Citation
From Pinus elliottii Plantation to Pinus palustris Ecosystem

Material Information

Title:
From Pinus elliottii Plantation to Pinus palustris Ecosystem The Role of Herbicide in Longleaf Pine Restoration
Creator:
Freeman, Johanna E.
Place of Publication:
[Gainesville, Fla.]
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (85 p.)

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Interdisciplinary Ecology
Committee Chair:
Jose, Shibu
Committee Members:
Miller, Deborah L.
Long, Alan J.
Graduation Date:
5/1/2008

Subjects

Subjects / Keywords:
Interdisciplinary Ecology -- Dissertations, Academic -- UF
diversity, fire, flatwoods, herbicide, hexazinone, imazapyr, longleaf, palustris, pine, pinus, restoration, sulfometuron
Herbicides ( jstor )
Shrubs ( jstor )
Seedlings ( jstor )
Genre:
Electronic Thesis or Dissertation
bibliography ( marcgt )
theses ( marcgt )
Interdisciplinary Ecology thesis, M.S.

Notes

Abstract:
Under a natural fire regime, the longleaf pine (Pinus palustris Mill.) ecosystem is characterized by an open, longleaf pine-dominated canopy and a diverse, grass-dominated understory. In the years since European colonization, the vast majority of longleaf pine acreage has been cut over, fire-suppressed, and fragmented, which has resulted in extensive understory invasion by shrubs and a concurrent decline in herbaceous species diversity. One of the biggest challenges to successful restoration of this ecosystem is the persistence of shrubs in the understory, which suppress longleaf pine seedlings as well as native herbaceous plants. Herbicide can be used as a supplement to fire in order to enhance shrub control, but must be studied carefully because of the potential for negative impacts on native plants. Information is lacking about the effects of herbicide on natural longleaf pine flatwoods communities. We used a banded application of three herbicides and one tank mix as shrub control treatments following complete removal of the slash pine canopy and replanting with containerized longleaf pine seedlings in a mesic-wet flatwoods. The herbicides tested were Arsenal? (imazapyr), Oust? (sulfometuron methyl), Velpar L? (hexazinone), and an Oust? + Velpar L? tank mix. Four years after application, no negative impacts on understory species richness, diversity, evenness, or community composition were evident in any of the herbicide treatments. Oaks (Quercus spp.), one of the dominant shrub genera on the study site, were resistant to sulfometuron methyl, and this herbicide was therefore ineffective both as a pine release treatment and for enhancing herbaceous species growth. Imazapyr was the most effective treatment overall, significantly improving longleaf pine seedling growth and also enhancing herbaceous species cover. Both hexazinone and the hexazinone + sulfometuron methyl tank mix provided some seedling growth and understory enhancement as well, though these effects were not as pronounced as those in the imazapyr treatment. Shrubs resprouted aggressively after a dormant-season prescribed fire in the fifth year after treatment, indicating that herbicide-related increases in cover of wiregrass and other herbaceous species may be lost in future fire cycles. Retention of overstory pines for shrub competition and as a source of fine fuels is recommended in addition to herbicide for sustainable silviculture in longleaf pine flatwoods. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2008.
Local:
Adviser: Jose, Shibu.
Statement of Responsibility:
by Johanna E. Freeman

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright by Johanna E. Freeman. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/11/2008
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:

freeman_j.pdf

freeman_j_Page_28.txt

freeman_j_Page_08.txt

freeman_j_Page_10.txt

freeman_j_Page_55.txt

freeman_j_Page_75.txt

freeman_j_Page_34.txt

freeman_j_Page_54.txt

freeman_j_Page_22.txt

freeman_j_Page_59.txt

freeman_j_Page_40.txt

freeman_j_Page_48.txt

freeman_j_Page_70.txt

freeman_j_Page_17.txt

freeman_j_Page_05.txt

freeman_j_Page_49.txt

freeman_j_Page_50.txt

freeman_j_Page_38.txt

freeman_j_Page_73.txt

freeman_j_Page_41.txt

freeman_j_Page_78.txt

freeman_j_Page_68.txt

freeman_j_Page_39.txt

freeman_j_pdf.txt

freeman_j_Page_25.txt

freeman_j_Page_06.txt

freeman_j_Page_18.txt

freeman_j_Page_03.txt

freeman_j_Page_80.txt

freeman_j_Page_12.txt

freeman_j_Page_71.txt

freeman_j_Page_07.txt

freeman_j_Page_74.txt

freeman_j_Page_60.txt

freeman_j_Page_53.txt

freeman_j_Page_58.txt

freeman_j_Page_33.txt

freeman_j_Page_29.txt

freeman_j_Page_72.txt

freeman_j_Page_16.txt

freeman_j_Page_23.txt

freeman_j_Page_32.txt

freeman_j_Page_24.txt

freeman_j_Page_01.txt

freeman_j_Page_45.txt

freeman_j_Page_04.txt

freeman_j_Page_79.txt

freeman_j_Page_15.txt

freeman_j_Page_64.txt

freeman_j_Page_65.txt

freeman_j_Page_36.txt

freeman_j_Page_02.txt

freeman_j_Page_69.txt

freeman_j_Page_43.txt

freeman_j_Page_13.txt

freeman_j_Page_81.txt

freeman_j_Page_47.txt

freeman_j_Page_30.txt

freeman_j_Page_21.txt

freeman_j_Page_11.txt

freeman_j_Page_14.txt

freeman_j_Page_76.txt

freeman_j_Page_56.txt

freeman_j_Page_51.txt

freeman_j_Page_85.txt

freeman_j_Page_66.txt

freeman_j_Page_62.txt

freeman_j_Page_82.txt

freeman_j_Page_20.txt

freeman_j_Page_63.txt

freeman_j_Page_52.txt

freeman_j_Page_77.txt

freeman_j_Page_42.txt

freeman_j_Page_61.txt

freeman_j_Page_46.txt

freeman_j_Page_19.txt

freeman_j_Page_67.txt

freeman_j_Page_37.txt

freeman_j_Page_31.txt

freeman_j_Page_57.txt

freeman_j_Page_09.txt

freeman_j_Page_84.txt

freeman_j_Page_83.txt

freeman_j_Page_35.txt

freeman_j_Page_44.txt

freeman_j_Page_26.txt

freeman_j_Page_27.txt


Full Text





adequate control, eventually leads to the exclusion of herbaceous plants (Gilliam and Platt 1999,

Glitzenstein et al. 2003). The release of shrubs following overstory harvest is likely to be more

extensive on a site with a past history of fire suppression, due to the preponderance of

underground hardwood rootstocks established during periods without fire (Olson and Platt 1995;

Drewa et al. 2006). Shrub invasion excludes herbaceous species not only through competition

for light and resources, but through the alteration of fire regimes away from the high-frequency

fire intervals perpetuated by wiregrass and pine needles. Hardwoods produce litter of lower

flammability than pines or wiregrass, which can lead to longer fire return intervals in gaps,

ultimately exerting a negative feedback on the system's pyrogenicity and perpetuating a

hardwood midstory (Williamson and Black 1981, Rebertus et al. 1993, Jacqmain et al. 1999).

When a large portion of the overstory is removed, this negative feedback cycle is likely to be

exacerbated by the subsequent lack of needlefall and the potential loss of wiregrass due to

harvest-related soil disturbance. These are the two fine fuel inputs essential to maintaining a

frequent fire return interval in a longleaf pine ecosystem, and without them the keystone species

in the system will not regenerate (Williamson and Black 1981, Clewell 1989).

The ability of longleaf pine seedlings to survive and emerge from the grass stage will also

be impacted if there are not adequate fine fuels to carry fire through the system and keep

hardwoods suppressed. Longleaf seedlings are highly intolerant of competition and will not

initiate height growth if an excessive amount of competing vegetation is present (Boyer 1963,

Haywood 2000, Ramsey et al. 2003).

One approach to these problems, which has received increasing attention in recent years,

is to develop an uneven-aged silvicultural system that replicates natural smaller-scale patterns of

disturbance and succession so as not to disrupt the overall continuity of fine fuels and understory









harvest is required, 2) lower management costs, 3) lower personnel skill level and inventory

information needs, 4) higher short-term timber output, and 5) better equipment access and

maneuverability for effective site preparation in an area with severe understory competition, such

as a saw palmetto-dominated flatwoods (Palik et al. 2002, Brockway et al. 2006). If desired,

subsequent management of the new longleaf pine stand can follow a single-tree or group-

selection prescription in order to perpetuate the structural and functional diversity of the system

over the long-term.

The biggest challenges facing this method will be 1) to achieve control of competing

shrubs without negatively impacting native understory vegetation, 2) to maintain or enhance

wiregrass cover, and 3) to successfully apply fire to the system in spite of the loss of overstory

fine fuel inputs. The studies described above indicate that supplemental hardwood control in the

form of herbicide or mechanical treatment will almost always be necessary after clearcutting a

pine overstory, because of the release of hardwoods and the lack of fine fuels for prescribed fire.

Even in a group- or single-tree selection system, additional hardwood control may be necessary

if pyrogenicity is already compromised due to a past history of fire suppression and an

entrenched shrub component. In the single-tree selection study described above, the herbicide

treatment was the only one in which hardwood stem density did not increase over the course of

the study (Kirkman et al. 2007).

Following overstory harvest, longleaf pine plantings must be carefully managed in order

to achieve survival and growth rates adequate for stand replacement. A reasonable range of

survival rates for containerized plantings is 60-75%, but in order to achieve this range of survival

rates, site preparation is necessary before seedlings are planted (Brockway et al. 2006). Grasses

and other herbaceous species are the most serious short-term competitors for newly established









treatments, leaving more space for other small herbaceous species. These gains in species

richness, diversity, and evenness are therefore unlikely to persist through future fire cycles

because it will be more difficult to apply prescribed fire to this treatment. In support of this

prediction, we observed that sulfometuron was the only treatment in which species richness,

diversity, and evenness did not increase following the prescribed fire in 2007.

The long-term sustainability of any longleaf pine ecosystem restoration project depends

on the manager's ability to apply prescribed fire with enough frequency and intensity to control

shrubs. The herbicides applied in this study provided short-term shrub control and had positive

impacts on longleaf pine growth, wiregrass cover, and overall herbaceous cover. However,

complete removal of the slash pine overstory at the start of the project resulted in a lack of fine

fuels and presumably also encouraged aggressive regrowth of shrubs due to the loss of

competition from mature pines (Brockway et al. 1998, Kush et al. 1999, McGuire et al. 2001,

Kirkman et al. 2007, Pecot et al. 2007). These hindrances were most likely exacerbated by the

length of time since herbicide application (five years since first application and four years since

second application), as well as the use of dormant-season fire due to the site's proximity to a

highway and residential areas (Waldrop et al. 1992, Drewa et al. 2002, Drewa et al. 2006).

Though three of the four herbicides we tested were successful at releasing longleaf pine

seedlings, the herbicide-related gains in wiregrass and overall herbaceous cover may be lost in

the coming years due to the aggressive resprouting of shrubs. We therefore make the following

recommendation to those managing flatwoods sites for both timber and biodiversity: when

possible, some overstory pines should be retained following harvest to serve as a source of fine

fuels. Banded application of imazapyr is recommended in order to improve longleaf pine









Drewa, P.B., Thaxton, J.M., and W.J. Platt. 2006. Responses of root-crown bearing shrubs to
differences in fire regimes in Pinuspalustris (longleaf pine) savannas: Exploring old- growth
questions in second growth systems. Applied Vegetation Science 9:27-36.

Dufrene, M. & P. Legendre. 1997. Species assemblages and indicator species: the need for a
flexible asymmetrical approach. Ecological Monographs 67:345-366.

DuPontTM. 2007. Velpar L Label. http://www.dupont.com/ag/vm/products/velparl.html.
E. I. du Pont de Nemours and Company,Wilmington, DE.

DuPontTM. 2007. Oust' XP Label.
http://www.dupont.com/ag/us/prodinfo/prodsearch/information/H65144.pdf
E. I. du Pont de Nemours and Company,Wilmington, DE.

Frost, C.C. 2006. History and Future of the Longleaf Pine Ecosystem. Pages 9-47 in Jose, S.,
Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture,
and Restoration. Springer Science + Business Media, LLC., New York.

Gagnon, J.L., Jokela, E.J., Moser, W.K., and D.A. Huber. 2003 Dynamics of artificial
regeneration in gaps within a longleaf pine flatwoods ecosystem. Forest Ecology and
Management 172:133-144.

Gilliam, F.S. and W.J. Platt. 1999. Effects of long-term fire exclusion on tree species
composition and stand structure in an old-growth Pinuspalustris forest. Plant Ecology
140:15-26

Glitzenstein, J.S., Streng, D.R., and D.D. Wade. 2003. Fire frequency effects on longleaf pine
(Pinuspalustris Mill.) vegetation in South Carolina and northeast Florida, USA. Natural
Areas Journal 23(1):22-37.

Grace, S.L., and W.J. Platt. 1995. Effects of adult tree density and fire on the demography off
pregrass stage juvenile longleaf pine (Pinuspalustris Mill.) Journal of Ecology 83:75-86.

Hay-Smith, L. and G.W. Tanner. 1994. Responses of wildlife food plants to a forest herbicide.
University of Florida Cooperative Extension Fact Sheet WRS-9.

Haywood, J.D. 2000. Mulch and hexazinone herbicide shorten the time longleaf pine seedlings
are in the grass stage and increase height growth. New Forests 19:279-290.

Haywood, J.D. 2005. Effects of herbaceous and woody plant control on Pinuspalustris growth
and foliar nutrients through six growing seasons. Forest Ecology and Management 214:384-
397.

Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., Orzell, S.L., E.L. Bridges. 1993. Longleafpine
vegetation of the Southern Atlantic and Eastern Gulf Coast regions: a preliminary
classification. Proceedings of the Tall Timbers Fire Ecology Conference 18:83-104









application, this effect was no longer evident, as shrub stem density no longer varied

significantly among treatments. Following the prescribed fire in February 2007, mean shrub

stem density increased again by around 50% in every treatment (Figure 2-7a). These increases

were significant in all treatments but the control. Variation in Ericaceae stem density among

treatments approached significance in June 2006 (p = .061) (Table 2-1), but this difference was

moderated by the smaller control sample size and uneven pre-treatment distributions. Ericaceae

stem density did not change significantly in any of the treatments following fire (Table 2-1).

Shrub height

Mean shrub stem height for Quercus, Ilex, and Serenoa did not differ significantly among

treatments in June 2006 (p = 0.2341). Stem height ranged from 14.6cm in the imazapyr

treatment to 17.1cm in the hexazinone treatment (Table 2-1). Following the prescribed fire in

February 2007, mean shrub height increased in the imazapyr, sulfo, and sulfo + hexa treatments

(Table 2-1). Differences among treatments following fire were not significant (p = 0.40). Mean

stem height of Ericaceae did not differ significantly among treatments in June 2006 (p = 0.304)

or in June 2007 (p = 0.900), though height increases within some treatments following fire were

significant (Table 2-1).

Understory responses

A total of 93 species were identified during the June 2006 and June 2007 plant surveys

(Appendix A). Weedy species were not prevalent, and we found numerous plant specimens

belonging to functional groups identified by Walker and Peet (1983) and Glitzenstein et al.

(2003) as high-risk for being lost in shrub-invaded flatwoods, such as bunch grasses (Poaceae),

sedges (Cyperaceae), sundews (Droseraceae), and basal rosette species.

Overall species richness four years after herbicide application varied significantly among

treatments (p = 0.01), ranging from 7.4 species/m2 in the control treatment to 9.3 species/m2 in









evenness increased in the imazapyr, sulfo + hexa, and hexazinone treatments following fire

(Table 2-3).

Community Ordination

Analysis of the plant community four years after treatment using NMDS and MRPP

revealed only a slight relationship between treatment and species composition (Fig 2-8), with a

small chance-corrected within-group agreement (A) of 0.02 (p = 0.01) for the pre-fire survey

quadrats. Indicator Species Analysis following the method of Dufrene and LeGendre (1997)

revealed significant relationships between some species and treatments (Monte Carlo test of

significance) (Table 2-2).

Species of Special Interest

Curtiss' sandgrass (Calamovilfa curtissii Scribn.), a florida panhandle endemic, was

identified in one of the imazapyr treatment plots during both the June 2006 and June 2007 plant

surveys. Prior to treatment, there was a very low density of legumes on our study site: out of 90

survey quadrats, only two legume specimens were identified, both of which were Mimosa

quadrivalvis var. angustata Barneby (Ranasinghe, unpublished data). After four years, we found

Desmodium lineatum DC, Desmodium strictum DC, and Tephrosia hispidula Pers. in addition to

M. quadrivalvis, again at a low density: a total of eight legume specimens were identified out of

156 survey quadrats. In the post-fire survey, one D. lineatum, four M quadrivalvis, and one

Baptisia lanceolata Elliott specimen were identified. These species were not significantly

associated with any of the treatments.

Fire Temperature

The prescribed fire was cool, fast-moving, and patchy. Mean maximum fire temperature

by treatment ranged from 2520C to 3060C at 30cm and 166C to 2570C at 80cm (Table 2-4), and

did not differ significantly among herbicide treatments at either height (p = 0.42 and p = 0.55,









fuels, presence of ladder fuels that may cause damage to crowns, and duff accumulation that can

kill overstory trees when ignited (Walker and Siletti 2006). Growth of shrub rhizomes and root

systems during periods of fire suppression can be extensive, making these shrubs recalcitrant

even when fire is reintroduced (Drewa et al. 2006). Additionally, social factors such as

proximity to residential areas or highways can limit the ability of land managers to use

prescribed fire effectively. Therefore, mechanical and herbicidal hardwood removal techniques

are often considered as supplements to prescribed fire (Walker and Siletti 2006). Hardwood

control is also of key importance for restoration projects that include planting of longleaf pine

seedlings, which are very intolerant of competition and can quickly succumb to competing

vegetation while in the grass stage (Harrington 2006).

Herbicide, used in conjunction with prescribed fire, can be an effective tool for reducing

mid- and understory hardwoods in longleaf pine systems without negatively impacting native

understory species (Brockway et al 1998, Brockway and Outcalt 2000, Provencher et al. 2001).

However, most studies of herbicide as a restoration tool in longleaf pine systems have been

conducted on xeric longleaf pine sites (Litt et al. 2001). Relative to xeric longleaf pine sandhills,

mesic and wet longleaf pine flatwoods are more productive, have different soil characteristics,

and are associated with different understory plant communities. Understory responses to

herbicide in these systems are therefore likely to differ from those documented in sandhills (Litt

et al. 2001). At present, there is only limited information about herbicidal shrub control in mesic

to wet longleaf pine ecosystems.

The purpose of this research was to test the effects of different herbicide treatments, used

in conjunction with prescribed fire, on target and nontarget vegetation in a flatwoods ecosystem

undergoing conversion from a planted slash pine (Pinus elliottii Engelm.) overstory to a longleaf










*


a a
ab

^rlb


D Pre-Fire
* Post-Fire


Hexa


-o- Control
-I- Imazapyr
- Sulfo
-- Sulfo+Hexa
-- Hexa


Post-fire


Figure 2-7. a) Mean shrub stem density (stems/m2) before (June 2006) and after (June 2007) fire
for Quercus spp., I glabra, and S. repens. Pre-fire means not sharing the same letter
are significantly different. (*) indicates a significant change following fire (Tukey-
Kramer HSD, a = 0.05). b) Changes in shrub stem density within each treatment over
the course of the 5-year experiment. First-year results from Ranasinghe (2003 and
unpublished data), and Ranasinghe et al. (2005).


4f1
a


a) I
E
=
VI

I


*l

IIl

ffll


Control


Imazapyr


Sulfo
Treatment


Sulfo+Hexa


4 100
80

b) 60
40
c 20


Pre-harvest


3 Mos


9 Mos
Treatment









Kush, J.S., R.S. Meldahl, and W.D. Boyer. 1999. Understory plant community response after
23 years of hardwood control treatments in natural longleaf pine (Pinuspalustris) forests.
Canadian Journal of Forest Research 29:1047-1054.

Lauer, D.K. and G.R. Glover. 1998. Early pine response to control of herbaceous and shrub
vegetation in flatwoods. Southern Journal of Applied Forestry 15(4):201-208.

