green briar (imp. values for 1,2,3 were 1.31, 0, and 1.14, respectively). Total forb cover Other forbs was a broad classification that included all forbs except dog-fennel, such as legumes and asters (Pityposis spp.). The total forb cover was greatest in Site 5, which is the most disturbed (imp=35.89). This variable may be a useful measure for monitoring, however more research is needed to determine which species are responsible for the large importance value of forbs at Site 5. The seasonality of herbaceous plants and diversity of species can make species identification a difficult process when compared to woody plants. Shrubs Winged sumac (Rhus copallinum) Winged sumac is a common species found in old fields and forest plantations (Radford et al 1968). Its presence may increase following prescribed burning (Miller and Miller 1999). Although common in reference areas, winged sumac may prove to be a very useful monitoring species. Sand blackberry (Rubus cuneifolius) Sand blackberry is a small shrub that is frequently found in great abundance in areas that have been disturbed by fire or mechanical site preparation (Godfrey 1998). Along with winged sumac, sand blackberry dominates many areas of the WPU that are old slash pine plantations that have been planted with longleaf pine. It was found in a significantly higher proportion of the nonreference sites (P<0.01) than the reference sites. Florida rosemary (Ceratiola ericoides) Florida rosemary is an evergreen shrub that is common in the sandy soils of the WPU. In some areas this species forms monotypic stands often referred to as "rosemary balds" (see Chapter 3). This species was found more often in the reference sites (P<0.01) than the nonreference sites. Patches of rosemary balds exist in a mosaic within xeric sandhill at the WPU, and the samples selected for this analysis included points in rosemary balds. Site 3 was the only reference site to contain Florida rosemary, but it was a very important component (imp.=134.1). A review of historical aerial photography revealed that these rosemary scrub balds have existed in Site 3, to a similar spatial extent, sat least since 1949. Glaucous Blueberry (Vaccinium darrowii) Glaucous blueberry is a small shrub similar in appearance and distribution to Vaccinium myrsinites, in the WPU. This species was found more often in the reference sites than the nonreference sites (P=0.03) and may serve as a useful indicator species. However, field observations seem to suggest that the distribution of glaucous blueberry is very patchy, and as such, its utility as an indicator may be better understood with increased sample size. Tree Seedlings and Saplings Three species of tree seedlings and saplings, turkey oak, black cherry (Prunus serotina), and laurel oak, were selected as focal species for monitoring. Turkey oak was found more frequently in the reference sites, whereas black cherry and laurel oak were more prevalent in the nonreference areas. Turkey oak The presence of turkey oak seedlings and saplings was more prevalent in reference sites (P<0.01) than in nonreference sites. Typically, longleaf pine sandhill has been described as having a large number of small oaks, such as turkey oaks (Glitzenstein et al. 1995). Therefore, this species may serve as an important indicator species of longleaf restoration. Black cherry (Prunus serotina) Black cherry is an early succession hardwood species that was not found in all the reference sites. However, seedlings and saplings were very common in the understory of nonreference sites. The use of prescribed bums would probably reduce the importance of this species in the understory. This species should be a useful target species to monitor the effectiveness of prescribed burns. Laurel oak Laurel oak is very similar to black cherry in that it is an invasive hardwood species prevalent in fire suppressed longleaf sandhill. Laurel oak is more common that black cherry, therefore it may be a better focal species for monitoring. However, one confounding factor for monitoring this species is the difficulty in identification. Laurel oak seedlings look taxonomically very similar to some other oak species, especially bluejack oak. Trees Table E-1 includes the estimates of tree basal area for each of the sites. These data reveal that there is a much higher basal area of longleaf pine and turkey oak in the reference site. These two species should be monitored in the future. In particular, an increase in longleaf pine basal area is a desired management objective. However, the lack of mature turkey oaks in the nonreference sites is probably due to the fact that site preparation for conversion to plantation involved removing most of the oaks by girdling between 1968-1972. Longleaf pine There are very few mature longleaf pine trees remaining in the WPU, and as such, they are an excellent indicator species of remaining areas of high quality longleaf. In general, areas that include mature longleaf pine had excellent ground cover. However, from a monitoring standpoint, longleaf pine basal area is not a robust monitoring metric because it is a slow growing long-lived species that is not likely to be affected by management actions over short periods of time Turkey oak Turkey oak is almost always present where there is longleaf pine (Gordon 2003), and therefore, it may be an excellent indicator of the historic extent of the longleaf pine sandhill. It is likely that fire suppression in recent history has led to an increase in turkey oaks at the reference sites . CHAPTER 6 SUMMARY AND CONCLUSIONS The ranking methodology used in our study allows the development of a preliminary list of target species that could be utilized for future monitoring. However, the most important accomplishment of our study is the development of a system for sampling the understory, mid-story, and tree layer, utilizing a network of geo-referenced permanent points. Although the temporal scale at which the data were collected limited the analysis, sample points can be re-visited in future studies. As discussed in Chapter 2, this would allow for much more powerful statistical analyses. Suggestions for Future Research Our study was the first to quantify the vegetative structure and composition of the WPU and, as such, it raised many questions that could be addressed by further research. The following are recommendations for future research: Further Exploratory Analysis of Field Data The use of ordination and classification (see Chapter 2) to examine the field data may reveal some patterns that have useful implications for management and restoration. Multi-dimensional methods may provide some insight into those points, or even sections of the property, that are closest to the reference conditions. An ordination of the comparison areas delineated in our study could provide further information about whether the sites chosen are ecologically similar. Such methods may prove more useful than the similarity measures employed in our study. Determination of Target Abundances Although our study identifies a list of target species, it would also be useful for land managers to establish target abundances for monitoring efforts. For example, having identified wiregrass as a species to monitor, what percent cover of this species would be desired for restoration sites? Effect of Centipede Grass on Longleaf Pine Restoration Centipede grass was found throughout many areas of the WPU that are slated for restoration. A review of the literature found limited information on the effects of this grass as an invasive species in natural areas. More research needs to be done to study: a) the effect of this species on fine fuels for fire, b) interspecific competition with native plants and c) potential changes in distribution and abundance over time. APPENDIX A SPECIES SIMILARITY Table A-1. Understory similarity among comparison sites based on Jaccard's Index R 4 5 6 7 R 50.0 34.6 54.2 54.5 4 50.0 40.0 66.7 55.6 5 34.6 40.0 51.7 51.9 6 54.2 66.7 51.7 72.0 7 54.5 55.6 51.9 72.0 Table A-2. Understory similarity among comparison sites R 4 5 6 Table A-3. Understory based on Sorensen's Index 7 similarity among comparison sites based on Curtis Bray Index. 4 5 6 7 66.7 51.4 70.3 70.6 66.7 55.8 80.0 35.7 51.4 55.8 46.2 68.3 70.3 80.0 46.2 54.5 70.6 35.7 68.3 54.5 50.3 32.4 57.8 56.7 50.3 18.1 69.9 68.8 32.4 18.1 45.1 56.9 57.8 69.9 45.1 64.1 56.7 68.8 56.9 64.1 Table A-4. Shrub similarity among comparison sites based on Jaccard's Index. R 4 5 6 7 R 35.7 55.6 50.0 50.0 4 35.7 46.2 53.3 26.7 5 55.6 46.2 50.0 71.4 6 50.0 53.3 50.0 45.5 7 50.0 26.7 71.4 45.5 Table A-5. Shrub R similarity among comparison sites based on Sorensen's Index. 4 5 6 7 Table A-6. Shrub similarity among comparison sites based on R 4 5 6 R 43.5 27.8 49.5 4 43.5 81.7 82.0 5 27.8 81.7 77.2 6 49.5 82.0 77.2 7 30.1 73.3 85.2 75.0 Curtis Bray Index. 7 52.6 52.6 63.2 42.1 52.6 63.2 69.6 47.1 52.6 63.2 66.7 83.3 63.2 69.6 66.7 62.5 42.1 47.1 83.3 62.5 30.1 73.3 85.2 75.0 47 Table A-7. Tree seedling and saplings similarity among comparison sites based on Jaccard's coefficient. R 4 5 6 7 R 75.00 40.00 66.67 60.00 4 75.00 60.00 70.00 60.00 5 40.00 60.00 54.55 50.00 6 66.67 70.00 54.55 72.73 7 60.00 63.64 50.00 72.73 Table A-8. Tree seedling and saplings similarity among comparison sites based on the Sorensen's coefficient. R 4 5 6 7 R 85.71 57.14 85.71 85.71 4 85.71 75.00 87.50 87.50 5 57.14 75.00 70.59 70.59 6 85.71 87.50 70.59 57.14 7 85.71 87.50 70.59 57.14 Table A-9. Tree seedling and saplings similarity among comparison sites based on Curtis Bray coefficient. R 4 5 6 7 R 52.1 26.0 43.7 51.2 4 52.1 25.7 26.0 49.6 5 26.0 25.7 25.2 27.9 6 43.7 73.3 25.2 48.8 7 51.2 49.6 27.9 48.8 APPENDIX B PROPORTIONAL DISTRIBUTION OF SPECIES AMONG REFERENCE AND NONREFERENCE SITES Table B-1. Distribution of understory species among reference and nonreference points. Based upon the proportion of plots a species is present. Reference (n=56) Nonreference (n=128) # of points # of points Species found Proportion found Proportion Rubus cuneifolius 10 0.18 83 0.65 Aristida beyrichiana 56 1.00 67 0.52 Quercus laevis 38 0.68 31 0.24 Rhus copallinum 24 0.43 97 0.76 Eremochloe ophiuroides 6 0.11 54 0.42 Eupatorium capillifolium 7 0.13 44 0.34 Smilax spp. 4 0.07 36 0.28 Quercus hemisphaerica 0 0.00 18 0.14 Ceratiola ericoides 11 0.20 12 0.09 Total Forbs* 52 0.93 124 0.97 Prunus serotina 2 0.04 13 0.10 Quercus incana 6 0.11 21 0.16 Pinus palustris 12 0.21 20 0.16 Quercus nigra 0 0.00 6 0.05 Vaccinium darrowii 6 0.11 8 0.06 Myrica cerifera 0 0.00 5 0.04 Diospyros virginiana 4 0.07 13 0.10 Other Grasses 51 0.91 111 0.87 Quercus geminata 5 0.09 11 0.09 Vitis spp. 0 0.00 13 0.10 Woody vine ^' 6 0.11 1 0.01 Toxicodendron toxicarium 5 0.09 1 0.01 Opuntia humifosa 0 0.00 4 0.03 Pinus elliotti 0 0.00 4 0.03 Quercus virginiana 0 0.00 2 0.02 Vaccinium myrsinites 0 0.00 2 0.02 Asimina spp. 1 0.02 1 0.01 Passiflora incarnata 1 0.02 1 0.01 Paspalum notatum 1 0.02 1 0.01 Crataegus spp. 0 0.00 1 0.01 Panicum hemitomum 0 0.00 1 0.01 Quercus X asheana 0 0.00 1 0.01 Rubus betulifolius 0 0.00 1 0.01 Toxicodendron radicans 0 0.00 1 0.01 Vaccinium stamineum 0 0.00 1 0.01 Zanthoxylum clava-herculis 0 0.00 1 0.01 Gelsemium sempervirens 2 0.04 5 0.04 Serenoa repens 1 0.02 2 0.02 Table B-2. Distribution of shrub species among reference and nonreference points. Based upon the proportion of plots a species is present. reference (n=56) Nonreference (n=128) S s # of points found Species Asimina spp. 4 Calicarpa americana 0 Ceratiola ericoides 14 Crateagus spp. 0 Hypericum spp. 2 Myrica cerifera 0 Opuntia humifosa 0 Rhus copallinum 26 Rubus cuneifolius 10 Serenoa repens 2 Toxicodendron toxicarium 5 Vaccinium arboreum 1 Vaccinium darrowii 9 Vaccinium myrsinites 1 Vaccinium stamineum 0 Zamiafloridana 1 Proportion 0.07 0.00 0.25 0.00 0.04 0.00 0.00 0.46 0.18 0.04 0.09 0.02 0.16 0.02 0.00 0.02 # of points found 2 2 22 1 1 3 15 113 95 2 1 3 9 2 2 1 Proportion 0.02 0.02 0.17 0.01 0.01 0.02 0.12 0.88 0.74 0.02 0.01 0.02 0.07 0.02 0.02 0.01 Table B-3 Distribution of tree seedling and saplings among reference and nonreference points. Based upon the proportion of plots a species is present. Reference (n=56) Nonreference (n=128) Species Albiziajulibrissin Diospyros virginiana Ilex opaca Pinus elliottii Pinus palustris Pinus taeda Prunus serotina Quercus geminata Quercus hemisphaerica Quercus incana Quercus laevis Quercus nigra Quercus virginiana Zamia clava-herculis # of points found 0 7 0 0 16 0 0 7 2 11 49 0 0 0 Reference Proportion 0.00 0.13 0.00 0.00 0.29 0.00 0.00 0.13 0.04 0.20 0.88 0.00 0.00 0.00 # of points found 1 22 1 8 42 1 20 19 27 22 55 3 1 5 Proportion 0.01 0.17 0.01 0.06 0.33 0.01 0.16 0.15 0.21 0.17 0.43 0.02 0.01 0.04 APPENDIX C SPECIES RANKING Table C-1 Understory species ranking. Species Focal Species* Rubus cuneifolius Aristida beyrichiana Quercus laevis Rhus copallinum Eremochloe ophiuroides Eupatorium capillifolium Smilax spp. Quercus hemisphaerica Ceratiola ericoides Total Forbs * Prunus serotina Nonfocal Species Pinus palustris Quercus nigra Vaccinium darrowii Myrica cerifera Diospyros virginiana Other Grasses Quercus geminata Z Value 6.13 5.52 5.24 4.69 4.37 3.43 3.32 2.57 2.03 1.95 1.71 1.42 1.41 1.29 1.28 1.19 0.26 0.21 Significance <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 0.01 0.02 0.03 0.04 0.08 0.08 0.10 0.10 0.12 0.40 0.42 * Selection of focal species based upon a significance level of P = 0.05 Table C-2 Shrub species ranking. Species Z value Significance Focal Species* Rhus copallinum 26.35 <0.01 Rubus cuneifolius 20.41 <0.01 Ceratiola ericoides 4.46 <0.01 Vaccinium darrowii 1.90 <0.03 Uncommon Species Opuntia humifosa 6.11 Myrica cerifera 2.64 Callicarpa americana 2.15 Vaccinium stamineum 2.15 Vaccinium arboreum 1.96 Crataegus spp. 1.52 Vaccinium myrsinites 1.38 Serenoa repens 0.86 Toxicodendron toxicarium 0.75 Zamiafloridana 0.60 Asimina spp. 0.16 Hypericum spp. 0.11 * Selection of focal species based upon a significance level of P = 0.05 Table C-3 Tree seedling and sapling species ranking. Species Z Value Focal Species* Quercus laevis 5.61 <0.01 Prunus serotina 3.13 <0.01 Quercus hemisphaerica 3.00 <0.01 Nonfocal Species Diospyros virginiana 0.80 0.21 Quercus geminata 0.42 0.34 Pinus palustris 0.57 0.28 Quercus incana 0.40 0.34 Uncommon Species Pinus elliotti 1.91 Zanthoxylum clava-herculis 1.50 Quercus nigra 1.16 Albiziajulibrissin 0.66 Ilex opaca 0.66 Pinus taeda 0.66 Quercus virginiana 0.66 * Selection of focal species based upon a significance level of P = 0.05 APPENDIX D IMPORTANCE VALUES Table D-1 Understory species importance value. Species 1 2 3 4 5 6 7 Aristida beyrichiana 14.42 23.09 29.3 9.6 0 31.24 20.34 Other Forbs* 14.24 19.47 10.12 10.72 35.89 21.09 25.34 Other Grasses* 13.46 15.64 10.79 8.79 48.05 18.23 24.97 Quercus laevis 11.1 7.88 17.56 5.54 0 12.39 14.87 Rhus copallinum 13.12 13.65 0 8.53 21.79 18.9 21.54 Ceratiola ericoides 0 0 21.03 1.64 1.68 4.19 0.69 Pinus palustris 5.5 2.36 2.12 0 4 2.26 7.96 Rubus cuneifolius 4.12 3.28 0 6.12 15.57 13.03 16.02 Quercus geminata 4.14 0.92 1.92 5.28 4.68 4.26 0 Eupatorium cappillifolium 3.29 1.84 0 5.35 13.9 4.92 15.24 Eremochloa ophiuroides 0 2.35 2.2 17.27 20.27 26.33 24.66 Quercus incana 0.82 3.68 0 5.1 0 8.06 5.47 Vaccinium darrowii 0 2.46 1.14 0 1.24 0.8 2.95 Toxicodendron toxicarium 3.34 0 0 0 0 0 0.49 Diospyros virginiana 2.09 1.03 0 0.74 2.62 1.93 3.96 Smilax spp.** 1.31 0 1.14 1.48 8.89 5.19 4.29 Gelsemium sempervirens 1.42 0 0 0.71 2.43 0.87 0.69 Paspalum notatum 0 1.32 0 0.37 0 0 0 Serenoa repens 0 0 1.18 0 1.68 0 0 Prunus serotina 0 1.03 0 0.68 1.24 3.19 0 Pteridium aquilinum 0 0 0.94 0.91 0 3.2 3.37 Asimina spp.** 0 0 0.57 0 0 0.33 0 Toxicodendron radicans 0 0 0 0 0 0.33 0 Vitis spp.** 0 0 0 1.11 6.82 3.59 0 Passiflora incarnata 0 0 0 0 0.62 0 0 Opuntia humifusa 0 0 0 0 1.24 0.33 0.49 Crataegus spp.** 0 0 0 0.48 0 0 0 Myrica cerifera 0 0 0 3.85 0 0.67 0 Rubus betulifolius 0 0 0 0 0.73 0 0 Vaccinium myrsinites 0 0 0 0 0 0.8 0 Vaccinium stamineum 0 0 0 0.88 0 0 0 Pinus elliottii 0 0 0 0 1.16 0 3.97 Prunus angustifolia 0 0 0 0 0 0 1.07 Quercus hemisphaerica 0 0 0 3.68 4.76 9.26 0.49 Quercus nigra 0 0 0 0.88 0 2.53 0.69 Quercus virginiana 0 0 0 0 0 1.67 0 Zanthoxylum clava-herculis 0 0 0 0.31 0 0 0 Panicum hemitomon 0 0 0 0 0.73 0 0 * Represents those species only identified to life form. ** Identified to genus. Table D-2 Tree seedling and sapling importance value. Site Species Albiziajulibrissin Diospyros virginiana Ilex opaca Pinus elliotti Pinus palustris Pinus taeda Prunus serotina Quercus geminata Quercus hemisphaerica Quercus incana Quercus laevis Quercus nigra Quercus virginiana Zamia clava-herculis 1 0.0 8.9 0.0 0.0 30.7 0.0 0.0 24.5 0.0 12.3 123.6 0.0 0.0 0.0 2 0.0 17.1 0.0 0.0 12.3 0.0 0.0 28.4 6.1 46.8 89.3 0.0 0.0 0.0 3 0.0 0.0 0.0 0.0 32.0 0.0 0.0 7.1 0.0 0.0 160.9 0.0 0.0 0.0 4 0.0 9.4 0.0 0.0 0.0 0.0 13.6 53.7 38.8 18.5 60.0 0.0 0.0 6.0 5 0.0 31.6 5.8 4.5 54.4 0.0 28.1 50.7 16.0 0.0 0.0 0.0 0.0 8.9 6 0.0 9.0 0.0 1.2 18.6 1.2 11.7 52.8 22.5 24.3 53.5 2.7 1.2 1.2 7 2.1 13.3 0.0 9.9 28.0 0.0 1.7 1.7 8.3 8.3 26.6 1.6 0.0 0.0 Table D-3 Shrub species importance value. Site Species 1 2 3 4 5 6 7 Asimina spp.