|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 WET FLATWOO DS RESTORATION AFTER DECADES OF FIRE SUPPRESION By DAVID K. MITCHELL A THESIS PRE SENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PA R TIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE UNIVERSI TY OF FLORIDA 2011
2 2011 David K. Mitchell
3 ACKNOWLEDGEMENT S I would like to thank my advisor Carl Fitz for his assistance and support throughout the course of my studies. I would also like to thank my committee members Mack Thetfo rd and Debbie Miller for years of academic support and mentorship. This project could not have been completed without the help of Penelope Bishop for her always patient assistance in many long field days Tim Baxely for assistance in the greenhouse, and D avid Clayton and the Northwest Florida Water Management District.
4 TABLE OF CONTENTS page ACKNOWLEDGEMENTS 3 6 LIST OF FIGURE 7 ABSTRACT 8 CHAPTER 1 10 10 11 12 13 15 17 18 2 21 21 24 Sampling Plots 25 26 Vegetation Sampling 28 Soil Disturbance 29 29 3 RESULT 38 Herbaceous Cover and Com 38 Woody Cover and Comp 40 Seed Bank Composi 42 Vegetation and Tr eatments 44 Similarity between Extant Vegetatio 45 4 56
5 56 Woody Cover and Comp 58 61 Restoration Implicat 63 6 6 APPENDIX A SPECIES PERCEN T COVER OF EXT 67 B SPECIES PERCENT COVER OF EXTANT VE 69 L IST OF REFER E NCE S 71 77
6 LIST OF TABLES Table page 1 1 Common vegetation species found in high quality wet flatwoods of the Florida panh 20 2 1 3 5 3 1 Cover means for treatments by 47 3 2 Mean percent cover for selected specie 49 3 3 See dling emergence species, lo 50 3 4 52
7 LIST OF FIGURES Figure page 2 1 Sandhill Lakes Mitigation Bank (SHLMB) in 1949, showing the overall heterogeneity of th 32 2 2 Ward Creek West (WCW) in 1953, showing the overall heterogeneity of the habitats 32 2 3 Sandhill Lakes Mitigation Bank (2007), showing the outlines of locations that were subsequently cleared of veg etat ion and site locations 1 33 2 4 Ward Creek West (2008), showing the outlines of locations that were subsequently cleared of vegetation a nd site locations 5 34 2 5 Effects of gyro trac A) I mmediately following ap plication B) T wo months ............................................................................................................... 36 2 6 Effects of disking t 37 3 1 Percent cover for cleared and 48 3 2 NMS ordination representing similari 53 3 3 PCA ordination representing similari 54 3 4 NMS ordination of extant vegetation similarity to seed 5 5
8 Abstract of Th esis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the R equirements for the Degree of Masters of Science WET FLATWOODS RESTORATION AFTER DECADES OF FIRE SUPPRESION By David K. Mitchell December 2011 Chair: H. Carl Fitz Major: Soil and Water Science Fire suppression leads to encroachment of woody vegetation into herbaceous ecosystems worldwide. Without fire, wet flatwoods of the southeastern United States have transformed from open grass dominated communities to shrub thickets. To examine the potential for restoration of wet flatwoods converted to shrub thickets by fire suppression we quantified the seed banks, using the seedling emergence technique, and the responses of total vegetation cover to th e following multi plot, multi site treatments: 1) Thicket shrub thickets with no modification; 2) Cleared above ground woody vegetation removal via GyroTrac + prescribed fire + herbicide applied to woody resprouts; and 3) Disked same as Cleared, but followed by soil disking. The Disked treatment significantly reduced woody ground and canopy cover, and showed little resprouting and increased herbaceous cover, while the Cleared treatment did not reduce the resprouting of woody species and reduced he rbaceous cover Ten months after treatment, four species ( Rhexia sp. Drosera capillaris, Xyris brevifolia, and Hypericum s p .) common to fire maintained wet flatwoods were found in vegetation surveys and seed banks of the Disked treatments. These four spec ies were
9 found in seed banks of the Thicket and Cleared treatments, but not found in the associated vegetation surveys. Thus, soil disking may have helped express the seed bank and reduce woody cover. However, in systems that have been fire suppressed for decades, the vegetation and seed bank analyses indicated that the seed bank will be insufficient to restore the understory community composition without additional management activities such as direct seeding or planting.
10 CHAPTER 1 IN TRODUCTION Overview Wet flatwoods are low flat pinelands that span across the sout heastern Coastal Plain of the United States. They are an important ecological community because they provide many unique ecosystem services including ground water recharge, w ater quality improvement, wildlife habitat, biomass production, carbon sequestration, and energy redistribution (Cl ark et al. 1999, 2004). Wet flatwoods across the Southeast United States have been drastically altered by the influences of agriculture, silv iculture, and development. These practices have led to a fire excluded system which has rapidly changed the structure and species composition of the plant communities. Created by years of fire suppression, this new degraded system no longer supports the ri ch diversity of flora and fauna that was typical of pre settlement wet flatwoods communities By 1900, 50 % of natural fire maintained wet flatwoods had bee n severely altered and by 1990 almost all of the remaining wet flatwoods had been altered (Ware et al 1993). In recent years land managers throughout the Coastal Plain have encountered many obstacles in their attempts to successfully restore wet flatwoods. Intact fire maintained wet flatwoods are dominated by an overstory of longleaf pine ( Pinus palustr is Mill. ) and an herbaceous dominated understory. In the absence of fire, wet flat woods become degraded and are dominated by shrubs and hardwood trees (Ainsl i e 2002 ) This degraded condition is common throughout the southeastern Coastal Plain and has becom e a focus for restoration projects in the Florida panhandle. The Northwest Florida Water Management District has initiated over 400 acres of wet flatwoods restoration and has plans for additional restoration sites for the future
11 (NWFWMD 2010). Numerous stu dies (Coffey and Kirkman 2004, Cox et al. 2004 and Welch et al. 2004) have led to very good insights into techniques to restore upland groundcover, however, there is comparatively lim ited understanding of groundcover restoration processes in fire suppress ed wet flatwoods. The broad goal of the study is to investigate methods of re storing wet flatwoods groundcover Vegetation Pristine w et flatwoods are pine forests with sp arse to no midstory and dense groundcover of hydrophytic grasses, herbs, and low shru bs. The pine canopy typically consists of one pine species or a combination of longleaf pine ( Pinus palustris ) and slash pine ( Pinus elliottii Engelm. )(FNAI 2010). Compared to longleaf pine dominated wet flatwoods, it is thought that those dominated by sla sh pine with more shrub cover may have had longer fire intervals of 5 7 years, or a few periods of such longer intervals (Landers 1991 ). The understory in wet flatwoods is composed of some of the most species rich communities in the western hemisphere and c ontains some of the highest concentrations of threatened and endangered plant species in the Southeast (Walker and Peet 1983 Peet & Allard 199 5 ). The Florida panhandle has many threatened, endangered, and/or endemic plant species that are found in fire m aintained grassy wet flatwoods. Some of these species include pine woods bluestem ( Andropogon arctatus ), southern milkweed ( Asclepias viridula Calamovilfa curtissii ), wiregrass gentian ( Gentiana pennelliana ), Panhandle spiderlily ( Hy menocallis henryae ), white birds in a nest ( Macbridea alba ), bog tupelo ( Nyssa ursina ), Apalachicola dragon head ( Physostegia godfreyi ), pinewoods wild petunia ( Ruellia pedunculata ssp. pinetorum ), and Florida
12 skullcap ( Scutellaria floridana )( FNAI 2010 ). T he high species richness and diversity of these sites make them a priority for conservation and restoration efforts. T able 1 1 represents commonly found species from nine high quality wet flatwoods sites in the Florida panhandle with supporting data from other site s throughout the panhandle (Kindell 1997, FNAI 2010 ). High quality fire maintained wet flatwoods have been altered in a variety of ways over the past century, resulting in many of the original flatwo ods to become dominated by less desirable woo dy shrubs and trees, as discussed in the sections below. Hydrology At one time wet flatwoods were an abundant community type throughout the coastal plain of the Southeast. Due to the colonization and management by Europeans less than two percent of fire ma intained wet flatwoods remain (Ainslie 2002). Wet flatwoods are easily susceptible to degradation by change s in hydrology and fire regime Proper hydrology is vital to the success of any wetland restoration effort. The hydrology of wet flatwoods is maintai ned by shallow ground water tables, poor drainage, and flat landscapes with low hydraulic gradients (<0.5 percent slope). Soils in wet flatwoods often have an argillic (clay) horizon, which slow drainage. Th ey rarely flood deeper than 10 15 cm yet the wate r table can drop to 1 m or more below ground when evapotranspiration ( ET ) is high and rainfall is low (Rheindhart et al. 2002) A disturbance such as ditching would have a dramatic affect on the hydrology of wet flatwoods by slowly draining the system caus ing a shift in flora and fauna. However, it has been found that hydrologic responses to silviculture are typically short term increases in the water table due to a reducti on in ET (Lockaby et al. 1997a ). Tree
13 removal also has been found to have minimal im pacts on the hydrology due to soil moisture and heat conditions that prompt vegetation recovery and allowing the disturbed hydrology to recover quickly (Sun et al. 2002). Recent models have suggested that future forest removal and climate change would have pronounced impact in the ground table during dry periods but limited impacts under wet conditions (Lu et al. 2009). The hydrologic impacts of disturbances such as ditching, silviculture, and road building, are well documented, as stated above, however t h e effects of woody invasion on the hydrology of wet flatwood s communities is not well documented Disturbance Wet flatwoods are fire dependant systems and much of their composition, structure and functions depen d on natural fire regimes. Nutrient recyclin g occurs in pulses following fires resulting in a rapid turnover of nutrients that enable wet pine flats to quickly recover their characteristic biomass and structure after f ires (Rheinhardt et al. 2002 ) The historic presettlement fire interval for all ty pes of pinelands across the Southeast is estimated to have been 1 3 years (Frost 1998). This fire interval is sufficient to control the invasion of shrubs and maintain an open herbaceous understory (Drewa et al. 2002). Fire not only ke eps wet flatwoods ope n for herbaceous cover it also plays a vital role in the reproductive strategies of many species. Some plant species can not only tolerate fire but require fire to promote flowering. Fire is the flowering stimulus for wiregrass ( Aristida stricta Michx.) a consistently dominant understory plant in longleaf savannas (Platt et al.1988) For those reasons, species composition and structure often reflects the underlying fire regimes of the habitats.