Leach, M.K. and T.J. Givnish. 1996. Ecological determinants of species loss in remnant
prairies. Science 273:1555-1558.

Litt, A.R., Herring, B.J. and L. Provencher. 2001. Herbicide effects on ground-layer vegetation
in southern pinelands, USA: a review. Natural Areas Journal 21(2):177-188.

McNease. L.L. and G.A. Hurst. 1991. Preharvest treatment with imazapyr for seed tree
regeneration and habitat improvement. Proceedings of the Southern Weed Science Society
44:265.

Mehlman, D.W. 1992. Effects of fire on plant community composition of North Florida second
growth pineland. Bulletin of the Torrey Botanical Club 119:376-383.

McCune, B. and Grace, J. 2002. Analysis of Ecological Communities. MJM Software Design.
Gleneden Beach, OR.

McGuire, J.P., Mitchell, R.J., Moser, E.B., Pecot, S.D., Gjerstad, D.H., and C.W. Hedman.
2001. Gaps in a gappy forest: plant resources, longleaf pine regeneration, and understory
response to tree removal in longleaf pine savannas. Canadian Journal of Forest Research
31:765-778.

Miller, J.H., R.S. Boyd, and M.B. Edwards. 1999. Floristic diversity, stand structure, and
composition 11 years after herbicide site preparation. Canadian Journal of Forest Research
29:1073-1083.

Mitchell, R.J., Hiers, J.K., O'Brien, J.J., Jack, S.B., and R.T. Engstrom. 2006. Silviculture that
sustains: the nexus between silviculture, frequent prescribed fire, and conservation of
biodiversity in longleaf pine forests of the southeastern United States. Canadian Journal of
Forest Research 36:2724-2736.

Neary, D.G., J.E. Smith, B.F. Swindel, and K.V. Miller. 1991. Effects of forestry herbicides on
plant species diversity. Proceedings of the Southern Weed Science Society 44:266-272.

NOAA Satellite and Information Service: National Climatic Data Center. 2007. Panama City,
FL Station 5N.
http://www4.ncdc.noaa.gov/cgiwin/wwcgi.dll?wwDI~StnSrch-StnID~20004593.

Noss, R.F. 1989. Longleaf pine and wiregrass: keystone components of an endangered
ecosystem. Natural Areas Journal 9:211-213.
















a abc ab
S50 abc

S40 bc
O Pre-Fire
U Post-Fire
30

~ 20

2010 i
0 -


Control Imazapyr Sulfo Slfo+Hexa Hexazinone
'I-eatment

Figure 2-4. Mean % herbaceous cover (including both forbs and graminoids) 4 years after
treatment (Pre-fire) and 5 years after treatment (Post-fire). Pre-fire means not sharing
the same letter are significantly different (Tukey-Kramer HSD, a = .05). (*) indicates
significant pre-fire / post-fire differences within treatments (Tukey-Kramer HSD,
a=.05).









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... .................. ............................................................................. 3

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

L IS T O F F IG U R E S ........ ................................................................................... .......... ...... 7

A B STR A C T ......... .. ................................... .......................................................................... 8

CHAPTER

1 IN T R O D U C T IO N ........ ............................................................................... ................ .. 10

2 USING HERBICIDE TO RESTORE THE UNDERSTORY FOLLOWING HARVEST .... 13

In tro d u c tio n ............................................................................................................................. 1 3
Methods .................................................................. 18
Study Area ...................................... .................... 18
Treatm ents ..................................... .............. 19
Data Collection and Analysis....................................................... 21
R e su lts ............................................................................................................... . ........... ..... 2 5
S h ru b C o v e r ........................................................................................................ 2 5
H erbaceous C over ..................................................... 25
W iregrass C over............................. .................... 26
Shrub Stem D density ........................................ 26
S h ru b h e ig h t ......... ... ..................................................................................... 2 7
Understory responses ..................................................................... .... ......... ........ ...... ........ 27
Species D diversity ........................................ 28
Species E venness ........................................ 28
C om m unity O ordination ................................29.............................
Species of Special Interest ............. ............. ................. 29
Fire Tem perature.......................... ........... .............. 29
D iscu ssio n ................................30.............................

3 HERBICIDE AND SUSTAINABLE LONGLEAF PINE SILVICULTURE ........................ 51

In tro d u c tio n ............................................................................................................................. 5 1
Methods .................................................................. 57
Study Area ...................................... .................... 57
T re a tm e n ts .................. ... .................................................................................................... 5 8
Data Collection and Analysis....................................................... 60





4









CHAPTER 2:
USING HERBICIDE TO RESTORE THE UNDERSTORY FOLLOWING HARVEST

Introduction

Longleaf pine (Pinuspalustris Mill.) ecosystems occupy a wide environmental gradient

in the southeastern United States, from xeric sandhills to mesic and wet flatwoods (Walker and

Peet 1983, Peet 2006). Under a natural fire regime, the understory vegetation associated with the

two dominant species, longleaf pine and wiregrass (Aristida strict Michx.), varies predictably

with site hydrology (Harcombe et al. 1993). Mesic longleaf pine flatwoods, which occur on

Ultisols and Spodosols, fall in the middle of the longleaf pine hydrologic gradient and are highly

productive plant communities (Peet 2006). Whereas the understories of frequently burned xeric

longleaf pine sandhills are truly dominated by grasses and herbaceous plants, in mesic flatwoods

a large shrub component is present even under natural fire conditions (Peet 2006). These shrubs

are kept low to the ground by frequent fire, but in the absence of fire they will quickly succeed

into a dense midstory, eventually suppressing both longleaf pine and herbaceous plant

regeneration (Gilliam and Platt 1999).

Wiregrass is a key structural and functional component of all longleaf pine ecosystems in

the Southern Coastal Plain ecoregion (Peet 2006), and the establishment or enhancement of

wiregrass populations is essential for successful ecological restoration in these systems. Living

and dead wiregrass leaves are highly flammable and act as fire-spreaders, moving surface fires

evenly through the understory of a longleaf pine ecosystem (Clewell 1989). Dead longleaf pine

needles are also very flammable and are a critical component of the system's pyrogenicity

(Williamson and Black 1981, Clewell 1989). Both longleaf pine and wiregrass can therefore be

considered keystone species. Together they perpetuate a frequent fire return interval, which

favors their own reproductive success, acts as a selective force against competing, less fire-









Provencher, L., Herring, B.J., Gordon, D.R., Rogers, H.L., Galley, K.E.M., Tanner, G.W.,
Hardesty, J.L., and L.A. Brennan. 2001. Effects of hardwood reduction techniques on
longleaf pine sandhill vegetation in northwest Florida. Restoration Ecology 9:13-27.

Ramsey, C.L., S. Jose, B.J. Brecke, and S. Merritt. 2003. Growth response of longleaf pine
(Pinuspalustris Mill.) seedlings to fertilization and herbaceous weed control in an old field
in southern USA. Forest Ecology and Management 172(2-3):281-289.

Ramsey, C.L., and S. Jose. 2004. Growth, survival, and physiological effects of hexazinone and
sulfometuron methyl applied overtop of longleaf pine seedlings. Southern Journal of
Applied Forestry 28(1):48-54.

Ranasinghe, S. 2003. Role of herbicides in longleaf pine flatwoods restoration: pine growth,
understory vegetation response, and fate of applied herbicides. Master's thesis. University
of Florida, Gainesville.

Ranasinghe, S., Jose, S., and C. Ramsey. 2005. Understory species dynamics following
herbicidal release treatments on a longleaf pine site in the Southern U.S.A. In New Research
on Forest Ecosystems, A.R. Burk, ed. Nova Science Publishers, Inc.

Rebertus, A.J., Williamson, G.B., and W.J. Platt. 1993. Impacts of temporal variation in fire
regime on savanna oaks and pines. Proceedings of the Tall Timbers Fire Ecology
Conference 18:215-225.

Shepard, J.P., Creighton, J., and H. Duzan. 2004. Forestry herbicides in the United States: an
overview. Wildlife Society Bulletin 32(4):1020-1027.

Varner III, J.M., and J.S. Kush. 2004. Remnant old-growth longleaf pine (pinus palustris Mill.)
savannas and forests of the Southeastern USA; status and threats. Natural Areas Journal
24:141-149.

Waldrop, T.A., White, D.L., and S.M. Jones. 1992. Fire regimes for pine-grassland
communities in the southeastern United States. Forest Ecology and Management 47:195-
210.

Walker, J., and R.K. Peet. 1983. Composition and species diversity of pine-wiregrass savannas
of the Green Swamp, North Carolina. Vegetatio 55:163-179.

Walker, J. and A.M. Silletti. 2006. Restoring the Ground Layer of LongleafPine Ecosystems.
Pages 297-333 in Jose, S., Jokela, E., and Miller, D.L., editors. The LongleafPine Ecosystem:
Ecology, Silviculture, and Restoration. Springer Science + Business Media, LLC., New York.

Welch, J.R., Miller, K.V., Palmer, W.E., and T.B. Harrington. 2004. Response of understory
vegetation important to the northern bobwhite following imazapyr and mechanical
treatments. Wildlife Society Bulletin 32(4): 1071-1076.









LIST OF REFERENCES


BASF Corporation. 2007. Arsenal' AC label. http://www.cdms.net/LDat/ld542007.pdf. BASF
Agricultural Products, Research Triangle Park, NC.

Behm, A.L., Duryea, M.L., Long, A.J., and W.C. Zipperer. 2004. Flammability of native
understory species in pine flatwood and hardwood hammock ecosystems and implications for
the wildland-urban interface. International Journal of Wildland Fire 13:355-365.

Boyd, R.S., Freeman, J.D., Miller, J.H., and M.B. Edwards. 1995. Forest herbicide influences
on floristic diversity seven years after broadcast pine release treatments in central Georgia,
U.S.A. New Forests 10:17-37.

Boyer, W.D. 1963. Development of longleaf pine seedlings under parent trees. USDA Forest
Service, Southern Forest Experiment Station, Research Paper SO-4, New Orleans, LA.

Brockway, D.G. and C.E. Lewis. 1997. Long-term effects of dormant season prescribed fire on
plant community diversity, structure, and productivity in a longleaf pine-wiregrass
ecosystem. Forest Ecology and Management 96:167-183.

Brockway, D.G. and K.W. Outcalt. 1998. Gap-phase regeneration in longleaf pine wiregrass
ecosystems. Forest Ecology and Management 106:125-139.

Brockway, D.G., Outcalt, K.W., and R.N. Wilkins. 1998. Restoring longleaf pine wiregrass
ecosystems: plant cover, diversity, and biomass following low-rate hexazinone application on
Florida sandhills. Forest Ecology and Management 103:159-175.

Brockway, D.G. and K.W. Outcalt. 2000. Restoring longleaf pine wiregrass ecosystems:
hexazinone application enhances effects of prescribed fire. Forest Ecology and Management
137:121-138

Brockway, D.G., Outcalt, K.W., and W.D. Boyer. 2006. Pages 95-133 in Jose, S., Jokela, E.,
and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and
Restoration. Springer Science + Business Media, LLC., New York.

Clewell, A.F. 1989. Natural history of wiregrass (Aristida strict Michx., Graminae). Natural
Areas Journal 9:234-245.

Croker, T.C. Jr. 1975. Seedbed preparation aids natural regeneration of longleaf pine. USDA
Forest Service Research Paper. Southern Forest Experiment Station No. SO-112.

Daubenmire, R.F. 1959. Canopy coverage method of vegetation analysis. Northwest Scientist
33:43-64.

Drewa, P.B., Platt, W.J., and E.B. Moser. 2002. Fire effects on resprouting of shrubs in
headwaters of southeastern longleaf pine savannas. Ecology 83(3):755-767.









recent studies is that small group selection and single-tree selection are recommended as

silvicultural treatments for longleaf pine systems in which both biodiversity conservation and

timber yield are desired. These methods enable new seedlings to establish at acceptably high

survival rates without a major disruption in the supply of fine fuels or excessive hardwood

release.

But what of a system in which there are no longleaf pines in the overstory, and the

desired outcome is a complete conversion of the overstory from slash pine (Pinus elliottii

Engelm.) to longleaf pine? The same issues with hardwood control and fine fuel supply must be

considered. It has recently been shown that a gradual replacement approach is also an effective

method for making the conversion from slash to longleaf. The single-tree selection method was

used to create plots with variable basal area densities, into which longleaf pine seedlings were

introduced in conjunction with various mechanical and herbicidal hardwood control treatments.

High rates of seedling survival were achieved at all densities, though rates of grass stage

emergence were low and were not expected to increase until the next harvest cycle (Kirkman et

al. 2007).

The second approach to slash-longleaf conversion, and the one employed in this study, is

to remove the entire slash pine canopy, replant with longleaf pine seedlings, and use mechanical

and herbicidal treatments to control competing vegetation. The natural analog of this type of

clearcut is a very large-scale disturbance, such as a hurricane (Mitchell et al. 2006). Since both

longleaf pine seedlings and native herbaceous species respond positively to full sunlight, this

method, too, has the potential to restore a longleaf pine canopy and a diverse, herbaceous

understory. Some of the potential benefits of this approach, as compared to single-tree and

group-selection methods, are: 1) less logging damage to the site over time because only one









prevalent selective force prior to European colonization of the Southeast, and are more effective

at controlling shrubs and maintaining herbaceous plant diversity in the understory (Glitzenstein

et al 1995, Streng et al. 1993). Others have reported shrub resprouting responses to dormant-

season fire in flatwoods similar to those we observed. On mesic and wet sites in Florida and

Lousiana, shrub stem densities were greater than pre-treatment levels after dormant-season fire,

and continued to increase with repeated dormant season fires, whereas long-term growing season

fires maintained stem densities at the same level (Drewa et al. 2002). In another recent study, it

was observed that stem densities of root-crown bearing shrubs (Quercus and Ilex spp.) after

repeated dormant-season fires were seven times greater than those subjected to growing-season

fires (Drewa et al. 2006). In one case, stem densities of root-crown bearing shrubs were 50 times

greater in South Carolina savannas managed with annual dormant season fires than in plots

managed with annual growing-season fires over a 30 year period (Waldrop et al 1992). Some of

the resprouting ability of Quercus can be attributed to a past history of fire suppression, during

which underground stems are likely to have grown extensively (Drewa et al. 2006). Given that

our site was previously a slash pine plantation, periods of fire suppression are likely to have

occurred in the past during seedling establishment and may have contributed to the recalcitrance

of the dominant woody shrubs.

However, results indicating that dormant-season fire increases shrub cover in flatwoods

are not universal. In one case, a long-term dormant season prescribed fire regime significantly

reduced shrub cover on an I. glabra-dominated flatwoods while increasing herbaceous species

cover and diversity (Brockway and Lewis 1997), suggesting that dormant-season fire is a viable

option for flatwoods management when growing-season fire cannot be used. Hiers et al. (2000)

offer experimental evidence, based on patterns of legume reproduction, that variable fire regimes














June 2006 (4 years after treatment)


3 40 -
[ Ericaceae
a) 30
SSerenoa
20- Quercus

10 Ilex


Control Imazapyr Sulfo Sulfo+Hexa Hexa
Treatment

45
June 2007 (post-fire)
40
35
S30 -O Ericaceae

b) 25 Serenoa
I QIerCuI
15 -
S10 E Ilex
15

0-
Control Imazapyr Sulfo Silfo+Hexa Hexa
'Ieatment
Figure 2-2. a) Mean percent shrub cover four years after treatment (June 2006) for Quercus spp.,
L glabra, S. repens and Ericaceous shrubs (Vaccinium spp., K. hirsuta, Gaylussacia
spp., and L. lucida). b) Mean percent shrub cover in June 2007, four months after fire.























a)


Grass Stage


Control Imazapyr


Sulfo Sulfo + Hexa Hexazinone


Treatment


Control Imazapyr Sulfo Sulfo + Hexa Hexazinone
Treatment

Figure 3-5. Mean stem-volume index (SVI) of a) grass stage seedlings and b) OOGS seedlings
by treatment. Means not sharing the same letter are significantly different (Tukey-
Kramer HSD, a = 0.05).


18000
16000
14000
12000
10000
b) so000
6000
4000
2000
0


Out of
Grass Stage









respectively). A significant negative relationship existed between percent Quercus cover and

maximum fire temperature at 30cm (r2= 0.06, p = 0.032) (Figure 2-9a), whereas a significant

positive relationship existed between percent I. glabra cover and maximum fire temperature at

30cm (r2 = 0.07, p = 0.017) (Figure 2-9b). The negative relationship observed with Quercus is

largely attributable to the fact that areas with high Quercus cover were often only partially

burned or did not ignite at all, whereas the I. glabra tended to ignite readily and burn completely.

Temperature was not significantly related to wiregrass cover (p = 0.73) or overall herbaceous

cover (p = 0.28).

Discussion

Four years after initial herbicide application, differential treatment effects on understory

composition were still evident. Sulfometuron methyl had a releasing effect on the two most

common oak species, Q. minima and Q. elliottii, indicating that this is not an appropriate

herbicide for flatwoods with a high density of oaks. However, in spite of having higher shrub

cover and lower herbaceous cover than the other herbicide treatments, sulfometuron plots also

had the highest fourth year herbaceous species richness, diversity, and evenness. One reason for

this may have been that wiregrass cover in the sulfometuron plots was somewhat lower than the

other treatments, leaving more space for other small herbaceous species. These gains in species

richness, diversity, and evenness are therefore unlikely to persist through future fire cycles

because it will be more difficult to apply prescribed fire to this treatment. Indeed, sulfometuron

was the only treatment in which species richness, diversity, and evenness did not increase

following the prescribed fire in 2007.

Woody shrub cover, height, and stem density were lowest in the imazapyr treatment, but

the differences between imazapyr and the other treatments were not as pronounced as those

observed in the first year after application. These results suggest that the increases in shrub stem









evident in any of the herbicide treatments. Oaks (Quercus spp.), one of the dominant shrub

genera on the study site, were resistant to sulfometuron methyl, and this herbicide was therefore

ineffective both as a pine release treatment and for enhancing herbaceous species growth.

Imazapyr was the most effective treatment overall, significantly improving longleaf pine

seedling growth and also enhancing herbaceous species cover. Both hexazinone and the

hexazinone + sulfometuron methyl tank mix provided some seedling growth and understory

enhancement as well, though these effects were not as pronounced as those in the imazapyr

treatment. Shrubs resprouted aggressively after a dormant-season prescribed fire in the fifth year

after treatment, indicating that herbicide-related increases in cover of wiregrass and other

herbaceous species may be lost in future fire cycles. Retention of overstory pines for shrub

competition and as a source of fine fuels is recommended in addition to herbicide for sustainable

silviculture in longleaf pine flatwoods.









effects in a univariate ANOVA model with two independent variables: 'treatment' with

'Block&Random' as a covariate. Data were log-transformed where necessary to meet the

assumptions of ANOVA. Significant differences between treatments were separated with

Tukey's HSD or Hsu's MCB. Post-fire treatment differences were analyzed with ANCOVA,

using pre-fire distributions as a covariate. ANCOVA was also used to analyze fourth year

abundance of Ericaceae, using pre-treatment distributions as a covariate. A significance level of

a = 0.05 was used to test all parameters.

Community patterns were analyzed in PC-ORD version 4, using Nonmetric Multi-

Dimensional Scaling (NMDS) and Indicator Species Analysis following the method of Dufrene

and LeGendre (1997). NMDS uses rank-transformed distances to linearize the degree of

difference between plant survey quadrats. Ordination of plant survey quadrats (containing both

species presence and percent cover data) was done using a Sorenson distance measure in PC-

ORD's autopilot mode. The Sorensen coefficient is a measure of percent dissimilarity that can

be applied to either presence-absence data or quantitative data (in this case percent cover values),

and is recommended for analysis of ecological communities (McCune and Grace 2002). In the

autopilot mode, NMDS conducts 40 runs with real data and 50 runs with randomized data.