** 0.0 6.6 14.7 2.0 3.3 4.0 0.0 Callicarpa americana 0.0 0.0 0.0 2.3 0.0 1.4 0.0 Ceratiola ericoides 0.0 0.0 134.1 8.4 5.1 16.1 1.8 Crateagus spp. 0.0 0.0 0.0 0.0 0.0 1.6 0.0 Hypericum spp. 4.7 4.9 0.0 2.0 0.0 0.0 0.0 Myrica cerifera 0.0 0.0 0.0 7.7 0.0 2.5 0.0 Opunitia humifusa 0.0 0.0 0.0 4.3 8.5 6.6 3.5 Rhus copallinum 93.2 93.9 11.3 107.9 100.8 87.0 89.9 Rubus cuneifolius 29.7 42.0 0.0 54.8 77.6 67.6 79.5 Serenoa repens 0.0 0.0 13.1 2.0 1.7 0.0 0.0 Toxicodendron toxicarium 62.1 0.0 0.0 0.0 0.0 0.0 2.2 Vaccinium arboreum 5.7 0.0 0.0 2.1 0.0 3.4 0.0 Vaccinium darrowii 4.7 46.1 26.9 0.0 3.0 6.5 23.1 Vaccinium myrsinites 0.0 3.3 0.0 0.0 0.0 3.2 0.0 Vaccinium stamineum 0.0 0.0 0.0 4.1 0.0 0.0 0.0 Zamia floridana 0.0 3.3 0.0 2.5 0.0 0.0 0.0 ** Identified to genus APPENDIX E TREE BASAL AREA Table E-1 Overstory tree basal area (ft.2/acre) Site Species R 4 5 6 7 Pinus palustris 27.50 1.25 0.00 1.67 0.42 Quercus laevis 14.17 0.42 0.00 1.67 0.42 Quercus geminata 0.42 1.25 0.00 1.67 0.00 Quercus virginiana 0.42 0.00 0.00 0.00 0.00 Prunus serotina 0.00 0.42 0.00 0.83 0.00 Quercus0.00 3.33 0.00 2.08 0.00 hemisphaerica Quercusincana 0.00 0.00 0.00 0.83 0.00 Quercus nigra 0.00 0.00 0.00 0.42 0.00 LIST OF REFERENCES Babbitt K. J. and Tanner G. W. 1997. Effective management for frogs and toads on Florida's ranches. University of Florida Cooperative Extension Service. WEC 16. Gainesville, Florida. Baker W. J. 1992. Effects of settlement and fire suppression on landscape structure. Ecology 73:1879-1887. Bonham C. D. 1989. Measurements for Terrestrial Vegetation. John Wiley and Sons, New York, New York. Brockway D. G. and Outcalt K. W. 1998. Gap-phase regeneration in wiregrass ecosystems. Forest Ecology andManagement 106: 125-139. Brockway D. G., Outcalt K. W., and Wilkins N. R. 1998. Restoring longleaf pine wiregrass ecosystems following low-rate hexazinone application on Florida sandhills. Forest Ecology and Management 103: 159-175. Croker T. C. Jr. 1987. LongleafPine: A History ofMan and a Forest. Atlanta, GA, USDA Forest Service.For. Rep. R8-FR7. Dickinson B., Huber W. C., Heaney J. P., and Brezonik, P.L. 1982. Gazetteer ofFlorida Lakes. Florida Water Resources Research Center. no. 63, Gainesville Florida. Digby P. G. N. and Kempton R. A. 1987. Multivariate Analysis of Ecological Communities. Chapman and Hall, London. Elzinga C. L., Salzer D. W., Willoughby J. W., and Gibbs J. P. 2001. Monitoring Plant andAnimal Populations. Blackwell Science, Malden, Mass. FDEP. 2003. Florida Department of Environmental Protection, Government Land Office Records. www.labins.org. Last accessed 3/02/03. The Florida Natural Areas Inventory (FNAI) 1990. Guide to the Natural Communities of Florida. Florida Department of Natural Resources. Gauch H. G. Jr. 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge. Gilliam F. S. and Platt W. J. 1999. Effects of long term fire exclusion on tree species composition and structure in an old-growth Pinus palustris (Longleaf pine) forest. Plant Ecology 140: 15-26. Glitzenstein J. S., Platt W. J., and Streng D. R. 1995. Effects of fire regime and habitat on tree dynamics in north Florida longleaf pine savannas. Ecological Monographs 65:441-476. Godfrey R. K. 1988. Trees .\ln uh\ and Woody Vines of Northern Florida and Adjacent Georgia and Alabama. University of Georgia Press. Athens, Georgia. Gordon D. 2003. Personal Communication. State Ecologist, The Nature Conservancy / University of Florida. Gainesville, FL. Griffith G. E. 1998. Lake Regions ofFlorida. Geological Survey. Reston, VA. Hedman C. W, Grace S. L., and King S. E. 2000. Vegetation composition and structure of southern coastal plain pine forests: an ecological comparison. Forest Ecology and Management 134: 233-247. Horton T. 1995. Longleaf pine: a southern revival. Audubon 97:74-82. Hum, J. 1989. GPS, a Guide to the Next Utility. Trimble Navigation, Sunnyvale, CA. Johnson R. and Gjerstad D. 1998. Landscape-scale restoration of the longleaf pine ecosystem. Restoration and Management Notes 16: 41-45. Jordan W. R., Gilpin M. E., and Aber J. D. 1987. Restoration ecology: ecological restoration as a technique for basic research. P. 3-23 In: Restoration Ecology, A Synthetic Approach to Ecological Research. Jordan, W. R., Gilpin, M.E., Aber, J.D. (eds.) Cambridge University Press. Kennedy K. A. and Addison P. A. 1987. Some considerations for the use of visual estimates of plant cover in biomonitoring. Journal of Ecology 75: 151-157. Krebs C. J. 1998. Ecological Methodology, Second edn. Addison -Wesley, Menlo Park, California. Lewis, R. 1997. Global positioning systems P. 251-258 In GISData Conversion: Strategies, Techniques, Management. Hohl, P. (eds.). Onword Press, Santa Fe, New Mexico. Long A. J. 2002. Benefits of prescribed burning.University of Florida Cooperative Extension Service.Gainesville, Florida. Ludwig J. A. and Reynolds J. F. 1988. Statistical Ecology. John Wiley and Sons, New York, New York. 61 McCoy E. D. and Mushinsky H. R. 2002. Measuring the success of wildlife community restoration. Ecological Applications 12:1861-1871. Miller J. H. and Miller K. V. 1999. Forest Plants of the Sn.,,ihei, and Their Wildlife Uses. Southern Weed Science Society. Champaign, Illinois. Monroe M. C., Babb G., and Heuberger K. A. 1999 Designing a prescribed fire demonstration area. Gainesville, Florida, Cooperative Extension Service, University of Florida. Myers R. E. 1990. Scrub and high pine. P. 150-193 In Ecosystems ofFlorida. University Presses of Florida. Gainesville, Florida. Myers R. E. and White D. L. 1987. Landscape and history changes in sandhill vegetation in north-central and south-central Florida. Bulletin of the Torrey Botanical Club 114: 21-32. Olmert B. 2001. Personal Communication. President, Loncala Corporation. High Springs, Florida. Ott, L. R. 1993. An Introduction to Statistical Methods and Data Analysis.Wadsworth, Inc. USA. Outcalt K. W. 2000. The longleaf pine ecosystem of the south. Native Plants Journal 1: 43-51. Philips, E. A. 1959. Methods of Vegetation Study. Henry Holt and Company, Inc. USA. Pielou E. C. 1984. The Interpretation of Ecological Data. John Wiley and Sons, New York, New York. Provencher L., A. R. Litt, and Gordon D. R. 2000. Compilation of Methods Used by the LongleafPine Restoration Project from 1994 1999 at Eglin Air Force Base, Florida. Product to Natural Resources Division, Eglin Air Force Base, Niceville, Florida. Science Division, The Nature Conservancy, Gainesville, Florida. Radford A.E., Ahles, H.E. and Bell C.R. 1968. Manual of the Vascular Flora of the Carolinas. Univ. N. Carolina Press: Chapel Hill, North Carolina. Reed D. D. and Mroz G. D. 1997. Resource Assesment in Forested Landscapes. John Wiley and Sons Inc., New York, New York. Rodgers L. H., and Provencher, L. 1999. Analysis of longleaf pine sandhill vegetation in northwest Florida. Castanea 64: 138-162. Seamon G. 1998 A longleaf pine sandhill restoration in northwest Florida. Restoration and Management Notes 16.46-50. 62 Thompson S.K. 1992. Sampling. John Wiley and Sons Inc. New York, New York. Wade D. D. 1988.A Guide for PrescribedFire in S. ,tuihei Forests. USDA, Forest Service. WolfP.R. and Brinker, R.C. 1994. Elementary Surveying. Harper Collins Publishers. New York, New York. Wunderlin R. P. 1998.Guide to the Vascular Plants ofFlorida. University of Florida Press, Gainesville, Florida. BIOGRAPHICAL SKETCH Michael Rasser was born February 11, 1975 in Bayshore, New York He spent several years working in the field of environmental education, including an internship in the Zoo of The Newark Museum, New Jersey, and a part-time instructor in the Miami- Dade Community College Environmental Center. He graduated Cum Laude with his B.A. in Environmental Studies (with a minor in Biology) from Florida International University. He developed an interest in ecological research while spending two summers as part of a research team examining the effects of global warming on Alaskan arctic plants. Mike plans to continue his scientific pursuits by pursuing a Ph.D. program at the University of Texas. Full Text PAGE 1 MONITORING AND ASSESSING LONGLEAF PINE ECOSYSTEM RESTORATION: A CASE STUDY IN NORTH-CENTRAL FLORIDA By MICHAEL K. RASSER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003 PAGE 2 Copyright 2003 by Michael K. Rasser PAGE 3 ACKNOWLEDGMENTS There are many people I would like to thank for their contributions to this manuscript. Loukas Arvanitis was a true mentor throughout my graduate student career here at the University of Florida. My committee members, Wendell Cropper, Alan Long, and George Tanner, always provided excellent advice and guidance in their particular areas of expertise. This project required extensive field data collection, in very hot and trying conditions. Without David Zabriskie, Christine Housel, Nik Chourey, Willie Wood, Brett Jestrow, and Louise Lundberg, this work would not have been possible. My co-workers here in the Forest Information Systems Lab, Balaji Ramachandran, Steve Moore, and Douglas Shipley, were essential in providing logistical support, especially when it came to using GIS and GPS. The Florida Division of Forestry provided the funding for this project. I would especially like to thank Charlie Marcus, DOF State Lands Coordinator, for his generous support and help throughout this project. Mpower3/Emerge donated high-resolution aerial photography, which proved to be invaluable in the field. Robin Boughton, and the Goethe State Forest Staff, were always very supportive. Throughout my research there were many people who provided their personal knowledge, especially Brian Olmert, President of the Loncala Corporation; Nancy Coile, Division of Plant Industry, DOACS; Michael Drummond, Alachua County iii PAGE 4 Environmental Protection, and Dr. Doria Gordon, Department of Botany, UF/The Nature Conservancy. Of course, I have to thank my friends and family who provided endless moral support. Soumya Mohan tirelessly reviewed this manuscript, and she and Arjun Mohan tolerated my long physical and mental absences, for which I am greatly indebted. iv PAGE 5 TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................ix LIST OF FIGURES...........................................................................................................xi ABSTRACT......................................................................................................................xii CHAPTER 1 THE PROBLEM...........................................................................................................1 Study Objectives...........................................................................................................1 Outline..........................................................................................................................2 2 LITERATURE REVIEW.............................................................................................3 History of the Longleaf Pine Ecosystem......................................................................3 Sandhill Ecology...........................................................................................................3 Restoration Ecology......................................................................................................4 Historical Information in Restoration Ecology.............................................................4 Sources of Historical Information................................................................................5 Aerial Photography................................................................................................5 General Land Office Survey..................................................................................5 Other Sources of Information................................................................................5 Challenges for Longleaf Pine Ecosystem Restoration..................................................6 Ground Cover........................................................................................................6 Fire.........................................................................................................................7 Monitoring and Assessing Restoration Success...........................................................8 What is Restoration Success..................................................................................9 How is Restoration Success Monitored and Assessed..........................................9 Reference Sites....................................................................................................10 Vegetation Sampling..................................................................................................10 Community Monitoring.......................................................................................10 Measuring Vegetation.........................................................................................10 Percent foliar cover......................................................................................11 Density.........................................................................................................11 Frequency.....................................................................................................12 v PAGE 6 Basal area.....................................................................................................12 Important Considerations for Sampling......................................................................12 Permanent or Temporary Sampling Locations....................................................12 Number of Sample Units.....................................................................................13 Sample Design............................................................................................................14 Simple Random Sampling...................................................................................14 Stratified Random Sampling...............................................................................15 Systematic Sampling...........................................................................................15 Statistical Analysis of Plant Communities.................................................................15 Indicators.............................................................................................................16 Measures of Similarity........................................................................................16 Classification.......................................................................................................17 Ordination............................................................................................................17 Statistical Analysis Based on Re-sampling.........................................................17 3 STUDY AREA...........................................................................................................19 Location......................................................................................................................19 Physiography.......................................................................................................19 Land Use..............................................................................................................19 Biological Communities......................................................................................20 Sandhill Upland Lakes........................................................................................20 Sandhill................................................................................................................22 Scrub....................................................................................................................23 Hardwood Hammock...........................................................................................23 Prairie..................................................................................................................23 Ephemeral Pond...................................................................................................24 Management Considerations......................................................................................24 4 METHODOLOGY.....................................................................................................25 Vegetation Data Collection.........................................................................................25 Ground Cover......................................................................................................26 Shrubs, Seedlings and Saplings..........................................................................26 Trees....................................................................................................................27 Species Sampled..................................................................................................27 Classification of Comparative Study Areas................................................................27 Nonreference Areas.............................................................................................27 Site 4....................................................................................................................28 Site 5....................................................................................................................28 Site 6....................................................................................................................28 Site 7....................................................................................................................29 Reference Sites....................................................................................................29 Field Data Interpretation.............................................................................................30 Importance Values...............................................................................................31 Basal Area of Trees.............................................................................................31 Similarity Analysis..............................................................................................31 vi PAGE 7 Determination of Targets for Monitoring............................................................32 Proportional Abundance of Species....................................................................33 5 RESULTS AND DISCUSSION.................................................................................35 Comparison Sites........................................................................................................35 Understory...........................................................................................................35 Shrubs..................................................................................................................35 Tree Seedlings and Saplings................................................................................36 Summary of Similarity Indices............................................................................36 Suggested Target Species...........................................................................................36 Understory...........................................................................................................37 Wiregrass (Aristida beyrichiana).................................................................37 Centipede grass (Eremochloe ophiurides)...........................................................38 Dog-fennel (Eupatorium capilifolium).........................................................38 