14 Folkerts ( 1982 ) noted that 20 years of fire suppression resu lted in the eliminat ion of herbaceous pine flatwoods in some areas of the Gulf Coast of the southeastern US Loss of the herbaceous understory component is attributed to the reduction in light and competition created by the production of shrubs and trees a nd from the elimination of germination sites due to litter layer buildup (Maliakal et al 2000). The encroachment of trees such as black titi ( Cliftonia monophylla L. ) and titi ( Cyrilla racemiflora Brit. ) can happen rapidly and permanently alter the site. I n the Apalachicola National Forest over 14% (80,000 acres) of the landscape was degraded due to titi invasion over the past several decades. It was also determined that fire in tervals exceeding 3 years were not adequate to control titi invasion (Hess and Laniray 2008). Afte r several years without fire woody species may encroach to a point where the reintroduction of prescribed fire will not effectively eliminate the shrubs and trees and consequently mechanical removal may be necessary (Drewa et al. 2002) Titi can be eliminated by extremely hot fire but that typically results in dangerous crown fires that have the potential to spot and lead to the mortality of pines (Ferguson 1998). L on g term shifts in fire regimes may produce changes to plant communities that are ir reversible (Drewa et al. 2002 ) showed that t he short term (<10 yr) introduction of pres cribed fire will not reverse the shift in plant communities Frequent fire used as the only restoration tool has not been successful in achieving diverse p lant communities in some ecosystems (Heslinga and Grese 2010). In Florida dry prairies, m echanical removal of woody material prior to the reintroduction of fire has been found to be more effective than prescribed burning alone (Watts and Tanner 2006). Fire suppression can also result in the accumulation of organic litter and
1 5 eventually lead to the formation of an organic layer that is not normally seen in wet flatwood systems In summary, f ire suppression has caused a major shift in plant community dynamics in all fire dependant systems in the Southeast resulting in the loss or degradation of the majority of wet fl atwoods throughout this region (Ware et al. 1993). When a disturbance regime such as fire has been altered it is possible for the ecosystem to r each an alternate stable state. Alternate states are combinations of ecosystem states and environmental conditions that may persist at a particular spatial and temporal scale (Gunderson 2000). In the case of some wet flatwoods the al ternate state can beco me a hardwood dominated wet flat. If so, this suggests that those wet flatwoods have shifted to a new state that cannot be restored to the previous condition solely by re establishing historical disturbance regimes (sensu Suding et al. 2004). Thus, r estora tion efforts may need to manipulate more than the single factor that led to the degradation (Beisner et al. 2003). Management decisi ons need to acknowledge that restoring a degraded system can be a complex effort, as the degraded system may represent a sta ble state resilient to change. Active M anagement Removal of invaded woody species can be the most arduous task of the restoration process. Typically, t he clearing of trees and shrubs results in the subsequent resprouting of those plants from their underg round biomass. Titi and black titi are aggressive resprouters after removal of aboveground biomass by clearing method s such as the use of Gyro trac machines Gyro trac machines are all terrain mulching machines that have shown to be effective at removing and mulching understory and mid story vegetation, including small to medium trees, while causing little physical
16 disturbance to the soils (Mitchell 2005). Control of resprouting has been attempted through fire. More r e sprouting is found to occur following burns in the dormant season than during the growing season regardless of habitat or region. However repeated growing season burns have not resulted in reduced densities of established trees and shrubs in pine savannahs (Drewa et al 2002). Fire can help control resprouting of woody species by the use of growing season burns but prescribed fire primarily disrupts the above ground biomass of the plant. When extensive resprou ting is a problem, it may be necessary to attempt methods that will have a greater impact on the below ground biomass. Disking is one such method that impacts the below ground biomass leading to diminished resprouting of shrubs and trees. The r esprouting capability of woody vegetation depends on both above and below ground plant reserv es, and on the possibility of making or maintaining the necessary storage organs during intervals between disturbances (Bellingham 2000). Disking, to over turn the soil with an implement such as a harrow or plow, has a massive impact on the reduction of ab ove ground and below ground biomass, whereas prescribed fire generally only impacts the above ground biomass. Fritzsch ( 2004 ) revealed that disking greatly diminished resprouting vigor in temperate grasslands The potential for disking to assist in control ling resprouting needs to be examined in wet fla twoods, where resilient woody vegetation species have encroached. Furthermore, disking m ay have a positive effect on the expression of a seed bank that has been in submission in the soil for years. Thus, it s hould be important to consider the affects of disking on the seed bank, soil, and vegetative community in wet flatwoods that are targeted for restoration.
17 Seed B ank Viable seed banks are imperative to the success of many restoration efforts. Planting veget ation in absences of a viable seed bank can be very expensive and often times the proper native species are not available In degraded wet flatwoods woody overgrowth can result in the accumulation of an organic layer that impedes the emergence of the hist oric seed bank. It is possible that a disking treatment could bring the buried see d bank to the surface and intermix the organic layer with lower mineral soil layers In a study of shallow marshlands disking increased vegetation div ersity and taxon richne ss and increased abundance of annual species but decreased that of perennials (Polesk et al.1995) Species richness was increased in a study of Carolina bays in the season following disking (Kirkman and Shiritz 1994). However, in order to know what affec ts disking may have on emergence of seedlings from a seed bank, it is first necessary to know whether or not there is a viable seed bank present. T he presence and composition of a seed bank should be studied before a restoration plan that utilizes the see d bank is attempted (van der Valk & Pederson 1989). When the natural understory v egetation has been eliminate d, as in the case of many decades of woody encroachment, reestablishment through a seed bank presents a valuable restoration potential (Simpson et al. 1989). However, reliance on seed banks to revegetate degraded systems has been a widely studied and debated topic in restoration ecology (Aronson et al. 2008, Middelton 1999, 2003, Smith et al. 2002, Wetzal et al. 2001). While many wetland restoration projects rely on the seed bank for revegetation, many researchers have suggested that the use of a seed bank will never lead to
18 historical diversity. For example, t he seed banks of restored wetlands were found to co ntain fewer species and fewer seeds than those of natural wetlands (Galatowitsh and van der Valk 2006, Aronson et al. 2008). In a review of global seed banks studies (Hopfensperger 2007), s pecies richness was not related to seed banks and the similarity between seed banks and extant vegetation w as low in all ecosystems. In fire suppressed flatwoods there is a potential to lose s pecies which re cover vegetatively after fires, but which are not present in the seed bank (M ailakal 200 0 ) F or several important species in longle af pine ecosystems, ther e is l ack of evidence of a p ersistent seed bank, and reintroduction of seed would likely be neces sary for successful complete restoration of groundcover species in these ecosystems (Coffey and Kirkman 2006). However, some target species such as white toppe d picture plant ( Sarracenia leucophylla Raf. ) are very long lived, and are capable of remaining dormant for decades in the seed bank, emerging only after the reintroduction of fire and elimination of woody vegeta tion (Folkerts 1990). I t is apparent that i t can be difficult to generalize about the presence and composition of seed banks across habitats and regions. Contrary to in severely disturbed longleaf pine ecosystems of the North Carolina coastal plain (Cohen et al 2004). Study Objectives R estoration of highly degraded ecosystems can be an ominous task for any restor ation ecologist Degraded wet flatwoods present numerous challenges given that the system has often shifted to an alternate resilient ecosystem. The absence of a natural fire regime creates the shift from an open herbaceous dominated flatwoods
19 system to a tree and shrub dom inated thicket The failure of land managers to transform dense woody wetlands into open herbaceous dominated wet flatwoods h as increased the necessity to better understand the factors influencing successful restoration The overall goal of this study is to examine the restoration potential of wet flatwoods sites in Florida, assessin g current restoration practices and determining the viability of new combinations of restoration treatments. Investigating several degraded wet flatwood sites in north Florida, the specific objectives of this study are to: 1) evaluate the vegetation respo nse to a combination of restoration treatments ; 2) determine the viability and species composition of th e existing seed bank ; and 3) f ollowing previous restoration treatments determine if soil disturbance via disking will a) reduce resprouting of undesira ble woody species and b) aid in seedling emergence from existing seed banks
20 Table 1 1 Common vegetation species found in high quality wet flatwoods of the Florida panhandle (FNAI 2010) Common Wet Flatwood Species Aletris lutea Hypericum brachyphyll. Polygala cymosa Andropogon arctatus Hypericum fasciculatum Polygala ramose Aristida beyrichiana Ilex coriacea Rhexia alifanus Balduina uniflora Ilex myrtifolia Rhexia lutea Bigelowia nudata Lachnanthes caroliniana Rhynchospora ch apmanii Carphephorus pseudolia. Liatris spicata Rhynchospora ciliaris Chaptalia tomentosa Lilium catesbaei Rhynchospora oligantha Coreopsis floridana Lophiola Americana Rhynchospora plumose Ctenium aromaticum Lycopodium spp. Sabatia bartramii Dichrom ena latifolia Muhlenbergia expansa Sabatia macrophylla Drosera capillaris Myrica heterophylla Sarracenia flava Drosera tracyi Oxypolis filiformis Sarracenia psittacina Eriocaulon compressum Panicum spretum Scleria baldw inii Eriocaulon decangulare Pin guicula lutea Scleria triglomerata Euphorbia inundata Platanthera nivea Taxodium ascendens Helenium vernale Pleea tenuifolia Tofieldia racemosa Helianthus heterophyllus Pogonia ophioglossoides Xyris ambigua Helianthus radula Polygala cruciata Xyri s baldwiniana
21 CHAPTER 2 METHODS Site Management H istory After sixty years of fire suppression on former wet flatwoods in the Florida panhandle, the overgrowth of shrubs and trees resulted in communities drastically altered from their former st ates. These areas were overgrown with undesirable vegetation, primarily Cliftonia monophylla and Cyrilla racemiflora that formed a dense thicket. The thicket had trees that reached 25 30 cm diameter breast height (dbh), stood 5 10 meters tall, and made a s olid wall across the landscape (Clayton 2010). A number of former wet flatwoods locations are currently being restored by the Northwest Florida Water Management District (NWFWMD), to mitigate for wetland losses due to activities by the Florida Department of Transportation. In 2004 the NWFWMD purchased 2,155 acres of land known as the Carter Tract, and established the Sand Hill Lakes Mitigation Bank (SHLMB). Prior to the purchase, the property was used since the early 1950 s as a fish camp, with active fire suppression. One of the primary goals of the SHLMB project was to restore wet flatwoods and savannah habitats in the region. The NWFWMD is also performing mitigation activities on state owned lands that are not official mitigation banks. Ward Creek West (WCW), a 719 acre parcel, was chosen for 145 acre s of wet flatwoods restoration. WCW is a coastal landscape comprised of wet flatwoods and savannah with pockets of gum and cypress swamps. Most of this land was converted to slash pi ne plantation in the earl y 1960 s and fire suppressed. Figures 2 1 and 2 2 show aerial photographs of the SHLMB and WCW landscapes in 1949 and 1953, respectively, that show a mosaic of ecosystems with open pinelands mixed throughout. The landscapes
22 have degraded since that time, and prior to the recent restoration activities (below), generally all of the SHLMB and WCW landscapes were densely forested with undesirable titi and other woody vegetation In 2 007, the NWFWMD started a five year effort to restore the wet flatwoods under specific mitigation requireme nts. These requirements constitute the overall restoration goals: 1) Titi and other woody species should be no taller than the coppice sprouts that may have risen from the most recent fire; 2) fire adapted, native, wet flatwoo ds herbaceous species shall average at least 55% cover; 3) the average cover of graminiods should be 60% or greater of the total herbaceous groundcover; 4) the collective cover of pioneer species, such as Andropogon spp. should not exceed 25% of those gra minoids; and 5) long leaf pine should average between 100 200 trees per acre (FDEP 2005). The primary management actions that were used to meet the restoration goals were vegetation removal, followed by a single prescribed fire and multiple herbicide appli cations to control resprouting of woody species (instituted at later times, as described below). In 2007 and 2008, 165 acres of degraded wet flatwoods located at SLMB were cleared by NWFWMD using a Gyro Trac TM machine which cut and mulched all trees and sh rubs at ea ch mitigation location (Figure 2 3 ). In 2008, 145 acres were cleared at WCW using the same method (Figure 2 4). Figure 2 5 shows an example of the site conditions before and after clearing. A single prescribed fire was then conducted at each loca tion dur ing the following winter season Table 2 1 shows the schedule of restoration activities for all sites. Wiregrass, Aristida stricta, and toothache grass, Ctenium aromaticum Walter plugs were planted in subsequent years by the