Species with less than 3 occurrences were dropped from the analysis to reduce the effects of rare

species (McCune and Grace 2002), reducing the total number of species in the analysis from 87

to 55. PC-ORD chooses the solution with the highest dimensionality that also meets the criterion

of having less stress than 95% of the random runs. Each point on the NMDS axes represents a

single plant survey quadrat, and the distance between points on the NMDS axes represents the

Sorenson distance between points in n dimensions (in this case a 3-dimensional solution was

chosen). A Multi-Response Permutation Procedure (MRPP) can then be used to determine the









LIST OF TABLES

Table page

2-1 Shrub cover, stem density, and height. ........................................ ........................ 37

2-2 Indicator species analysis............................................................... .................... ...... 38

2-3 Understory species richness, diversity, and evenness.................................................... 39

2-4 Mean maximum fire temperature (C) by treatment......... .................................... 40

3-1 Pine seedling growth after five seasons ................................................................ 66









will maintain a broader suite of native species than either growing-season or dormant-season fire

alone, and cite studies suggesting that significant variation in the frequency and season of

lightning may have occurred during the 7000 years BP that longleaf pine communities have been

established in the southern coastal plain.

Though dormant season fire may be effective and even desirable in some cases, our

results suggest that dormant season fire should be avoided when species of low flammability are

abundant, there is a lack of fine fuels, and there are no overstory pines to compete with shrubs

for light and soil resources. Ilex glabra is known to be one of the most flammable flatwoods

species, while oaks are among the least flammable (Behm et al. 2004), and this may be another

part of the reason for the success of dormant-season fire on an I. glabra-dominated site reported

by Brockway and Lewis (1997). In our study, though stem density of both Quercus and Ilex

increased following fire, only Quercus increased in mean percent cover across treatments; mean

I. glabra cover remained roughly the same or decreased. It may be that in the coming growing

seasons, some of the initial flush of basal I. glabra sprouts will die back, and this species will not

show an overall cover increase. Since Quercus increased in cover as well as stem density, it

seems more likely that the fire-related increase will have long-term effects.

In summary, only positive herbicide impacts on wiregrass and herbaceous cover were

observed in this study, and understory species diversity, richness, evenness, and community

composition were largely unaffected. The imazapyr treatment, which had the most dramatic

shrub control effects in the first year after application, still had the lowest shrub levels and the

highest herbaceous cover after four years, indicating that this was the most successful treatment

overall. The poor vegetation control offered by sulfometuron, which was largely attributable to

the resistance of Q. elliottii and Q. minima, indicates that it is not an appropriate herbicide for









use in an oak-dominated flatwoods community. However, the sulfometuron + hexazinone tank

mix showed promise as a flatwoods restoration treatment, offering better control over Quercus

spp. and significantly increasing wiregrass cover relative to the control. Though hexazinone

plots had significantly higher overall herbaceous cover than the control, our results suggest that

this herbicide, at the rate applied, did not provide much control over the dominant shrubs in the

long term. Because we applied a very low rate of hexazinone, it is difficult to make any absolute

comparisons between hexazinone and the other herbicides based on the results of this study.

However, given the success others have had with hexazinone (Brockway et al 1998, Brockway

and Outcalt 2000, Provencher et al 2001), it seems likely that we would have seen greater shrub

control and attendant increases in herbaceous cover had we used a higher rate.

As of June 2007, it was too early to tell whether the herbicide-related increases in

wiregrass and overall herbaceous cover will be maintained through future fire cycles. The

Florida Division of Forestry normally burns this site on a three-year fire interval, and results

from more frequent burning may be more successful than those we observed five years after the

initial herbicide application and four years after the second application. Because the entire

overstory was harvested, it may be several fire cycles before the new longleaf pine canopy

provides sufficient needlefall to sustain the level of prescribed fire necessary for shrub control.

In the meantime, this limitation is likely to be exacerbated by the Florida Division of Forestry's

inability to use growing-season fire on this site due to its proximity to a highway and several

beach housing communities. If the trends of increasing shrub cover, height, and density

continue, the herbicide-related gains in wiregrass and overall herbacous cover may be lost. The

aggressive post-fire resprouting of shrubs observed in this study underscores the utility of

herbicide as a supplemental restoration tool, especially in flatwoods ecosystems where undesired










Table 3-1. Root collar diameter (mm), height (mm), and stem volume index
(basal area2 x height) of grass stage and OOGS seedlings five years after
initial herbicide treatment.
Root Collar
Diameter (mm) Height (mm) Stem Volume Index
Grass Out of Grass Out of Grass Out of
Treatment Stage Grass Stage Stage Grass Stage Stage Grass Stage
Control 25.4a 39.5abc 63.1ab 389.1c 492ab 6332b

Imazapyr 23.6ab 41.8a 70.2a 663.6a 467ab 14,542a

Sulfometuron 22.3b 38.6c 62.2b 490.8bc 388b 9437b

Sulfo + Hexa 25.0a 39.4bc 71.0a 513.1b 519a 10,042b

Hexazinone 25.1a 40.6ab 65.4ab 484.1bc 499a 6332b
ANOVA p< p= p= p< p= P<
0.0001 0.0013 0.003 0.0001 0.0015 0.0001
*Letters denote significant differences between means in the same column at
a = 0.05 based on Tukey-Kramer HSD test. P-values are for randomized
block ANOVA within columns (significant at a = 0.05).









longleaf seedlings, while woody species control is important for the development of fine fuels to

perpetuate the ecosystem in the longer-term (Haywood 2005). On a frequently-burned site

dominated by herbaceous species, prescribed fire may provide adequate site preparation, but on

sites with heavy woody plant competition, mechanical or herbicide treatments are necessary

(Brockway et al. 2006). Intensive mechanical site preparation techniques such as bedding,

disking, and harrowing are effective for longleaf pine establishment (Johnson and Gjerstad

2006), but are very detrimental to wiregrass (Clewell 1989) and are therefore not options on sites

where understory restoration is a goal. Therefore, herbicide, which removes competing

vegetation without disturbing the soil, has potential utility as a component of a sustainable forest

management strategy.

In order to gain a better understanding of the role of herbicide in sustainable longleaf pine

silviculture, we compared the effects of five post-planting, banded herbicide treatments on the

growth and survival of containerized longleaf pine seedlings in a coastal flatwoods in the Florida

panhandle: 1) Sulfometuron methyl (Oust', E.I. du Pont de Nemours and Company,

Wilmington, Del.), 2) Imazapyr (Arsenal', BASF Corporation, Research Triangle Park, NC), 3)

Hexazinone (Velpar L, E.I. du Pont de Nemours and Company, Wilmington, Del.), 4) A tank

mix of Oust' and Velpar L, and 5) a no-herbicide control. These herbicide treatments were

chosen because they are commonly used for loblolly (Pinus taeda L.) and slash pine release in

the Southeast, but their effects on longleaf pine growth and flatwoods community composition

are not well studied. A recent survey by Shepard et al. (2004) reported that the forestry herbicide

most commonly used in the Southeast was Arsenal', followed by tank mixes of Arsenal +

Oust' and Oust' + Velpar. These tank mixes are used to achieve broader spectrum control, but

as a result they may also have more negative implications for native plant communities (Keyser









R e su lts ................... ................... ...................6.........1
S e e d lin g S u rv iv al ................................................................... ................................... 6 1
P percent O ut of G rass Stage ................................................................ ....................... 6 1
Growth: Out of Grass Stage Seedlings ........................................ ........................ 62
G row th: G rass Stage Seedlings................................................... ........... .............. 62
D iscu ssio n ................... ................... ...................6.........2

4 SUMMARY AND CONCLUSIONS ....................... ......................... ..................................... 72

APPENDIX

SPECIES LIST............................................ .............. 76

L IST O F R E F E R E N C E S .................................................................................... .................... 78

B IO G R A PH IC A L SK E T C H ......................................................................... ......................... 85









changes in understory community composition can be rapid and difficult to reverse. It also

underscores the need for effective prescribed fire in order to maintain the early gains in wiregrass

and herbaceous cover conferred by herbicide application.









density evident in the first year (Figure 2-7b) were a short-term phenomenon, and subsequent

dieback of sprouts occurred as these shrubs matured and grew taller.

The responses of Ericaceae to our treatments were surprising, given that in other studies

members of this family have been highly hexazinone-resistant (Wilkins et al. 1993, Outcalt et al.

1999, Miller et al. 1999), and hexazinone is commonly used as a release treatment for blueberry

crops. If anything, we expected to see higher Ericaceae cover in the hexazinone treatment, when

in fact this treatment had the lowest Ericaceae cover of all. This result was particularly notable

given that the hexazinone plots had the highest percent cover of Ericaceous shrubs prior to

treatment (Figure 2-1). The sensitivity of Ericaceous shrubs to hexazinone may have been a

function of the timing of application and the poorly drained, sandy soil; Velpar L is labeled for

blueberry release only prior to budbreak in the spring and in conditions where there is no

standing water (DuPont, 2007). Control of these species may be desirable for restoration

scenarios in which the goal is to increase herbaceous cover on a site dominated by Ericaceous

shrubs (Outcalt et al. 1999). However, many of these shrubs are important wildlife food plants

(Hay-Smith and Tanner 1994). If the goal is to avoid negative impacts to these populations, our

results suggest that caution may be in order with regard to soil type, soil water, and timing of

application.

Both imazapyr and hexazinone significantly increased overall herbaceous cover relative

to the control, and the sulfo + hexa treatment significantly increased wiregrass cover relative to

the control. Though the shrub control initially provided by the herbicide treatments was

relatively short-lived, these results suggest that the short-term shrub control allowed herbaceous

species (and wiregrass in particular) to gain more of a foothold in the understory.









ACKNOWLEDGMENTS


I would like to thank Dr. Shibu Jose, my supervisory committee chair, for the opportunity

to conduct this research, and for his cheerful support and advice. I would also like to thank Drs.

Alan Long and Debbie Miller for serving on my supervisory committee, providing instruction on

the identification and ecology of North Florida vegetation, and assisting me in developing fire

ecology research methods. Thanks also go to Dr. Eric Holzmueller for extensive assistance with

field data collection and analysis throughout the project; to Don Hagan, Michael Morgan, and

Kelly Thayer for assisting me with field data collection; and to Brian Hinton for assistance with

fire methods. I thank the University of Florida herbarium staff, Dr. Walter Judd, and especially

Dr. Susan Carr for invaluable help with the identification of plant specimens. I also thank

Meghan Brennan of the IFAS statistical consulting unit, who assisted me greatly with data

analysis. Thanks go to the Florida Division of Forestry for funding this project and providing the

study site, and especially to site administrator Tom Beitzel and his crew for their assistance in

collecting fire temperature data. Finally, I would like to thank my parents and David Kaplan for

their constant support and encouragement.









Wilkins, R.N., W.R. Marion, D.G. Neary, and G.W. Tanner. 1993. Vascular plant community
dynamics following hexazinone site preparation in the lower Coastal Plain. Canadian
Journal of Forest Research 23:2216-2229

Williamson, G.B. and E.M. Black. 1981. High temperature of forest fires under pines as a
selective advantage over oaks. Nature 293:643-644.









herbicide application in March of 2003, three rows received a second application in April of

2003, and four rows were left with only the first-year treatment. The results from these subplots

were pooled with the first year data in the final analysis because of a lack of significant

differences between one and two-year applications in any of the parameters.

Data Collection and Analysis

The pre-treatment vegetation survey was conducted in June 2001, prior to the overstory

harvest. Post-treatment vegetation surveys were conducted at three and nine months after

treatment, and the results of these first-year surveys have been published elsewhere (Ranasinghe

2003, Ranasinghe et al. 2005). The fourth-year data were collected in June of 2006. Within

each treatment plot, six 1m2 quadrats were randomly selected for understory sampling (two in

each subplot). In all, 36 randomly selected quadrats were surveyed in each herbicide treatment

and 12 were surveyed in the control, for a total of 156 quadrats. The smaller sample size in the

control treatment was due to the fact that a portion of each control plot was sprayed in the second

year, and these rows were not included for analysis. All species were identified and percent

cover of each species was estimated visually using the cover classes defined by the modified

Daubenmire scale (Daubenmire 1959, Peet et al. 1998): 0-1%, 1-2%, 2-5%, 5-10%, 10-25%, 25-

50%, 50-75%, 75-95%, and 95-100%. For each woody species, average stem height was also

estimated and all stems were counted. I. glabra, S. repens, and Quercus spp. (Q. minima and Q.

elliottii), were uniformly distributed across all study plots prior to treatment (Ranasinghe 2003)

and were grouped together for analysis. Woody species belonging to the family Ericaceae

(heaths) were analyzed separately from the other dominant shrubs because they were not evenly

distributed across plots prior to treatment, and also because members of this family are tolerant

to hexazinone and may react differently to treatment than other shrub groups (Wilkins et al 1993,



















E 500-
o
C-

1 400-
0)

1 300-
0)
S -
200-
--

0)
L100-


0-






600-


E 500-
0)
o
16 400-

C -



200-


1-00-


0 10 20 30

Quercus % Cover


I I5
40 50


U -I* I I T I i- "1 i -I I
0 5 10 15 20 25 30
Ilex % Cover



Figure 2-10. a) relationship between mean Quercus cover and mean fire temperature (C) (r2

0.06, p = 0.03) in treatment subplots, and b) relationship between mean Ilex cover and

mean fire temperature (C) (r2 = 0.13, p = 0.002) in treatment subplots.


I


rr\n


I







~



I










Table 2-4. Mean maximum fire
temperature (C) by
treatment.
Mean Fire Temp. (oC)
Treatment 30cm 80cm
Control 276 (57) 181(36)

Imazapyr 275 (32) 190 (27)

Sulfometuron 306 (29) 197 (21)

Sulfo + Hexa 260(26) 166 (22)

Hexazinone 252 (30) 257 (30)

ANOVA p = 0.420 p =0.546
* P-values for randomized block
ANOVA are significant at a = 0.05.
Standard error in parentheses.












Table A-1. Continued.
Species name
Ilex coriacea
Ilex glabra
Ilex opaca
Kalmia hirsuta
Lachnanthes caroliniana
Lachnocaulon auceps
Liatris gracilis
Liatris tenuifolia
Lycania michauxii
Lyonia lucida
Mimosa quadrivalvis
Mitreola sessilifolia
Muhlenbergia capillaris
Opuntia humifusa
Orchidaceae
Osmunda cinnamomea
Photinia pyrifolia
Pinus elliottii
Piteopsis graminifolia
Polygala lutea
Polygala nana
Polygonella polygama
Pteridium aquilinum
Pterocaulon pychnostachium
Quercus elliottii
Quercus incana
Quercus minima
Rhexia alifanus
Rhexia petiolata
Rhynchospora Jlitlia,
Rhynchospora plumosa
Sabatia brevifolia
Scleria ciliata
Serenoa repens
Seriocarpus tortifolius
Smilax auriculata
Smilax glauca
Smilax laurifolia
Smilax pumila
Solidago odora
Sporobolusjunceus
\ihll,,, 1 sylvatica
Stylisma patens
Stylisma villosa
Symphiotrichum adnatum
Tephrosia hispidula
Tragia urens
Vaccinium darrowii
Vaccinium elliottii
Vaccinium myrsinites
Vernonia angustifolia
Vitis rotundifolia
Xyris caroliniana


Common name
Giant Gallberry
Gallberry
American Holly
Hairy Wicky
Carolina Redroot
Whitehead Bog Button
Slender Blazing Star
Shortleaf Blazing Star
Gopher Apple
Fetterbush
Sensitive Vine
Swamp Hompod
Cutover Muhly
Prickly Pear
Orchid sp.
Cinnamon Fern
Red Chokeberry
Slash Pine
Grassleaf Goldenaster
Orange Milkwort
Candyroot
October Weed
Bracken Fern
Blackroot
Runner Oak
Bluejack Oak
Dwarf Live Oak
Savannah Meadow-Beauty
Fringed Meadow-Beauty
Threadleaf Beaksedge
Plumed Beaksedge
Shortleaf Rosegentian
Fringed Nutrush
Saw Palmetto
Whitetop Aster
Earleaf Greenbrier
Wild Sarsparilla
Laurel-leaf Greenbrier
Sarsparilla Vine
Sweet Goldenrod
Pineywoods Dropseed
Queen's Pleasure
Coastalplain Dawnflower
Hairy Dawnflower
Scaleleaf Aster
Sprawling Hoarypea
Wavyleaf Noseburn
Darrow's Blueberry
Elliott's Blueberry
Shiny Blueberry
Tall Ironweed
Muscadine Grape
Carolina Yelloweyed Grass


Family
Aquifoliaceae
Aquifoliaceae
Aquifoliaceae
Ericaceae
Haemodoraceae
Eriocaulaceae
Asteraceae
Asteraceae
Chrysobalanaceae
Ericaceae
Fabaceae
Loganiaceae
Poaceae
Cactaceae
Orchidaceae
Osmundaceae
Rosaceae
Pinaceae
Asteraceae
Polygalaceae
Polygalaceae
Polygonaceae
Dennstaedtiaceae
Asteraceae
Fagaceae
Fagaceae
Fagaceae
Melastomataceae
Melastomataceae
Cyperaceae
Cyperaceae
Gentianaceae
Cyperaceae
Aracaceae
Asteraceae
Smilacaceae
Smilacaceae
Smilacaceae
Smilacaceae
Asteraceae
Poaceae
Euphorbiaceae
Convolvulaceae
Convolvulaceae
Asteraceae
Fabaceae
Euphorbiaceae
Ericaceae
Ericaceae
Ericaceae
Asteraceae
Vitaceae
Xyridaceae


Life Form
Shrub
Shrub
Shrub
Shrub
Monocot
Monocot
Forb
Forb
Forb
Shrub
Legume
Forb
Grass
Forb
Monocot
Fern
Shrub
Tree
Forb
Forb
Forb
Forb
Fern
Forb
Shrub
Tree
Shrub
Forb
Forb
Sedge
Sedge
Forb
Sedge
Shrub
Forb
Monocot
Monocot
Monocot
Monocot
Forb
Grass
Forb
Forb
Forb
Forb
Legume
Forb
Shrub
Shrub
Shrub
Forb
Vine
Monocot









Litt et al. (2001) also concluded that the effects of herbicide on ground-layer vegetation

in natural flatwoods systems had rarely been measured. They reviewed only two studies

documenting herbicide impacts on groundcover in non-plantation flatwoods ecosystems, both of

which reported decreases in species richness. Decreases in richness ranged from 5.1% for

herbaceous species following hexazinone application (Wilkins et al. 1993), to 71.8% for all

species following treatment with a mix of sulfometuron methyl, glyphosate, and triclopyr (Neary

et al. 1991).

The effects of herbicide on sandhill communities are better documented. Four years after

treatment, hexazinone herbicide was effective in eliminating woody species and increasing

diversity relative to the control, though plots restored with fire alone had even higher species

diversity than those treated with herbicide (Provencher et al. 2001). On another north Florida

sandhill site, an initial decrease in species diversity was documented following hexazinone

application, followed by a recovery and then an increase in the second growing season

(Brockway et al. 1998). A subsequent study on the same site revealed that prescribed fire further

enhanced species diversity in conjunction with hexazinone (Brockway and Outcalt 2000). Three

herbicide release treatments (imazapyr, glyphosphate, and hexazinone) were tested on loblolly

pine plantations at three hilly sites in central Georgia. Seven years after treatment, there were no

significant difference in understory species diversity between any of the treatments including the

control (Boyd et al. 1995). Eleven years after treatment, Miller et al. (1999) arrived at the same

conclusion. Imazapyr, used alone or in conjunction with fire, was effective for restoring grass-

dominated bobwhite quail habitat on a loblolly pine sandhill in the Florida panhandle,

significantly reducing hardwood encroachment while enhancing the growth of native herbaceous

species (Welch et al. 2004). Beneficial results were also reported for imazapyr in a quail habitat









a covariate. Data were log-transformed where necessary to meet the assumptions of ANOVA.

Significant differences between treatments were separated with the Tukey-Kramer HSD test.

Post-fire survival was analyzed with ANCOVA, using pre-fire survival as a covariate. All

parameters were analyzed in JMP IN version 5 (SAS Institute, Inc.).

Results

Seedling Survival

Five growing seasons after treatment (November 2006), survival rate varied significantly

among herbicide treatments (a = 0.05, p < 0.0001) (Figure 3-1). The highest survival rates

occurred in the hexazinone and control treatments (80% and 78.5%, respectively) (Figure 3-1).