Green briar (Smilax spp.).............................................................................38 Total forb cover............................................................................................39 Shrubs..................................................................................................................39 Winged sumac (Rhus coppallinum).............................................................39 Sand blackberry (Rubus cuneifolius)............................................................39 Florida rosemary (Ceratiola ericoides)........................................................40 Tree Seedlings and Saplings................................................................................40 Turkey oak...........................................................................................................41 Black cherry (Prunus serotina)....................................................................41 Laurel oak.....................................................................................................41 Trees....................................................................................................................41 Longleaf pine................................................................................................42 Turkey oak....................................................................................................42 6 SUMMARY AND CONCLUSIONS.........................................................................43 Suggestions for Future Research................................................................................43 Further Exploratory Analysis of Field Data...............................................................43 Determination of Target Abundances.........................................................................44 Effect of Centipede Grass on Longleaf Pine Restoration...........................................44 ABSTRACT A SPECIES SIMILARITY.............................................................................................45 B PROPORTIONAL DISTRIBUTION OF SPECIES AMONG REFERENCE AND NONREFERENCE SITES.........................................................................................48 C SPECIES RANKING.................................................................................................51 D IMPORTANCE VALUES..........................................................................................55 E TREE BASAL AREA................................................................................................58 vii PAGE 8 LIST OF REFERENCES...................................................................................................59 BIOGRAPHICAL SKETCH.............................................................................................63 viii PAGE 9 LIST OF TABLES Table page A-1 Understory similarity among comparison sites based on Jaccards Index...............45 A-2 Understory similarity among comparison sites based on Sorensens Index............45 A-3 Understory similarity among comparison sites based on Curtis Bray Index...........45 A-4 Shrub similarity among comparison sites based on Jaccards Index.......................46 A-5 Shrub similarity among comparison sites based on Sorensens Index.....................46 A-6 Shrub similarity among comparison sites based on Curtis Bray Index....................46 A-7 Tree seedling and saplings similarity among comparison sites based on Jaccard's coefficient.................................................................................................................47 A-8 Tree seedling and saplings similarity among comparison sites based on Sorensen's coefficient...............................................................................................47 A-9 Tree seedling and saplings similarity among comparison sites based on Curtis Bray coefficient........................................................................................................47 B-1 Distribution of understory species among reference and nonreference points........49 B-2 Distribution of shrub species among reference and nonreference points.................50 B-3 Distribution of tree seedling and sapling species among reference and nonreference points..................................................................................................50 C-1 Understory species ranking......................................................................................52 C-2 Shrub species ranking...............................................................................................53 C-3 Tree seedling and sapling species ranking...............................................................54 D-1 Understory species importance value.......................................................................56 D-2 Tree seedling and sapling importance value............................................................57 ix PAGE 10 D-3 Shrub species importance value...............................................................................57 E-1 Overstory tree basal area (ft2/acre)...........................................................................58 x PAGE 11 LIST OF FIGURES Figure page 3-1 Study area location...................................................................................................21 4-1 Sampling architecture...............................................................................................26 4-2 Comparative study areas..........................................................................................30 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 MONITORING AND ASSESSING LONGLEAF PINE ECOSYSTEM RESTORATION: A CASE STUDY IN NORTH-CENTRAL FLORIDA By Michael K. Rasser August 2003 Chair: Loukas G. Arvanitis Major Department: School of Forest Resources and Conservation In a cooperative project with the Florida Division of Forestry, a methodology was developed to monitor longleaf pine restoration. This project may serve as a model for future management of the sandhill longleaf pine community at Goethe State Forests Watermelon Pond Unit (WPU). A permanent network of sample points was established throughout the 4,778 acres (1934 ha) of the WPU. Baseline data were collected on vegetative structure and composition at the understory, midstory, and tree levels. Data collected included cover, density, and frequency, of over 80 species. Sample points were permanently marked in the field so that they could be relocated in the future for monitoring changes over time. A series of comparison areas were identified throughout the property, that were representative of the various ecological conditions of sandhill longleaf pine community, based on anthropogenic influences. These sites include areas that had been severely disturbed in the past, as well as reference sites representing desirable future conditions. xii PAGE 13 Site selection was based on available information regarding the presence of past disturbances, such as timber production, farming, and grazing. A statistical analysis was conducted to compare all study areas. The analysis included determining importance values of the different species as well as measuring similarity among sites. Species were ranked to determine which ones were responsible for the greatest difference among the restored and reference sites. From this list, a series of indicator species was chosen to allow land managers who are concerned about future monitoring to measure the level of success of their restoration efforts. xiii PAGE 14 CHAPTER 1 THE PROBLEM Florida legislation, such as Preservation 2000 and Florida Forever, has placed a large amount of land into public ownership for ecological restoration. Land managers who are challenged with the task of restoring dynamic ecosystems must have a thorough understanding of the current state of the vegetation and the desired future conditions in order to establish realistic management goals. However, vegetation monitoring programs are often hampered by the high cost of collecting, analyzing and interpreting field data. Restoration of the longleaf pine (Pinus palustris) sandhill community at the Watermelon Pond Unit (WPU) of Goethe State Forest provides an example of such a project. Once part of an extensive longleaf pine ecosystem, it has since been degraded through anthropogenic influences such as clear cuts, mining, farming and ranching. Current land managers are charged with the difficult task of returning this land to its earlier, more natural state. Study Objectives This pilot study aims to address some of the challenges associated with monitoring and measuring restoration success of the longleaf pine community at the WPU. The specific objectives were to Establish a protocol for a statistical sampling design that allows adaptive management of restoration success. Collect baseline field data. 1 PAGE 15 2 Conduct a comparative statistical analysis of the vegetative condition of the longleaf pine community based on past land use. Develop a list of indicator species for future monitoring. Outline Chapter 2 reviews the literature on the science of restoring the longleaf pine ecosystem. It focuses on the ecology of the longleaf pine ecosystem, obstacles to restoration, and important aspects of sampling and monitoring such plant communities. Chapter 3 describes the study site, the Watermelon Pond Unit of Goethe State Forest in Florida. Chapter 4 outlines the methodologies used in our study and the rationale behind their use. Chapter 5 examines the results obtained from using these methods. Chapter 6 summarizes the results with a discussion of the various indicators chosen for monitoring restoration success. Suggestions are also made for future research and analysis related to this area. PAGE 16 CHAPTER 2 LITERATURE REVIEW History of the Longleaf Pine Ecosystem Longleaf pine was once the dominant tree species throughout 24.3 million ha of forest land in the southeastern United States (Outcalt 2000). A small fraction, perhaps as little as 1.2 million ha, remains today (Hedman et al. 2000). The vast majority of this forest is fragmented secondary growth that regenerated after the railroads expedited the logging and clearing of these forests (Croker 1987). The largest remaining tract of old growth longleaf is only about 200 acres (81 ha) in size (Johnson and Gjerstad 1998). The degradation of this ecosystem may be attributed to a variety of factors including land conversion, fire suppression, logging, and preference for other pine species in silvicultural practices (Croker 1987, Gilliam and Platt 1999). Sandhill Ecology Large differences exist in the vegetative structure and composition of the longleaf pine ecosystem. These may be attributed to fire regimes, physical soil variability, anthropogenic soil disturbances, and geographic location (Myers and White 1987, Rodgers and Provencher 1999). Myers (1990) lists clayhills, turkey oak (Quercus laevis) barrens, longleaf pine/turkey oak, and upland pine forests among the many plant communities that have been described for the longleaf pine ecosystem. Sandhill pine, also referred to as High Pine, is a typical plant community on the sandhill ridges of peninsular Florida. This community typically contains a sparse canopy dominated by longleaf pine, low-stature midstory composed of xerophytic oaks and 3 PAGE 17 4 clonal shrubs, and a herbaceous ground cover dominated by wiregrass (Aristida beyrichiana) (Brockway et al. 1998, Myers 1990, Rodgers and Provencher 1999). Long periods of fire suppression increase the importance of hardwood species and decrease the forbs and graminoids component of the ground cover (Rodgers and Provencher 1999). Sandhill diversity may decrease after fire suppression. However, high levels of regional variation have made it very difficult to understand the community dynamics of these systems. Restoration Ecology Ecological restoration is an emerging science, tracing its roots to the arboretum at the University of Wisconsin, Madison (Jordan et al. 1987). It was there that a civilian conservation group undertook a painstaking restoration of a prairie on several acres of farmed land. The project was a landmark attempt to reverse damages that had been caused by anthropogenic disturbances, and to return the land to an earlier, ecologically more natural state. Although the desired objective of ecological restoration may vary, the idea of returning the land to historically natural conditions is ubiquitous in all restoration projects. For example, in the case of reclaimed mine lands, the desired result may simply be to vegetate an area in order to prevent erosion. However, more emphasis is now being placed on overall ecosystem restoration, which involves a comprehensive process of returning an ecological system to its naturally functioning state. Historical Information in Restoration Ecology One difficulty in restoring natural systems is the lack of knowledge about past conditions. This is particularly evident in plant communities where few representative PAGE 18 5 examples still exist. To establish realistic management objectives, managers need adequate information. Sources of Historical Information Aerial Photography Aerial photography is a valuable tool for assessing past conditions of land by providing a record of anthropogenic and natural disturbances at any given point in time. The United States National Archives maintain a large collection of aerial photographs taken before 1950. Local libraries, tax assessment offices, county offices and agricultural units often make aerial photographs available to interested individuals, students, and land managers. General Land Office Survey The United States General Land Office was responsible for creating land township plat maps for the entire country. This survey was primarily completed in the mid 19th century. In addition to installing survey markers at property corners, contractors usually recorded the condition of the property (Wolf and Brinker 1994). These notes are often readily available. For example, the Florida Department of Environmental Protection has electronic versions of the original survey notes available online (FDEP 2003). Other Sources of Information Landowners and others intimately associated with an area may be valuable sources of information regarding past land use and conditions of the land. For example, Brian Olmert, President of the Loncala Corporation (High Springs, Florida), provided extensive details on the companys prior land uses such as extent of cattle grazing, location and details of timber operations, as well as personal observations regarding past ecological PAGE 19 6 conditions of the land. These details are extremely useful when assessing current land conditions and establishing future restoration goals. Challenges for Longleaf Pine Ecosystem Restoration Restoration of the longleaf pine ecosystem is a challenging task for a variety of reasons. The two major ones are returning the natural fire regime to the land and recovering adversely affected ground cover (Glitzenstein et al. 1995). Fairly extensive research has been conducted in recent years to surmount these obstacles. Ground Cover With the exception of tropical rainforests, longleaf pine forests are among the most diverse ecosystems, sometimes containing over 40 species of plants in a square meter (Hedman et al. 2000). However, this diversity poses additional problems in restoring this unique ecosystem. Although much is known about planting trees, sufficient evidence is not available about restoring the diverse understory (Horton 1995). The presence of wiregrass is probably the most important component of the understory in these forests (Brockway et al. 1998). In fact, longleaf pine and wiregrass may be considered keystone species in the longleaf pine forest (Brockway and Outcalt 1998). The reason is that wiregrass both facilitates, and is tolerant of fire. An abundance of wiregrass creates a continuous layer of fine fuels in the understory that encourages and sustains frequent low intensity fires that were historically part of the fire regime. However, there is a lack of knowledge about the historical extent of wiregrass, which is very sensitive to soil disturbance (Rodgers and Provencher 1999). During restoration, the actual planting of longleaf trees is relatively inexpensive. Due to the economic value of the species, a fair amount of information is known about how to plant and raise the seedlings to maturity. The difficulty lies in reconstructing the PAGE 20 7 diverse understory to include wiregrass and other important species. The Nature Conservancy is currently conducting an ambitious restoration of a longleaf pine ecosystem on a sandhill in northwest Florida (Seamon 1998). The project seeks to restore both longleaf and the wiregrass dominated understory. Seeds are collected from nearby areas that contain wiregrass, and are distributed in the area being restored. Initially, manual methods were used to collect the seeds of wiregrass. This method was very expensive (about$3000 /acre). As a result, more cost-effective mechanical methods are being developed. The seeds are now distributed with a commercial hay blower. More research is needed to improve and refine these methods so that they may be applicable to a larger scale. Fire The most important, yet missing, disturbance in many current or potential sites of longleaf restoration, is fire, which is necessary to maintain forest diversity and structure. Fire suppression leads to an increase in hardwood trees and shrubs and eventually reduces plant diversity (Glitzenstein et al. 1995). To address this issue prescribed fire has become an invaluable tool for returning natural fire regimes (Long 2002, Wade 1988). However, due to smoke, local residents are often concerned about the use of prescribed fires (Monroe et al. 1999). In addition, there are dangers of burns escaping control and causing health problems and damage to nearby lands. As more people move into rural areas, the smoke shed, which is the area affected by smoke during a fire, impacts more people. Also, a return to a natural fire regime after decades of fire suppression, can be ecologically damaging (Baker 1992). Build up of fuel loads up can cause fires to burn more intensively than they would have when there were frequent low intensity fires. This