23 NWFWMD By 2009, a tot al of 800,000 wiregrass plugs were planted at SHLMB (with plugs separated by roughly 1 m in many areas), and 2010 survival monitoring data showed mixed results with some areas having high survival, while other areas revealed survival as low as 50% (NWFMD 2010). Our study areas we re planted with wiregrass and t oothache grass, however neither were found in our study's random vegetation sampling (see Results section). Several years after the wet flatwoods locations were cleared and burned, there was very lit tle progress towards general restoration goals (personal observation). Ground cover recovery was sparse, which suggested one, or a combination, of the following: lack of bud bank, lack of seed bank of desirable species, unsuccessful seed expression (germin ation), and/or a lack of recruitment from seeds dispersed by seed rain onto the sites. The resprouting of branches of Cliftonia monophylla and Cyrilla racemiflora from old stumps and from roots also occurred rapidly following clearing and/or fire (Figure 2 5). Although the clearing operations mulched the trees to the ground level, operations did not disrupt the root systems sufficient ly to prevent resprouting of woody species from the roots. In response, the NWFWMD established an herbicide program that spra yed the resprouting trees and shrubs, according to a regular schedule (Table 2 1 ). The herbicide applications were aimed to control resprouting of all woody species. Herbicide applications, at a minimum, were applied in the spring after first vegetative fl ush and in the fall after the conclusion of the growing season. The herbicide Triclopyr at 10% concentration was used during all applications. General observations indicated that the herbicide was not effe ctive in preventing resprouting Moreover, decades of dense titi (and other woody vegetation) growth led to the
24 accumulation of a thick organic soil surface layer, ranging from 5 15 cm. Such a significant organic layer is uncommon in pristine wet flatwoods, which are usually classified as having mineral s oils (Rheindhart 2002). The NWFMWD conducts annual vegetative sampling at the conclusio n of each growing seaon, at all wet flatwoods restoration sites. The percent vegetative cover is monitored at set transect locations throughout the wet flatwoods resto ration areas. The results are published in annual reports (NWFWMD 2010). The results of the vegetative sampling conducted by the NWFWMD may not be consistant with the results of our study due to different sampling techniques and primarily due to the contin ued use of herbicide applications in the areas sampled by NWFWMD, wheras our study areas discontinued the use of herbicide application at the start of our study. Study S ites Six field study sites were chosen within the broader locations of SHLMB and WCW d escribed above. General visual surveys determined all potential sites within these locations. Sites were sought that resembled each other with respect to elevation and hydrologic regimes, which was generally assumed to be reflected by the surrounding veg etation types. The landscape at SHLMB is a mix of sandhills, wet flatwoods, wet prairies, cypress domes and lakes. SHLMB was historically comprised of wet flatwoods or wet savannahs that transitioned to wet prairies before entering the cypress edges. Stu dy sites 1 4 are located at SHLMB b etween the sandhills and edges of the cypress lakes (Figure 2 3) Sites 1, 2, and 3 are dominated by occasionally ponded Clara a nd Plummer soils, while site 4 is dominated by ponded Pantego and Clara soils (USDA
25 2010) C lara, Plummer, and Pantego soils are very deep, very poorly drained, moderately to rapidly permeable soils that formed in thick loamy sediments on the Coastal Plain (USDA 1999,2004) The landscape at Ward Creek West is typical of the outer Coastal Plain o f the Florida panhandle. It is a mosaic of pine flatwoods communities mixed with depressional wetlands. Ward Creek West contains research sites 5 and 6 (F igure 2 4 ). WCW contains soils from the Rutledge series with Rutledge sands consisting of very deep an d poorly drained soils with rapid permeability formed on sandy Coastal Plain sediments ( USDA 1995). Sampling P lots At each of the six study sites, three 15 m X 15 m plots wer e established: two plots (Cleared and Disked ) were within the area previously cle ared of vegetation, and one plot (Thicket ) was w ithin the closest available woody thicket outside of the cleared area. The Cleared and Disked plots were chosen such that they generally appeared to have consistent vegetative and/or soil conditions within e ach site, and their borders were within at least 2 10 m of each other. The Thicket plot for each site was in an area that was estimated to represent the pre cleared condition, based on proximi ty and relative elevation. The N WFWMD had previously selected th e locations for vegetation removal, and those locations were not alway s in close proximity to titi thicket habitats. Th us, the Thicket plot associated with each study site was not always in the imm ediate vicinity of Cleared and Disked plots Rather than b eing considered experimental control treatments, the primary purpose of the Thicket plots was to determine if the clearing
26 method itself had an impact on the seed bank and to examine species composition of the titi thicket Seed B ank The seedling emergenc e method was considered to be the most suitable method for assessing aspects of the existing seed bank in the study sites. Rather than observing successful emergence of seedlings, the alternative of extracting individua l seeds from soils can lead to diffic ulties in seed identification, and to unknown viabilit y of the extracted seeds. Poiani and Johnson ( 1988 ) found that while emerged seedlings were well ( 93% ) correlat ed with the seeds found in the soil samples, some plant species may require specific condit ions (e.g., drought, cold, or stratification ) to germinate Such conditions may not be provided by the seedling emergence method (Galinato and van der Valk 1986). Nevertheless, the seedling emergence method provides useful indications of the presence of ma ny (but not all) of the viable seeds within a location, which is the simple objective of this part of the study. Seedling emergence was observed in the lab, using soil core samples collected from each site. Because of the existence of large woody debris not all soil locations within a plot were available for sampling. Within each plot, all possible sample locations were identified and separated into two microrelief categories, either high or low. The elevation difference from low to high locations was estimated to be approximately 30 50 cm, and was generally associated with long term vegetation growth prior to clearing the plot. Five high locations and five low locations were randomly chosen within each plot. Random selection was achieved by locating al l available core sample locations, flagging
27 and numbering them, and then randomly selecting until five high and five low locations were chosen. In May June 2010, a three inch diameter coring device was used to sample all locations, for a total of 180 sampl es. The coring device contained a stainless steel sharp ened head attached to a 60 cm clear acrylic tube. Every core was taken to the depth to include 10 cm of mineral soil. This was accomplished by taking deep cores, removing and measuring the organic laye r and then placing the first 10 cm of mineral soil in a tube and cutting it away from the remaining core sample. An organic layer is not typical in w et flatwoods (Rheindhart 2002). T herefore, it was imperative to set a standard depth of mineral soil and al low the organic layer to fluctuate per sample. Taking a core to a standard depth of 10 cm total would have produced many samples that were only organic material. All soil cores were transported to the University of Florida greenhouse in Milton, Florida fo r the seedling emergence experiment. Once the core samples air dried for at least 10 days, it was evident that many of the samples contained a mixed organic/mineral zone between the normally distinct organic and mineral layers, with the mixed layer indisti nguishable in the field. Therefore, the cores were then divided into three subsamples, organic, mixed and mineral. The mineral fraction remained as the lower 10 cm of the core. The samples were then sifted through a 2 mm sieve to remove any roots or large organic materials. The volume and mass of the subsamples from each core was recorded. In July 2010, a 30 mL sample was taken from each subsample and spread evenly across 12 inch round pots that contained 300 ml of farfard media mix. The pots were then la beled and placed randomly on greenhouse benches with an automatic
28 irrigation system that supplied a fine mist of water via an intermittent mist system. The irrigation was controlled by automated timer (QCom Environmental Control System) that provided a min imum of five seconds of misting every 30 seconds from dawn to dusk The greenhouse was enclosed and climate controlled, allowing for germination conditions that approximated summer months to continue through the fall and winter. Once a month during the dur ation of the experiment, all emerged species were identified and counted individually, then removed from the pots. Plants that could not be indentified were removed from the twelve inch round pot and potted into a separate pot until species identification could be determined. After approximately seven months of observation, germination became limited, and the seedling emergence experiment was terminated. Vegetation S ampling Existing vegetation was sampled in the Cleared and Disked plots at each site in Jun e 2010, and then repeated in November 2010 and May 2011. The Thicket plot at each site was sampled in November 2010 and again in May of 2011. It was not sampled in June of 2010 because those plots primarily contained dense woody species, and seasonal varia bility in species composition had been previously observed to be minimal. Vegetation was s urveyed in the Cleared and the Disked plots multiple times to ensure that most seasonal species were recorded, and to provide some indication of short term trends in vegetation structure and species composition. T he vegetation was sampled at the same (randomized) core sampling locations. A 1 x 1 m quadrat was placed directly over each core sample location, and the percent cover of each species was estimated visually wi thin each of 20 sub quadrats within a grid of the 1 m 2 quadrat For each tree
29 species that provided overhead canopy cover, the relative cover of each was estimated by standing inside the quadrat and visually assessing the percent cover. Soil D isturbance Af ter the initial vegetation surveys and soil sampling the Disked plot at each site was disked in July of 2010 using a ire 15 m X 15 m plot was disked affectively removing and breaking up woody roots, and also bringing mineral soil t o the surface. The disking disturbed the top 20 30 cm of soil (Figure 2 6). Data A nalysis The ex (S) was used to compare relationships between the seed bank s and the extant vegetation among all treatments, dates, and elevations. I t was also utilized to compare the relationships among extant vegetation of all treatments. S= 2 W a + b where a = # of species in sample a b = # of species in sample b W = # of species shared by the two samples Data were organized usi ng a combination of ordination techniques. Ordination aims at arranging species and treatments in a low dimensional space such that similar entities are in close proximity and dissimilar entities are more distant Ordination was used to show patterns of si milarity between the seed bank and extant vegetation using relative cover of the extant vegetation and relative abundance of the seed bank applying
30 nonmetric multid imensional scaling (NMS ), with also utilized to show sim ilarity in vegetation between treatments. NMS a nalyses were completed in PC ORD 6 (McCune and Medford 2006) based on guide lines by Peck ( 2010 ) using the autopilot function and setting the distance measure to Sorenson (Bray Curtis). Based on the stress ca lculations (difference in rank order distances of data and ordination), PC ORD recommended two dimensions for the final run. Principal co mponents analysis (PCA) was an additional ordination tool used ( in PC ORD 6) to display the relationship between treatm ents as well as the relationship of certain species to the various treatments. The effects of treatment, date, and elevation were analyzed for the following variables; herbaceous cover, woody groundcover, woody canopy cover, total woody cover, bare ground Cliftonia monophylla, Rhexia s pp., Hypericum sp., and Xyris brevifolia using repeated me asure ANOVA (PROC MIXED in SAS 9.2 ) with Bonferroni corrections to separate means. Herbaceous groundcover was the total combined percent cover of all herbaceous speci es found in extant vegetation surveys. Woody groundcover was the combined percent cover of all woody species found in extant surveys. Woody canopy was the total combined percent cover of woody canopy species in extant veg etation surveys. Total woody equale d the combination of woody groundcover and woody canopy cover. Bare ground included all areas w here vegetation was absent The variables above were chosen to determine the impacts of restoration activities on the broad goals of removing woody vegetation an d increasing herbaceous vegetation. Cliftonia monophylla was chosen because it is the p rimary woody species of concern Rhexia spp ., Xyris brevifolia Michx. and Hypericum sp were chosen because
31 they are all species indicative of pristine wet flatwoods. R esults higher than p > 0.05 were not considered significantly different. Tests for normality were run in (SAS 9.2) and it was determined that the data met the assumptions of normality and did not need transformation.