Survival rates were significantly lower in the sulfometuron + hexazinone (sulfo + hexa) and

imazapyr treatments (64.2% and 63.2%, respectively). It should be noted that the mortality in

the imazapyr and sulfo + hexa treatments occurred shortly after herbicide application

(Ranasinghe 2003), and there was little subsequent mortality in any of the treatments.

None of the treatment plots experienced significant mortality following fire (Figure 3-2).

However, there were some slight differences in fire-related mortality among the different

treatments, and as a result, survival rate no longer varied significantly by treatment following fire

(p = 0.238). In the imazapyr treatment, the survival rate dropped by only .7% following fire, as

compared to decreases of 3% to 3.5% in the other treatments.

Percent Out of Grass Stage

Five growing seasons after herbicide application (November 2006), the mean grass-stage

release rate varied significantly among treatments (p < 0.0001). The highest level of grass-stage

release after four years was observed in the imazapyr treatment, though both sulfo + hexa and

hexazinone also significantly raised the release rate relative to the control (Figure 3-3).









Noss, R.F., LaRoc, III, E.T., and J.M. Scott. 1995. Endangered ecosystems of the United States:
a preliminary assessment of loss and degradation. USDI National Biological Service,
Washington, D.C., Biological Report No. 28.

Olson, M.S. and W.J. Platt. 1995. Effects of habitat and growing season fires on resprouting of
shrubs in longleaf pine savannas. Vegetatio 119:101-118.

Outcalt, K.W., Williams, M.E., and 0. Onokpise. 1999. Restoring Aristida strict to Pinus
palustris ecosystems on the Atlantic Coastal Plain, U.S.A. Restoration Ecology 7(3):262-
270.

Peet, R.K., Wentworth, T.R., and P.S. White. 1998. Method for recording vegetation
composition and structure. Castanea 63:262-274.

Palik, B.J., Mitchell, R.J., Houseal, G., and N. Pederson. 1997. Effects of canopy structure on
resource availability and seedling responses in a longleaf pine ecosystem. Canadian Journal
of Forest Research 27:1458-1464.

Palik, B.J., Mitchell, R.J., J.K. Hiers. 2002. Modeling silviculture after natural disturbance to
sustain biodiversity in the longleaf pine (Pinuspalustris) ecosystem: balancing complexity
and implementation. Forest Ecology and Management 155:347-356.

Pecot, S.D., Mitchell, R.J., Palik, B.J., Moser, E.B., and J.K. Hiers. 2007. Competitive
responses of seedlings and understory plants in longleaf pine woodlands: separating canopy
influences above and below ground. Canadian Journal of Forest Research 37:634-648.

Peet, R.K. 2006. Ecological classification of longleaf pine woodlands. Pages 51-93 in Jose, S.,
Jokela, E., and Miller, D.L., editors. The LongleafPine Ecosystem: Ecology, Silviculture,
and Restoration. Springer Science + Business Media, LLC., New York.

Peet, R.K. and D.J. Allard. 1993. Longleaf pine-dominated vegetation of the southern Atlantic
and eastern Gulf Coast region, USA. Proceedings of the Tall Timbers Fire Ecology
Conference 18:45-81.

Platt, W.J. and S.L. Rathbun. 1993. Dynamics of an old-growth longleaf pine population.
Proceedings of the Tall Timbers Fire Ecology Conference 18:275-297.

Platt, W.J., Beckage, B., Doren, R.F. & H.H. Slater. 2002. Interactions of large-scale
disturbances: prior fire regimes and hurricane mortality of savanna pines. Ecology 83:1566-
1572.

Platt, W.J., Carr, S.M., Reilly, M., and J. Fahr. 2006. Pine savanna overstorey influences on
ground-cover biodiversity. Applied Vegetation Science 9:37-50.









blocks laid out in mesic-dry areas. A pre-treatment vegetation survey identified wetter areas by

the presence of indicator species such as sedges (Cyperaceae). The soils of the study site are

sandy, siliceous, thermic, aeric alaquods of the Leon series (Spodosols consisting of deep, poorly

to very poorly drained soils derived from marine parent material (Ranasinghe 2003)). Yearly

weather patterns during the study period were variable. Total annual precipitation during the

study period was 103cm in 2002, 200cm in 2003, 135cm in 2004, 146cm in 2005, and 106cm in

2006 (NOAA 2007). Rainfall in the months following the prescribed fire was very low, with

totals of .5cm in March 2007, 4cm in April 2007, and .3cm in May 2007 (NOAA 2007).

Prior to August 2001, the 20-acre study site was a planted slash pine stand with a mean

stand age of about 26 years, which had been burned on a 3-year rotation by the Division of

Forestry since its acquisition of Point Washington State Forest in 1992. A pre-harvest, baseline

data survey was conducted in the areas slated to become study blocks in June 2001. In August

2001, the entire overstory was harvested and the site was prepared for the study with a single,

light roller chopping followed by a prescribed burn in October 2001. In December 2001, six

study blocks, each divided into five 36.6m x 24.4m treatment plots, were established according

to the aforementioned randomized complete block design. Each treatment plot was hand-planted

with 100 containerized longleaf pine seedlings laid out in 10 rows, with 3.1m between rows and

1.8m between the seedlings in each row.

Treatments

Within each block, each 100-seedling plot was randomly selected to receive one of five

herbicide treatments in March of 2002:

1) Sulfometuron methyl (Oius ") at 0.26 kg a.i./ha. This is a moderate application rate

for sulfometuron methyl; other studies have used 0.16 kg a.i./ha (Lauer and Glover, 1998) and

0.21 kg ai/ha (Shiver and Martin 2002, Keyser and Ford 2006), while the manufacturer-









CHAPTER 3
HERBICIDE AND SUSTAINABLE LONGLEAF PINE SILVICULTURE

Introduction

Perhaps more than any other forested ecosystem in the United States, the longleaf pine

(Pinus palustris Mill.) savanna lends itself to simultaneous management for timber and

biodiversity. The understory communities associated with longleaf pine are dominated by

herbaceous plants adapted to frequent ground layer disturbance in the form of fire, and regular

overstory disturbance by lightning, windthrow, and hurricanes (Platt and Rathbun 1993, Platt et

al. 2002). In longleaf pine systems with a frequent fire regime and an intact seedbank, most

native herbaceous species respond positively to the increased light levels in newly-opened

canopy gaps (Keddy et al. 2006, Platt et al. 2006), and it has been hypothesized that many of

these species are such weak competitors for light that they could be considered fugitive or

ephemeral (Keddy et al. 2006). In one case, removal of the overstory resulted in a 20% increase

in the number of understory species in gaps within five years, and surveys of older gaps indicate

that similar increases have been maintained over several decades (Platt et al. 2006). Wiregrass

(Aristida strict Michx.), the dominant bunchgrass in most longleaf pine systems and a keystone

species with regard to pyrogenicity, is also released in response to overstory removal (McGuire

2001). Since the renowned plant biodiversity of the longleaf pine ecosystem resides largely in

the herbaceous component (Walker and Peet 1983, Kirkman et al. 2001), the removal of large

overstory trees for timber harvest has the potential to benefit the ecosystem.

However, there are several important caveats to the analogy between silviculture and

natural disturbance in longleaf pine systems (Mitchell et al. 2006). One such caveat is that the

removal of overstory pines across a large area also releases shrubs (Brockway et al. 1998, Kush

et al. 1999, McGuire et al. 2001, Kirkman et al. 2007, Pecot et al. 2007), which, in the absence of









recommended rate for herbaceous weed control in longleaf pine stands is 0.10 to 0.42 kg a.i./ha.

Sulfometuron methyl is a selective herbicide primarily effective for controlling herbaceous

species (Lauer and Glover 1998), and mixing with hexazinone is recommended if broader

spectrum control is desired (DuPont 2007).

2) Hexazinone: (Velpar L) at 0.56 kg a.i./ha. Recommended application rates for

hexazinone range from 2.0 to 6.7 kg a.i./ha (E.I. Du Pont De Nemours), while rates successfully

used to control hardwoods in sandhills ecosystems have ranged from 1.1 kg a.i./ha to 2.4 kg a.i

/ha (Brockway et al. 1998, Provencher et al. 2001). The hexazinone application rate in our study

was therefore very low. This rate was chosen because hexazinone had been deemed effective at

the aforementioned rates, and we were interested in whether an even lower application rate

would provide vegetation control.

3) Sulfometuron + Hexazinone mix (0.26 kg a.i./ha Sulfometuron methyl + 0.56 kg

a.i./ha Hexazinone). This is a common tank mix used throughout the Southeast (Shepard et al.

2004) with the potential to control a broad spectrum of woody and herbaceous plants.

4) Imazapyr (Arsenal) at 0.21 kg a.i./ha. For a restoration project, this was a

moderate application rate for imazapyr. Significant hardwood control and habitat improvements

have been obtained using imazapyr at 0.42 kg a.i./ha (Jones and Chamberlain 2004), and at all

rates from 0.08 kg a.i./ha to 0.24 kg a.i..ha (though higher rates did not offer significant

improvements over lower rates) (Keyser and Ford 2006). Arsenal' is not labeled for use over

longleaf pine at the recommended rates of 0.32 to 0.42 kg a.i./ha until after the end of the second

growing season (BASF 2007), and has caused stunted growth in loblolly pine seedlings (Barber

1991). We chose the rate of 0.21 kg a.i./ha in order to determine whether a lower-than-

recommended application rate would prove less injurious to pine seedlings.









Results


Shrub Cover

Four years after initial herbicide application, percent cover of glabra, S. repens, and

Quercus spp. ranged from 13.5% in the sulfometuron + hexazinone (sulfo + hexa) treatment to

22.9% in the sulfometuron treatment, but did not vary significantly among treatments overall (p

= 0.2154) (Table 2-1). Indicator species analysis revealed a significant association between the

sulfometuron treatment and increased cover of both rhizomatous oak species, Q. elliottii and Q.

minima, four years after treatment (p = 0.040 and p = 0.031) (Table 2-2). Since these species

were not significantly associated with the control plots, this result can be interpreted to mean that

the sulfometuron treatment enhanced oak growth.

Woody species belonging to the family Ericaceae were abundant on the study site, and

included blueberries (V. myrsinites and V. darrowii), huckleberries (G. dumosa and G. frondosa),

hairy wicky (K. hirsuta), and Lyonia lucida Koch fetterbushh). Ericaceae cover four years after

treatment was four to five times higher in the control plots than in any of the herbicide

treatments, a difference which approached significance (p = 0.06), but was moderated by the

uneven pre-treatment distributions and the smaller control sample size (Table 2-1, Figure 2-1).

Overall percent cover of woody species was highest in the control and lowest in the sulfo + hexa

treatment (Figure 2-2). In June 2007, four months after the February 2007 prescribed fire,

percent cover of Quercus spp., I. glabra, and S. repens did not vary significantly among

treatments (p = 0.1740) (Figure 2-3a, Table 1), nor did Ericaceae cover (p = 0.097).

Herbaceous Cover

Herbaceous cover varied significantly among treatments (p = 0.002) (Figure 2-4a). In

both the imazapyr and hexazinone treatments, mean herbaceous cover (46.5% and 46.1%,

respectively) was significantly higher than the control (22.1%). Following fire, however,











APPENDIX
SPECIES LIST


Table A-1. Species list.
Species name
Andropogon virginicus var. glaucus
Andropogon virginicus var. virginicus
Aristida spiciformis
Aristida stricta
Asclepias cinerea
Balduina uniflora
Baptisia lanceolata
Calamovilfa curtissii
Carphephorus odoratissimus
C( I, % gossypina subsp. cruseana
(C ,Ii. -'.' mariana
Cladoniaceae
Ctenium aromaticum
Cynanchum i,, .I'4, h ,iii. tl
Cyrilla racemiflora
Desmodium lineatum
Desmodium strictum
Dichanthelium aciculare
Dichanthelium ensifolium
Dichanthelium ovale
Dichanthelium erectifolium
Dichanthelium strigosum
Drosera capillaris
Eryngium yuccifolium
Eupatorium capillifolium
Eupatorium compostifolium
Eupatorium mohrii
Eupatorium pilosum
Eupatorium rotundifolium
Euphorbia inundata
Eurybia eryngiifolia
Euthamia graminifolia
Gaylussacia dumosa
Gaylussaciafrondosa
Gelsemium sempervirens
Gratiola hispida
Houstonia procumbens
Hypericum brachyphyllum
Hypericum hypericoides
Hvpoxis sessilus


Common name
Chalky Bluestem
Broomsedge Bluestem
Bottlebrush Threeawn
Wiregrass
Carolina Milkweed
Oneflower Honeycombhead
Gopherweed
Curtiss's Sandgrass
Deer's Tongue
Cruise's Goldenaster
Maryland Goldenaster
Deer Moss
Toothache Grass
Gulf Coast Swallowwort
Swamp Titi
Sand Ticktrefoil
Pinebarren Ticktrefoil
Needleleaf Witchgrass
Cypress Witchgrass
Eggleaf Witchgrass
Erectleaf Witchgrass
Roughhair Witchgrass
Pink Sundew
Button Rattlesnakemaster
Dog Fennel
Yankeeweed
Mohr's Thoroughwort
Rough Boneset
Roundleaf Thoroughwort
Florida Pineland Spurge
Thistleleaf Aster
Flattop Goldenrod
Dwarf Huckleberry
Dangleberry
Yellow Jessamine
Rough Hedgehyssop
Roundleaf Bluet
Coastalplain St. John's Wort
St. Andrew's Cross
Yellow Star Grass


Family
Poaceae
Poaceae
Poaceae
Poaceae
Apocynaceae
Asteraceae
Fabaceae
Poaceae
Asteraceae
Asteraceae
Asteraceae
Cladoniaceae
Poaceae
Apocynaceae
Cyrillaceae
Fabaceae
Fabaceae
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Droseraceae
Apiaceae
Asteraceae
Asteraceae
Asteraceae
Asteraceae
Asteraceae
Euphorbiaceae
Asteraceae
Asteraceae
Ericaceae
Ericaceae
Gelsemiaceae
Veronicaceae
Rubiaceae
Clusiaceae
Clusiaceae
Hypoxidaceae


Nomenclature follows Wunderlin (2004). Species are divided into the following life forms: Grass =
member of the Poaceae family, Sedge = member of the Cyperaceae family, Monocot = other monocot,
Legume = member of the Fabaceae family, Carnivorous = carnivorous forb, Forb = other forb, Fern =
member of the order Filicophyta, Shrub = woody understory species, Tree = woody overstory species.


Life Form
Grass
Grass
Grass
Grass
Forb
Forb
Legume
Grass
Forb
Forb
Forb
Moss
Grass
Forb
Shrub
Legume
Legume
Grass
Grass
Grass
Grass
Grass
Carnivorous
Forb
Forb
Forb
Forb
Forb
Forb
Forb
Forb
Forb
Shrub
Shrub
Vine
Forb
Forb
Shrub
Shrub
Monocot









dynamics (Palik et al. 2002, Mitchell et al. 2006, Brockway et al. 2006). It has long been

observed that longleaf pine seedlings regenerate only where disturbances have created canopy

openings (Boyer 1963). Several recent studies have looked at patterns of regeneration in natural

gaps as well as gaps created by group- or single-tree selection. Naturally-regenerating longleaf

pine seedlings on a xeric sandhill were clustered near the center of gaps but were absent around

the outer edge of each gap, an observation which led Brockway and Outcalt (1998) to propose

the existence of a 12-16m "seedling exclusion zone" resulting from 1) higher fire intensity due to

high litter accumulation at the bases of trees around the gap edge, and 2) competition with the

root systems of adult pines around the gap edge. Some studies of longleaf pine regeneration

have supported this hypothesis. As basal area in a mature longleaf pine stand increased, fire-

related seedling mortality also increased and seedling size decreased (Grace and Platt 1995),

while higher light intensity and N availability in gap centers were correlated with seedling

survival and growth (Palik et al. 1997). Though subsequent research has not found evidence for

a seedling exclusion zone per se, suppressed growth with increasing proximity to adults has been

shown in all studies. Container-grown seedlings planted in gaps exhibited maximum growth in

response to high light intensity at the center of gaps as small as .1-ha, but no strong response to

nutrient availability (McGuire et al. 2001). Others have reported that seedling survival was

actually higher toward the outer edges of gaps, though the seedlings able to survive in the center

were significantly larger than those closer to the edges (Gagnon et al. 2003, Pecot et al. 2007).

Results from a trenching experiment suggest that this tradeoff is due to underground facilitation

by adult longleaf pines at gap edges vs. greater light availability at the gap center (Pecot et al.

2007). Trenching also revealed that hardwoods were immediately released when belowground

competition from mature canopy trees was removed (Pecot et al. 2007). The upshot of these










Table 2-1. Mean percent cover, stem density, and height of shrubs 4 years after treatment
(YAT) (June 2006) and following prescribed fire (June 2007).
% Cover Stem Density (stems/m2) Stem Height (cm)
Q, S, I Ericaceae Q, S, I Ericaceae Q, S, I Ericaceae
2006 Post- 2006 Post- 2006 Post- 2006 Post- 2006 Post- 2006 Post-


Treatment 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire
Control 18.7 22.5 21.3 2.8 31.8ab 56.3 42.9 10.3 16.1 21.1 11.6 12.5

Imazapyr 13.6 24.3* 5.1 2.3 20.8b 47.8* 8.4 8.9 14.6 19.7* 12.3 12.1

Sulfometuron 22.9 34.0* 4.4 5.6 34.7a 64.9* 17.2 25.0 16.9 22.3* 12.2 15.2*

Sulfo+Hexa 13.5 21.5 5.8 3.6 25.1ab 50.9* 16.5 12.8 15.9 21.6* 12.0 17.0*

Hexazinone 19.9 26.6 4.0 3.3 32.9a 59.1* 13.0 9.6 17.1 19.7 11.9 16.1*

ANOVA/ P= P= P= P= P= P= P= P= P= P= P= P=
ANCOVA 0.215 0.174 0.060 0.097 0.012 0.331 0.061 0.108 0.234 0.400 0.304 0.900


*Letters denote significant differences between means in the same column at a = 0.05
based on Tukey-Kramer HSD test. (*) denotes significant post-fire change within
treatment (Tukey-Kramer HSD). P-values are for randomized block ANOVA / ANCOVA
within columns (significant at a = 0.05). Q, S, I stands for Quercus spp., S. repens and .
glabra.









restoration project on hilltops in southern Lousiana (Jones and Chamberlain 2004). However, no

differences in herbaceous species cover and richness relative to the control were found on a site

in the Virginia piedmont treated with three different levels of imazapyr for pine release (Keyser

and Ford 2006). Other pine release studies have also reported beneficial or neutral results for

imazapyr, with wildlife forage vegetation showing recovery and in some cases improvement

after treatment (Hurst 1987, McNease and Hurst 1991).

In order to gain a better understanding of herbicide as a restoration tool for longleaf pine

flatwoods ecosystems, we compared the effects of five herbicide treatments on the understory of

a coastal flatwoods in the Florida panhandle: 1) Sulfometuron methyl (Oust', E.I. du Pont de

Nemours and Company, Wilmington, Del.), 2) Imazapyr (Arsenal', BASF Corporation,

Research Triangle Park, NC), 3) Hexazinone (Velpar L, E.I. du Pont de Nemours and

Company, Wilmington, Del.), 4) A tank mix of Oust' and Velpar L, and 5) a no-herbicide

control. These herbicide treatments were chosen because they are known to be effective

vegetation control treatments for loblolly and slash pine release in the Southeast, but their effects

on flatwoods community composition are not well studied. A recent survey reported that the

forestry herbicide most commonly used in the southeast was Arsenal', followed by tank mixes

of Arsenal' + Oust' and Oust' + Velpar (Shepard et al. 2004).

We examined the longer-term effects of these herbicide treatments on flatwoods

vegetation four years after initial herbicide application (see Ranasinghe 2003 for first year

results), after which we conducted a dormant-season prescribed fire on all treatment plots and

assessed post-fire vegetation composition.