PAGE 21

8 unnatural fuel loading can create intense fires that may cause mortality in plants and animals that would have otherwise survived. An alternative approach to removing hardwood trees that are encroaching in areas where longleaf pine is under restoration is the use of herbicides. Recent study have used the herbicide hexazinone to eliminate the hardwoods (Brockway et al. 1998, Provencher et al. 2000). Broadly applied, the herbicide killed a diverse amount of plants and was not very effective in restoring the natural vegetation present in the understory. However, when applied more selectively in a spot application, particularly at levels of 2.2kg/ha, it was effective in eliminating hardwoods and allowing wiregrass to regain dominance. Monitoring and Assessing Restoration Success Restoration ecology is an experimental science. Much is to be learned about restoring longleaf pine communities. Long term monitoring allows restoration efforts to take place within the framework of a scientific experiment. The success or failure of management methods can be tested and assessed. This allows for adaptive management and may lead to a better understanding of the ecological dynamics of the system. Monitoring is conducted for two primary purposes: a) to allow land managers to monitor the relative success or failures of their efforts, and b) for adaptive management purposes. In order to accomplish these objectives, monitoring efforts must establish criteria for determining restoration success. For example, land managers may wish to reduce hardwood tree dominance in a longleaf community. Monitoring would allow land managers to examine the effects of prescribed burns on reducing hardwoods. If it were determined that a change in prescribed fire frequency or seasonality was more effective, management could be adapted accordingly.

PAGE 22

9 Whether or not a restoration plan fulfills all its desired objectives, valuable knowledge can be attained. Ecological restoration projects allow for basic research on how various components of ecosystems and plant communities function and interact. This is accomplished through a mechanistic approach to the restoration process. According to Jordan et al. (1987) . . one of the most valuable and powerful ways of studying something is to attempt to reassemble it, to repair it, and to adjust it so it works properly. Monitoring efforts allow us to measure the effects of our tinkering with the ecological communities. What is Restoration Success Restoration success is a relative term that largely depends upon the stated objectives. Restoration objectives can be clearly stated prior to the start of the project, such as to increase the population of a particular species, or group of species. For example, a wetland restoration objective may be to increase the number of wading birds. Sometimes goals are vague, such as the restoration of historic ecosystem conditions. Ill defined projects run the risk of not being successful due to the inherent complexity of natural systems. Therefore, a restoration project must have clearly defined goals in order to establish measures of restoration success. How is Restoration Success Monitored and Assessed Monitoring vegetative community structure is one ways to assess restoration success. Restoration success is usually measured in relation to one or more representative reference sites that represent the ideal future condition of the system being restored. A variety of univariate, multivariate, and qualitative methods are used to compare and monitor areas being restored with reference sites (Elzinga et al. 2001).

PAGE 23

10 Reference Sites Choosing the correct reference site or sites is an important part of any monitoring program for restoration efforts. Reference sites should represent the ideal future conditions of the site that needs to be restored. In the case of ecosystems that have been reduced in extent, finding an ideal representative site is difficult. In addition, ecological communities can vary widely primarily, due to edaphic factors. Therefore, it is essential that reference sites be located as near as possible to the potential restoration site. Once the reference site, or sites have been chosen, monitoring restoration progress is accomplished by comparing the reference site(s) with those that are being restored. Monitoring efforts are usually administered at the ecological community level. As such, they focus upon the assemblage of plants and or animals at a particular site. Vegetation Sampling Community Monitoring Monitoring plant communities can be accomplished through statistical sampling. Vegetation data are collected from a representative sample of plants. Characteristics that may be measured include: foliar cover, density, frequency, and basal area. Following vegetation measurements, statistical analyses can be used to characterize and describe the plant community. In addition, there are a variety of indices that may be used to assess the similarities among reference sites and the sites to be restored. Measuring Vegetation There are a multitude of methods for measuring vegetation (Bonham 1989, Phillips 1959). The variables measured during the course of our study and discussed below include percent cover, density, and basal area. Each of these variables, including some of the associated sampling errors, are discussed below.

PAGE 24

11 Percent foliar cover Percent cover is one of the most commonly used measurements. Its popularity is probably attributed to the ease with which it can be ocularly estimated. One of the most common method for conducting this type of sampling is to use a fixed area quadrat within which species are identified along with their percent cover. However, there are many errors that must be taken into account when using ocularly estimated measurements of vegetative cover. Kennedy and Addison (1987) conducted a study that quantified some of the biases in such measurements. The authors found that species morphology, distribution of species, and species identification, were the most common sources of error. Measurements of plants with compound leaves, proportionally large amounts of woody material, and variable leaf size were more variable than others. Rare species or inconspicuous species are often missed because they are widely dispersed. In addition, misidentification of species accounted for a large amount of errors. Another method for assessing vegetative cover is the line-intercept method, which compares the length of each species along a line with total line length (Phillips 1959). This methods was not used in our study because it was found to be less efficient to permanently locate a line, as compared to a quadrat, due to an abundance of shrubs at the study site. These non-sampling errors can be reduced in several ways. Having more than one observer examine each quadrat may reduce the error due to species being identified incorrectly or from simply missing a plant. In addition, if the same work team is used throughout the sampling process, biases among individuals can be reduced. Density Density is the number of individuals per unit area. In vegetation sampling, density is usually determined by utilizing fixed area plots. Plots of all sizes and shapes are used.

PAGE 25

12 However, the ability to detect plants, and boundary errors, are important considerations for selecting appropriate size, shape, and number of plots (sample size). Quadrats that are too large may prevent the observer from locating all species. A significant source of error associated with using quadrats for measuring density occurs at the edges, as plants may falsely be identified as being in or out of the quadrat. Frequency Plant frequency may be defined as the percentage of total plots within which a species of interest can be found (Phillips 1959). Frequency is solely determined by the presence of the species, without regard to abundance. Obviously, larger quadrats may have a greater chance of including plants, and as such, frequency is strongly affected by plot size. Basal area Basal area is the cover of a plants stem expressed as square feet per acre or square meters per hectare. The basal area of trees can be estimated in several ways, one of the most common being a properly calibrated prism. As opposed to fixed area plots, prisms deal with probabilities proportional to diameter at breast height of trees. The procedure is known as point sampling. Larger trees have a greater probability of being included in point sampling than smaller trees. Shrubs and other obstructions can obscure the observers view of the tree stem, reducing the ability to detect the critical distance to determine whether a tree is included in a plot (Reed and Mroz 1997) Important Considerations for Sampling Permanent or Temporary Sampling Locations One decision that must be made when establishing a sampling protocol is whether the sample units should be located permanently or temporarily. If plots are temporary, it

PAGE 26

13 will not be possible to assess changes over time as easily as with permanent plots, points, transects or strips. This is because permanent sampling units eliminates the confounding effects of spatial heterogeneity inherent in plant populations. Overall, the main advantage of permanent sampling units is that it is easier to statistically prove changes over time than with temporary samples (Elzinga et al. 2001). The primary reason for not always using permanent sample units is the higher cost compared to temporary samples. Devices for permanent sample units include metal wire, posts, and other materials, all of which contributes substantially to the cost of the project. The advent of the Global Positioning System (GPS) in the 1990s has led to an increase in the cost effectiveness of utilizing permanent sampling units. GPS was developed by the U.S. Department of Defense to allow accurate location of nuclear submarines (Hurn 1989). It has since gained widespread use for civilian applications. The system is based on a network of 24 satellites orbiting the earth at an altitude of 12,600 miles (Lewis 1998). Using a method similar to tri-lateration, distance from the ground and the orbital location of three or more satellites is used to accurately locate positions on earth. During the course of our study GPS reliably placed sample units less than one-meter from their intended location. Number of Sample Units The number and size of sampling units is an important consideration for any sampling design. The most important consideration when determining sample size is the objective of the sampling (Elzinga et al. 2001). A larger sample size increases precision, however the gain in precision lessens as sample size increases. Collecting samples is generally very expensive; therefore the purpose should be to collect enough samples in a

PAGE 27

14 cost effective manner that will estimate the variable of interest with a predetermined level of confidence. For example, suppose the management objective is to detect change in basal cover of wiregrass as an indicator of understory conditions in a longleaf pine sandhill (Brockway et al. 1998) The goal of the land managers may be to have wiregrass cover in a restoration site with 10% error of sampling. The role of the investigator is to determine the sample size and the appropriate method of allocation of samples that will yield the reliable and acceptable sample estimates. Non-Sampling Errors Errors that are made during the process of measuring a variable are considered non-sampling errors (Thompson 1992). One source of error is transcription and recording errors. Field data are often recorded in a notebook, later entered into a computer, and transformed and analyzed. During each step of this process human errors can occur such as errors in measuring, recording, and eliminating or failing to include all sample units. In sampling vegetation, these errors can be significant. Sample Design Simple Random Sampling Simple random sampling requires that each sample unit be independent of each other and have the same probability of being selected. A variety of methods are employed to create random points such as the simple random-coordinate method (Elzinga et al. 2001). Using this method, a Cartesian plane is transposed onto a map of the sample area, and random X and Y coordinates can be selected. With the advent of geographic information systems (GIS) technologies, there are computer programs