32 Figure 2 1. Sandhill Lakes Mit igation Bank (SHLMB) in 1949, showing the overall heterogeneity of the habitats. ( USDA 1949 ) Figure 2 2 Ward Creek West (WCW) in 1953, showing the overall heterogeneity of the habitats. ( USDA 1953 )
33 Figure 2 3 Sandhill Lakes Mitigation Bank (2007) s howing the outlines of locations that were subsequently cleared of vegetation and site locations 1 4 (Photo courtesy of NWFWMD )
34 Figure 2 4. Ward Creek West (2008) showing the outlines of locations that were subsequently cleared of vegetation and site l ocations 5 6 (Photo courtesy of NWFWMD )
35 Table 2 1 Schedule of management activities Location Veg. Clearing Fire Herbicide Site 1 Sandhill Lakes April 2007 December 2007 Sept. 08,May, July, Oct. 09 & May 10 Site 2 Sandhill Lakes Marc h 2007 December 2007 May, July, & Sept. 09, & May 10 Site 3 Sandhill Lakes May 2007 December 2007 May, July, & Sept. 09, & May 10 Site 4 Sandhill Lakes July 2008 December 2008 May, July, & Sept. 09, & May 10 Site 5 Ward Creek West August 2008 December 2008 May, July, & Sept. 09, & May 10 Site 6 Ward Creek West August 2008 December 2008 May, July, & Sept. 09, & May 10 Schedule of vegetation management activities conducted by NWFWMD at the six site locations used in this study. Veg. Clearing is vegeta tion removal with Gyro Trac; Fire is prescribed burning; Herbicide is herbicide application. The herbicide Triclopyr at 10% concentration was used for all applications and all woody species were targeted.
36 A B Figure 2 5. An example of the t horough vegetation removal at a W ard C reek W est mitigation location. A ) The foreground shows the remaining vegetative debris, and the background shows the dense thickets of Cliftonia monophylla and associated vegetation. The photograph was taken in Augus t 2008, immediately after Gyro Trac vegetation removal. B ) Same site 2 months later. (Photos courtesy of NWFWMD).
37 Figure 2 6. WCW July 2010. Immediately following disking treatment. Mineral soils are observed mixed on surface with organics. Beyond the d isked plot there is visible Cliftonia monophylla resprout cover throughout the C leared plot. (Photo courtesy of author).
38 CHAPTER 3 RESULTS Herbaceous C over and C omposition The influence of the restoration tre atments on herbaceous cover was cons iderable, yet variable among treatment and through time. The Thicket plots had little to no herbaceous cover during any sampling date. Total h erbaceous cover was significantly affected by treatment (p= <0.001, F= 33.29 df= 2,10) date (p= 0.0003, F=20.40, df= 2,10 ) microrelief (p= 0.0096, F=16.63, df=1,5 ), and there was a significant interaction of date with treatment (p=<0.0001, F= 20.40, df=4,20) Mean herbaceous cover for the untreated Thicket plots did not signifi cantly vary through time and had signif icantly less herbaceou s cover than the Cleared and Disked treatments. Prior to disking, herbaceous cover of the Cleared plots and the plots to be disked (June 2010) did not differ significantly (Table 3 1) By November (4 months after disking) herbaceous c over was significantly greater in Cleared plots compared to Disked or Thicket plots Herbaceous cover in the C leared plots increased significantly from June to Novemb er, but decreased by the May sampling date. D isked plots increased in herbaceous cover fro m June to Nov ember, however, not significantly. Eleven months after disking, the greatest mean herbaceous cover 27.2%, occurred in Disked plots and differed significantly from all other treatments and dates Microrelief significant ly affected the cover of herbaceous species (p= 0.0096 F= 16.63, df =1, 5). Microrelief by treatment was found close to significant (p= 0.056, F= 3.89, df =2, 10). Mean cover was higher in the lower elevation locations, increasing from 10 % to 16 % and from 6 % to 15 % for Disked and Cleared treatments, respectively. Eleven months after di sking (May 2011), mean cover was 7 % and 14 % in the high and
39 low sites (respectively) in C leared plots and 21 % and 33 % in high and low sites in D isked plots Lachnanthes caroliana Lam. and Rhyncho spora sp. were the two species that showed the greatest percent increase from hi gh to low elevations (Appendix B ). There were 14 herbaceo us species found in the plots during the course of the study. The most common species were Rhynchospora sp ., Dicanthel ium spp., Andropogon spp., and Lach nanthes caroliana (Appendix A ). Five of the 14 documented species were listed on the (FNAI 2010) list of common wet flatwoods species of high quality panhandle sites (Table 1 1). One o f those species was located in Thicke t plots two of tho se species were located in the C leared treatment plots, and all five in the D isked treatment plots. The interaction of date and treatment was significant for Rhexia sp p. (P=0.0037, F=5.51, df=4,20). Rhexia spp were located in D isked plo ts four months after disking (November) but none were found in the C leared plots (Table 3 2). Rhexia spp cover increased significantly from the November to May in the D isked plots. The interaction of date and treatment was significant for cover of Xyris b revifolia ( P=0.0093, F=4.5, df= 4,20). X. brevifolia was found solely in the D isked plots in May 2011 and was not found in any treatment at any other date. Drosera capillaris Poir. Bidens mitis Michx. Hedyotis sp. and Syngonanthus flavidulus Michx. were also only found in D isked plots in May 2011 (Appendix A). In May 2011, there were no herbaceous species documented in the T hicket plots, six herbaceous species in the C leared plots, and 11 herbaceous species in the Disked plots Rhynchospora sp., L. carol iana, and multiple Dicanthelium species had the greatest increase in percent cover in the disked plots following disking.
40 Woody C over and C omposition Woody species cover and composition were significantly altered by the restoration treatments The effe ct of treatment (P=<0.0001, F= 54.65, df=2,10), microrelief ( P=0.0066, F=19.95, df=1,5 ) and the interaction of date and treatment were significant (P=0.0007, F= 7.58, df= 4,20) (Table 3 1). Thicket plots showed little to no variation in woody groundcover throughout the three sampling dates. The C leared plots showed a trend of increasing woody groundcover over time. Woody cover increased from 17% in June to 27% in November 2010, and 40% in May 2011 The woody groundcover in the Cleared plots did not differ significantly from the Thicket plots Contrary to the increases in woody cover in the Cleared plots, the D isked treatment showed a trend of decreased woody groundcover. Prior to di sking, the mean cover was 14% in June of 2010, decreased to 2% in November 2010, and 3% in May 2011. After disking, the mean woody groundcover of the D isked plots was significantly lower tha n the means of the Thicket and C leared plots (Table 3 1), with the trends evident in Figure 3 1. Woody groundcover was combined with woody canopy cover (which was measured independently from groundcover) to quantify the affects of the treatments on total woody material that covers and shades the plots The interaction of date and treatment was significant for woody cover (P=0.0007, F=7.58, d f=4,20). Mean woody cover was greater than 124% for Thicket plots on all dates and was significantly greater than that found in Cleared and Disked plots for all dates (Table 3 1) There was no woody canopy cover in the Cleared or Disked plots.
41 Microrelief significantly affected total woody cover (ground cover + canopy cover) (P=0.0066,F=19.95,df=1,5) Woody ground cover of Thicket plots was 37% for high sites and 24% for low sites and total woody cover was 132% for high and 119% for low sites in the Thicket plots Total woody cover for Cleared plots was 33% on high sites and 23% on low sites. Woody cover of Disked plots was 9% on high sites and 4% on low sites. There we re a total of 19 woody species found within all the plots The most common species were Cl iftonia monophylla, Lyonia lucida Lam. and Cyrilla racemiflora. Cliftonia monophylla was the dominant woody species at all sites and treatments. Treatment and the interaction of date with treatment were significant for the cover of Cliftonia monophylla (P = <0.0001, f= 30.89, df= 2,10 a nd P= 0 .0034, F= 5.62, df= 4,20), but microrelief was not a significant factor The mean cover values of this species in the Cleared plots in November 2010 and May 2011 were significantly higher than the mean percent cover in all other dates and treatments (Table 3 2), reaching a maximum of 27% in May of 2011. A contrary trend was seen in the D isked plots, in which the mean cover of Cliftonia monophylla decreased significantly from 9% in June 2010 to 1% in November 2010 and Ma y of 2011. Lyonia lucida was the second most dominant species by cover at all treatment sites and increased p roportionally in percent cover more than any species (Appendix A). L Lucida had the lowest decrease in mean cover of all wo ody species in the Disked plots. Hypericum sp. were documented in November 2 010 and May 2011 solely in the D isked treatment plots. Treatment and the interaction of date with treatment were significant ( P= 0.0004, F= 8.30, df= 4,20). In November, the mean cover of Hypericum
42 s p was 0.03% and significantly increased to 0.73% in May of 2011. Pinus elliottii was only documented in May 2011. The emergence of the above two new species, common species of maintained wet flatwoods, contributed 1.15% of the total 3.4% total woody mean cover for the D isked treatment plots in May of 2011 While, these changes may seem insignificant, they show a trend in a shifting plant community Seed B ank C omposition During the course of the seven month seedling emergence study a total of 1,600 seedlin gs emerged. Among these 1,075 seedlings of 13 different species were considered weedy/pioneer sp ecies, 39 seedlings of one invasive species ( Lygodium japonicum Thunb. ) were found and the re maining seedlings represented 12 different species that were cons idered common to m aintained wet flatwoods (Table 3 3 ). Five of the 12 common wet flatwoo ds species were listed in the (FNAI 2010) list of common species of high quality wet flatwoods of the Florida panhandle (Table 1 1 ). Thirty five tax a germinated from t he seed bank of which 18 were indentified to the species level, 11 to genera, and a group of ferns to order. Many species were difficult to indentify to the species level, due to the timeframes of the experiment and lack of flowering. Four of the species t hat were identified to genera in Table 3 3 contained more than one species as presented below. The ge nus Xyris was represented by Xyris brevifolia and Xyris bal d winiana Schult., with Xyris brevifolia being the majority of seedlings that emerged The Cy perus genus was represented by at least two species, Cyperus erythrorhizon Muhl. and Cyperus sp. Rhexia spp. was predominately Rhexia virginica L. but also contained Rhexia m arian a L. and Rhexia alifanus Walter Dicanthe l i um spp represented at least
43 thr ee different species. Pteridophyta, ferns, were represented by several different species. The most common of all species were Hedyotis sp ( 829 ) seedlings Xyris sp p .( 188 ) seedlings, Hypericum sp with ( 98 ) seedlings, and Erigeron sp with ( 83 ) seedlings. The largest number (740) of seedlings occurred in the mineral soil layer, with 442 found in the mixed soil horizon, and 417 in the organic layer. This numeric predominance in the mineral layer was strongly influenced by the presence of Hedyotis sp ., repre senting three fourths of the emerged seedlings of the mineral layer. If the pioneer/weedy species Hedyotis sp. were not considered in the totals, the distribution of seedling by soil layer switched to dominance by the organic horizon (with 160 in the miner al layer, 255 in the mixed horizon, and 355 in the organic layer ) Lygodium japonicum an exotic invasive species, had 39 seedlings emerge. While there is a possibility the spores were introduced in the greenhouse, it is unlikely due to the fact that the seedlings were only found in mixed and organic soil samp les and they were not found in t he control pots that were intermixed in the seed bank study There were no occurrences of L. japonicum in any of the extant vegetation surveys. The distribution of se edlin g by treatments was 499 in the Cleared plots and 1,100 in the thicket. In this cas e, "cleared" includes both the Disked and Cleared plots combined (because the soils were not disked until after the soil cores were taken). Again, the presence of Hedyot is sp. had a significant influence on the distribution of seedlings with 789 occurring in the T hicket plots. If Hedyotis sp is ignored, t he totals are 459 seedlings in Cleared plots and 311 in Thicket plots. Microrelief did not have a
44 strong influence on the distribution of seedlings, with 772 in high locations and 827 in low locations (Table 3 3). Vegetation and Treatments The NMS ordination of extant vegetation cover of all treatments from May 2011 revealed an obvious separation between treatments (st ress= 13.79, P=0.004) (Figure 3 2). The Disked treatments showed a clear pattern of breaking away from the Cleared treatments and complete separation from the Thicket treatments. The Thicket plots also showed some separation from the central areas that had the most commonality, however most of the thicket points accumulated in the center of the two dimensional space, clumped with Cleared plots. The Cleared treatment plots remained in the central area of the ordination and were intermixed with the thicket pl ots. These ordination results support the earlier data that suggests the C leared plots became more similar to the Thickets plots, while the Disked plots increasingly differed from the other two treatments. The PCA ordination of the final sampling date (F igure 3 3) showed very similar results to the NMS ordination. The species highlighted in Figure 3 3 are considered species of highest concern. The three most common woody species, Cliftonia monophylla, Lyonia luci d a, and Cyrilla racemiflora were all assoc iated with the thicket and cleared treatment plots. The species common to "maintained" wet flatwoods, Rhexia sp., Xyris sp., Rhynchospora sp., Drosera capillaris and Dicanthelium sp., were all associated with the disked treatment plots. This PCA ordinatio n further supports the above results, all indicating that common wet flatwoods species were most closely
45 associated with Disked treatments, while problematic woody species were most closely associated with the Cleared and Thicket treatments. Similarity be tween Vegetation and Seed B ank Comparisons between the treatment areas revealed little similarity between the 4). This index ranks similarity on a scale from 0.0 to 1.0, with 1.0 indicating the most similarity The extant vegetation of the C leared treatment s became less similar to the seed banks from June of 2010 to May 2011, starting at 0.34 and ending at 0.29. However the extant vegetation of the Disked treatments revea led the opposite, and increased in similarity to the seed bank from June 2010 to May 2011, starting at 0.23 and ending at 0.41. The highest similarity value between extant vegetation and seed banks was 0.48, for the comparison between the seed banks of the Cleared treatments an d the extant vegetation of the Disked treatments in May of 2011. The lowest similarity values were 0.06 for the comparison of the seed bank of the Cleared and the extant vegetation of the Thickets in May of 2011. ty index was also used to calculate the similarity in species the composition of both the disked and cleared treatment becoming less similar through time when compared to sp ecies composition in June of 2010 (Table 3 4). Species composition remained similar through time for the thicket (0.84). The second NMS ordination showed a clear distinction between the seed banks and the extant vegetation (Figure 3 4). There were no overl apping points in the two dimensional space, and the separation between the seed bank and extant vegetation
46 was significant (stress= 7.95, P= 0.004). These results provided further evidence that the composition of the seed banks was dissimilar from that of the extant vegetation.