Growth: Out of Grass Stage Seedlings

The mean RCD, height, and SVI of seedlings out of the grass stage (OOGS) varied

significantly among treatments (p = 0.0013, p < 0.0001, and p < 0.0001, respectively). In all

parameters, imazapyr showed the greatest improvement relative to the control, though sulfo +

hexa also significantly improved height relative to the control (Table 3-1, Figure 3-4, Figure 3-

5). RCD of OOGS seedlings varied less among treatments than height or SVI, which is to be

expected due to the preferential investment in height growth characteristic of longleaf pine

seedlings in the "bolting stage."

Growth: Grass Stage Seedlings

Root collar diameter (RCD), height, and SVI of grass-stage seedlings varied significantly

among treatments (p < 0.0001, p = 0.003, and p = 0.0015, respectively) (Table 3-1, Figure 3-5),

with the smallest seedlings in the sulfometuron treatment. Among seedlings that were still in the

grass stage after five growing seasons, none of the herbicide treatments showed significant

improvements relative to the control. This pattern has been observed since the first growing

season after treatment (Ranasinghe 2003), though it was unclear whether the lower seedling size

was due to seedling injury from sulfometuron or the poor shrub control evident in the

sulfometuron treatment since the first growing season (Chapter 2).

Discussion

Five growing seasons after treatment, imazapyr plots had a grass stage release rate more

than twice that of the control, and the mean stem volume index of OOGS seedlings in the

imazapyr treatment was also more than double that in the control. Although direct application of

imazapyr to recently-planted seedlings caused significant mortality in the first year after

planting, the overall post-fire survival rate of 62.5% in this treatment is still within the acceptable

range of 60-75% recommended for adequate overstory restocking (Brockway et al. 2006).









became less pronounced over the years, differences in biomass accumulation rates between

suppressed seedlings probably also decreased. In contrast, the earlier in the process a pine

seedling was released, the better its chances may have been of gaining a competitive advantage

for light, water, and soil resources. The early gains made by seedlings in the imazapyr plots

(Ranasinghe 2003) apparently translated into an increasing competitive advantage over the years,

leading to much larger seedlings in these plots even as vegetation differences between plots

became less pronounced. In a scenario such as this, where the dual goals of silviculture and

ecosystem restoration dictate light site preparation and banded herbicide application in spite of

the heavy flatwoods competition, early release from the grass stage may be critical to the

ultimate success of the planted pines.

The site preparation burn conducted following overstory harvest in 2001 consumed the

last of the accumulated slash pine needles, and, though none of the herbicide treatments

decreased wiregrass cover, distribution of wiregrass across the site was patchy (Chapter 2). Five

years after initial herbicide treatment, we had difficulty applying a prescribed fire of sufficient

intensity to control shrubs-a phenomenon which was undoubtedly exacerbated by length of

time since the application of herbicide and the need to apply fire during the dormant season

(Chapter 2). The lack of fine fuels was also probably a major factor limiting our ability to apply

fire effectively, given the key role played by dead pine needles in the pyrogenicity of a longleaf

pine system (Williamson and Black 1981, Noss 1989). Our results support the theory proposed

by Kirkman et al. (2007) that clearcutting, even when the goal is a total canopy conversion from

slash pine to longleaf pine, may ultimately hinder the success of a restoration project.

The overstory and understory effects of imazapyr, hexazinone, and the sulfometuron +

hexazinone tank mix lined up well with the restoration goals for this site. These treatments,




















r rlr*


Sulfo+Hexa


--- Control
- Imazapyr
- Sulfo
-- Sulfo+Hexa
-- Hexa


Pre-treat 3 Mos 9 Mos 4yrs Post-Fire
Time
Figure 2-8. a) Mean species diversity (Shannon Index) by treatment four years after treatment
(Pre-fire) and five years after treatment (Post-fire). Pre-fire differences among
treatments are not significant (ANOVA, p = 0.13). (*) indicates a significant pre-
fire/post-fire difference (Tukey-Kramer HSD, a = 0.05). b) Changes in species
richness within each treatment over the course of the 5-year experiment. First-year
results from Ranasinghe (2003 and unpublished data), and Ranasinghe et al. (2005).


a)


Control


O Pre-Fire
* Post-Fire


Imazapyr


Sulfo


Treatment


2.6
2.4
2 2.2
a 2
b) 1.8
S1.6
1.4









CHAPTER 4
SUMMARY AND CONCLUSIONS


One of the biggest challenges to successful restoration of the longleaf pine (Pinus

palustris Mill.) ecosystem is the persistence of shrubs in the understory, which suppress longleaf

pine seedlings as well as native herbaceous plants. Herbicide can be used as a supplement to fire

in order to enhance shrub control, but must be studied carefully because of the potential for

negative impacts on native plants. Information is lacking about the effects of herbicide on

natural longleaf pine flatwoods communities. We used a banded application of three herbicides

and one tank mix as shrub control treatments following harvest of a slash pine stand and

replanting with containerized longleaf pine seedlings in a mesic-wet flatwoods. The herbicides

tested were Arsenal' (imazapyr), Oust' (sulfometuron methyl), Velpar L (hexazinone), and an

Oust' + Velpar L tank mix.

Imazapyr herbicide significantly improved longleaf pine seedling growth, due to its

effectiveness at controlling competing shrubs during the first growing seasons after planting.

Though these shrub control effects were short-lived, the short term shrub control allowed

herbaceous species to gain more of a foothold in the understory, and significant increases in

herbaceous cover relative to the control were recorded in imazapyr-treated plots four years after

initial treatment. In a longleaf pine silviculture project with both timber and conservation goals,

this was a very desirable result, because of the high level of biodiversity in the herbacous

component of longleaf pine ecosystems (Walker and Peet 1983, Kirkman et al. 2001). No

negative impacts on understory species richness, diversity, evenness, or community composition

resulted from the application of this herbicide. Though over-the-top application of imazapyr

caused significant longleaf pine seedling mortality, the mean survival rate of seedlings in

imazapyr-treated plots was 62.5% after five growing seasons and 61.8% after a fifth-year









CHAPTER 1
INTRODUCTION


Two hundred years ago, longleaf pine (Pinuspalustris Mill.) ecosystems covered an

estimated 37 million hectares of the southeastern United States (Frost 2006). Based on fuel

accumulation patterns and plant life histories it is believed that light to moderate intensity surface

fires burned throughout the range of the longleaf pine every two to three years prior to European

colonization (Frost 1996). These frequent fires maintained an open, longleaf pine-dominated

canopy and a grass-dominated understory containing some of the most species-rich plant

communities ever documented outside the tropics (Peet and Allard 1993). However, in the years

following European colonization, fires were suppressed and old-growth longleaf pine systems

were logged almost out of existence. Today, it is estimated that only 841,000 hectares of

naturally regenerating longleaf pine systems still remain, the great majority of which are second

growth and fire-suppressed (Frost 2006). This area contains only about 5,095 hectares of old-

growth longleaf pine (Varner and Kush 2004). Fire suppression has resulted in extensive

understory invasion by hardwood shrubs and a concurrent regional decline in herbaceous species

diversity (Walker and Peet 1983, Mehlman 1992, Gilliam and Platt 1999). In 1995, a U.S.

Biological Survey report listed the longleaf pine ecosystem as the third most endangered

ecosystem in the United States (Noss et al. 1995).

In recent years, recognition of the value of the longleaf pine ecosystem has motivated

widespread restoration efforts throughout the southeast. One of the biggest obstacles to

successful longleaf pine ecosystem restoration is the persistence of hardwoods in the understory

even after the reintroduction of fire (Walker and Siletti 2006). In some cases the reintroduction

of fire alone may restore the desired two-layered savanna structure (Kush et al. 1999), but many

factors can limit the efficacy of fire after a long period of suppression, including insufficient fine









FROM PINus ELLIOTTII PLANTATION TO PINUS PALUSTRIS ECOSYSTEM:
THE ROLE OF HERBICIDE IN LONGLEAF PINE RESTORATION



















By

JOHANNA E. FREEMAN


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

UNIVERSITY OF FLORIDA

2008









Hiers, J.K., Wyatt, R., and R.J. Mitchell. 2000. The effects of fire regime on legume
reproduction in longleaf pine savannas: is a season selective? Oecologia 125:521-530.

Hurst, G.A. and W.E. Palmer. 1987. Vegetation responses to imazapyr for pine release.
Proceedings of the Southern Weed Science Society 40:247.

Hobbs, R.J., Currall, J.E.P., and C.H. Gimingham. 1984. The use of"thermocolor" pyrometers
in the study of heath fire behavior. Journal of Ecology 72:241-250.

Iverson, L.R., Yaussy, D.A., Rebbeck, J., Hutchinson, T.F., Long, R.P., and A.M. Prasad. 2004.
A comparison of thermocouples and temperature paints to monitor spatial and temporal
characteristics of landscape-scale prescribed fires. International Journal of Wildland Fire.
13(3):311-322.

Jacqmain, E.I., Jones, R.H., and R.J. Mitchell. 1999. Influences of frequent cool-season burning
across a soil moisture gradient on oak community structure in longleaf pine ecosystems.
American Midland Naturalist 141:85-100.

Jones, J.D., and M.J. Chamberlain. 2004. Efficacy of herbicides and fire to improve vegetative
conditions for northern bobwhites in mature pine forests. Wildlife Society Bulletin
32(4):1077-1084.

Johnson, R. and D.J. Gjerstad. 2006. Restoring the overstory of longleaf pine ecosystems.
Pages 271-295 in Jose, S., Jokela, E., and Miller, D.L., editors. The LongleafPine
Ecosystem: Ecology, Silviculture, and Restoration. Springer Science + Business Media,
LLC., New York.

Keyser, P.D., and V.L. Ford. 2006. Wildlife habitat and herbicides: an evaluation of a widely
applied tank mix. Southern Journal of Applied Forestry 30(1):46-51.

Keddy, P.A., Smith, L., Campbell, D.R., Clark, M., and G. Montz. 2006. Patterns of herbaceous
plant diversity in a Louisiana pine savanna. Applied Vegetation Science 9:17-26.

Kirkman, L.K., Mitchell, R.J., Helton, R.C., and M.B. Drew. 2001. Productivity and species
richness across and environmental gradient in a fire-dependent ecosystem. American Journal
of Botany 88(11):2119-2128.

Kirkman, L.K., Mitchell, R.J., Kaiser, M.J., Pecot, S.D., and K.L. Coffey. 2007. The perpetual
forest: using undesirable species to bridge restoration. Journal of Applied Ecology 44:604-
614.

Knapp, B.O., Wang, G.G., Walker, J.L., and S. Cohen. 2006. Effects of site preparation
treatments on early growth and survival of planted longleaf pine (Pinuspalustris Mill.)
seedlings in North Carolina. Forest Ecology and Management 226:122-128.






















80 June 2006 (4 years after treatment)
70
60
50

a) b40
30
20
10 5- Herbaceous
0O Sh ubs
Control Imazapyr Sulfo Sulfo +Hex Hexzinone
80 h Ericaceous
Shrubs
70 June 2007 (Following Feb. 2007 fire)

60

50

b) 40
2 30

20

10


Control Imazapyr Sulfo Sulfo+Hexa Hexazinone
Treatment


Figure 2-5. Overall vegetation cover in a) June 2006 (4 years after treatment) and b) June 2007
(4 months after fire). Cover is separated into herbaceous, shrub, and Ericaceous
shrub components.

























a

ab










Sulfo Hexa Hexa


l Pre-Fire
* Post-Fire


--- Control
--- Imazapyr
- Sulfo
-*- Sulfo+Hexa
-X Hexa


Pre-harvest 3 Mos 9 Mos 4yrs Post-fire
Time


Figure 2-6. a) Mean % cover of wiregrass four years after initial treatment (pre-fire) and five
years after treatment (post-fire). Pre-fire means not sharing the same letter are
significantly different (Hsu's MCB, a = 0.05). (*) indicates a significant pre-
fire/post-fire difference (Tukey-Kramer HSD, a = 0.05). b) Mean % wiregrass cover
within treatments over the course of the 5-year experiment. First-year results from
Ranasinghe (2003 and unpublished data), and Ranasinghe et al. (2005).


40
35
o 30
I 25


15
10
5
0


b





Control


b








Sulfo
Treatment


a

Imazapyr


35
S30
S25

b) 20
S15

10
a









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

FROM PINUS ELLIOTTII PLANTATION TO PINUS PALUSTRIS ECOSYSTEM:
THE ROLE OF HERBICIDE IN LONGLEAF PINE RESTORATION

By

Johanna E. Freeman

May 2008
Chair: Shibu Jose
Major: Interdisciplinary Ecology

Under a natural fire regime, the longleaf pine (Pinuspalustris Mill.) ecosystem is

characterized by an open, longleaf pine-dominated canopy and a diverse, grass-dominated

understory. In the years since European colonization, the vast majority of longleaf pine acreage

has been cut over, fire-suppressed, and fragmented, which has resulted in extensive understory

invasion by shrubs and a concurrent decline in herbaceous species diversity. One of the biggest

challenges to successful restoration of this ecosystem is the persistence of shrubs in the

understory, which suppress longleaf pine seedlings as well as native herbaceous plants.

Herbicide can be used as a supplement to fire in order to enhance shrub control, but must be

studied carefully because of the potential for negative impacts on native plants. Information is

lacking about the effects of herbicide on natural longleaf pine flatwoods communities. We used

a banded application of three herbicides and one tank mix as shrub control treatments following

complete removal of the slash pine canopy and replanting with containerized longleaf pine

seedlings in a mesic-wet flatwoods.

The herbicides tested were Arsenal' (imazapyr), Oust' (sulfometuron methyl), Velpar L

(hexazinone), and an Oust' + Velpar L tank mix. Four years after application, no negative

impacts on understory species richness, diversity, evenness, or community composition were










Table 2-3. Species Richness (species/m2), Diversity (Shannon Index), and Evenness
(Simpson's Index, 1/D) 4 years after treatment (YAT) (June 2006) and following fire
(June 2007).
Species Richness (spp/m2) Shannon Diversity Evenness (1/D)
Overall Herbaceous Overall Herbaceous Overall Herbaceous
2006 Post- 2006 Post- 2006 Post- 2006 Post- 2006 Post- 2006 Post-
Treatment 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire 4YAT Fire
Control 7.36c 9.75* 4.58 6.83 1.34 1.67 1.10ab 1.59 3.54ab 4.24 3.02ab 4.36

Imazapyr 8.97ab 9.0 5.75 5.97 1.48 1.70 1.01ab 1.40* 3.51ab 4.43 2.45ab 3.67*

Sulfometuron 9.25a 8.19 5.41 5.05 1.57 1.51 1.17a 1.30 4.15a 3.80 3.03a 3.53

Sulfo +Hexa 8.03bc 9.0 5.13 6.14 1.34 1.69* 0.85b 1.40* 3.12b 4.54* 2.11b 3.79*

Hexazinone 8.22bc 8.86 5.33 5.94 1.35 1.65* 0.88b 1.45* 3.01b 4.07 2.13b 3.95*

ANOVA/ P= P= P= P P= P= P= P= P= P P P
ANCOVA 0.014 0.151 0.498 0.114 0.066 0.131 0.036 0.293 0.116 0.249 0.014 0.533
*Letters denote significant differences among means in the same column at a = 0.05 based
on Tukey-Kramer HSD test. (*) denotes significant post-fire change within treatment
(Tukey-Kramer HSD). P-values are for randomized block ANOVA / ANCOVA within
columns (significant at a = 0.05).









LIST OF FIGURES

Figure page

2-1 Ericaceae cover .............................. ........ .................. 41

2-2 O overall shrub cover ..................................... ............................. ............ .. 42

2-3 F ire effects on shrub cover .................................................... ...................................... 43

2-4 H erbaceous plant cover..................... ................................................ .......................... 44

2-5 Changes in cover of plant functional groups before and after fire ................................. 45

2-6 W iregrass cover ......... ................................ ...................... .. ......... .. ............ 46

2-7 Shrub stem density ..................................... ................. ............ .............. .. 47

2-8 Shannon D diversity ............................................................... .. .... ..... .. .............. 48

2-9 O ordination of plant com m unities ..................................................................... ........ 49

2-10 Relationship of fire temperature to dominant shrub species.......................................... 50

3-1 P ine seedling survive al ............. .................................................... ............................. 67

3-2 Post-fire pine seedling survival...................................................... .......................... 68

3-3 Rates of grass stage release by treatm ent................................... ...................... .. ........ 69

3-4 Pine seedling height ................................................................... ........ ..... 70

3-5 Stem volume index of grass stage and out of grass stage seedlings............................. 71









The mean fire temperatures we recorded were low for a dormant-season fire in a mesic

flatwoods: mean maximum fire temperatures for several other dormant season fires on

flatwoods sites ranged from 3830C to 5880C (Drewa et al. 2002), whereas our mean maximum

temperatures were 2000C at 80cm and 2740C at 30cm. Across the study site, we made the

general observation that I. glabra burned more readily and thoroughly than Quercus spp., which

often burned only partially, and, in some highly oak-dominated patches, failed to ignite at all.

Since dead pine needles are a critical component of pyrogenicity in longleaf pine systems

(Williamson and Black 1981, Clewell 1989), it is likely that the complete removal of overstory

pines contributed to our inability to apply fire effectively to the site.

In our study, stem counts of both Quercus spp. and I. glabra more than doubled in all

treatments in the four months following fire. Because we did not have an unburned control, we

cannot say to what degree the observed post-fire increases in shrub cover and density were

greater than the increases expected from normal spring growth of these shrubs in the absence of

overstory competition. However, it is safe to say that the prescribed fire was ineffective at

controlling the shrubs, because it did not decrease shrub cover or even maintain it at pre-fire

levels. This result can be attributed to aggressive shrub regrowth due to the lack of competition

from overstory pines, the loss of needlefall as a fine fuel supply, and the length of time since the

previous fire and subsequent herbicide applications (four and five years). Shrub control, and by

extension floral diversity, are maximized by fire return intervals of one to three years (Kirkman

et al. 2001; Glitzenstein et al. 2003). If future fires are conducted on a shorter interval, better

prescribed fire results may be obtained.

The season of burn may also have contributed to the difficulty we had in applying

prescribed fire effectively. It is generally believed that summer fires were probably the most

































2008 Johanna Freeman














80%
70% a
7 ab
60% b
50% -
40% C
30% -
o 20% -
10% -
0%
Control Imazapyr Sulfo Sulfo + Hexa Hexa
Treatment

Figure 3-3. Mean grass-stage release rate by treatment after five growing seasons, calculated as #
released / # surviving in each treatment row. Means not sharing the same letter are
significantly different (Tukey-Kramer HSD).
















AL
0


A m, A
0 AI





A A



x U
C;


t
U


A 0


*L,
A


* Control
* Imazapyr
Sulfo
+ Sulfo + Hexa
A Hexazinone


-1.5. *
-1 5 -05 05 15
Axis 1

Figure 2-9. NMDS ordination of plant communities. Each point represents one plant survey
quadrat. Chance-corrected within-group agreement (A) = 0.02 (a = 0.05, p = 0.01).



mm









pine canopy. The study included both overstory and understory components and had the

following objectives:

* To determine whether differences in the level of shrub control provided by sulfometuron
methyl (Oust'), hexazinone (Velpar L), a sulfometuron methyl + hexazinone tank mix, and
imazapyr (Arsenal') were evident in the fourth year after application;

* To determine whether differential treatment effects of sulfometuron, hexazinone, a
sulfometuron + hexazinone mix, and imazapyr on native understory species richness,
diversity, evenness, and community composition were evident in the fourth year after
application;

* To compare the effects of over-the-top applications of sulfometuron methyl, hexazinone, a
sulfometuron methyl + hexazinone mix, and imazapyr on the survival and growth of planted
longleaf pine seedlings;

* To observe how the herbicide treatment plots responded to a prescribed fire administered five
years after herbicide application.


The understory component of the study is described in Chapter 2 of this thesis, the overstory

component is described in Chapter 3, and overall conclusions are discussed in Chapter 4.









the sulfometuron treatment (Table 2-3). Herbaceous species richness, however, did not vary

significantly among treatments (p = 0.498) (Table 2-3). Following fire, overall species richness

no longer varied significantly among treatments (p = 0.15). Within treatments, responses to the

fire varied (Table 2-3). Overall species richness increased significantly in the control, but did

not change significantly in the other treatments. Herbaceous species richness did not vary

among treatments before or after fire (p = 0.498 and p = 0.114, respectively).