PAGE 28

15 available that can easily create a number of random points throughout the sampling frame. Stratified Random Sampling Stratified sampling is a type of sampling where the sampling frame is first divided into strata, or relatively homogenous regions (Reed and Mroz 1997). Samples can be allocated to each stratum in different numbers. For example, if one stratum contains more variation in the attribute of interest, more samples can be allocated to it. As such, when the sample area can easily be allocated to strata, it usually results in greater sampling efficiency. Systematic Sampling In systematic sampling the starting point for sampling is chosen randomly, and successive samples are chosen at intervals from the starting point (Thompson 1992). Since the first sample unit is the only one chosen at random, each of the sampling units is not independent of each other. Estimates of means, totals and percents are usually similar to simple random sampling. The same cannot be said for their respective sample variances. An example of systematic sampling is a transect in which the starting point is chosen at random and samples are selected at regular 100 m intervals thereafter. An advantage of systematic sampling is the ability to increase sampling efficiency by reducing travel time between sample units. In addition, systematic sampling can be good at distributing samples uniformly throughout the sample area (Reed and Mroz 1997). Statistical Analysis of Plant Communities There are several methods to analyze plant communities that do not involve quantitative measures (Elzinga 2001). One qualitative method for monitoring plant communities the use of photography. By taking a series of photographs, over time in the

PAGE 29

16 same location, land managers can keep track of changes. Other methods include the use of site assessments and species checklists where vegetative communities can be categorized based on an individuals observations. Although qualitative methods can be valuable one must recognize the limitations of such methods given individual biases. Indicators Sometimes a single species or variable can be used as an indicator for monitoring changes in community structure and composition. Indicators are surrogate measures for other variables that are more difficult or more expensive to measure. For example, cattails (Typha spp.), that form the dominant vegetative component in areas of the oligotrophic everglades are affected by eutrophication from agricultural runoff. Rather than measure phosphorous levels in water, perhaps land managers could use cattails as an indicator of eutrophication. This could perhaps be more economical, but the effectiveness of all indicators is based upon a strong correlation with the attribute of interest. Measures of Similarity Ecologists have developed many methods to measure how closely one vegetative community resembles another. One of the most common is the use of similarity indices. Some similarity measures rely on the presence or absence of species and do not account for the difference in abundance among species (species evenness) (Digby 1987, Ludwig 1988). Two examples of such indices are that of Jaccards and Sorensens. Jaccards index is simply the proportion of species common to two sites divided by the total number of species found in each site. Sorensens index divides the total number of common species by the mean, rather than total number of species (Reed and Mroz 1997). Other methods, such as the Curtis-Bray, incorporate species abundance, accounting for

PAGE 30

17 species evenness. Each of these measures has been used in our study and the respective equations can be found in Chapter 4. Classification Classification is the process of grouping items based on similar properties. Early classification methods relied primarily on the subjectivity of expert ecologists to place communities into classes (Pielou 1982). Today, the use of computers allows more sophisticated and quantitatively rigorous classifications to be made. To classify plant communities, it is necessary to have sample data, usually in the form of species presence and some measure of abundance. Since plant communities do not often occur in discrete classes, interpretation can be problematic. Ordination Unlike classification, which clusters data into groups, ordination arranges samples along one or more coordinate axes in multi-dimensional space (Pielou 1984). Axes are usually based upon environmental variables. As plant communities are often continuous, ordination is a natural way to represent data over gradients (Gauch 1982). However, the multi-dimensional aspect of ordination can make interpretation of results more complex than classification. Statistical Analysis Based on Re-sampling Recently, a variety of methods based on re-sampling data have become popular in analyzing ecological data. The value of these methods is that one does not have to adhere to the assumptions of normality required for parametric statistics. The major utility in these methods is the ability to construct confidence intervals and conduct significance tests that might not be possible with other statistical methods (Elzinga et al. 2001). Two popular methods are the bootstrap and the jackknife.

PAGE 31

18 The bootstrap technique involves re-sampling the samples as if it was an entire population. All the original data are pooled and recombined. Bootstrap sampling is conducted by taking a sample of these data and calculating the parameter of interest for each sample. After repeated sampling, often in the several thousands, one can calculate the mean and standard error from the bootstrap estimates (Krebs 1998). From these intervals confidence values and significance values can be assigned. The jackknife is similar to the bootstrap, except that re-sampling is done without replacement. The data are recombined by removing one random sample each time. Every time the samples are recombined, the mean and standard error are calculated. The jacknife, due to the limited amount of combinations inherent with sampling without replacement does not require intensive calculations (Krebs 1998).

PAGE 32

CHAPTER 3 STUDY AREA Location Our study was conducted at Goethe State Forest, in Floridas Levy and Alachua Counties. The 4,778-acre (1911 ha) Watermelon Pond Unit (WPU) is a disjoint collection of parcels primarily to the south and west of Watermelon Pond (latitude 29 32 42, longitude 82 36 20), on the Brooksville Ridge in north-central Florida (Fig. 3-1). Physiography The WPU is on the northern part of the Brooksville Range, an area characterized by deep marine deposited soils of the Candler-Apopka-Astatula associations (Dickinson et al. 1982). The elevation is quite variable for this region of Florida, ranging from 45 to 125 feet (14 to 38 m) above sea level. The areas of lowest elevation may contain sandhill upland lakes or depressional marshes. These lakes are usually acidic and fairly oligotrophic (Griffith 1998) Land Use During the past century, a variety of land uses have impacted the natural communities of the WPU. These include phosphate mining, timber removal, and cattle grazing. General land office survey records indicate that much of the upland areas of the WPU were pine forest in 1840s (FDEP 2003). Logging operations cleared most of the longleaf pine trees by the 1930s and they were apparently left to regenerate (Olmert 2001). Very few mature longleaf trees are present on the property today. More than 19

PAGE 33

20 1000 acres (405 ha) of the property were planted with slash pine (Pinus elliotti) between 1969-1971 and cleared before purchase of the land by the state of Florida in 1995. Previously planted areas are primarily located on the Watermelon Pond North tracts (Fig. 3-1). The trees were planted in the rough with site preparation consisting of girdling oak trees. The Florida Division of Forestry has now planted these areas with longleaf pine with the intention of restoring the sandhill community. Phosphate mining was a profitable business at the turn of the century the profits from exporting phosphates to foreign markets created an economic boon in some of the small towns near WPU. There is evidence of three such mines on the property. Phosphate was removed from the open mines without the benefit of machinery, using mules and forced labor (Olmert 2001). Biological Communities According to descriptions of the Florida Natural Areas Inventory (1990) there are eight basic community types present in the WPU. These community types exist in mosaics throughout the landscape, often grading into one another. Presently there are no accurate maps available to describe the extent and location of these communities. However, there is a complete set of high resolution (0.3m ground sampling distance) aerial photographs taken in the 2002 by Emerge (Andover, Massachusetts). Sandhill Upland Lakes Watermelon Pond is a complex system of interconnected bodies of water known as sandhill upland lakes (The Florida Natural Areas Inventory 1990). Seasonal fluctuations of water levels can be quite extreme. Watermelon pond contains 551 acres (223 ha) of wetlands and surface water (Dickinson et al. 1982). However, at the time of our study there was no water present in any of these lakes due to an extended drought. Watermelon

PAGE 34

21 Figure 3-1 Study area location

PAGE 35

22 Pond has been used recreationally for boating and fishing. The County of Alachua operates a boat ramp on the northeastern portion of the pond. Prior to drought of recent years, it was well known as an excellent venue for large-mouth bass (Micropterus salmoides) fishing. Sandhill Sandhill is by far the most widespread vegetative community in the WPU. Fire suppression, soil disturbances, and other anthropogenic impacts, have created large differences in the structure and composition of this community. Typically a sandhill community is described as containing a sparse canopy of longleaf pine with a midstory of small xerophytic oaks, such as turkey oak (Quercus laevis), bluejack oak (Q. incana) and sand post oak (Q. stelleta). The understory is dominated by wiregrass (Aristida beyrichiana), pineywoods-dropseed (Sporobolus junceus), and a diverse number of herbaceous species and several species of small shrubs. A very small part of the WPU sandhill contains all of the characteristics described above. The primary missing component is a mature longleaf canopy, apparently as a result of historic logging. There is also evidence of prolonged fire suppression, which has resulted in an increase in hardwood species. This is also evidenced by the high density and large sizes of xeropyhtic oaks throughout much of the property. In some cases Turkey Oak Barrens have formed where large turkey oaks dominate the canopy. Soil disturbance has also had a profound impact upon the ground cover in some areas. In particular, wiregrass, a key component of the understory, does not regenerate after severe soil disturbances, such as plowing. There is a 150-acre previously farmed section that contains virtually no wiregrass.

PAGE 36

23 Scrub At higher elevations and in well-drained sandy soils, rosemary/oak scrub can be found. Dominant species in this community include the Florida rosemary (Ceratiola ericoides) and sand live oak (Q. geminata). The ground area typically contains patches of bare soil and wiregrass, as well as lichen (primarily Cladonia and Cladinia spp.). There is an ongoing debate as to whether many of these areas are naturally occurring scrub community or have been formed as the result of sandhill being degraded by poor logging practices, followed by decades of fire. Some of the scrub dominated by Florida rosemary still contains longleaf pine. However, longleaf pine is conspicuously absent in some areas of higher elevation. This suggests that perhaps a more typical scrub community may have always existed in areas of higher elevation and deep sands. Hardwood Hammock At lower elevations and on less permeable soils, hardwood hammocks frequently occur. In particular, the eastern most section of WPU east, which lies on an Otela-Candler-Taveras soil association, supports a n extensive area of hardwood hammock. Much of the hardwood hammocks within the study area are dominated by laurel oak (Quercus hemisphaerica) with relatively low tree diversity. Historical aerial photography indicates that these hammocks of low diversity are the result of recent succession on cleared land. In more diverse areas, commonly occurring tree species include red bay (Persea borbonia), pignut hickory (Carya glabra), and eastern redcedar (Juniperus virginiana). Prairie The land surrounding Watermelon Pond is seasonally inundated by water, forming a prairie community, which is dominated by maidencane (Panicum hemitomum), sand

PAGE 37

24 cord grass (Spartina bakeri) and other species of grass. There are scattered trees and shrubs located on the prairie such as slash pine, gallberry (Ilex glabra), wax myrtle (Myrica cerifera) and St. Johns wort (Hypericum spp.). Ephemeral Pond Depressional areas within the prairie that have long hydroperiods are called ephemeral ponds. They contain a different assemblage of plants, including spatterdock (Nuphar luteum), redroot (Lachnanthes caroliniana), and pennywort (Hydrocotyle spp.). Ephemeral ponds are unable to sustain populations of fish, however they are particularly important breeding habitat for amphibians (Babitt and Tanner 1997). Management Considerations By far the greatest challenge to management is the fact that the property is a disjoint collection of parcels, with little connectivity among them (see Fig. 3-1). Problems associated with small nature reserves include genetic isolation, and detrimental edge effects, such as those caused by invasive species, and anthropogenic influences. Numerous ranchettes and other residential structures border many of the property boundaries. These may complicate the use of prescribed fire, which is necessary to maintain the fire adapted plant communities present in the WPU. Other negative impacts caused by humans are illegal dumping, all-terrain vehicle use, and feral dogs. The main office at Goethe State Forest, which maintains regulatory and administrative control, is more than 20 miles (8 kilometers) away from the main tract of the forest. It is therefore difficult, if not impossible, to maintain a continuous presence in the WPU. Despite these challenges, protecting the WPU is essential, since it is one of the last remaining examples of the diverse upland and lacustrine communities of the Northern Brooksville Ridge.

PAGE 38

CHAPTER 4 METHODOLOGY Vegetation Data Collection The initial vegetation data collection consisted of 321 sample points located systematically throughout the property that were marked with metal stakes, identified with aluminum tags and geo-referenced using a Trimble (Sunnyvale, California) Pro-XRS GPS with real time differential correction (Fig. 4-1). The distance between sample points in each transect is 100 meters. Systematic sampling was chosen for this part of the study as it allows a large number of samples to be distributed evenly throughout the entire property. Each permanent sample point is the focal point from which all sampling is conducted. Measurements were taken on three categories of the vegetation: a) ground cover b) shrubs, seedlings and saplings, and c) trees. The methodology and observed variables varied among these three classes to ensure reliable estimates. Seven areas were chosen for a comparative analysis. These sites included three reference sites and four nonreference sites. These areas were delineated using aerial photographs in a geographic information system (GIS). To increase sampling intensity in areas selected for comparative analysis, random sampling was used within the seven selected sites. The methodology for data collection was the same as for the systematic sampling. Random points were generated in ArcView 3.2 GIS software (ESRI, Redlands, CA). These points were navigated to in the field using GPS. There were a 25

PAGE 39

26 total of 184 sample points located in the 7 test areas, 128 in nonreference sites and 56 in reference sites. Ground Cover Ground cover was ocularly assessed in two 1m2 quadrats at each sample point. Estimates were recorded in 5% categories from 0 to 100. Location of quadrats was determined by a random azimuth between 0and 360 relative to the permanent point. Quadrat locations were marked with a metal stake and embossed aluminum tags so that they could be relocated in the future. Each random quadrat was positioned 2 m away from the permanent point (Fig. 4-1). Percent cover of each species was ocularly estimated and recorded. In addition, the number of stems of each species was recorded. Figure 4-1 Sampling architecture Shrubs, Seedlings and Saplings All shrubs, seedlings and saplings within a 2m radius of the permanent plot center were counted. Saplings were defined as trees less than 10 cm DBH but greater than 1.37

PAGE 40

27 m in height. The seedlings were less than 1.37 m in height. Every shrub, sapling, and seedling within the plot was classified by species. Trees A 10-factor prism (variable radius plot or point sampling) was used to sample trees greater than 10cm DBH to estimate the basal area per acre in each sample point. The species and diameter at breast height (1.37 m) was recorded for each tree in the plot. Species Sampled Almost all woody plants that were encountered were identified to species in the sampling process. Woody species whose identification proved taxonomically difficult, such as those requiring fertile specimens when none were available, were identified to genus. Nomenclature in our study followed that of Godfrey (1988) for woody plants and Wunderlin (1998) for herbaceous species. Classification of Comparative Study Areas Several study sites were selected to provide additional information on the sandhill community, targeted for ecological restoration on this property. The purpose of these sites was to sample as much of the vegetative diversity in the sandhill longleaf pine community as possible. Both high quality sites (reference) and highly disturbed areas were chosen to represent the wide range of vegetation at the WPU. Nonreference Areas Sites 4, 5, 6, and 7 have a similar land-use history. The nonreference sites were all former plantation that were established between 1969-1972 using small tractors with a drag-type setter. The only site preparation at the time of planting was hand girdling of the oaks, which were primarily turkey oak and bluejack oak.