47 T able 3 1 Cover means for treatments by date and variable. Date Management June 2010 November 2010 May 2011 Bare Ground Thicket -70.1 b 67.4 b Cleared 73.6 ab 55.9 bc 49.7 c Disk 80.6 a 88.9 a 67.8 b Herbaceous Thicket -0.5 de 0.3 e Cleared 5.0 cd 16.6 b 10.7 bc Disk 3.0 cd 9.5 cd 27.2 a Woody Groundcover Thicket -29.2 abc 32.3 ab Cleared 17.4 bc 27.2 abc 40.1 a Disk 13.9 cd 1.7 d 3.4 d Woody Groundcover +Canopy Thicket -124.2 a 127.3 a Clear ed 17.4 c 27.2 bc 40.1 b Disk 13.9 cd 1.7 d 3.4 d Mean percent c over of species lumped into cover type categories for all treatments and dates The repeated measures ANOVA Bonferroni corrections m eans within cover type category wit h differences between sample means followed by the same letter not being significantly different Canopy cover was an independent >100. The Thicket plots were not sampled in June 2010.
48 Figure 3 1 Mean percent cover for cleared and disked treatments for June 2010, November 2010, and May 2011. Herb aceous = herbaceous species; Woody= woody groundcover species Cleared above ground woody vegetation removal via GyroTrac + p rescribed fire + herbicide Triclopyr applied to woody resprouts; and Disked same as Cleared, but followed by soil disking.
49 Table 3 2 Mean percent cover for selected species by treatments and dates. Date Management June 2010 November 2010 May 2011 Cliftonia monophylla Thicket -5.6 b 6.5 b Cleared 10.9 b 18.4 a 26.7 a Disk 8.7 b 0.7 b 1.3 b Hypericum sp. Thicket -0.0 b 0.0 b Cleared 0.0 b 0.0 b 0.0 b Disk 0.0 b 0.03 b 0.73 a Rhexia sp. Thicket -0.0 b 0.0 b Cleared 0.0 b 0.0 b 0.05 b Disk 0.0 b 0.03 b 0.6 a Xyris brevifolia Thicket -0.0 b 0.0 b Cleared 0.0 b 0.0 b 0.0 b Disk 0.0 b 0.0 b 0.8 a Mean c over (not including canopy) of selected species for all treatments and dates Th e repeated measures ANOVA Bonferroni corrections m eans within a species with differences between sample means followed by the same letter not being significantly different The Thicket plots were not sampled in June 2010.
50 Table 3 3 Total Seedling emergence from seed bank Species Species description Total # Soil Fraction Treatment Elevation Mineral Mixed Organic Cleared Thicket High Low Centella asiatica HC 2 0 0 2 1 1 1 1 Conzya canadensis HP 57 14 12 31 40 17 30 27 Cyperus spp. HC 13 1 4 8 7 6 7 6 Dichanthelium spp. HC 31 13 6 12 24 7 13 18 Drosera capillaris HC 17 0 17 0 0 17 1 16 Ere chtites hieraciifol ia HP 3 0 2 1 2 1 3 0 Erigeron sp. HP 83 38 16 29 59 24 49 34 Eupatorium capillaris HP 23 0 1 22 9 14 11 12 Pteridophyta HC 24 0 1 23 13 11 10 14 Fimbristylis sp. HC 2 1 0 1 1 1 0 2 Gnaphalium pensylvanica HP 2 1 0 1 2 0 2 0 Gnaphalium spicat um HP 6 0 2 4 3 3 2 4 Hedyotis corymbosa HP 19 0 3 16 18 1 16 3 Hypericum sp. WC 98 17 28 53 20 78 36 62 Juncus sp. HC 8 3 0 5 3 5 2 6 Lachnanthes ca roliana HC 1 0 0 1 1 0 1 19 Ludwigia sp. HP 4 2 0 2 1 3 2 2 Linaria canadensis HP 1 0 1 0 1 0 1 0 Lygodium japonicum HI 39 0 9 30 23 16 20 0 Lyonia lucida WC 20 3 1 16 8 12 12 8 Oxalis corniculata HP 52 8 27 17 42 10 21 31 Rhexia spp. HC 47 2 32 13 4 0 7 31 16
51 Table 3 3 Continued. Species Species description Total # Soil Fraction Treatment Elevation Mineral Mixed Organic Cleared Thicket High Low Rhynchospora spp. HC 36 3 14 19 25 11 19 17 Hedyotis sp. HP 829 580 187 62 40 789 449 380 Xyris spp. HC 188 56 78 54 117 71 36 152 Total Species 25 14 19 21 22 21 23 19 Total Seedlings 1599 740 442 417 499 1100 772 827 Total s eedling emergence of individual species from seed bank by Soil Fraction, Treatment, and Microrelief ca tegory. Species description; H=herbaceous, W= woody species, C= common wet flatwoods species, P= pioneer/weedy species, and I= invasive exotic species.
52 T able 3 4 Treatment Cl June Cl Nov Cl May Dk June Dk Nov Dk May Th Nov Th May Sd Cl Sd Dk Sd Th Cl June 1 Cl Nov 0.92 1 Cl May 0.79 0.85 1 Dk June 0.87 0.79 0.73 1 Dk Nov 0.67 0.72 0.81 0.69 1 Dk May 0.63 0.69 0.68 0.59 0.71 1 Th Nov 0.47 0.44 0.47 0.44 0.48 0.48 1 Th May 0.43 0.54 0.43 0.53 0.44 0.34 0.84 1 Sd Cl 0.34 0.36 0.29 0.32 0.35 0.48 0.15 0.06 1 Sd Dk 0.25 0.27 0.25 0.23 0.32 0.41 0.11 0.06 0.82 1 Sd Th 0.29 0.31 0.29 0.27 0.3 0.48 0.15 0.12 0.88 0.84 1 Similarity relationships between extant vege tation of treatments and between seed bank and extant vegetation of 1.0, with 0=least similar, 1= most similar). Cl= Cleared treatments, Dk= Disked treatments, Th Thicket treatments, and Sd= Seed bank.
53 Figure 3 2 NMS ordination representing similarity between treatments. Mean species cover values from the last sampling date of May 2011 were used f or the ordination. Disked plots show the greatest separation from the Thicket plots Thicket shrub thickets with no modification; Cleared above ground woody vegetation removal via GyroTrac + prescribed fire + herbicide Triclopyr applied to woody resprouts; and Disked same as Cleared, but followed by soil disking.
54 Figure 3 3 PCA ordina tion representing similarity between treatments. Mean species cover values from the last sampling date of May 2011 were used for the ordination. Woody and herbaceous species of concern were selected to show species associations with treatments. Thicket s hrub thickets with no modification; Cleared above ground woody vegetation removal via GyroTrac + prescribed fire + herbicide Triclopyr applied to woody resprouts; and Disked same as Cleared, but followed by soil disking.
55 Figure 3 4 NMS ordination of extant vegetation similarity to seed bank. Clear separation between extant vegeation and seed bank is shown above. Seed bank data and the extant vegetation data (May 2011) were derived from total emergence of the combination of the Cleared and Disked tr eatments.