Species Diversity

Overall Shannon diversity (woody and herbaceous species combined) four years after

herbicide application did not vary significantly among treatments (p = 0.07) (Figure 2-8a, Table

2-3). However, herbaceous species diversity did vary significantly among treatments (p =

0.036), with the highest diversity in the sulfometuron treatment and the lowest in the sulfo +

hexa treatment (Table 2-3). Within treatments, responses to the prescribed fire varied (Figure 2-

8b). Overall diversity increased following fire in the sulfo + hexa and hexazinone treatments,

but did not change significantly in any of the other treatments (Figure 2-8b). Herbaceous

diversity increased in the imazapyr, sulfo + hexa, and hexazinone treatments following fire

(Table 2-3). Initial gains in overall Shannon diversity during the first year after treatment

(Ranasinghe 2003) were no longer evident after four years (Figure 2-8c).

Species Evenness

Mean species evenness (1/D) four years after herbicide application, calculated for woody

and herbaceous species combined, did not vary significantly among treatments overall (p = 0.12)

(Table 2-3). Herbaceous species evenness, however, did vary significantly among treatments (p

= 0.014), with the highest evenness in the sulfometuron treatment and the lowest in the sulfo +

hexa treatment (Table 2-3). Following fire, only the sulfo + hexa treatment experienced a

significant change in overall evenness (from 3.12 to 4.54) (Table 2-3) Herbaceous species









prescribed fire, which is still within the acceptable range of 60-75% recommended for adequate

overstory restocking (Brockway et al. 2006). We therefore recommend imazapyr for use in

sustainable longleaf pine silviculture, as it enhanced both longleaf pine seedling growth and

herbaceous cover in the understory.

Hexazinone, both alone and in a tank mix with sulfometuron, also showed promise as

both a longleaf pine release treatment and as an understory restoration tool. Four years after

treatment, hexazinone plots had significantly higher herbaceous cover than the control, and plots

treated with a sulfometuron + hexazinone tank mix had significantly higher wiregrass cover than

the control. Significant improvements in the grass stage release rate were achieved with both of

these treatments, and released seedlings in plots treated with the tank mix were significantly

taller than those in the control. An absolute comparison between these two formulations versus

imazapyr is not warranted because we used a very low rate of hexazinone. However, based on

success others have had using hexazinone for longleaf pine release (Haywood 2000, Ramsey and

Jose 2004, Haywood 2005) and restoration (Brockway et al. 1998, Brockway and Outcalt 2000,

Provencher et al. 2001) we can expect that at a higher rate-both alone and in the tank mix-this

herbicide would have compared more favorably with imazapyr.

Sulfometuron methyl did not improve pine release relative to the control, and instead had

a releasing effect on the two most common oak species (Quercus minima and Quercus elliottii),

indicating that this is not an appropriate herbicide for flatwoods with a high density of oaks.

However, in spite of having higher shrub cover and lower herbaceous cover than the other

herbicide treatments, sulfometuron plots also had the highest herbaceous species richness,

diversity, and evenness four years after initial treatment. One reason for this may have been that

wiregrass cover in the sulfometuron plots was somewhat lower than the other herbicide









degree of agreement within groups with regard to an environmental variable, which in this case

was a categorical variable (treatment). The weighting method for the MRPP was n/sum(n), and

we again used a Sorenson distance measure. MRPP yields a test statistic (A) and a p-value (A =

1 when all items are identical within groups (the groups in this case were herbicide treatments),

and A=0 when heterogeneity within groups equals expectation by chance). A therefore expresses

the degree to which variation in position between plant survey quadrats on the ordination axes is

"explained" by the herbicide treatments. Indicator species analysis using the Indicator Value

method defines an indicator species as a species found mostly in a single group and present in

the majority of sites belonging to that group (DuFrene and LeGendre 1997). The indicator value

(IV) is maximum (100%) when the individuals of species I are observed in all sites of only one

group (in this case, the each "group" was an herbicide treatment). IV is obtained by combining

relative abundance and relative frequency values for each species in each group. Relative

abundance of a species in a group is the average abundance of a given species in a given group

of plots over the average abundance of that species in all plots (expressed as a percent), and

relative frequency is the average number of plots in a given group where a given species is

present. Significance of each Indicator Value was established using a 1000-permutation MRPP

which compared the observed maximum IV for each species to the IV obtained from randomized

groups, and generated a p-value for this comparison using the Monte Carlo test of significance.

Species identified as indicators for one of the herbicide treatments are assumed to have been

enhanced by that treatment relative to the others, whereas species identified as indicators for the

control are assumed to have been negatively impacted by the herbicide treatments.









BIOGRAPHICAL SKETCH

Johanna Freeman is a native of western New York state. She received a B.S. in natural

resources from Cornell University in 2001, where her studies focused on animal behavior and

environmental policy. She worked as a land-use planner at the New York City Department of

Parks and Recreation for three years prior to entering the masters program in interdisciplinary

ecology at the University of Florida.






















Control Imazapyr Sulfo Sulfo+Hexa
Treatment


Control


Imazapyr


Sulfo


Sulfo + Hexa


Non-Ericaceae
5 Pre-Fire
a Post-Fire


Hexa


Ericaceae
O Pre-Fire
* Post-Fire


Hexa


Treatment


Figure 2-3. Mean percent shrub cover before (June 2006) and after fire (June 2007) for a)
Quercus spp., I. glabra, and S. repens, and b) Ericaceae. Differences between and
within treatments are not significant on either graph.


0

a)


30
25
20
b) S 15
10


L Ai I FT









herbaceous cover no longer differed significantly among treatments (p = 0.302). Herbaceous

cover decreased significantly following fire in all but the control and sulfometuron plots (Figure

2-4b). Much of this decrease can be attributed to the decrease in wiregrass crown size following

fire (see below). Overall vegetation cover before and after fire is summarized in Figure 2-5.

Wiregrass Cover

Mean wiregrass cover four years after initial herbicide application differed significantly

among treatments (p = 0.04) (Figure 2-6a). Wiregrass cover in the sulfo + hexa treatment was

three times greater than the control (Figure 2-6). In all treatments except the control, wiregrass

cover 4 years after treatment was higher than pre-treatment levels (Figure 2-6c). Following fire,

wiregrass cover decreased in all treatments (Figure 2-6b). The overall decrease in wiregrass

cover following fire is not attributed to wiregrass mortality, but to the decrease in wiregrass

crown size following fire. Wiregrass cover was surveyed only four months after the February

2007 fire, and it was observed that individual wiregrass clumps were alive and regenerating, but

greatly reduced in size from pre-fire levels. A similar effect was documented after the herbicide

treatment in 2002, with an initial decrease in wiregrass cover followed by a rebound over the

next nine months (Figure 2-6c).

Shrub Stem Density

Four years after treatment, mean shrub stem density for Quercus, Ilex, and Serenoa

varied significantly among treatments (p = 0.012), displaying a similar pattern to percent cover

(Table 2-1). Mean stem density in the imazapyr treatment (20 stems/m2) was significantly lower

than sulfometuron (34.7 stems/m2) and hexazinone (32.9 stems/m2). None of the treatments

varied significantly from the control, in which we measured a mean stem density of 31.75

stems/m2. In the first nine months after herbicide application, imazapyr was the only treatment

that did not experience an increase in shrub stem density (Figure 2-7b). In the fourth year after









The reasons for using a banded post-planting herbicide application on unshielded

seedlings were both operational and environmental. Prior to harvest, the study site already

contained many desirable understory species (including wiregrass) and only a few weedy

species, none of which were exotic (Ranasinghe 2003). The banded application allowed us to

pinpoint the delivery of herbicide to an area directly around planted seedlings, while native

vegetation remained intact in the strips between rows. This application method also lowered the

cost of herbicide treatment relative to a broadcast application.

The herbicides were applied in a 1.2m band over the seedling rows using a backpack

sprayer. Within each 10-row treatment plot, three rows received a second herbicide application

in March of 2003, three rows received a second application in April of 2003, and four rows were

left with only the first-year treatment. The results from these subplots were pooled with the first

year data for most parameters because of a lack of significant differences between one and two-

year applications.

Data Collection and Analysis

In November 2006, the root collar diameter (RCD) and height to the top of bud were

measured on every seedling. Following a prescribed fire in February of 2007 (Chapter 2), post-

fire seedling mortality was assessed in June 2006. Survival rate was calculated as a percentage

of each 10-tree, 56m2 treatment row. Rate of grass-stage release was calculated as a percentage

of surviving seedlings in each treatment row with a bud height greater than 12cm, following

Haywood (2000). Height and RCD comparisons were made separately for seedlings in the grass

stage (GS) and out of the grass stage (OOGS). The study addressed only the main effects of

herbicide treatment, and tests of these effects were not dependent on the assumption of no

treatment x block interaction. Block effects were therefore treated as random effects in a

univariate ANOVA model with two independent variables: 'treatment' with 'Block&Random' as









Forestry since its acquisition of Point Washington State Forest in 1992. A pre-harvest, baseline

data survey was conducted in the areas slated to become study blocks in June 2001. In August

2001, the entire overstory was harvested and the site was prepared for the study with a single,

light roller chopping followed by a prescribed burn in October 2001. In December 2001, six

study blocks, each divided into five 36.6m x 24.4m treatment plots, were established according

to the aforementioned randomized complete block design. Each treatment plot was hand-planted

with 100 containerized longleaf pine seedlings laid out in 10 rows, with 3.1m between rows and

1.8m between the seedlings in each row.

Treatments

Within each block, each 100-seedling plot was randomly selected to receive one of five

herbicide treatments in March of 2002:

1) Sulfometuron methyl (Oius ") at 0.26 kg a.i./ha. This is a moderate application rate

for sulfometuron methyl; other studies have used 0.16 kg a.i./ha (Lauer and Glover, 1998) and

0.21 kg ai/ha (Shiver and Martin 2002, Keyser and Ford 2006), while the manufacturer-

recommended rate for herbaceous weed control in longleaf pine stands is 0.10 to 0.42 kg a.i./ha

(DuPont 2007). Sulfometuron methyl is a selective herbicide primarily effective for controlling

herbaceous species (Lauer and Glover, 1998), and mixing with hexazinone is recommended if

broader spectrum control is desired (DuPont, 2007).

2) Hexazinone (Velpar L) at 0.56 kg a.i./ha. Recommended application rates for

hexazinone range from 2.0 to 6.7 ai kg/ha (Du Pont 2007), while rates successfully used to

control hardwoods in sandhills ecosystems have ranged from 1.1 kg a.i./ha to 2.4 kg a.i/ha

(Brockway et al. 1998, Provencher et al. 2001). The hexazinone application rate in our study

was therefore very low. This rate was chosen because hexazinone had been shown to be









tolerant species, and allows a host of herbaceous understory species to survive and reproduce in

the open, high-light understory (Williamson and Black 1981, Clewell 1989, Rebertus et al. 1993,

Jacqmain et al. 1999).

With high productivity and an abundance of shrubs in the understory, flatwoods are likely

to lose herbaceous species more quickly with fire exclusion than drier longleaf pine sites. Even

slight fire suppression in a flatwoods system leads to increased sprouting and vegetative

reproduction among shrubs, reducing available space for herbaceous plants and lowering species

richness (Glitzenstein et al. 2003). Different plant functional groups experience different rates of

species loss with reduced fire frequency. Mesic sites, with more species to lose and more rare

species, are likely to experience the highest rates of species loss (Leach and Givnish 1996). In

mesic to wet longleaf pine savannas, the species most likely to be reduced or lost are the

dominant bunch grasses, basal rosette species, sedges, other small monocots, and insectivorous

species (Walker and Siletti 2006).

Because the understory species composition and soil properties vary markedly with

hydrology in longleaf pine ecosystems, it is necessary to test restoration methods in a variety of

site conditions in order to develop appropriate restoration protocols. Mesic flatwoods in the

Florida panhandle are generally co-dominated by wiregrass (A. stricta, saw palmetto (Serenoa

repens W. Bartram), runner oak (Quercus elliottii Wilbur), gallberry (Ilex glabra A. Gray), and

hairy wicky (Kalmia hirsuta Walter). The overstory in a mesic longleaf pine flatwoods is

typically an open canopy of longleaf pine (basal area of 2.5 to 12.5m2/ha), with slash pine (Pinus

elliottii Engelm.) and pond pine (Pinus serotina Michx.) codominant in the transitional areas

near wetlands (Peet 2006). Fire is a key management tool for meeting restoration goals in these









Seedlings in this treatment experienced a fire-related mortality rate of only .06%, whereas fire

mortality was around 3% in all of the other treatments. As a result, the differences in survival

rate between imazapyr and the other treatments were no longer significant following fire.

Though the imazapyr treatment provided the best overall pine release results, both

hexazinone and the sulfometuron + hexazinone tank mix also showed promise as over-the-top

longleaf pine release treatments for flatwoods ecosystems. Significant improvements in the

grass-stage release rate were achieved with both of these treatments, and the OOGS seedlings in

plots treated with the tank mix were significantly taller than those in the control. An absolute

comparison between the effectiveness of these two formulations versus imazapyr is not

warranted because we used a very low rate of hexazinone. However, based on success others

have had using hexazinone for longleaf pine release (Haywood 2000, Ramsey and Jose 2004,

Haywood 2005), we can expect that at a higher rate-both alone and in the tank mix-this

herbicide would have compared more favorably with imazapyr. Contrary to the results

previously observed with longleaf pine seedlings planted in an old field (Ramsey and Jose 2004),

sulfometuron alone did not provide any pine release benefits over the control. This was most

likely due to the higher degree of woody competition on this site, against which sulfometuron

was significantly less effective than the other treatments (Chapter 2).

For grass stage seedlings, variation among treatments was much less pronounced. Grass

stage seedlings in the sulfometuron treatment were significantly smaller than those in the control,

but none of the other treatments differed significantly from the control in any of the grass stage

parameters. A possible explanation for this pattern is that as vegetation regrew in the years

following application, any seedlings not released relatively early in the process were suppressed

and remained in the grass stage. As differences in understory vegetation between treatments









Outcalt et al 1999, Miller et al 1999). The same procedure was repeated in June of 2007,

following a prescribed fire on February 16, 2007.

Pyrometers were constructed using aluminum forestry tags and Tempilaq heat-sensitive

paints, following a method employed in other fire ecology studies (Hobbs et al. 1984, Iverson et

al. 2004). Thirteen Tempilaq paints (designed to melt at 930 C, 1490C, 2040C, 260C, 3160C,

3710C, 4270C, 4820C, 5380C, 593C, 6490C, 7040C, 760C, and 871C) were applied to each

tag. A second tag was paper-clipped over the front of the first to protect the paints from

charring. Six stakes were randomly placed in each herbicide treatment plot (two in each subplot)

and two in each control, with two pyrometers attached to each stake at 30cm and 80cm, for a

total of 36 pyrometers at each height for each herbicide treatment and 12 at each height for each

control. Each pyrometer gave a measure of the maximum fire temperature at that location: all of

the paint dots below that temperature were melted, while all of the paint dots above it remained

unchanged. When analyzing species cover vs. temperature relationships, the two pyrometer

readings at the 30cm level in each subplot were averaged and plotted against the average of the

two understory quadrats in that subplot. Since the pyrometers were placed randomly, they were

not necessarily in the same locations as the quadrats. Averaging them gave a more accurate

overall picture of the cover and fire intensity in a given plot. Pyrometers on which none of the

paints were burned (i.e. temperatures did not reach 930C) were assumed to be unburned and were

assigned a value of 100C to represent ambient air temperature.

All parameters were analyzed in JMP IN version 5 (SAS Institute, Inc.), using ANOVA

within the framework of a randomized complete block design. The study addressed only the

main effects of herbicide treatment, and tests of these effects were not dependent on the

assumption of no treatment x block interaction. Block effects were therefore treated as random









systems, but social and environmental factors can often hinder the ability of land managers to use

fire effectively (Chapter 1).

When a restoration project includes conversion of the overstory from a planted slash pine

or loblolly pine (Pinus taeda L.) stand, complete removal of overstory pines across a large area

has the counterproductive effect of releasing shrubs (Brockway et al. 1998, Kush et al. 1999,

McGuire et al. 2001, Kirkman et al. 2007, Pecot et al. 2007). The release of shrubs following

overstory harvest is likely to be more extensive on a site with a past history of fire suppression,

due to the preponderance of underground hardwood rootstocks established during periods

without fire (Olson and Platt 1995; Drewa et al. 2006). Herbicide treatment and mechanical site

preparation can therefore be important components of restoration projects in which a slash or

loblolly pine plantation is being converted to longleaf (Walker and Siletti 2006). A potential

negative impact of herbicide, however, is a reduction in nontarget, desirable understory

vegetation.

Litt et al. (2001) conducted an extensive literature review on the effects of herbicide on

understory vegetation in southern pinelands. They examined 125 studies from which 21 were

used for review. Throughout the herbicide literature, they found a lack of experimental rigor and

inconsistent reporting standards, which made it difficult to reach any general conclusions about

the effects of herbicide on ground-layer vegetation. They were able to conclude only that

widespread use of herbicide to control unwanted vegetation may have undesirable effects on

nontarget plant species, and that additional studies of herbicide impacts are needed before

treating large, diverse landscapes. They recommend that future herbicide studies follow rigorous

experimental designs and make a deliberate effort to track species of special concern.









especially imazapyr, released pines while also maintaining or improving wiregrass cover and

overall cover of native herbaceous species. While tank mixes have been widely studied in

silvicultural applications, they have understandably been avoided as habitat treatments thus far,

out of concern that a broad spectrum mix will negatively impact native species (Keyser and Ford

2006). However, our results indicate that this tank mix not only released pines, but had positive

impacts on the understory, which might have been even more significant if a higher level of

hexazinone had been included in the mix (Chapter 2). An important caveat to this conclusion is

that the use of a banded application method may have been a key to the positive result. On pine

plantations in the Virginia piedmont, broadcast application of an imazapyr + sulfometuron

methyl tank mix significantly lowered herbaceous species diversity relative to the control and to

imazapyr-only treatments, and banded application was suggested as an alternative for sites with

wildlife habitat concerns (Keyser and Ford 2006).

Recent studies of sustainable silviculture in longleaf pine systems have cited the potential

negative impacts of herbicide as a justification for the use of group or single-tree selection, since

these harvesting methods should enhance the effectiveness of prescribed fire as a management

alternative to herbicide (Mitchell et al. 2006, Pecot et al. 2007, Kirkman et al. 2007). However,

we observed only beneficial impacts on understory species following chemical treatment,

indicating that herbicide can in fact play a positive supporting role in conservation-oriented

longleaf pine silviculture. Based on the results of this study, we recommend banded application

of imazapyr as an ecological restoration tool for flatwoods sites with aggressive understory

shrubs.
















800 -

700 a

S600 -

500 be

400 -

300 I

200

100
Control Imazapyr Sulfo Sulfo + Hexa Hexazinone
Treatment

Figure 3-4. Mean height (mm) of out of grass stage seedlings by treatment. Means not sharing
the same letter are significantly different (Tukey-Kramer HSD, a = 0.05)









effective at the aforementioned rates, and we were interested in whether an even lower

application rate would be effective.

3) Sulfometuron methyl (0.26 kg a.i./ha) + Hexazinone (0.56 kg a.i./ha) mix. This is a

common tank mix used throughout the Southeast (Shepard et al. 2004) with the potential to

control a broad spectrum of woody and herbaceous plants. Tank mixes are used to achieve more

complete vegetation control than either herbicide alone, but as a result they may also have more

negative implications for native plant communities (Keyser and Ford 2006). Banded herbicide

application, the method used in this study, may help ameliorate these negative impacts (Keyser

and Ford 2006).

4) Imazapyr (Arsenal) at 0.21 kg a.i./ha. For a restoration project, this was a

moderate application rate for imazapyr. Significant hardwood control and habitat improvements

have been obtained using imazapyr at 0.42 kg a.i./ha (Jones and Chamberlain 2004), and at all

rates from 0.08 kg a.i./ha to 0.24 kg a.i..ha (though higher rates did not offer significant

improvements over lower rates) (Keyser and Ford 2006).

5) Control. (No herbicide).