PAGE 41

28 The comparison regions (Fig. 4-2), classified for the purposes of our study, are as follows : Site 4 Before 1969 this land was leased for cattle. Centipede grass was planted as cattle forage, and in many areas this species dominates the ground cover in lieu of wiregrass. Site 4 was planted in slash pine in 1969-70. Samples for this comparative analysis were stratified so as to avoid inclusion of vegetation in and around a historic phosphate mine, that is a xeric hardwood community. All harvestable timber was extracted in 1995 by the Loncala Corporation before sale of the land to the Florida Division of Forestry. Although prescribed burning was carried out in the fall of 2000, it has not yet been planted with longleaf pine seedlings. Site 5 This site was apparently plowed for agriculture at one point due to the state of the existing ground cover. For example, wiregrass, a species highly susceptible to soil disturbance, was not found on this site. Site 5 was planted in slash pine 1968-1969, in the same manner as Section 4. This timber was clear cut prior to acquisition by DOF, and later planted twice with bare-root longleaf seedlings in 1996/1997 and 1997/1998. Site 6 Site 6 is a former slash pine plantation that was established between 1968 and 1970. After being clearcut in 1996 it was planted with longleaf pine seedlings in December 1999 at a density of 590 trees per acre (1458 per ha). However, only 30% of them survived. There are some scattered remnant mature longleaf pines in this area. The ground cover is in better condition than Section 5.

PAGE 42

29 Site 7 Site 7 is also a former slash pine plantation, established in 1969/1970, and cut in 1996. This section was planted in January 1999 with longleaf pine seedlings, utilizing scalping as a ground treatment. A prescribed burn took place in 2000. Reference Sites Three reference sites were chosen. These areas were smaller in size than the nonreference sites because there are few tracts of quality sandhill left. Criteria used for choosing these sites were the presence of mature longleaf pine trees and intact ground cover. Field reconnaissance was conducted on the ground to verify the condition of these sites prior to their selection as reference sites. In addition, assessments done in 1995 by the Florida Natural Areas Inventory (FNAI) have identified these areas as being in excellent condition. The general land office records from the mid-19th century (FDEP 2003) state that these areas sustained pineland, sometimes referring to them in their notes as first rate pineland. It is likely that all reference sites were historically logged in the 1930s and low intensity cattle grazing was undertaken from the 1930s through the 1990s. Although ranching may have taken place on reference sites, the extent of centipede grass present is much less than in the nonreference sites. However, due to the extensive conversion of the northern Brooksville range to agriculture and housing development, it is unlikely that a more representative portion of longleaf sandhill community for the WPU exists today.

PAGE 43

30 Figure 4-2 Comparative study areas. Sites 1, 2, and 3 are reference sites. Sites 4,5,6 and 7 are non reference sites. Field Data Interpretation The purpose of this data analysis was to determine a list of indicator species for monitoring restoration efforts. This process involved comparing values for species importance and tree basal area. It was done in a preliminary assessment that attempted to sites. In addition, ecological similarity was measured between sites.

PAGE 44

31 The second stage of the data analysis involved ranking those species that accounted for the greatest difference between the reference and nonreference sites. McCoy and Mushinsky ( 2002) used statistical analysis of binomial probability to effectively determine which fauna were responsible for differences in community structure of wildlife species in restored phosphate mine lands. A similar methodology was used in examining the vegetative composition in our study. Following ranking, a list of indicator species was developed based upon statistical significance. The methods used for the purpose of our study are described in detail below. Importance Values For understory species importance values are the sum of the relative frequency and relative percent cover of each of the species. For shrubs, seedlings and saplings importance value is the sum of the relative frequency and relative density. For example, the relative frequency of Florida rosemary (Ceratiola ericoides) at site 3 is 1.67% and the relative density is 0.09%, so the importance value is 1.76. Basal Area of Trees Basal area of trees was estimated with probability proportional to tree diameters at breast height using a 10factor prism. A random sample of 24 plots was drawn from each of the areas of interest for this analysis. Similarity Analysis Similarity indices were used to compare the vegetative structure of each of the study areas. The three measures of similarity used in our study included Jaccards coefficient of similarity, Sorensens index, and the Curtis Bray index. To make comparison easier all similarity measures were expressed in percentages. The Jaccards Index is a percentage of the species that are common between two sites. Sorensens

PAGE 45

32 index is the number of common species between both sites. The Curtis Bray Index incorporates measures of abundance, in this case density and foliar cover, to account for differences in species evenness. The equations for these measures are as follows: Jaccards Coefficient of Similarity is shown in Equation 4-1 Js = Species ofNumber TotalSpeciesCommon ofNumber *10 (4-1) Sorensens Index is shown in Equation 4-2 Ss = 100)/2 (SSpeciesCommon ofNumber abS (4-2) Sa = The number of species in community A Sb = The number of species in community B Curtis Bray Index is shown in equation 4-3 Cs = iciiCiciiYXYX111),min(2 i (4-3) where (Xi,Yi) are the abundance measures for a given species in a population for community A and B. Determination of Targets for Monitoring. In order to determine indicator species for monitoring, a ranking system was employed. It relies on comparing the binomial proportion of species occurrence between the reference and nonreference sites (McCoy and Mushinsky 2002). Species were ranked according to their statistical significance of occurrence between reference and nonreference sites.

PAGE 46

33 Proportional Abundance of Species The first step for ranking the species was to determine the proportion of sample points at which each species was present. This was done by aggregating all samples for the 56 plots in the reference sites, and all samples for the 128 nonreference sites. Appendix B lists the proportional abundance of each species. It was calculated based on the respective number of plots at which specific species was found, without regard to abundance (Appendix B). For example, sand blackberry (Rubus cuneifolius), was found in 10 of the 56 reference plots and in 83 of the 128 nonreference plots. Therefore, the proportional abundance of this species was 0.18 in the reference and 0.65 in the nonreference plots. The statistical differences between the two binomial proportions was determined by approximating the normal distribution. In this case, a two-tailed test was performed with the null hypothesis that the proportion between the two sites was equal. The formula used for the statistical test of comparison of two binomial proportions (Ott, 1993) was: 2121Z where 2111121nn and 2121nnyy 1 = Proportion of sample points at which a species is observed for the reference sites

PAGE 47

34 2 = Proportion of sample points at which a species is observed for the nonreference sites. = Proportion of sample points at which a species is observed for both the reference and nonreference sites. 1y = Number of occurrences of a species at the nonreference sites. 2y = Number of occurrences of species at the reference sites. n1 = Number of sample points in the reference sites. n2 = Number of sample points in the nonreference sites. Since this is a two-tailed test, the null hypothesis 21 is rejected for any given value of if: 2/ Next, the species are ranked according to this value (Appendix C). Those species with a statistically significance level of p=0.05 or greater were considered focal species. Those not found to possess a statistical significance were defined as non-focal species. In addition, there was a number of species that were not found in a sufficient number of the plots for the normal approximation to be valid in this test, that is or 1 were not greater than or equal to 5. Species included in this category were listed as un-common.

PAGE 48

CHAPTER 5 RESULTS AND DISCUSSION Comparison Sites Appendix A includes the results of the Jaccard's (J), Sorensen's (S), and Curtis Bray (CB) similarity indices, as calculated for the understory, shrubs, and tree seedlings and saplings, which compare ecological similarity among the sites. Importance values (Appendix D) are also incorporated into the following discussion. Understory Site 5 is least similar to the reference site for all three indices. The similarity was only 32.4% with the CB measure, and 34.6% and 51.4%, respectively, using the J and S measure of similarity. This is an obvious finding considering that Site 5, having been previously farmed, has sustained the most anthropogenic disturbance of all the sites. Tillage of the soil has eliminated wiregrass, which is present in all of the other sites. In addition, there is a greater abundance of forbs and grasses, compared to the other sites (Table D-1). Sites 6 (CB=57.8%, J=54.2%, S=70.3%), and 7 (CB=56.7%, J=54.5%, S=70.6%), were the most similar to the reference site. Site 4 and 5 were least similar according to the CB, likely due to the relatively greater importance of forbs and grasses at Site 5 (Table D-1). Shrubs For the shrub layer, the CB showed significantly less similarity than the J and S, among most sites. This can be attributed to differences in species evenness, relative to the number of species. For example, site 5 proved to be most similar to the reference site, 35

PAGE 49

36 using the J (55.6%) and S (63.2%), but least similar according to the CB (27.8%). J and S do not account for the abundance of species. They merely account for the presence or absence of common species. Site 5 and 7 were the most similar of all measures (CB= 85.2, J=71.4, S=83.3). This may be attributed to the fact that these sites have been disturbed the most, Site 5 being tilled historically and Site 7 prescribed burned in 2000. Tree Seedlings and Saplings Site 5 showed the least similarity with the reference site. This may be partially attributed to the absence of turkey oak in Site 5, which is very common in the understory of all the other sites, especially the reference areas. Sites 4 and 7, which have been prescribed burned in 2000 and 1998, respectively, are most similar to the reference sites. Summary of Similarity Indices The results of the similarity indices seem to suggest that although similar species composition may exist between reference and nonreference sites, there may be large differences in evenness. This is especially evident with mid-story shrubs and tree seedlings whose abundance and sizes may be a function of fire history. For example, a more recently burned site may contain a higher abundance of small oaks, such as turkey oak seedlings. This suggests that the Curtis Bray similarity index may be the most appropriate for measuring ecological similarity between reference and nonreference sites. Suggested Target Species The following is an outline of the suggested target species based upon species ranking, importance value, and basal area. The ranking methods employed in our study provide a clear indication of the significance of the proportion of sampling points that contained species without regard to abundance. Appendix C contains a list of the corresponding statistical significance.

PAGE 50

37 Appendix D includes a complete list of all of the importance values for each of these ranked species. These importance values serve as a useful tool to discuss the reasons why some species may have been found more often at the reference or the nonreference sites. In addition, importance values give relative measures of abundance in each of the comparison areas. It should be noted that there is some overlap in the sampling of species in the understory, shrub, and tree seedling and sapling layer. For example, the presence of sand-blackberry is noted in both the understory quadrats, and in the 2m radius fixed area plots intended to measure shrubs, saplings and tree seedlings. In such situations, the observations from the 2m radius plot are used, as the larger sampling area provides a better estimate. Understory Five targets for monitoring were determined for the understory. Among these, wiregrass was found more often in the reference than in the nonreference sites. The others, centipede grass, dog-fennel, green briar, and total forb cover, were more often detected in the nonreference areas. Wiregrass (Aristida beyrichiana) Wiregrass was identified in the early planning stages of this project as a potential target species. There is a significantly higher proportion of wiregrass in the reference sites than in the nonreference sites (P<0.01). However, Site 6, a nonreference site, had the highest importance value for wiregrass (31.24). The difference in proportional abundance between reference and nonreference sites (Appendix B) may be due to the fact that Site 5, a nonreference site, does not contain any wiregrass due to past tillage (Table D-1). This suggests that even though the nonreference sites were converted to plantation for about a

PAGE 51

38 25-year period, the impact on wiregrass might be negligible. This corroborates research by Hedman et al. (2000) that past agriculture use has a greater impact on ground cover than many forest management activities. Centipede grass (Eremochloe ophiurides) Centipede grass is a rhizomotous perennial grass (Godfrey 1988) introduced in the WPU to improve grazing (Olmert 2001). Centipede grass was observed more often in the nonreference plots (P<0.01 ) than the reference plots and has become established as a dominant ground cover species throughout many areas of the nonreference sites. The effects of centipede grass on restoration of longleaf pine ecosystems have not been extensively studied. Two possible ways in which centipede grass could inhibit restoration efforts are to compete with native plants and to alter the ground cover fuel composition. Centipede grass should continue to be monitored over time so that its persistence in the restoration sites may be measured. For example, it may be that once a tree canopy is established, this species will be competitively excluded by trees or more shade tolerant understory species. Dog-fennel (Eupatorium capilifolium) Dog-fennel is a ruderal species that is often found in disturbed woods and fields (Radford et al 1968) and, as such, is very common in the nonreference sites. Expectedly, dog fennel was found significantly more often at the nonreference sites (P<0.01). This species is easy to identify, and should be a useful target species for monitoring. Green briar (Smilax spp.) Green briar is a woody vine that was most prevalent in the nonreference sites (P<0.01). The increase of woody vines with fire suppression has been documented (Rodgers and Provencher. 1999), and the importance values for this shrub seem to

PAGE 52

39 indicate a similar finding. Green briar was less important in Site 4 (imp. = 1.48), prescribed burned in 2000 and Site 7 (imp =4.29), than Sites 5 (imp=8.89) and 6 (imp=5.19), which have not been prescribed burned. In comparison, the reference sites contained relatively little green briar (imp. values for 1,2,3 were 1.31, 0, and 1.14, respectively). Total forb cover Other forbs was a broad classification that included all forbs except dog-fennel, such as legumes and asters (Pityposis spp.). The total forb cover was greatest in Site 5, which is the most disturbed (imp=35.89). This variable may be a useful measure for monitoring, however more research is needed to determine which species are responsible for the large importance value of forbs at Site 5. The seasonality of herbaceous plants and diversity of species can make species identification a difficult process when compared to woody plants. Shrubs Winged sumac (Rhus copallinum) Winged sumac is a common species found in old fields and forest plantations (Radford et al 1968). Its presence may increase following prescribed burning (Miller and Miller 1999). Although common in reference areas, winged sumac may prove to be a very useful monitoring species. Sand blackberry (Rubus cuneifolius) Sand blackberry is a small shrub that is frequently found in great abundance in areas that have been disturbed by fire or mechanical site preparation (Godfrey 1998). Along with winged sumac, sand blackberry dominates many areas of the WPU that are

PAGE 53

40 old slash pine plantations that have been planted with longleaf pine. It was found in a significantly higher proportion of the nonreference sites (P<0.01) than the reference sites. Florida rosemary (Ceratiola ericoides) Florida rosemary is an evergreen shrub that is common in the sandy soils of the WPU. In some areas this species forms monotypic stands often referred to as rosemary balds (see Chapter 3). This species was found more often in the reference sites (P<0.01) than the nonreference sites. Patches of rosemary balds exist in a mosaic within xeric sandhill at the WPU, and the samples selected for this analysis included points in rosemary balds. Site 3 was the only reference site to contain Florida rosemary, but it was a very important component (imp.=134.1). A review of historical aerial photography revealed that these rosemary scrub balds have existed in Site 3, to a similar spatial extent, sat least since 1949. Glaucous Blueberry (Vaccinium darrowii ) Glaucous blueberry is a small shrub similar in appearance and distribution to Vaccinium myrsinites, in the WPU. This species was found more often in the reference sites than the nonreference sites (P=0.03) and may serve as a useful indicator species. However, field observations seem to suggest that the distribution of glaucous blueberry is very patchy, and as such, its utility as an indicator may be better understood with increased sample size. Tree Seedlings and Saplings Three species of tree seedlings and saplings, turkey oak, black cherry (Prunus serotina), and laurel oak, were selected as focal species for monitoring. Turkey oak was found more frequently in the reference sites, whereas black cherry and laurel oak were more prevalent in the nonreference areas.