56 CHAPTER 4 DISCUSSION Herbaceous Cover and Composition The results suggest very little progress towards restoration of the herbaceous plant community in these north Florida wet flatwoods FNAI (2009) recommended that wet fla twoods her baceous cover should exceed 25% with le ss than 2% cover of bare ground. F urthermore a reference community for Florida panhandle wet flatwoods had herbaceous cover of 74% and bare gro und of less than 2.5% (FNAI 2009). However, t he restoration techniqu es ut ilized by NWFWMD on our wet flatwoods sites resulted in less than 11% herbaceous cover after four years of implementation. The addition of the soil disturbance via disking resulted i n increase s in herbaceous cover to 27%, which was still much lower tha n th e reference community but suggests a potential for herbaceous cover restoration However, the study did not reveal whether that trend towards increased herbaceous cover will continue. To the contrary, the C leared treatments showed trends of decreasing herb aceous cover suggesting that without further management activities restoration of the herbaceous community is unlikely. Herbaceous species common to wet flatwoods were found in our study, but major components of the reference plant community were missing. Peet and Walker (1984) documented over 50 herbaceous species in unimpacted wet pinelands of North Carolina (that are similar to Florida wet flatwoods), with Muhlebergia expansa, Ct enium aromaticum, and Aristida stricta being the dominant grasses. This is consistent with Florida reference communities and the species list of Table 1 1. In a classific ation of Florida wet pinelands Carr et al. ( 2010) found that herbaceous species were 95% of indicat or species for that community. The herbaceous composition foun d in our study
57 varied gr eatly from that of ref erence wet flatwoods, with the C leared treatments containing only six herbaceous species. Dominant warm season grasses were not found in either our vegetation surveys or the seed bank emergence. S arre cenia spp. an iconic species of wet flatwoods and prairies of the Florida panhandle, were absent. There were also major families, such as Asteraceae, absent from our surveys. Site preparation techniques have been previously explored to d etermine the effects on und erstory plant communities A study on coastal plain pine plantations, examined several methods of site preparation including chopping, burning and disking, finding little change to the herba ceous biomass (Brockway et al. 1998). Our study revealed that cho pping, burning, and d isking led to an increase in h erbaceous cove r and species richness The D isk ed treatments had eleven herbaceous species, compared to six in the Cleared treatments. The increase in species was likely the result of the disking treatment expressing the seed bank Three of those species, Drosera capillaris, Rhexia sp., and Xyris brevifolia are considered common herbaceous species of reference wet flatwoods and were all found in the seedling emergence study. Fire suppression results in the proliferation of a midstory component that leads to the accumulation of a litter and root layer that diminishes the herbaceous groundcover (Maliaikal et al. 2000). The organic layer in our study ranged from 0 20 cm in depth (Appendix C) and had the potenti al to suppress emergence of seeds and reduce the necessary mineral germination locations. The results suggest that microrelief, elevated microsites suppressed the emergence of herbaceous species due to a buildup of organic material that remained from remn ant titi mounds The depressed microsites had higher cover and richness of herbaceous species. The increased herbaceous cover and
58 richness that we found in the D isked treatments could be a result of the disking process that breaks down and mixes the inhibi tory organic layer or otherwise expose the mineral soils. The disking did not show any immediate negative effects to the herbaceous community, with no reduction in species compared to the pre disking surveys, and increased percent cover of all species Th ere are several factors that could have influenced the he rbaceous cover results. During the first 1 2 years after gyro tracking and burning there remained large deposits of woody debris that could have potentially suppressed germination sites and prevented emergence from the seed bank or from seed rain entering the site. There also have been irregular rain pa tterns during the study period In 2010, there was a wet spring followed by a drought in June (personal observation) which could have decimated any se edlings that had emerged in the spring. In 2011, an extreme drought took place throughout the spring and into the summer which could have prevented new emergence of herbaceous species. The brief (1 yr) time period of the study was also a limiting factor. A dditional monitoring over the next few years would be beneficial to confirm the trends we saw in our research. Woody Cover and Composition After removal of woody vegetation by mechanical means or burns resprouting of woody species has proven to be one of the most difficult obstacles in the restoration of all types of pinelands. Dormant season burns, as used in our research area, are not effective in controlling woody resprouting and in some cases have led to vigorous resprouting that exceeded the woody bi omass prior to the burn (Drewa et al. 2002) Growing season burns were also determined to be ineffective at eliminating resprouting,
59 but the practice did control resprouting more effectively than dormant season burns (Drewa et al. 2002). Our results were c onsistent with the above findings and confirmed that Cliftonia monophylla and Lyonia lucida are especially aggressive resprouters. Moreover, significantly higher woody cover was found at the elevated microrelief sample locations which is likely due to the f act that those sites were generally remnant mounds of Cliftonia monophylla which resprouted Herbicide application is a common management tool to control resprouting of woody vegetation in pinelands and many other fire dependant herbaceous ecosystems. Selective herbicides have been effective in controlling woody resprouts in many longleaf pine c ommunities (Jack et al. 2005, Brockway et al. 2000). However, t hese results differ ed dramatic ally from our study. Four to five herbicide applications, of Triclop yr at 10% concentration, over the course of three years proved to be ineffective in controlling the resprouting of Cliftonia monophylla or Lyonia lucida In the Cleared (but non disked plots), w oody cover increased significantly throughout the study and di d not seem inhibited by herbicide. It is likely that without further management actions, the woody groundcover will continue to expand and prevent restoration of the herbaceous groundcover. results show we're pr ogressing extremely well towards diverse wet (NWFWMD 2010). These results differ significantly from the results of our Cleared plots in the same region, which had 27% woody cover in November of 2010. NWFWMD applied one more herbicide application in the fall of 2010, prior to sampling, which could be a reason for their comparatively lower woody cover values. Our study indicates that if
60 herbicide in discontinued in cleared (but undisked) areas, woody species will be able to res prout. The NWFWMD woody cover results may differ if some moderate time is allowed for resprouting after herbicide application. Furthermore, our Cleared plot results did not show a trend towards diverse wet flatwoods species richness did not increase durin g our study, and the observed increase in woody cover suppressed the ability for herbaceous cover to increase. However, as noted below, our experimental plots in which the soil was disked had significant decreases in woody cover Site preparation treatm ents for pine plantations have been studied thoroughly, w ith many of the treatments similar to the restoration treatments of this study. A combination of treatments including disking has been found to be effective at controlling woody sprouts and increasin g species richness, evenness, and diversity (Nilsson and Allen, 2003, Fredrickson et al. 1991). In our study, disking treatments appeared to significantly help reduce woody groundcover and prevent resprouting. The disk harrow was able to break and disrupt root systems (personal observation), which prevented most resprouting during the course of this experiment. The primary species of concern, Cliftonia monophylla had significantly decreased cover in the disking treatment. While there was a slight increase in cover from November of 2010 to May of 2011, the resprouts were limited to small areas where the disking was relatively ineffective, and did not sufficiently disrupt the remaining roots. There did not appear to be any new areas resprouting from undergrou nd root structures. Lyonia lucida a woody species of less concern, was not suppressed by the disking treatment, but it is unlikely that it will have a significant impact on the overall woody cover in the long term. Two new woody species were found after t he disking treatment: Hypericum fasculatum and Pinus elliottii
61 The emergence of Pinus elliottii is most likely due to the breakdown of surface organic material, exposing the necessary mineral soils for germination. Hypericum fasculatum was also found in t he seedling emergence s tudy and was only found in the D isked treatment areas, suggesting that its emergence was a result of the disking treatment. The ability to control the rapid occurrence of woody resprouting is a key element to successful restoration of these systems I f successful, i t opens up or maintains germination locations for new seeds, decreases competition with herbace ous species, and allows for fuel continuity and fire carry Overall, our results suggest that a disking treatment following cl earing and burning could be beneficial for a number of reasons including the suppression of woody resprouts. However, it is important to acknowledge that disking treatments may not be appropriate in systems that are not as degraded as those in our study s ites. Here, f ire suppression led to the complete absence of groundcover and the conversion of the habitat to a stable, densely fores ted system When restoring s ites with remnants of healthy groundcover, disking could be detrimental to any existing, desirab le warm season perennial grasses and other native groundcover species Seed Bank Via seed banks and other recruitment processes, natural recolonization of plants has been the primary source for revegetation in many wetland restoration efforts worldwide. Wh ile this method of natural recolonization is popular, it has rarely led to plant communities that resemble the original diversity of unimpacted wetlands (Aronson and Galatowitsch 2008, Allen 1997, Middleton 2003, Smith et al. 2002). The results of our stud y support those findings and suggest that the seed banks of degraded wet
62 flatwoods would not be sufficient in restoring the plant communities to reference conditions. Almost half of the species found in the seed bank were considered weedy species and are n ot characteristic of reference wet flatwoods communities. Nevertheless, some species that are common to reference wet flatwoods, such as Rhynchospora sp., Rhexia sp., Xyris sp., and Dicanthelium spp ., were represented in the seed bank. Rhynchospora sp. an d Dicanthelium spp. were also found in the seedling emergence study, and were two of the common herbaceous species found in the vegetation surveys. In contrast, Andropogon sp p were prevalent in the surveys, yet were not found in seed bank. The seed bank also did not contain any common perennial warm season grasses, asters, or legumes (common to reference wet flatwoods), all of which suggests that sole reliance on the existing seed banks is a poor method for restoration of these habitats. The seed bank a nalysis did reveal several species common to wet flatwoods that were not found in the vegetation surveys. Such dissimilarity between seed banks and extant vegetation occurs in many ecosystems worldwide (Thompson and Grime, 1979). x (0 1 range) is a common tool used to analyze the similarity review of seed bank studies concluded that wetland communities had an average value of 0.47 similarity between t he seed bank and vegetation. The seed bank similarity to extant vegetation in our study ranged from 0.25 to 0.48. The highest values of similarity were found between the seed banks and vegetation found in our last sampling of May
63 2011. This suggests that t he disk treatment successfully promoted seedling emergence from seed bank. However, the trend of increasing similarity in this relatively short term study does not necessarily imply that the system is closer to being restored. The extant vegetation in t his case is dissimilar from reference wet flatwoods, and therefore the similarity relationships show seed bank expression potential, but not necessarily restoration to intact wet flatwoods communities. Seedling e mergence of woody species was low consid ering that the site has been dominated by woody trees and shrubs for the past several decades. There was only one Cliftonia monophylla individual and twenty Lyonia lucida individuals the tw o dominant woody species at these locations The absence of a woo dy seedling presence could be a result of the limitations of the seed ling emergence technique (Brown 1992). There also could be a connection with the timing of the soil sampling in summer. Collections made in the late spring would have been affected by a n atural cold stratification prior to entering the greenhouse, which could have stimulated germination. In addition, many species did not flower during the experiment, making it difficult to identify them beyond the level of genus. Starting such experiments in early spring may allow more time for species to flower under natural light conditions and perhaps provide a greater distinction among species of genera such as Carex, Cyperus, Rhynchospora, Dichanthelium, and Fimbrystylis Restoration Implications Com mon management practices that include a combination of burning, clearing woody vegetation, and herbicide treatments are insufficient to restore native
64 groundcover in severely degraded pinelands that have been fire suppressed for decades. For example, thirt y year old unmanaged pine stands in South Carolina still had hardwood resprouting issues after twenty years of prescribed burning, and the only prescribed burn treatment that controlled resprouting was annual summer burns (Lewis and Harshbarger, 1976). Bur ning and chopping was ineffective at removing woody species in north Florida flatwoods (Lewis et al. 1988). On many sites, the ability to burn every summer is difficult due to rapid resprouting of shrubs, lack of herbaceous cover to carry fire, economic fe asibility, and drought leading to bans on burning. In addition to showing that simply clearing and burning was not effective in controlling titi and other woody shrubs, our study revealed that herbicide may not be effective in controlling that vegetation, even after several applications. Herbicide can temporarily reduce woody material, but it did not appear to control resprouting sufficiently to allow herbaceous species to germinate and spread. This study suggests that a disking treatment may help solve se veral problems encountered when trying to restore wet flatwoods. The disking treatment significantly reduced woody resprouting and increased herbaceous cover. Due to the relatively short duration of the study (relative to multi year vegetation succession s cales), we could not determine whether the woody reduction was permanent, but results did suggest that disking will at least control resprouting in the short term. Even if the control of woody vegetation and increase in herbaceous cover is of limited durat ion, this could allow for prescribed burns to be executed more frequently. The ability to burn annually would reduce the need for constant herbicide application to control woody resprouting.