The herbicides were applied in a 1.2m band over the seedling rows using a backpack

sprayer. The reasons for using a banded post-planting herbicide application on unshielded

seedlings were both operational and environmental. Prior to harvest, the study site already

contained many desirable understory species (including wiregrass) and few weedy species

(Ranasinghe 2003). The banded application allowed us to pinpoint the delivery of herbicide to

an area directly around planted seedlings, while native vegetation remained intact in the strips

between rows. This application method also lowered the cost of herbicide treatment relative to a

broadcast application. Within each 10-row treatment plot, three rows received a second
















35 0 OPre-Harvest
S30 4 yrs post-treatment
0 25
20
15
10



Control Imazapyr Sulfo Sulfo + Hexa Hexazinone
Treatment

Figure 2-1. Pre-harvest (June 2001) and June 2006 percent cover of Ericaceae by treatment.
Pre-harvest Ericaceae cover was not uniform among treatment plots (ANOVA, p =
0.009). 4 years post-treatment, variance in Ericaceae cover by treatment approached
significance (ANCOVA, p = 0.060).









seedling growth and enhance herbaceous cover in the understory, especially on sites where

growing-season fire is not an option.









Methods


Study Area

The study was conducted at the Point Washington State Forest, which is located on the

gulf coast of Walton County, Florida (300 20'16" N, 860 04'19" W). The 20-acre study site is

part of a larger mesic flatwoods/wet flatwoods matrix, with an understory dominated by saw

palmetto (S. repens), gallberry (I. glabra), rhizomatous oaks (Quercus minima Small and

Quercus elliottii Wilbur), blueberries (Vaccinium myrsinites Lam. and V darrowii Camp.),

huckleberries (Gaylussacia dumosa Torr. and A. Gray and G. frondosa L. Torr and A. Gray ex.

Torr), bluestem grasses (Andropogon virginicus var. glaucus (L.) Heck and Andropogon

virginicus L.) and wiregrass (Aristida stricta. The majority of the site can be classified as

mesic, but due to subtle topographical variation across the site some areas are at the wetter end of

the mesic flatwoods spectrum and some are at the drier end. For this reason, a randomized

complete block design was used, with three blocks laid out in mesic-wet areas and three blocks

laid out in mesic-dry areas. A pre-treatment vegetation survey identified wetter areas by the

presence of indicator species such as sedges (Cyperaceae). The soils of the study site are sandy,

siliceous, thermic, aeric alaquods of the Leon series (Spodosols consisting of deep, poorly to

very poorly drained soils derived from marine parent material (Ranasinghe, 2003)). Yearly

weather patterns during the study period were variable, with hurricanes in some years and

droughts in others. Total annual precipitation during the study period was 103cm in 2002,

200cm in 2003, 135cm in 2004, 146cm in 2005, and 106cm in 2006 (NOAA 2007). Rainfall in

the months following the prescribed fire was very low, with totals of 0.5cm in March 2007, 4cm

in April 2007, and 0.3cm in May 2007 (NOAA 2007).

Prior to August 2001, the 20-acre study site was a planted slash pine stand with a mean

stand age of about 26 years, which had been burned on a 3-year rotation by the Division of













90.0%
80.0%
70.0%
2 60.0%
50.0% -O Pre-Fire
S40.0% Post-Fire
S30.0%
t 20.0%
10.0%
0.0%
Control Imazapyr Sulfo Sulfo+Hexa Hexa
Treatment

Figure 3-2. Percent seedling survival (trees/row), before and after fire. None of the treatments
experienced significant mortality due to fire (Tukey-Kramer HSD, a = 0.05).
















90%
a a
5 80% a

S70% b
60%

50%
40%

30% -
20% .
Control Imazapyr Sulfo Sulfo + Hexa Hexa
Treatment
Figure 3-1. Mean survival rate by treatment after five growing seasons, calculated as # surving /
# planted in each treatment row. Means not sharing the same letter are significantly
different (Tukey-Kramer HSD).











Table 2-2. Indicator species analysis following the method of Dufrene and LeGendre (1997).
IV from randomized groups
Observed
Species Treatment Indicator Value (IV) Mean S. Dev P*
Asclepias cineria Sulfo + Hexa 15.5 5.4 2.91 0.010
Pteridium aquilinum Sulfometuron 20.3 11.7 3.4 0.020

Lachnocaulon auceps Imazapyr 12.2 5 2.86 0.025

Lachnanthes caroliniana Hexazinone 16.7 4.1 2.5 0.010

Quercus elliottii Sulfometuron 19.1 12.6 3.27 0.040

Quercus minima Sulfometuron 29.7 21.7 3.56 0.031

Smilaxpumila Control 12.5 3.7 2.39 0.029

* Monte Carlo test of significance, a =0.05









and Ford 2006). Sulfometuron methyl and hexazinone are effective as over-the-top pine release

applications on longleaf pine seedlings (Haywood 2000, Ramsey and Jose 2004, Haywood

2005), but effects of a mixed application on longleaf seedlings or on the understory of a natural

flatwoods system have not been studied. Imazapyr mixed with triclopyr was effective as a pre-

planting application to release longleaf pine seedlings on a poorly-drained site (Knapp et al.

2006), but information on imazapyr as an over-the-top release treatment for longleaf is also

limited.

We examined the longer-term effects of these herbicide treatments on longleaf pine

seedlings five growing seasons after initial application (see Ranasinghe 2003 for first year

results). We then conducted a dormant-season prescribed fire on all treatment plots and assessed

post-fire survival at the beginning of the sixth growing season.

Methods

Study Area

The study was conducted at the Point Washington State Forest, which is located on the

gulf coast of Walton County, Florida (300 20'16" N, 860 04'19" W). The 20-acre study site is

part of a larger mesic flatwoods/wet flatwoods matrix, dominated by saw palmetto (Serenoa

repens W. Bartram), gallberry (Ilex glabra A. Gray), rhizomatous oaks (Quercus minima Small

and Q. elliottii Wilbur), blueberries (Vaccinium myrsinites Lam. and V. darrowii Camp.),

huckleberries (Gaylussacia dumosa Torr. and A. Gray and G. frondosa L. Torr and A. Gray ex.

Torr), bluestem grasses (Andropogon virginicus var. glaucus (L.) Heck and Andropogon

virginicus L.) and wiregrass (Aristida stricta). The majority of the site can be classified as

mesic, but due to subtle topographical variation across the site some areas are at the wetter end of

the mesic flatwoods spectrum and some are at the drier end. For this reason, we used a

randomized complete block design, with three blocks laid out in mesic-wet areas and three




Full Text

PAGE 1

FROM PINUS ELLIOTTII PLANTATION TO PINUS PALUSTRIS ECOSYSTEM: THE ROLE OF HERBICIDE IN LONGLEAF PINE RESTORATION By JOHANNA E. FREEMAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Johanna Freeman 2

PAGE 3

ACKNOWLEDGMENTS I would like to thank Dr. Shibu Jose, my s upervisory committee chair, for the opportunity to conduct this research, and for hi s cheerful support and advice. I would also like to thank Drs. Alan Long and Debbie Miller for serving on my supervisory committee, providing instruction on the identification and ec ology of North Florida vegetation, and assisting me in developing fire ecology research methods. Thanks also go to Dr. Eric Holzmueller for extensive assistance with field data collection and anal ysis throughout the pr oject; to Don Hagan, Michael Morgan, and Kelly Thayer for assisting me with field data co llection; and to Brian Hinton for assistance with fire methods. I thank the University of Florid a herbarium staff, Dr. Walter Judd, and especially Dr. Susan Carr for invaluable help with the id entification of plant sp ecimens. I also thank Meghan Brennan of the IFAS statistical consulti ng unit, who assisted me greatly with data analysis. Thanks go to the Florida Division of Forestry for funding this project and providing the study site, and especially to site administrator Tom Beitzel and hi s crew for their assistance in collecting fire temperature data. Finally, I woul d like to thank my parents and David Kaplan for their constant support and encouragement. 3

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS .............................................................................................................. 3 LIST OF TABLES .......................................................................................................................... 6 LIST OF FIGURES ........................................................................................................................ 7 ABSTRACT .................................................................................................................................... 8 CHAPTER 1 INTRODUCTION .................................................................................................................. 10 2 USING HERBICIDE TO RESTORE THE UNDERSTORY FOLLOWING HARVEST .... 13 Introduction ............................................................................................................................. 13 Methods ................................................................................................................................... 18 Study Area ........................................................................................................................ 18 Treatments ......................................................................................................................... 19 Data Collection and Analysis ............................................................................................ 21 Results ..................................................................................................................................... 25 Shrub Cover ...................................................................................................................... 25 Herbaceous Cover ............................................................................................................. 25 Wiregrass Cover ................................................................................................................ 26 Shrub Stem Density .......................................................................................................... 26 Shrub height ...................................................................................................................... 27 Understory responses ........................................................................................................ 27 Species Diversity .............................................................................................................. 28 Species Evenness .............................................................................................................. 28 Community Ordination ..................................................................................................... 29 Species of Special Interest ................................................................................................ 29 Fire Temperature ............................................................................................................... 29 Discussion ............................................................................................................................... 30 3 HERBICIDE AND SUSTAINABLE L ONGLEAF PINE SILVICULTURE ........................ 51 Introduction ............................................................................................................................. 51 Methods ................................................................................................................................... 57 Study Area ........................................................................................................................ 57 Treatments ......................................................................................................................... 58 Data Collection and Analysis ............................................................................................ 60 4

PAGE 5

Results ..................................................................................................................................... 61 Seedling Survival .............................................................................................................. 61 Percent Out of Grass Stage ............................................................................................... 61 Growth: Out of Grass Stage Seedlings ............................................................................. 62 Growth: Grass Stage Seedlings ........................................................................................ 62 Discussion ............................................................................................................................... 62 4 SUMMARY AND CONCLUSIONS ..................................................................................... 72 APPENDIX SPECIES LIST........................................................................................................................ 76 LIST OF REFERENCES .............................................................................................................. 78 BIOGRAPHICAL SKETCH ........................................................................................................ 85 5

PAGE 6

LIST OF TABLES Table page 2-1 Shrub cover, stem density, and height. ............................................................................. 37 2-2 Indicator species analysis .................................................................................................. 38 2-3 Understory species rich ness, diversity, and evenness ....................................................... 39 2-4 Mean maximum fire temperature (C) by treatment ......................................................... 40 3-1 Pine seedling growth after five seasons ............................................................................ 66 6

PAGE 7

LIST OF FIGURES Figure page 2-1 Ericaceae cover ........................................................................................................ ....... 41 2-2 Overall shrub cover. .......................................................................................................... 42 2-3 Fire effects on shrub cover. ............................................................................................... 43 2-4 Herbaceous plant cover ................................................................................................... .. 44 2-5 Changes in cover of plant func tional groups before and after fire ................................... 45 2-6 Wiregrass cover ................................................................................................................ 46 2-7 Shrub stem density ....................................................................................................... ..... 47 2-8 Shannon Diversity ............................................................................................................. 48 2-9 Ordination of plant communities ...................................................................................... 49 2-10 Relationship of fire temperat ure to dominant shrub species. ............................................ 50 3-1 Pine seedling survival. ...................................................................................................... 67 3-2 Post-fire pine seedling survival. ........................................................................................ 68 3-3 Rates of grass stage release by treatment. ......................................................................... 69 3-4 Pine seedling height .......................................................................................................... 70 3-5 Stem volume index of grass st age and out of grass stage seedlings ................................. 71 7

PAGE 8

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FROM PINUS ELLIOTTII PLANTATION TO PINUS PALUSTRIS ECOSYSTEM: THE ROLE OF HERBICIDE IN LONGLEAF PINE RESTORATION By Johanna E. Freeman May 2008 Chair: Shibu Jose Major: Interdisciplinary Ecology Under a natural fire regime, the longleaf pine ( Pinus palustris Mill.) ecosystem is characterized by an open, longleaf pine-dominated canopy and a diverse, grass-dominated understory. In the years since Eu ropean colonization, the vast majo rity of longleaf pine acreage has been cut over, fire-suppressed, and fragmented, which has resulted in extensive understory invasion by shrubs and a concurrent decline in herbaceous species diversity. One of the biggest challenges to successful restoration of this eco system is the persistence of shrubs in the understory, which suppress longleaf pine seedlin gs as well as native herbaceous plants. Herbicide can be used as a supplement to fire in order to enhance shrub control, but must be studied carefully because of the potential for ne gative impacts on native plants. Information is lacking about the effects of herbicide on natural longleaf pine flatwoods communities. We used a banded application of three he rbicides and one tank mix as sh rub control treatments following complete removal of the slash pine canopy and replanting with contai nerized longleaf pine seedlings in a mesic-wet flatwoods. The herbicides tested were Arsenal (imazapyr), Oust (sulfometuron methyl), Velpar L (hexazinone), and an Oust + Velpar L tank mix. Four years after application, no negative impacts on understory species richness, diversity, evenness, or community composition were 8

PAGE 9

evident in any of the herbicide treatments. Oaks ( Quercus spp.), one of the dominant shrub genera on the study site, were resi stant to sulfometuron methyl, a nd this herbicide was therefore ineffective both as a pine release treatment and for enhancing herbaceous species growth. Imazapyr was the most effective treatment ove rall, significantly improving longleaf pine seedling growth and also enhancing herbaceous species cover. Both hexazinone and the hexazinone + sulfometuron methyl tank mix pr ovided some seedling growth and understory enhancement as well, though these effects were not as pronounced as those in the imazapyr treatment. Shrubs resprouted a ggressively after a dormant-season pr escribed fire in the fifth year after treatment, indicating that herbicide-relate d increases in cover of wiregrass and other herbaceous species may be lost in future fire cycles. Retention of overstory pines for shrub competition and as a source of fine fuels is reco mmended in addition to herbicide for sustainable silviculture in longleaf pine flatwoods. 9

PAGE 10

CHAPTER 1 INTRODUCTION Two hundred years ago, longleaf pine ( Pinus palustris Mill.) ecosystems covered an estimated 37 million hectares of the southeastern United States (Frost 2006). Based on fuel accumulation patterns and plant life hi stories it is believed that light to moderate intensity surface fires burned throughout the range of the longleaf pine every two to three years prior to European colonization (Frost 1996). These frequent fire s maintained an open, longleaf pine-dominated canopy and a grass-dominated understory contai ning some of the most species-rich plant communities ever documented outside the tropics (P eet and Allard 1993). However, in the years following European colonization, fires were supp ressed and old-growth longleaf pine systems were logged almost out of existence. Toda y, it is estimated that only 841,000 hectares of naturally regenerating longleaf pine systems still remain, the great majority of which are second growth and fire-suppressed (Frost 2006). This area contains only about 5,095 hectares of oldgrowth longleaf pine (Varner and Kush 2004). Fire suppression has resulted in extensive understory invasion by hardwood shrubs and a conc urrent regional decline in herbaceous species diversity (Walker and Peet 1983, Mehlman 1992, Gilliam and Platt 1999). In 1995, a U.S. Biological Survey report liste d the longleaf pine ecosystem as the third most endangered ecosystem in the United States (Noss et al. 1995). In recent years, recognition of the value of the longleaf pine ec osystem has motivated widespread restoration efforts throughout the southeast. One of the biggest obstacles to successful longleaf pine ecosystem restoration is the pe rsistence of hardwoods in the understory even after the rein troduction of fire (Walker and Siletti 2006 ). In some cases the reintroduction of fire alone may restore the de sired two-layered savanna struct ure (Kush et al. 1999), but many factors can limit the efficacy of fire after a long period of suppr ession, including insufficient fine 10

PAGE 11

fuels, presence of ladder fuels that may cause damage to crowns, and duff accumulation that can kill overstory trees when ignited (Walker and Siletti 2006). Growth of shrub rhizomes and root systems during periods of fire suppression can be extensive, making these shrubs recalcitrant even when fire is reintroduced (Drewa et al 2006). Additionally, social factors such as proximity to residential areas or highways can limit the ability of land managers to use prescribed fire effectively. Therefore, mechan ical and herbicidal hardwood removal techniques are often considered as supplements to prescr ibed fire (Walker and Siletti 2006). Hardwood control is also of key importance for restorati on projects that include planting of longleaf pine seedlings, which are very intolerant of competition and can quickly succumb to competing vegetation while in the gr ass stage (Harrington 2006). Herbicide, used in conjunction with prescrib ed fire, can be an eff ective tool for reducing midand understory hardwoods in longleaf pine systems without negatively impacting native understory species (Brockway et al 1998, Broc kway and Outcalt 2000, Provencher et al. 2001). However, most studies of herbicide as a restora tion tool in longleaf pine systems have been conducted on xeric longleaf pine si tes (Litt et al. 2001). Relative to xeric longleaf pine sandhills, mesic and wet longleaf pine flatwoods are more productive, have different soil characteristics, and are associated with differe nt understory plant communities Understory responses to herbicide in these systems are therefore likely to differ from those documented in sandhills (Litt et al. 2001). At present, there is only limited information about herbicidal shrub control in mesic to wet longleaf pine ecosystems. The purpose of this research was to test the effects of different herbicide treatments, used in conjunction with prescribed fire, on target and nontarget vege tation in a flatwoods ecosystem undergoing conversion from a planted slash pine ( Pinus elliottii Engelm.) overstory to a longleaf 11

PAGE 12

pine canopy. The study included both overstory and understory components and had the following objectives: To determine whether differences in the leve l of shrub control provided by sulfometuron methyl (Oust), hexazinone (Velpar L), a sulfometuron methyl + hexazinone tank mix, and imazapyr (Arsenal) were evident in the fourth year after application; To determine whether differential treatm ent effects of sulfometuron, hexazinone, a sulfometuron + hexazinone mix, and imazapyr on native understory species richness, diversity, evenness, and community composition were evident in the fourth year after application; To compare the effects of over-the-top applic ations of sulfometuron methyl, hexazinone, a sulfometuron methyl + hexazinone mix, and imaza pyr on the survival and growth of planted longleaf pine seedlings; To observe how the herbicide treatment plots res ponded to a prescribed fi re administered five years after herbicide application. The understory component of the study is describe d in Chapter 2 of this thesis, the overstory component is described in Chapter 3, and overa ll conclusions are discussed in Chapter 4. 12

PAGE 13

CHAPTER 2: USING HERBICIDE TO RESTORE THE UNDERSTORY FOLLOWING HARVEST Introduction Longleaf pine ( Pinus palustris Mill.) ecosystems occupy a wide environmental gradient in the southeastern United States, from xeric sandhills to mesic and wet flatwoods (Walker and Peet 1983, Peet 2006). Under a natural fire regime the understory vegetati on associated with the two dominant species, longl eaf pine and wiregrass (Aristida stricta Michx.), varies predictably with site hydrology (Harcombe et al. 1993). Me sic longleaf pine flatwoods, which occur on Ultisols and Spodosols, fall in th e middle of the longleaf pine hydr ologic gradient and are highly productive plant communities (Peet 2006). Whereas the understories of fre quently burned xeric longleaf pine sandhills are truly dominated by gra sses and herbaceous plants, in mesic flatwoods a large shrub component is present even under na tural fire conditions (Peet 2006). These shrubs are kept low to the ground by frequent fire, but in the absence of fire th ey will quickly succeed into a dense midstory, eventually suppressing both longleaf pine and herbaceous plant regeneration (Gilliam and Platt 1999). Wiregrass is a key structural and functional component of all longleaf pine ecosystems in the Southern Coastal Plain eco region (Peet 2006), and the estab lishment or enhancement of wiregrass populations is essentia l for successful ecological restora tion in these systems. Living and dead wiregrass leaves are highly flammable and act as fire-spreaders, moving surface fires evenly through the understory of a longleaf pine ecosystem (Cle well 1989). Dead longleaf pine needles are also very flammable and are a cri tical component of the systems pyrogenicity (Williamson and Black 1981, Clewell 1989). Both l ongleaf pine and wiregrass can therefore be considered keystone species. Together they perpetuate a frequent fire return interval, which favors their own reproductive success, acts as a selective force against competing, less fire13