PAGE 54

41 Turkey oak The presence of turkey oak seedlings and saplings was more prevalent in reference sites (P<0.01) than in nonreference sites. Typically, longleaf pine sandhill has been described as having a large number of small oaks, such as turkey oaks (Glitzenstein et al. 1995). Therefore, this species may serve as an important indicator species of longleaf restoration. Black cherry (Prunus serotina) Black cherry is an early succession hardwood species that was not found in all the reference sites. However, seedlings and saplings were very common in the understory of nonreference sites. The use of prescribed burns would probably reduce the importance of this species in the understory. This species should be a useful target species to monitor the effectiveness of prescribed burns. Laurel oak Laurel oak is very similar to black cherry in that it is an invasive hardwood species prevalent in fire suppressed longleaf sandhill. Laurel oak is more common that black cherry, therefore it may be a better focal species for monitoring. However, one confounding factor for monitoring this species is the difficulty in identification. Laurel oak seedlings look taxonomically very similar to some other oak species, especially bluejack oak. Trees Table E-1 includes the estimates of tree basal area for each of the sites. These data reveal that there is a much higher basal area of longleaf pine and turkey oak in the reference site. These two species should be monitored in the future. In particular, an increase in longleaf pine basal area is a desired management objective. However, the

PAGE 55

42 lack of mature turkey oaks in the nonreference sites is probably due to the fact that site preparation for conversion to plantation involved removing most of the oaks by girdling between 1968-1972. Longleaf pine There are very few mature longleaf pine trees remaining in the WPU, and as such, they are an excellent indicator species of remaining areas of high quality longleaf. In general, areas that include mature longleaf pine had excellent ground cover. However, from a monitoring standpoint, longleaf pine basal area is not a robust monitoring metric because it is a slow growing long-lived species that is not likely to be affected by management actions over short periods of time Turkey oak Turkey oak is almost always present where there is longleaf pine (Gordon 2003), and therefore, it may be an excellent indicator of the historic extent of the longleaf pine sandhill. It is likely that fire suppression in recent history has led to an increase in turkey oaks at the reference sites

PAGE 56

CHAPTER 6 SUMMARY AND CONCLUSIONS The ranking methodology used in our study allows the development of a preliminary list of target species that could be utilized for future monitoring. However, the most important accomplishment of our study is the development of a system for sampling the understory, mid-story, and tree layer, utilizing a network of geo-referenced permanent points. Although the temporal scale at which the data were collected limited the analysis, sample points can be re-visited in future studies. As discussed in Chapter 2, this would allow for much more powerful statistical analyses. Suggestions for Future Research Our study was the first to quantify the vegetative structure and composition of the WPU and, as such, it raised many questions that could be addressed by further research. The following are recommendations for future research: Further Exploratory Analysis of Field Data The use of ordination and classification (see Chapter 2) to examine the field data may reveal some patterns that have useful implications for management and restoration. Multi-dimensional methods may provide some insight into those points, or even sections of the property, that are closest to the reference conditions. An ordination of the comparison areas delineated in our study could provide further information about whether the sites chosen are ecologically similar. Such methods may prove more useful than the similarity measures employed in our study. 43

PAGE 57

44 Determination of Target Abundances Although our study identifies a list of target species, it would also be useful for land managers to establish target abundances for monitoring efforts. For example, having identified wiregrass as a species to monitor, what percent cover of this species would be desired for restoration sites? Effect of Centipede Grass on Longleaf Pine Restoration Centipede grass was found throughout many areas of the WPU that are slated for restoration. A review of the literature found limited information on the effects of this grass as an invasive species in natural areas. More research needs to be done to study: a) the effect of this species on fine fuels for fire, b) interspecific competition with native plants and c) potential changes in distribution and abundance over time.

PAGE 58

APPENDIX A SPECIES SIMILARITY Table A-1. Understory similarity among comparison sites based on Jaccards Index R 4 5 6 7 R 50.0 34.6 54.2 54.5 4 50.0 40.0 66.7 55.6 5 34.6 40.0 51.7 51.9 6 54.2 66.7 51.7 72.0 7 54.5 55.6 51.9 72.0 Table A-2. Understory similarity among comparison sites based on Sorensens Index R 4 5 6 7 R 66.7 51.4 70.3 70.6 4 66.7 55.8 80.0 35.7 5 51.4 55.8 46.2 68.3 6 70.3 80.0 46.2 54.5 7 70.6 35.7 68.3 54.5 Table A-3. Understory similarity among comparison sites based on Curtis Bray Index. R 4 5 6 7 R 50.3 32.4 57.8 56.7 4 50.3 18.1 69.9 68.8 5 32.4 18.1 45.1 56.9 6 57.8 69.9 45.1 64.1 7 56.7 68.8 56.9 64.1 45

PAGE 59

46 Table A-4. Shrub similarity among comparison sites based on Jaccards Index. R 4 5 6 7 R 35.7 55.6 50.0 50.0 4 35.7 46.2 53.3 26.7 5 55.6 46.2 50.0 71.4 6 50.0 53.3 50.0 45.5 7 50.0 26.7 71.4 45.5 Table A-5. Shrub similarity among comparison sites based on Sorensens Index. R 4 5 6 7 R 52.6 52.6 63.2 42.1 4 52.6 63.2 69.6 47.1 5 52.6 63.2 66.7 83.3 6 63.2 69.6 66.7 62.5 7 42.1 47.1 83.3 62.5 Table A-6. Shrub similarity among comparison sites based on Curtis Bray Index. R 4 5 6 7 R 43.5 27.8 49.5 30.1 4 43.5 81.7 82.0 73.3 5 27.8 81.7 77.2 85.2 6 49.5 82.0 77.2 75.0 7 30.1 73.3 85.2 75.0

PAGE 60

47 Table A-7. Tree seedling and saplings similarity among comparison sites based on Jaccard's coefficient. R 4 5 6 7 R 75.00 40.00 66.67 60.00 4 75.00 60.00 70.00 60.00 5 40.00 60.00 54.55 50.00 6 66.67 70.00 54.55 72.73 7 60.00 63.64 50.00 72.73 Table A-8. Tree seedling and saplings similarity among comparison sites based on the Sorensen's coefficient. R 4 5 6 7 R 85.71 57.14 85.71 85.71 4 85.71 75.00 87.50 87.50 5 57.14 75.00 70.59 70.59 6 85.71 87.50 70.59 57.14 7 85.71 87.50 70.59 57.14 Table A-9. Tree seedling and saplings similarity among comparison sites based on Curtis Bray coefficient. R 4 5 6 7 R 52.1 26.0 43.7 51.2 4 52.1 25.7 26.0 49.6 5 26.0 25.7 25.2 27.9 6 43.7 73.3 25.2 48.8 7 51.2 49.6 27.9 48.8

PAGE 61

APPENDIX B PROPORTIONAL DISTRIBUTION OF SPECIES AMONG REFERENCE AND NONREFERENCE SITES

PAGE 62

Table B-1. Distribution of understory species among reference and nonreference points. Based upon the proportion of plots a species is present. Reference (n=56) Nonreference (n=128) Species # of points found Proportion # of points found Proportion Rubus cuneifolius 10 0.18 83 0.65 Aristida beyrichiana 56 1.00 67 0.52 Quercus laevis 38 0.68 31 0.24 Rhus copallinum 24 0.43 97 0.76 Eremochloe ophiuroides 6 0.11 54 0.42 Eupatorium capillifolium 7 0.13 44 0.34 Smilax spp. 4 0.07 36 0.28 Quercus hemisphaerica 0 0.00 18 0.14 Ceratiola ericoides 11 0.20 12 0.09 Total Forbs* 52 0.93 124 0.97 Prunus serotina 2 0.04 13 0.10 Quercus incana 6 0.11 21 0.16 Pinus palustris 12 0.21 20 0.16 Quercus nigra 0 0.00 6 0.05 Vaccinium darrowii 6 0.11 8 0.06 Myrica cerifera 0 0.00 5 0.04 Diospyros virginiana 4 0.07 13 0.10 Other Grasses 51 0.91 111 0.87 Quercus geminata 5 0.09 11 0.09 Vitis spp. 0 0.00 13 0.10 Woody vines* 6 0.11 1 0.01 Toxicodendron toxicarium 5 0.09 1 0.01 Opuntia humifosa 0 0.00 4 0.03 Pinus elliotti 0 0.00 4 0.03 Quercus virginiana 0 0.00 2 0.02 Vaccinium myrsinites 0 0.00 2 0.02 Asimina spp. 1 0.02 1 0.01 Passiflora incarnata 1 0.02 1 0.01 Paspalum notatum 1 0.02 1 0.01 Crataegus spp. 0 0.00 1 0.01 Panicum hemitomum 0 0.00 1 0.01 Quercus X asheana 0 0.00 1 0.01 Rubus betulifolius 0 0.00 1 0.01 Toxicodendron radicans 0 0.00 1 0.01 Vaccinium stamineum 0 0.00 1 0.01 Zanthoxylum clava-herculis 0 0.00 1 0.01 Gelsemium sempervirens 2 0.04 5 0.04 Serenoa repens 1 0.02 2 0.02 49

PAGE 63

50 Table B-2. Distribution of shrub species among reference and nonreference points. Based upon the proportion of plots a species is present. reference (n=56) Nonreference (n=128) Species # of points found Proportion # of points found Proportion Asimina spp. 4 0.07 2 0.02 Calicarpa americana 0 0.00 2 0.02 Ceratiola ericoides 14 0.25 22 0.17 Crateagus spp. 0 0.00 1 0.01 Hypericum spp. 2 0.04 1 0.01 Myrica cerifera 0 0.00 3 0.02 Opuntia humifosa 0 0.00 15 0.12 Rhus copallinum 26 0.46 113 0.88 Rubus cuneifolius 10 0.18 95 0.74 Serenoa repens 2 0.04 2 0.02 Toxicodendron toxicarium 5 0.09 1 0.01 Vaccinium arboreum 1 0.02 3 0.02 Vaccinium darrowii 9 0.16 9 0.07 Vaccinium myrsinites 1 0.02 2 0.02 Vaccinium stamineum 0 0.00 2 0.02 Zamia floridana 1 0.02 1 0.01 Table B-3 Distribution of tree seedling and saplings among reference and nonreference points. Based upon the proportion of plots a species is present. Reference (n=56) Nonreference (n=128) Species # of points found Reference Proportion # of points found Proportion Albizia julibrissin 0 0.00 1 0.01 Diospyros virginiana 7 0.13 22 0.17 Ilex opaca 0 0.00 1 0.01 Pinus elliottii 0 0.00 8 0.06 Pinus palustris 16 0.29 42 0.33 Pinus taeda 0 0.00 1 0.01 Prunus serotina 0 0.00 20 0.16 Quercus geminata 7 0.13 19 0.15 Quercus hemisphaerica 2 0.04 27 0.21 Quercus incana 11 0.20 22 0.17 Quercus laevis 49 0.88 55 0.43 Quercus nigra 0 0.00 3 0.02 Quercus virginiana 0 0.00 1 0.01 Zamia clava-herculis 0 0.00 5 0.04

PAGE 64

APPENDIX C SPECIES RANKING

PAGE 65

Table C-1 Understory species ranking. Species Z Value Significance Focal Species* Rubus cuneifolius 6.13 <0.01 Aristida beyrichiana 5.52 <0.01 Quercus laevis 5.24 <0.01 Rhus copallinum 4.69 <0.01 Eremochloe ophiuroides 4.37 <0.01 Eupatorium capillifolium 3.43 <0.01 Smilax spp. 3.32 <0.01 Quercus hemisphaerica 2.57 0.01 Ceratiola ericoides 2.03 0.02 Total Forbs* 1.95 0.03 Prunus serotina 1.71 0.04 Nonfocal Species Pinus palustris 1.42 0.08 Quercus nigra 1.41 0.08 Vaccinium darrowii 1.29 0.10 Myrica cerifera 1.28 0.10 Diospyros virginiana 1.19 0.12 Other Grasses 0.26 0.40 Quercus geminata 0.21 0.42 Selection of focal species based upon a significance level of P = 0.05 52

PAGE 66

53 Table C-2 Shrub species ranking. Species Z value Significance Focal Species* Rhus copallinum 26.35 <0.01 Rubus cuneifolius 20.41 <0.01 Ceratiola ericoides 4.46 <0.01 Vaccinium darrowii 1.90 <0.03 Uncommon Species Opuntia humifosa 6.11 Myrica cerifera 2.64 Callicarpa americana 2.15 Vaccinium stamineum 2.15 Vaccinium arboreum 1.96 Crataegus spp. 1.52 Vaccinium myrsinites 1.38 Serenoa repens 0.86 Toxicodendron toxicarium 0.75 Zamia floridana 0.60 Asimina spp. 0.16 Hypericum spp. 0.11 Selection of focal species based upon a significance level of P = 0.05

PAGE 67

54 Table C-3 Tree seedling and sapling species ranking. Species Z Value Focal Species* Quercus laevis 5.61 <0.01 Prunus serotina 3.13 <0.01 Quercus hemisphaerica 3.00 <0.01 Nonfocal Species Diospyros virginiana 0.80 0.21 Quercus geminata 0.42 0.34 Pinus palustris 0.57 0.28 Quercus incana 0.40 0.34 Uncommon Species Pinus elliotti 1.91 Zanthoxylum clava-herculis 1.50 Quercus nigra 1.16 Albizia julibrissin 0.66 Ilex opaca 0.66 Pinus taeda 0.66 Quercus virginiana 0.66 Selection of focal species based upon a significance level of P = 0.05