65 Controlling resprouting by a onetime disking treatment could be significantly more economical than spraying herbicide multiple times over several years. The disking treatment also appeared to enhance germination from the seed bank and increase species richness Recruitment from the seed bank was not sufficient to comp letely restore the native groundcover community, yet the method can be a valuable tool to facilitate restoration efforts. The lessons learned from our results could be easily applied to management practices by simply disking an area after clearing and burn ing a site, in order to determine to the viability, composition, and emergence potential of the existing seed bank. This would provide restoration practitioners with early knowledge of the potential species composition, and thus information on what species may need to be planted at the site on a later date. With this information, the lengthy and labor intensive process of the greenhouse seedling emergence method could be avoided. Vigor of resprouting could also be tested by an early disking treatment that w ould help determine the need for herbicide applications to treat resprouting. It is likely that seeding or transplanting will be required to successfully restore the groundcover community to reference conditions. Further research is needed into economica lly feasible methods for transplanting or seeding. Direct seeding of wet flatwoods sites is unproven at this time, seed purchases are very expensive, and many seeds are n ot available for local ecotypes. Purchasing large quantities of containerized seedling s to plant across large sites is likewise very expensive. Further research into the establishment of seed collection areas, estimates of seed dispersal rates and
66 distances for individual species and effective protocols for direct seeding are all subjects t hat could aid in successful restoration of wet flatwoods in the future. Conclusion The restoration of degraded wet flatwoods communities can be a complex and difficult process that may have more obstacles than restoration of other pineland communities. R esprouting of aggressive woody wetland species has limited the restoration success in these habitats. Removal of the undesir able woody vegetation via Gyro T rac mulching, prescribed burning, and multiple herbicide treatments had mediocre success in preventi ng woody vegetation from resprouting. Restoration of the herbaceous understory was inhibited by the presence of woody cover, accumulation of organic matter, lack of suitable germination substrates, and an insufficient or suppressed seed bank. The additio n of a soil disturbance via disking appears to be an important aid to overcome many of these obstacles. Disking significantly reduced resprouting of woody vegetation, increased herbaceous vegetation, helps break down organic litter, and has the potential t o express, or enhance germination from, the seed bank. However, after decades of habitat degradation caused by fire suppression, the limited diversity of the seed bank was insufficient to restore the native groundcover plant community to reference conditio ns. Therefore, the seed bank should be examined early on in the restoration process, as understanding seed bank composition can guide managers tasked with restoration of these important habitats
67 APPENDIX A SPECIES PERCENT COVER OF EXTANT VEGETATION BY T REATMENT Appendix A 1. Species Cleared June Cleared November Cleared May Disked June Disked November Disked May Thicket November Thicket May Groundcover Andropogon spp. 0.9 5.15 4.93 0.78 1.07 3.27 0 0 Bidens mitis 0 0 0 0 0 0.25 0 0 Centella asiatica 0.17 0.08 0.05 0 0 0.033 0 0 Clethra alnifolia 0 0 0 0.27 0 0 0.5 0.33 Cliftonia monophylla 11.14 18.38 26.7 8.72 0.67 1.28 5.58 6.5 Cyrilla racemiflora 2.75 2.75 4.5 1.95 0.25 0.116 3.33 2.92 Dichanthelium spp. 0.41 1.33 1.1 0.75 0.73 4.22 0 0 Drosera capillaris 0 0 0 0 0 0.17 0.03 0 Eriocaulon spp. 0 0 0 0 0 0 0.08 0 Eupatorium capilliolium 0.22 0.08 0 0.33 0 0.1667 0 0 Gaylussacia moseri 0.66 1.25 1.5 0.58 0 0 0.37 5.8 Gordonia lasianthus 0 0 0 0 0 0 0.33 0.333 Hypericum sp. 0 0 0 0 0.03 0.73 0 0 Ilex coriacea 1.1 0.62 0.98 0.35 0.08 0.05 1 0.917 Ilex myrtifolia 0 0 0 0 0 0 0.51 0.5 Lachnanthes caroliana 3.19 1.92 1.5 1.1 0.77 5.3 0.13 0 Lyonia lucida 1.63 3.77 6.08 1.75 0.47 1.17 6.7 9.03 Magnolia virginiana 0 0 0 0 0 0 0.08 0.116 Morella heterophylla 0 0 0 0 0.02 0 0.25 0.5 Persea palustris 0 0 0 0 0 0.08 4 4 Pieris phillyreifolia 0 0 0 0 0 0 4.73 4.91
68 Appendix A 1 cont. Species Cleared June Cleared November Cleared May Disked June Disked November Disked May Thicket November Thicket May Pinus elliottii 0 0 0.167 0 0 0.417 0 0 Photinia pyrifolia 0.03 0 0 0.03 0 0 0.27 0.25 Polygala cruciata 0 0 0 0 0 0 0.17 0 Rhynchospora sp. 4.02 8.05 2.95 2.73 6.93 12 0.33 0 Rhexia spp. 0 0 0.05 0 0. 03 0.6 0 0 Rubiaceae 0 0 0 0 0 0.267 0 0 Rubus sp. 0.03 0 0 0.05 0 0 0 0 Smilax laurifolia 0.14 0.38 0.133 0.05 0.12 0.033 0 0 Syngonanthus flavidu s 0 0 0 0 0 0.1667 0 0 Vaccinium corymboum 0 0 0.05 0.17 0.02 0.021 1.58 1.416 Vitus rotundifolia 0 0 0 0.03 0 0 0 0 Xyris spp. 0 0 0 0 0 0.77 0.05 0 Canopy Cliftonia monophylla 0 0 0 0 0 0 88.25 89.75 Cyrilla racemiflora 0 0 0 0 0 0 9.23 8.12 Ilex myrtifolia 0 0 0 0 0 0 2.7 2.13 Total Herbaceous 9.03 16.65 10.58 5.12 9.74 27.05 0. 79 0 Total woody 17.48 27.18 40.11 14.28 1.66 4.05 128.71 135.5 Litter cover 73.49 55.85 49.7 80.6 88.6 67.8 70.5 67.5 Complete species list and mean cover values of extant vegetation at all dates and treatments. Thicket shrub thickets with no modif ication; Cleared above ground woody vegetation removal via GyroTrac + prescribed fire + herbicide applied to woody resprouts; and Disked same as Cleared, but followed by soil disking.
69 APPENDIX B SPECIES PERCENT COVER OF EXTANT VEGE TATION BY MICROREL IEF Appendix B 1. Species High June Cl. + Ds. Low June Cl.+ Ds High Nov. Thicket Low Nov. Thicket High May Thicket Low May Thicket High May Cleared Low May Cleared High May Disked Low May Disked Groundcover Andropogon spp 1.05 0.78 0 0 0 0 4.5 5.37 4 2.53 Bidens mitis 0 0 0 0 0 0 0 0 0.5 0 Centella asiatica 0 0.17 0 0 0 0 0 0.1 0.07 0 Clethra alnifolia 0.27 0 0.33 0.67 0.33 0.33 0 0 0 0 Cliftonia monoph. 12.73 6.93 8.5 2.67 8.67 4.33 28.6 24.83 1.17 1.4 Cyrilla racemiflo. 3 .55 1.1 2.83 3.83 3 3 5.67 3.33 0.233 0 Dichanthelium sp. 0.7 0.45 0 0 0 0 1.67 1.03 3.8 4.63 Drosera capillaris 0 0 0.07 0 0 0 0 0 0.17 0.17 Eriocaulon sp. 0 0 0 0.17 0 0 0 0 0 0 Eupatorium capill ifolium 0.55 0 0 0 0 0 0 0 0.33 0 Gaylussacia moseri 0 .87 0.37 0.33 0.4 0.5 0.67 2.17 0.83 0 0 Gordonia lasianus 0 0 0 0.67 0 0.67 0 0 0 0 Hedyotis sp. 0 0 0 0 0 0 0 0 .13 .4 Hypericum sp. 0 0 0 0 0 0 0 0 0.73 0.73 Ilex coriacea 0.82 0.62 0.33 1.67 0.33 1.5 1.7 0.267 0 0.1 Ilex myrtifolia 0 0 1 0 1 0 0 0 0 0 Lachnanthes car 0.08 4.15 0.1 0.17 0 0 0.33 2.67 2.93 7.7 Lyonia lucida 2.45 1.07 7 6.4 8.9 9.17 7.23 4.93 1.57 0.77 Magnolia virginana 0 0 0 0.17 0 0.233 0 0 0 0
70 Appendix B 1 cont. Species High June Cl. + Ds. Low June Cl.+ Ds Hi gh Nov. Thicket Low Nov. Thicket High May Thicket Low May Thicket High May Cleared Low May Cleared High May Disked Low May Disked Morella heterophyla. 0 0 0.33 0.17 0.5 0.5 0 0 0 0 Persea palustris 0 0 7.33 0.67 7.17 0.83 0 0 0.1 0.067 Pieris phillyre if. 0 0 6.47 3 6.73 3.1 0 0 0 0 Pinus elliottii 0 0 0 0 0 0 0 0.33 0.13 0.7 Photinia pyrifolia 0.03 0.03 0 0.53 0 0.5 0 0 0 0 Polygala cruciata 0 0 0.33 0 0 0 0 0 0 0 Rhynchospora sp 1.22 5.47 0.33 0.33 0 0 0.77 5.13 8.37 15.37 Rhexia spp. 0 0 0 0 0 0 0.1 0 0.67 0.53 Rubiaceae 0 0 0 0 0 0 0 0 0.13 0.4 Rubus sp. 0 0.83 0 0 0 0 0 0 0 0 Smilax laurifolia 0.08 0.1 0 0 0 0 0 0.267 0 0.067 Syngonanthus fl. 0 0 0 0 0 0 0 0 0 0.33 Vaccinium corymbosum 0.17 0 1.17 2 1.5 1.33 0.1 0 0.43 0 Vitus rotundifoli a 0.03 0 0 0 0 0 0 0 0 0 Xyris spp. 0 0 0 0.1 0 0 0 0 0.43 1.1 Canopy Cliftonia monop. 0 0 89.13 87.66 0 0 0 0 0 0 Cyrilla racemif. 0 0 9.55 8.88 0 0 0 0 0 0 Ilex myrtifolia 0 0 2.59 2.85 0 0 0 0 0 0 Total H erbaceous 3.42 11.44 0.68 0.65 0 0 7.37 14.3 21.07 32.76 Total woody 20.98 10.26 135.62 122.95 138.63 126.16 45.47 34.8 4.36 3.83 Litter cover 75.6 78.3 63.7 76.4 61.5 73.3 48.5 50.9 73.6 62 Complete species list and mean cover values of extant vegeta tion at all dates and treatments, separated by microrelief. Thicket shrub thickets with no modification; Cleared above ground woody vegetation removal via GyroTrac + prescribed fire + herbicide applied to woody resprouts; and Disked same as Cleared, but followed by soil disking.
71 LIST OF REFERENCES Ainslie, W. B. 2002. Forested Wetlands. The Southern Forest Resource Assessment. Southern Research Station, USDA Forest Service, Asheville, NC, pp. 479 99, Chapter 20. Allen, J.A. 1997. Reforestation of bottomland hardwoods and the issue of woody species d iversity. Restoration Ecology 5 : 125 134. Aronson M., S. Galatowitsch 2008. Long term vegetation development of restored prairie potholes w etlands. Wetlands 28 : 883 895 Bellingham, P.J. 2000. Resprouti ng as a life history strategy in woody plant communities. Oikos 82 : 409 416. Beisner B.E., D.T. Haydo n, and K. Cuddington. 2003. Alternative stable states in ecology. Frontiers in Ecology and the Environment 1 : 376 382. Bissett, N., and R. Garren. 2005. F loristic quality a ssessment. Wildlife Resources and Ecological Restoration: A Joint Meeting of the Coastal Plain Chapter of the Society for Ecological Restoration and the Florida Chapter of the Wildlife Society. Brockway, D.G. K.W. Outcalt. 2000. Restori ng longleaf pine wiregrass ecosystems: hexazinone application enha nces effects of prescribed fire. For est Ecology M anagement 103 : 159 175 Brockway D.G., G.L. Wolters H.A. Pearson R.E. Thill C.V. Baldwin and A. Martin 1998. Understory plant response t o site preparation and fertilization of loblolly and shortleaf pine f orests Journal of Range Management 51 : 47 54 Brown, D. 1992. Estimating the composition of a forest seed bank: a comparison of the seed extraction and seedling emergence methods. Canad ian Journal of Botany 70 :1603 1612. Clark, K. L., H. L. Gholz, and M. S. Castro. 2004. Carbon dynamics along a chronosequence of slash pine plantations in North Florida. Ecological Applications 14 :1154 11 71. Clark, K. L., H. L. Gholz, J. B. Moncrieff, F. Croppley, and H. W. Loescher. 1999. Environmental controls over net exchanges of carbon dioxide from contrasting Florida ecosystems. Ecological Applications 9 : 936 9 48. Clayton, D. 2010. Personal Communication. Restoration Ecologist. Northwest Florida Wat er Management District.