PAGE 68

APPENDIX D IMPORTANCE VALUES

PAGE 69

Table D-1 Understory species importance value. Species 1 2 3 4 5 6 7 Aristida beyrichiana 14.42 23.09 29.3 9.6 0 31.24 20.34 Other Forbs* 14.24 19.47 10.12 10.72 35.89 21.09 25.34 Other Grasses* 13.46 15.64 10.79 8.79 48.05 18.23 24.97 Quercus laevis 11.1 7.88 17.56 5.54 0 12.39 14.87 Rhus copallinum 13.12 13.65 0 8.53 21.79 18.9 21.54 Ceratiola ericoides 0 0 21.03 1.64 1.68 4.19 0.69 Pinus palustris 5.5 2.36 2.12 0 4 2.26 7.96 Rubus cuneifolius 4.12 3.28 0 6.12 15.57 13.03 16.02 Quercus geminata 4.14 0.92 1.92 5.28 4.68 4.26 0 Eupatorium cappillifolium 3.29 1.84 0 5.35 13.9 4.92 15.24 Eremochloa ophiuroides 0 2.35 2.2 17.27 20.27 26.33 24.66 Quercus incana 0.82 3.68 0 5.1 0 8.06 5.47 Vaccinium darrowii 0 2.46 1.14 0 1.24 0.8 2.95 Toxicodendron toxicarium 3.34 0 0 0 0 0 0.49 Diospyros virginiana 2.09 1.03 0 0.74 2.62 1.93 3.96 Smilax spp.** 1.31 0 1.14 1.48 8.89 5.19 4.29 Gelsemium sempervirens 1.42 0 0 0.71 2.43 0.87 0.69 Paspalum notatum 0 1.32 0 0.37 0 0 0 Serenoa repens 0 0 1.18 0 1.68 0 0 Prunus serotina 0 1.03 0 0.68 1.24 3.19 0 Pteridium aquilinum 0 0 0.94 0.91 0 3.2 3.37 Asimina spp.** 0 0 0.57 0 0 0.33 0 Toxicodendron radicans 0 0 0 0 0 0.33 0 Vitis spp.** 0 0 0 1.11 6.82 3.59 0 Passiflora incarnata 0 0 0 0 0.62 0 0 Opuntia humifusa 0 0 0 0 1.24 0.33 0.49 Crataegus spp.** 0 0 0 0.48 0 0 0 Myrica cerifera 0 0 0 3.85 0 0.67 0 Rubus betulifolius 0 0 0 0 0.73 0 0 Vaccinium myrsinites 0 0 0 0 0 0.8 0 Vaccinium stamineum 0 0 0 0.88 0 0 0 Pinus elliottii 0 0 0 0 1.16 0 3.97 Prunus angustifolia 0 0 0 0 0 0 1.07 Quercus hemisphaerica 0 0 0 3.68 4.76 9.26 0.49 Quercus nigra 0 0 0 0.88 0 2.53 0.69 Quercus virginiana 0 0 0 0 0 1.67 0 Zanthoxylum clava-herculis 0 0 0 0.31 0 0 0 Panicum hemitomon 0 0 0 0 0.73 0 0 Represents those species only identified to life form. ** Identified to genus. 56

PAGE 70

57 Table D-2 Tree seedling and sapling importance value. Site Species 1 2 3 4 5 6 7 Albizia julibrissin 0.0 0.0 0.0 0.0 0.0 0.0 2.1 Diospyros virginiana 8.9 17.1 0.0 9.4 31.6 9.0 13.3 Ilex opaca 0.0 0.0 0.0 0.0 5.8 0.0 0.0 Pinus elliotti 0.0 0.0 0.0 0.0 4.5 1.2 9.9 Pinus palustris 30.7 12.3 32.0 0.0 54.4 18.6 28.0 Pinus taeda 0.0 0.0 0.0 0.0 0.0 1.2 0.0 Prunus serotina 0.0 0.0 0.0 13.6 28.1 11.7 1.7 Quercus geminata 24.5 28.4 7.1 53.7 50.7 52.8 1.7 Quercus hemisphaerica 0.0 6.1 0.0 38.8 16.0 22.5 8.3 Quercus incana 12.3 46.8 0.0 18.5 0.0 24.3 8.3 Quercus laevis 123.6 89.3 160.9 60.0 0.0 53.5 26.6 Quercus nigra 0.0 0.0 0.0 0.0 0.0 2.7 1.6 Quercus virginiana 0.0 0.0 0.0 0.0 0.0 1.2 0.0 Zamia clava-herculis 0.0 0.0 0.0 6.0 8.9 1.2 0.0 Table D-3 Shrub species importance value. Site Species 1 2 3 4 5 6 7 A simina spp.** 0.0 6.6 14.7 2.0 3.3 4.0 0.0 Callicarpa americana 0.0 0.0 0.0 2.3 0.0 1.4 0.0 Ceratiola ericoides 0.0 0.0 134.1 8.4 5.1 16.1 1.8 Crateagus spp. 0.0 0.0 0.0 0.0 0.0 1.6 0.0 H ypericum spp. 4.7 4.9 0.0 2.0 0.0 0.0 0.0 Myrica cerifera 0.0 0.0 0.0 7.7 0.0 2.5 0.0 Opunitia humifusa 0.0 0.0 0.0 4.3 8.5 6.6 3.5 R hus copallinum 93.2 93.9 11.3 107.9 100.8 87.0 89.9 R ubus cuneifolius 29.7 42.0 0.0 54.8 77.6 67.6 79.5 Serenoa repens 0.0 0.0 13.1 2.0 1.7 0.0 0.0 Toxicodendron toxicarium 62.1 0.0 0.0 0.0 0.0 0.0 2.2 Vaccinium arboreum 5.7 0.0 0.0 2.1 0.0 3.4 0.0 Vaccinium darrowii 4.7 46.1 26.9 0.0 3.0 6.5 23.1 Vaccinium myrsinites 0.0 3.3 0.0 0.0 0.0 3.2 0.0 Vaccinium stamineum 0.0 0.0 0.0 4.1 0.0 0.0 0.0 Zamia floridana 0.0 3.3 0.0 2.5 0.0 0.0 0.0 ** Identified to genus

PAGE 71

APPENDIX E TREE BASAL AREA Table E-1 Overstory tree basal area (ft.2/acre) Site Species R 4 5 6 7 Pinus palustris 27.50 1.25 0.00 1.67 0.42 Quercus laevis 14.17 0.42 0.00 1.67 0.42 Quercus geminata 0.42 1.25 0.00 1.67 0.00 Quercus virginiana 0.42 0.00 0.00 0.00 0.00 Prunus serotina 0.00 0.42 0.00 0.83 0.00 Quercus hemisphaerica 0.00 3.33 0.00 2.08 0.00 Quercus incana 0.00 0.00 0.00 0.83 0.00 Quercus nigra 0.00 0.00 0.00 0.42 0.00 58

PAGE 72

LIST OF REFERENCES Babbitt K. J. and Tanner G. W. 1997. Effective management for frogs and toads on Floridas ranches. University of Florida Cooperative Extension Service. WEC 16. Gainesville, Florida. Baker W. J. 1992. Effects of settlement and fire suppression on landscape structure. Ecology 73:1879-1887. Bonham C. D. 1989. Measurements for Terrestrial Vegetation. John Wiley and Sons, New York, New York. Brockway D. G. and Outcalt K. W. 1998. Gap-phase regeneration in wiregrass ecosystems. Forest Ecology and Management 106: 125-139. Brockway D. G., Outcalt K. W., and Wilkins N. R. 1998. Restoring longleaf pine wiregrass ecosystems following low-rate hexazinone application on Florida sandhills. Forest Ecology and Management 103: 159-175. Croker T. C. Jr. 1987. Longleaf Pine: A History of Man and a Forest. Atlanta, GA, USDA Forest Service.For. Rep. R8FR7. Dickinson B., Huber W. C., Heaney J. P., and Brezonik, P.L. 1982. Gazetteer of Florida Lakes. Florida Water Resources Research Center no. 63, Gainesville Florida. Digby P. G. N. and Kempton R. A. 1987. Multivariate Analysis of Ecological Communities. Chapman and Hall, London. Elzinga C. L., Salzer D. W., Willoughby J. W., and Gibbs J. P. 2001. Monitoring Plant and Animal Populations. Blackwell Science, Malden, Mass. FDEP. 2003. Florida Department of Environmental Protection, Government Land Office Records. www.labins.org. Last accessed 3/02/03. The Florida Natural Areas Inventory (FNAI) 1990. Guide to the Natural Communities of Florida. Florida Department of Natural Resources. Gauch H. G. Jr. 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge. 59

PAGE 73

60 Gilliam F. S. and Platt W. J. 1999. Effects of long term fire exclusion on tree species composition and structure in an old-growth Pinus palustris (Longleaf pine) forest. Plant Ecology 140: 15-26. Glitzenstein J. S., Platt W. J., and Streng D. R. 1995. Effects of fire regime and habitat on tree dynamics in north Florida longleaf pine savannas. Ecological Monographs 65:441-476. Godfrey R. K. 1988. Trees Shrubs and Woody Vines of Northern Florida and Adjacent Georgia and Alabama. University of Georgia Press. Athens, Georgia. Gordon D. 2003. Personal Communication. State Ecologist, The Nature Conservancy / University of Florida. Gainesville, FL. Griffith G. E. 1998. Lake Regions of Florida. Geological Survey. Reston, VA. Hedman C. W, Grace S. L., and King S. E. 2000. Vegetation composition and structure of southern coastal plain pine forests: an ecological comparison. Forest Ecology and Management 134: 233-247. Horton T. 1995. Longleaf pine: a southern revival. Audubon 97:74-82. Hurn, J. 1989. GPS, a Guide to the Next Utility. Trimble Navigation, Sunnyvale, CA. Johnson R. and Gjerstad D. 1998. Landscape-scale restoration of the longleaf pine ecosystem. Restoration and Management Notes 16: 41-45. Jordan W. R., Gilpin M. E., and Aber J. D. 1987. Restoration ecology: ecological restoration as a technique for basic research. P. 3-23 In: Restoration Ecology, A Synthetic Approach to Ecological Research. Jordan, W. R., Gilpin, M.E., Aber, J.D. (eds.) Cambridge University Press. Kennedy K. A. and Addison P. A. 1987. Some considerations for the use of visual estimates of plant cover in biomonitoring. Journal of Ecology 75: 151-157. Krebs C. J. 1998. Ecological Methodology, Second edn. Addison -Wesley, Menlo Park, California. Lewis, R. 1997. Global positioning systems P. 251-258 In GIS Data Conversion: Strategies, Techniques, Management. Hohl, P. (eds.). Onword Press, Santa Fe, New Mexico. Long A. J. 2002. Benefits of prescribed burning.University of Florida Cooperative Extension Service.Gainesville, Florida. Ludwig J. A. and Reynolds J. F. 1988. Statistical Ecology. John Wiley and Sons, New York, New York.

PAGE 74

61 McCoy E. D. and Mushinsky H. R. 2002. Measuring the success of wildlife community restoration. Ecological Applications 12:1861-1871. Miller J. H. and Miller K. V. 1999. Forest Plants of the Southeast, and Their Wildlife Uses. Southern Weed Science Society. Champaign, Illinois. Monroe M. C., Babb G., and Heuberger K. A. 1999 Designing a prescribed fire demonstration area. Gainesville, Florida, Cooperative Extension Service, University of Florida. Myers R. E. 1990. Scrub and high pine. P. 150-193 In Ecosystems of Florida. University Presses of Florida. Gainesville, Florida. Myers R. E. and White D. L. 1987. Landscape and history changes in sandhill vegetation in north-central and south-central Florida. Bulletin of the Torrey Botanical Club 114: 21-32. Olmert B. 2001. Personal Communication. President, Loncala Corporation. High Springs, Florida. Ott, L. R. 1993. An Introduction to Statistical Methods and Data Analysis.Wadsworth, Inc. USA. Outcalt K. W. 2000. The longleaf pine ecosystem of the south. Native Plants Journal 1: 43-51. Philips, E. A. 1959. Methods of Vegetation Study. Henry Holt and Company, Inc. USA. Pielou E. C. 1984. The Interpretation of Ecological Data. John Wiley and Sons, New York, New York. Provencher L., A. R. Litt, and Gordon D. R. 2000. Compilation of Methods Used by the Longleaf Pine Restoration Project from 1994 at Eglin Air Force Base, Florida. Product to Natural Resources Division, Eglin Air Force Base, Niceville, Florida. Science Division, The Nature Conservancy, Gainesville, Florida. Radford A.E., Ahles, H.E. and Bell C.R. 1968. Manual of the Vascular Flora of the Carolinas. Univ. N. Carolina Press: Chapel Hill, North Carolina. Reed D. D. and Mroz G. D. 1997. Resource Assesment in Forested Landscapes. John Wiley and Sons Inc., New York, New York. Rodgers L. H., and Provencher, L.1999. Analysis of longleaf pine sandhill vegetation in northwest Florida. Castanea 64: 138-162. Seamon G. 1998 A longleaf pine sandhill restoration in northwest Florida. Restoration and Management Notes 16.46-50.

PAGE 75

62 Thompson S.K. 1992. Sampling. John Wiley and Sons Inc. New York, New York. Wade D. D. 1988.A Guide for Prescribed Fire in Southern Forests. USDA, Forest Service. Wolf P.R. and Brinker, R.C. 1994. Elementary Surveying. Harper Collins Publishers. New York, New York. Wunderlin R. P. 1998.Guide to the Vascular Plants of Florida. University of Florida Press, Gainesville, Florida.

PAGE 76

BIOGRAPHICAL SKETCH Michael Rasser was born February 11, 1975 in Bayshore, New York He spent several years working in the field of environmental education, including an internship in the Zoo of The Newark Museum, New Jersey, and a part-time instructor in the Miami-Dade Community College Environmental Center. He graduated Cum Laude with his B.A. in Environmental Studies (with a minor in Biology) from Florida International University. He developed an interest in ecological research while spending two summers as part of a research team examining the effects of global warming on Alaskan arctic plants. Mike plans to continue his scientific pursuits by pursuing a Ph.D. program at the University of Texas. 63