72 Coffey, K. L., and L. K. Kirkman. 2004. Native ground cover partners: reestablishing native ground cover in longleaf pine savannas. Ecological Restoration 22 : 288 289. Coffey, K.L. and K.L. Ki rkman 2006. Seed germination strate gies of s pecies with restoration potential in a fire maintained pine s avanna. Natural Areas Journal 26 :289 299. Cohen, S., R. Braham, and F. Sanchez. 2004. Seed bank viability in disturbed longleaf pine sites. Restor ation Ecology 12 : 503 515 Cox, A.C., D.R. Gordon, J.L. Slapcinsky, and G.S. Seamon. 2004. Understory restoration in longleaf pine sandhills. Natural Areas Journal 24 :4 14. Drewa, P.B., W.J. Platt and B.E. Moser 2002 a Fire effects on resprouting of shrubs in h eadwaters of southeastern long leaf pine s avannas Ecology 83 : 755 767 Drewa, P.B., W.J. Platt and E.B. Mo ser 2002b Community structure along elevation gradients in headwater regions of longleaf pine savannas. Plant Ecology 160 : 61 78. Ferguson, P.J 1998. Prescribed fire on the Apa lachicola Ranger District: the shift from dormant s eason to growing season and effects on wildfire suppression. Pages 120 126 in Teresa L. Pruden and Leonard A. Br ennan (eds.). Fire in ecosystem management: shifting the paradigm from suppression to prescri ption. Tall Timbers Fire Ecology Conference Procccding, No. 20. Tall Timbers Research Station, Tallahassee, FL. Florida Department of Environmental Protection. 2005. Northwest Florida Water Management District. Mitigation Bank Permit. Sand Hill Lakes Miti gation Bank. Florida Natural Areas Inventory. 2009. Reference natural community. Wet Flatwoods: Apalachicola National Forest. Florida Natural Areas Inventory (FNAI). 2010. Guide to the natural communities of Florida: 2010 edition. Florida Natural Areas I nventory, Tallahassee, FL. Folkerts, G.W. 1982. The Gulf Coast pitcher plant bogs. American Scientist 70 :260 267. Frederic ksen, T.S., H.L Allen and T.R. Wentworth. 1991. Competing v egetation and pine growth response to silvicultural t reatme nts in a six ye ar old piedmont loblolly pine p lantation Southern Journal of Applied Forestry 15 : 138 144 Fritzsch, Katrin. 2004. Plant responses to changes in disturbance magnitude. PhD thesis. Carl on Ossietzky University Oldengurg: 220 pp.
73 Frost, C.C. 1998. Presettle ment fire frequency regimes of the United States: a first approximation. Pages 70 81. Fire in Ecosystem Management: Shifting the Paradigm from Suppression to Prescription Tall Timbers Fire Ecology Conference Proceedings, No. 20. Tall Timbers Research Stati on, Tallahassee, Florida. Ga latowitsch, S. M., and A. G. van der Valk. 1996. The Vegetation of restored and natural prairie w etlands. Ecological Applications 6 :102 112 Galianto, M.I., and A.G. v an der Valk 1986. Seed g ermination traits of annuals and eme rgents during drawdowns in a Delta Marsh, Manitoba, Canada. Aquatic Botany 26 : 89 102 Gunderson, L.H. 2000. Ecological resilience in theory and application. An nual Review Ecology Systematics 32 : 765 781 Heslinga, J.L., and R.E.Grese 2010. Assessing plant community changes over sixteen years of restoration in a remnant Michigan tallgrass prairie. The Amer ican Midland Naturalist 16 : 322 336. Hess, C.A and A. Laniray 2008. Landscape changes on the Apalachicola National Forest, a fire dependent ecosystem Pro ceedings from the Ecological Society of America annual meeting 2008. Hopfensperger, K. N. 2007. A review of similarity between seed bank and standing vegetatio n across ecosystems. Oikos 116 : 1438 1448. Jack S.B., J.L. Gagnon and M.A. Melvin 2005. Chem ical control of woody sprouts following mechanical removal of large hardwoods within a longleaf pine matrix. Proceedings of the Fifth Longleaf Alliance Regional Conferenc e Longleaf Alliance Report No. 8 Kindell, C 1997. Historic distribution of wet s ava nnas in Tate's Hell State Forest Final Report for the U. S. Fish and Wildlife Service and Northwest Florida Water Management District Florida Natural Areas Inventory. Kirkman, L.K., and R.R. Sharitz 1994. Vegetation disturbance and maintenance of diver sity in intermittently flooded Carolina bays in South Carolina. Ecology Applications 4 : 177 188. Landers, J.L. 1991. Disturbance influences on pine traits in the southeastern United States. In: Proceedings, Tall Timbers f ire ecology conference 17. Tall Tim bers, Tallahassee, FL. Lewis, C. E., and T. J. Harshbarger. 1976. Shrub and herbaceous vegetation after 20 years of prescribed burning in the South Carolina Coastal Plain. Journal of Range Management 29 : 13 18.
74 Lewis, C.E., G.W. Tanner, and S.W. Terry. 198 8. Plant responses to pine m anag ement and deferred rotation grazing in n orth Florida Journal of Range Management 41 : 460 465 Lockaby, B. G., J.A. Stanturf, and M.G. Messina. 1997a. Effects o f silvicultural activity on ecological processes in floodplain forests of the southern United States: a review of existing reports Forest Ecology and Management 90 : 93 100 Lu, J G. Sun S.G. Mcnulty and N.B. Comefors 2009. Sensitivity of pine flatwoods hydrology to climate change and f orest m anagement in Florida USA. Wetlands 29 : 826 836. Maliakal, E.S., E.S. Menges and J.S. Denslow 2000. Community composition and r e generation of Lake Wales Ridge wiregrass f la twoods in r e lation to time since fire. Journal of the Torrey Botanical Society 127 :125 138 McC une B., and M. J. Mefford. 2011 PC ORD. Multivariate analysis of ecological d ata. Version 6. MjM Software, Gleneden Beach, Oregon, U.S.A. Middleton B. 1999 Hydrochory, seed banks, and regeneration dynamics along the landscape boundaries of a forested w etl and. Plant Ecology 146 : 169 184. Middleton B. 2003 Soil seed banks and the p otentia l restoration of f orested wetlands after f arming. Journal of Applied Ecology 40 : 1025 1034. Mitc hell, D 2005. Fuel reduction treatments with a gyrotrac GT 25. Fores t Opera tions Review 7 : 21 22 Northwest Florida Water Manag ement District. 2010. Sand Hill Lakes Mitigation Bank Fifth Annual Report. http://www.nwfwmdwetlands.com/MonitoringReports.php accesse d September 27, 2011. Nilsson U. and H.E. Allen 2003. Short and long term effects of site preparation, fertilization and vegetation control on growth and stand development of planted loblolly pine f orest Ecology and Manag eme nt 175 : 367 377 Peck, J.E. 2010. Multivariate analysis for community ecologist: step by step using PC ORD. MjM Software Desgin, Gleneden Beach, OR. Peet, R.K. and D.J. Allard. 1995. L ongleaf pine vegetation of the southern Atlantic and e astern Gulf Coast regions: a preliminary clas sification. Proceedings of the Tall Timbers Fire Ecology Conference 18 : 45 81. Platt, W. J., G. W. Evans, and M. M. Davis. 1988. Effects of fire season on flowering of forbs and shrubs in lon gleaf pine forests. Oecologia 76 : 353 363
75 Poiani, K.A., and W.C. Johnson 1988. Evaluation of the emergence method in estimating seed bank composition of prairie wetlands. A quatic Botany 32 : 91 97. Polasek, L.G., M.W. Weller and K.C. Je nsen 1995. Management of shallow impoundments to provide e mergent and submergent veg etation for w aterfowl. Army Corp of Engineers. Wetlands Research Pro gram Technical Report WRP SM 8. Rheinhardt, R. D., M.C. Rheinhardt, and M.M. Brinson. 2002 Regional guidebook for applying the hydrogeomorphic approach to wet pine flats on mineral soils in the Atlantic and Gulf Coastal Plains. P.220. Technical report WRP DE Vicksburg, MS: United States Army Corps of Engineers, Waterways Experiment Station. Carolina, USA. Simpson, R. L., M. A. Leck, and V. T. Parker. 1989. Seed banks: general concepts and methodological issues. Pages 3 8 in R. L. Simpson, M. A. Leck, and V. T. Parker, editors. Ecology of soil seed banks. Academic Press, San Diego, California Sm ith S ., P. McCormick J. Leeds and P. Garrett. 2002 Constraints of seed bank s peci es composition and water depth for restoring v egetation in the Florida Everglades, U.S.A. Restoration Ecology 10 : 138 145. Suding, K.N., K.L Gross and G.R. Houseman 2004. Alte rnate states and positive feedback s in restoration ecology. Trends in Eco logy and Evolution 19 : 46 53 Sun, G., S.G. McNulty DM. Amatya R.W. Skaggs, L. W. Swift, J. P. Shepard, and H. Riekerk. 2002. A comparison of the hydrology of the coastal forested w etlands/pine flatwoods and the mountainous uplands in the southern US. Journal of Hydrology 263 : 92 104 Thompson, K., and J. P. Grime. 1979. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology 67 : 893 921. Walker, J., and R. K. Peet. 1983. Composition and species diversity of pine wiregrass savanna of the Green Swam p, North Carolina. Vegetatio 55 : 163 179 Ware, S. A., C. Fr ost and P.D. Doerr. 1993 the former longleaf Biodiversity of the southeast: Lowland terrestrial communities W. H. Martin, S. G. Boyce, and A. C. Echternacht, ed., John Wiley and Sons, New York, 447 93. Watts A. G. Tanner and R. Dye 2006. Restoration of dry prairie using fire and r oller chopping. Land of Fire and Water: The Florida Dry Prairie Ecosystem. Proceedings of the Florida Dry Prairie Conference
76 Welch, J.R., K.V. Miller, W.E. Palmer, and T.B. Harrington. 2004. Response of understory vegetation important to the northern bob white following imazapyr and mechanical treatment s. Wildlife Society Bulletin 32 : 1071 1076. Wetzal P., A. Valk and L. Toth. 2001 Restoration on wetland v eg etation on the Kissimmee River floodplain: Potential role of seed b anks. Wetlands 21 : 189 198. U.S. Department of Agriculture 1949. Aerial photographs of Washington County flight 2F. http://ufdc.ufl.edu /UF00071792/00009 U.S. Department of Agriculture 1953. Aeri al photographs of Bay County f light 6L h ttp://ufdc.ufl.edu/UF00071792/00003 U.S. Department of Agriculture. 1995. Soil Survey of Okaloosa County, Florida. Soil Conservation Service. U.S. Department o f Agriculture. 1999. http://w ww2.ftw.nrcs.usda.gov/osd/dat/P/PANTEGO.html U.S. Department o f Agriculture. 2004. http://www2.ftw.nrcs.usda.gov/osd/dat/C/CLARA.html van der Valk, A. G., and R. L. Pederson. 1989. Se ed banks and the management and restorati on of natural vegetation. 329 346 in R. L. Simpson, M. A. Leck, and V. T. Parker, editors. Ecology of soil seed banks. Academic Press, San Diego, California.
77 BIOGRAPHICAL SKETCH David K. Mitchell was born in 1977 in Pensacola, Florida. He received his B.S. in Horticulture, with a minor in Wildlife Ecology from the University of Florida in 2008. He went on to earn his M.S. in Soil and Water Science from the University of Florida i n 2011. David worked for himself in different fields before entering the horticulture field in 2005. In 2008, he was hired as Nursery Manager of the Ecosystem Restoration Section with the Florida Department of Environmental Protection where he managed v egetative restoration efforts throughout the Florida panhandle until 2011. In September of 2011, David left Florida to work with various international ecological restoration efforts.