PAGE 14

tolerant species, and allows a host of herbaceous understory species to survive and reproduce in the open, high-light understory (Williamson and Black 1981, Clewell 1989, Rebertus et al. 1993, Jacqmain et al. 1999). With high productivity and an abundance of shru bs in the understory, flatwoods are likely to lose herbaceous species more quickly with fire exclusion than drier longl eaf pine sites. Even slight fire suppression in a fl atwoods system leads to incr eased sprouting and vegetative reproduction among shrubs, reducing available space for herbaceous plants and lowering species richness (Glitzenstein et al. 2003) Different plant functional groups experience different rates of species loss with reduced fire fre quency. Mesic sites, with more species to lose and more rare species, are likely to experience the highest rates of species loss (Leach and Givnish 1996). In mesic to wet longleaf pine savannas, the species most likely to be re duced or lost are the dominant bunch grasses, basal rosette species, sedges, other sm all monocots, and insectivorous species (Walker and Siletti 2006). Because the understory species composition and soil properties vary markedly with hydrology in longleaf pine ecosystems, it is necessary to test restoration me thods in a variety of site conditions in order to deve lop appropriate restor ation protocols. Mesic flatwoods in the Florida panhandle are generally co-dominated by wiregrass ( A. stricta ), saw palmetto ( Serenoa repens W. Bartram), runner oak (Quercus elliottii Wilbur), gallberry ( Ilex glabra A. Gray), and hairy wicky ( Kalmia hirsuta Walter). The overstory in a mesic longleaf pine flatwoods is typically an open canopy of longleaf pine (basal area of 2.5 to 12.5m2/ha), with slash pine ( Pinus elliottii Engelm.) and pond pine ( Pinus serotina Michx.) codominant in the transitional areas near wetlands (Peet 2006). Fire is a key management tool for m eeting restoration goals in these 14

PAGE 15

systems, but social and environmental factors can often hinder the ability of land managers to use fire effectively (Chapter 1). When a restoration project includes conversion of the overstory from a planted slash pine or loblolly pine ( Pinus taeda L.) stand, complete removal of overstory pines across a large area has the counterproductiv e effect of releasing shrubs (Bro ckway et al. 1998, Kush et al. 1999, McGuire et al. 2001, Kirkman et al. 2007, Pecot et al. 2007). Th e release of shrubs following overstory harvest is likely to be more extensive on a site with a past history of fire suppression, due to the preponderance of underground hardw ood rootstocks estab lished during periods without fire (Olson and Platt 1995 ; Drewa et al. 2006). Herbicide treatment and mechanical site preparation can therefore be important components of restoration projects in which a slash or loblolly pine plantation is being converted to longleaf (Wal ker and Siletti 2006). A potential negative impact of herbicide, however, is a reduction in nontarget desirable understory vegetation. Litt et al. (2001) conducted an extensive litera ture review on the effects of herbicide on understory vegetation in southern pinelands. They examined 125 studies from which 21 were used for review. Throughout the herbicide literature they found a lack of experimental rigor and inconsistent reporting standards, which made it di fficult to reach any ge neral conclusions about the effects of herbicide on ground-layer vegetati on. They were able to conclude only that widespread use of herbicide to control unwan ted vegetation may have undesirable effects on nontarget plant species, and that additional studies of herbicid e impacts are needed before treating large, diverse la ndscapes. They recommend that future herbicide studies follow rigorous experimental designs and make a deliberate effort to track sp ecies of special concern. 15

PAGE 16

Litt et al. (2001) also concluded that the effects of herbicide on ground-layer vegetation in natural flatwoods systems had rarely been measured. They reviewed only two studies documenting herbicide impacts on groundcover in nonplantation flatwoods ecosystems, both of which reported decreases in species richness. Decreases in richness ranged from 5.1% for herbaceous species following hexazinone applic ation (Wilkins et al. 1993), to 71.8% for all species following treatment with a mix of sulfometuron methyl, glyphosate, and triclopyr (Neary et al. 1991). The effects of herbicide on sandhill communitie s are better documented. Four years after treatment, hexazinone herbicide was effectiv e in eliminating woody species and increasing diversity relative to the control, though plots restored with fire alone had even higher species diversity than those treated with herbicide (Provencher et al. 2001). On another north Florida sandhill site, an initial decrease in species diversity was documented following hexazinone application, followed by a recovery and then an increase in the second growing season (Brockway et al. 1998). A subseque nt study on the same site revealed that prescribed fire further enhanced species diversity in conjunction with hexazinone (Brockway a nd Outcalt 2000). Three herbicide release treatments (imazapyr, glyphosphat e, and hexazinone) were tested on loblolly pine plantations at three hilly sites in central Ge orgia. Seven years after treatment, there were no significant difference in understory species diversity between any of the treatments including the control (Boyd et al. 1995). Eleven years after tr eatment, Miller et al. (1999) arrived at the same conclusion. Imazapyr, used alone or in conjuncti on with fire, was effective for restoring grassdominated bobwhite quail hab itat on a loblolly pine sandhi ll in the Florida panhandle, significantly reducing hardwood encr oachment while enhancing the growth of native herbaceous species (Welch et al. 2004). Beneficial results were also reported for imazapyr in a quail habitat 16

PAGE 17

restoration project on hilltops in southern Lousiana (Jones and Chamberlain 2004). However, no differences in herbaceous species cover and richness relative to the control were found on a site in the Virginia piedmont treated with three diffe rent levels of imazapyr for pine release (Keyser and Ford 2006). Other pine releas e studies have also reported be neficial or neutral results for imazapyr, with wildlife forage vegetation showing recovery a nd in some cases improvement after treatment (Hurst 19 87, McNease and Hurst 1991). In order to gain a better unders tanding of herbicide as a rest oration tool for longleaf pine flatwoods ecosystems, we compared the effects of five herbicide treatments on the understory of a coastal flatwoods in the Florida panhandle: 1) Sulfometuron methyl (Oust, E.I. du Pont de Nemours and Company, Wilmington Del.), 2) Imazapyr (Arsenal, BASF Corporation, Research Triangle Park, NC), 3) Hexazinone (Velpar L, E.I. du Pont de Nemours and Company, Wilmington, Del.), 4) A tank mix of Oust and Velpar L, and 5) a no-herbicide control. These herbicide treatments were chosen because they are known to be effective vegetation control treatments for l oblolly and slash pine release in the Southeast, but their effects on flatwoods community composition are not well studied. A recent survey reported that the forestry herbicide most commonly used in the southeast was Arsenal, followed by tank mixes of Arsenal + Oust and Oust + Velpar (Shepard et al. 2004). We examined the longer-term effects of these herbicide treatments on flatwoods vegetation four years after in itial herbicide application (s ee Ranasinghe 2003 for first year results), after which we conducted a dormant-seas on prescribed fire on all treatment plots and assessed post-fire vegetation composition. 17

PAGE 18

Methods Study Area The study was conducted at the Point Washi ngton State Forest, which is located on the gulf coast of Walton County, Florida (30 20 N, 86 04 W). The 20-acre study site is part of a larger mesic flatwoods/wet flatwoods matrix, with an understory dominated by saw palmetto (S. repens ), gallberry ( I. glabra ), rhizomatous oaks ( Quercus minima Small and Quercus elliottii Wilbur), blueberries ( Vaccinium myrsinites Lam. and V. darrowii Camp.), huckleberries ( Gaylussacia dumosa Torr. and A. Gray and G. frondosa L. Torr and A. Gray ex. Torr), bluestem grasses ( Andropogon virginicus var. glaucus (L.) Heck and Andropogon virginicus L.) and wiregrass ( Aristida stricta ). The majority of the s ite can be classified as mesic, but due to subtle topographi cal variation across the site some areas are at the wetter end of the mesic flatwoods spectrum and some are at the drier end. For th is reason, a randomized complete block design was used, with three blocks laid out in mesic-wet areas and three blocks laid out in mesic-dry areas. A pre-treatment vegetation survey identified wetter areas by the presence of indicator species su ch as sedges (Cyperaceae). The soils of the study site are sandy, siliceous, thermic, aeric alaquods of the Leon se ries (Spodosols consisting of deep, poorly to very poorly drained soils derive d from marine parent material (Ranasinghe, 2003)). Yearly weather patterns during the study period were va riable, with hurricanes in some years and droughts in others. Total annual precipitat ion during the study period was 103cm in 2002, 200cm in 2003, 135cm in 2004, 146cm in 2005, and 106cm in 2006 (NOAA 2007). Rainfall in the months following the prescribed fire was ve ry low, with totals of 0.5cm in March 2007, 4cm in April 2007, and 0.3cm in May 2007 (NOAA 2007). Prior to August 2001, the 20-acre study site was a planted slash pine stand with a mean stand age of about 26 years, which had been burned on a 3-year rotation by the Division of 18

PAGE 19

Forestry since its acquisition of Point Washington State Forest in 1992. A pre-harvest, baseline data survey was conducted in the areas slated to become study blocks in June 2001. In August 2001, the entire overstory was harves ted and the site was prepared for the study with a single, light roller chopping followed by a prescribed burn in October 2001. In December 2001, six study blocks, each divided into five 36.6m x 24.4m treatment plots, were established according to the aforementioned randomized complete block design. Each treatment plot was hand-planted with 100 containerized longleaf pine seedlings laid out in 10 rows, with 3.1m between rows and 1.8m between the seedlings in each row. Treatments Within each block, each 100-seedling plot was randomly selected to receive one of five herbicide treatments in March of 2002: 1) Sulfometuron methyl (Oust) at 0.26 kg a.i./ha. This is a moderate application rate for sulfometuron methyl; other studies have us ed 0.16 kg a.i./ha (Lauer and Glover, 1998) and 0.21 kg ai/ha (Shiver and Mart in 2002, Keyser and Ford 2006), while the manufacturerrecommended rate for herbaceous weed control in longleaf pine stands is 0.10 to 0.42 kg a.i./ha (DuPont 2007). Sulfometuron methyl is a selectiv e herbicide primarily effective for controlling herbaceous species (Lauer and Glover, 1998), an d mixing with hexazinone is recommended if broader spectrum control is desired (DuPont, 2007). 2) Hexazinone (Velpar L) at 0.56 kg a.i./ha. Recommended application rates for hexazinone range from 2.0 to 6.7 ai kg/ha (Du Pont 2007), while rates successfully used to control hardwoods in sandhills ecosystems have ranged from 1.1 kg a.i./ha to 2.4 kg a.i/ha (Brockway et al. 1998, Provencher et al. 2001). The hexazinone application rate in our study was therefore very low. This rate was chosen because hexazinone had been shown to be 19

PAGE 20

effective at the aforementioned rates, and we were interested in whether an even lower application rate woul d be effective. 3) Sulfometuron methyl (0.26 kg a.i./ha) + Hexazinone (0.56 kg a.i./ha) mix. This is a common tank mix used throughout the Southeast (S hepard et al. 2004) with the potential to control a broad spectrum of woody and herbaceous pl ants. Tank mixes are used to achieve more complete vegetation control than either herbicide alone, but as a result they may also have more negative implications for native plant commun ities (Keyser and Ford 2006). Banded herbicide application, the method used in this study, may he lp ameliorate these ne gative impacts (Keyser and Ford 2006). 4) Imazapyr (Arsenal) at 0.21 kg a.i./ha. For a restoration project, this was a moderate application rate for imazapyr. Signifi cant hardwood control and habitat improvements have been obtained using imazapyr at 0.42 kg a. i./ha (Jones and Chamberlain 2004), and at all rates from 0.08 kg a.i./ha to 0.24 kg a.i..ha (t hough higher rates did not offer significant improvements over lower rates) (Keyser and Ford 2006). 5) Control. (No herbicide). The herbicides were applied in a 1.2m ba nd over the seedling rows using a backpack sprayer. The reasons for using a banded post-planting herbicide application on unshielded seedlings were both operational and environmenta l. Prior to harvest, the study site already contained many desirable understory species (i ncluding wiregrass) a nd few weedy species (Ranasinghe 2003). The banded application allowed us to pinpoint the delivery of herbicide to an area directly around planted seedlings, while na tive vegetation remained intact in the strips between rows. This application method also lowered the cost of herbicide treatment relative to a broadcast application. Within each 10-row treatment plot, three rows received a second 20

PAGE 21

herbicide application in March of 2003, three rows received a se cond application in April of 2003, and four rows were left with only the firstyear treatment. The results from these subplots were pooled with the first year data in the final analysis because of a lack of significant differences between one and two-year appl ications in any of the parameters. Data Collection and Analysis The pre-treatment vegetation survey was conducted in June 2001, prior to the overstory harvest. Post-treatment vegetation surveys we re conducted at three and nine months after treatment, and the results of these first-year surveys have been published elsewhere (Ranasinghe 2003, Ranasinghe et al. 2005). The fourth-year data were collect ed in June of 2006. Within each treatment plot, six 1m2 quadrats were randomly selected for understory sampling (two in each subplot). In all, 36 randomly selected quadr ats were surveyed in each herbicide treatment and 12 were surveyed in the cont rol, for a total of 156 quadrats. The smaller sample size in the control treatment was due to the fact that a port ion of each control plot was sprayed in the second year, and these rows were not included for anal ysis. All species were identified and percent cover of each species was estimat ed visually using the cover cl asses defined by the modified Daubenmire scale (Daubenmire 1959, Peet et al 1998): 0-1%, 1-2%, 2-5%, 5-10%, 10-25%, 2550%, 50-75%, 75-95%, and 95-100%. For each woody species, average stem height was also estimated and all stems were counted. I. glabra, S. repens and Quercus spp. ( Q. minima and Q. elliottii ), were uniformly distributed across all study plots prior to treatment (Ranasinghe 2003) and were grouped together for analysis. Woody species belonging to the family Ericaceae (heaths) were analyzed separately from the other dominant shrubs because they were not evenly distributed across plots prior to treatment, and also because members of this family are tolerant to hexazinone and may react differently to trea tment than other shrub groups (Wilkins et al 1993, 21

PAGE 22

Outcalt et al 1999, Miller et al 1999). The same procedure was repeated in June of 2007, following a prescribed fire on February 16, 2007. Pyrometers were constructed using aluminum forestry tags and Tempilaq heat-sensitive paints, following a method employed in other fire ecology studies (Hobbs et al. 1984, Iverson et al. 2004). Thirteen Tempilaq paints (designed to melt at 93 C, 149C, 204C, 260C 316C 371C, 427C, 482C, 538C, 593C, 649C, 704C, 760C, and 871C) were applied to each tag. A second tag was paper-clipped over the fron t of the first to protect the paints from charring. Six stakes were randomly placed in each herbicide treatment plot (two in each subplot) and two in each control, with two pyrometers attached to each stake at 30cm and 80cm, for a total of 36 pyrometers at each height for each herbic ide treatment and 12 at each height for each control. Each pyrometer gave a measure of the ma ximum fire temperature at that location: all of the paint dots below that temperature were melted, while all of the paint dots above it remained unchanged. When analyzing species cover vs. temperature relationships, the two pyrometer readings at the 30cm level in e ach subplot were averaged and pl otted against the average of the two understory quadrats in that subplot. Since the pyrometers were placed randomly, they were not necessarily in the same locations as the qua drats. Averaging them gave a more accurate overall picture of the cover and fire intensity in a gi ven plot. Pyrometers on which none of the paints were burned (i.e. temperat ures did not reach 93C) were assumed to be unburned and were assigned a value of 10C to represent ambient air temperature. All parameters were analyzed in JMP IN version 5 (SAS Institute, Inc.), using ANOVA within the framework of a randomized complete block design. The study addressed only the main effects of herbicide treatment, and test s of these effects were not dependent on the assumption of no treatment x block interaction. Block effects were therefore treated as random 22

PAGE 23

effects in a univariate ANOVA model with tw o independent variable s: treatment with Block&Random as a covariate. Data were log-transformed where necessary to meet the assumptions of ANOVA. Significant differences between treatments were separated with Tukeys HSD or Hsus MCB. Post-fire treatme nt differences were analyzed with ANCOVA, using pre-fire distributions as a covariate. ANCOVA was also used to analyze fourth year abundance of Ericaceae, using pre-treatment distribu tions as a covariate. A significance level of = 0.05 was used to test all parameters. Community patterns were analyzed in PC-ORD version 4, using Nonmetric MultiDimensional Scaling (NMDS) a nd Indicator Species Analysis following the method of Dufrene and LeGendre (1997). NMDS uses rank-transformed distances to linearize the degree of difference between plant survey quadrats. Ordi nation of plant survey quadrats (containing both species presence and percent cover data) was don e using a Srenson distance measure in PCORDs autopilot mode. The Srensen coefficient is a measure of percent dissimilarity that can be applied to either presence-absence data or quantitative data (in this case percent cover values), and is recommended for analysis of ecological communities (McCune and Grace 2002). In the autopilot mode, NMDS conducts 40 runs with real data and 50 runs with randomized data. Species with less than 3 occurrences were dropped from the analysis to reduce the effects of rare species (McCune and Grace 2002), reducing the total number of species in the analysis from 87 to 55. PC-ORD chooses the solution with the highes t dimensionality that also meets the criterion of having less stress than 95% of the random runs. Each point on the NMDS axes represents a single plant survey quadrat, and the distance be tween points on the NMDS axes represents the Srenson distance between points in n dimensions (in this case a 3-dimensional solution was chosen). A Multi-Response Permutation Procedur e (MRPP) can then be used to determine the 23

PAGE 24

degree of agreement within groups with regard to an environmental variable, which in this case was a categorical variable (treatment). The weighting method for the MRPP was n/sum(n), and we again used a Srenson distance measure. MRPP yields a test statistic (A) and a p-value (A = 1 when all items are identical within groups (the groups in this case were herbicide treatments), and A=0 when heterogeneity within gr oups equals expectation by chance). A therefore expresses the degree to which variation in position between plant survey quadrats on the ordination axes is explained by the herbicide treatments. Indicat or species analysis using the Indicator Value method defines an indicator species as a specie s found mostly in a sing le group and present in the majority of sites belonging to that group (DuFrene and LeGendre 1997). The indicator value (IV) is maximum (100%) when the individuals of species I are observed in a ll sites of only one group (in this case, the each group was an herb icide treatment). IV is obtained by combining relative abundance and relative frequency va lues for each species in each group. Relative abundance of a species in a group is the averag e abundance of a given species in a given group of plots over the average abundance of that spec ies in all plots (expres sed as a percent), and relative frequency is the average number of plots in a given group where a given species is present. Significance of each Indicator Valu e was established using a 1000-permutation MRPP which compared the observed maximum IV for each species to the IV obtained from randomized groups, and generated a p-value for this comparison using the Monte Carlo test of significance. Species identified as indicators fo r one of the herbicide treatmen ts are assumed to have been enhanced by that treatment relati ve to the others, whereas species identified as indicators for the control are assumed to have been negatively impacted by the herbicide treatments. 24

PAGE 25

Results Shrub Cover Four years after initial herbic ide application, pe rcent cover of I. glabra, S. repens and Quercus spp. ranged from 13.5% in the sulfometuron + hexazinone (sulfo + hexa) treatment to 22.9% in the sulfometuron treatment, but did not vary significantly among treatments overall (p = 0.2154) (Table 2-1). Indicator species analysis revealed a significant association between the sulfometuron treatment and increased c over of both rhizomatous oak species, Q. elliottii and Q. minima four years after treatment (p = 0.040 and p = 0.031) (Table 2-2). Since these species were not significantly associated with the control pl ots, this result can be in terpreted to mean that the sulfometuron treatment enhanced oak growth. Woody species belonging to the family Er icaceae were abundant on the study site, and included blueberries ( V. myrsinites and V. darrowii ), huckleberries (G. dumosa and G. frondosa ), hairy wicky ( K. hirsuta ), and Lyonia lucida Koch (fetterbush). Ericac eae cover four years after treatment was four to five times higher in th e control plots than in any of the herbicide treatments, a difference which approached si gnificance (p = 0.06), but was moderated by the uneven pre-treatment distributions and the smaller control sample size (Table 2-1, Figure 2-1). Overall percent cover of woody speci es was highest in the control a nd lowest in the sulfo + hexa treatment (Figure 2-2). In June 2007, four m onths after the February 2007 prescribed fire, percent cover of Quercus spp., I. glabra, and S. repens did not vary significantly among treatments (p = 0.1740) (Figure 2-3a, Tabl e 1), nor did Ericaceae cover (p = 0.097). Herbaceous Cover Herbaceous cover varied signi ficantly among treatments (p = 0.002) (Figure 2-4a). In both the imazapyr and hexazinone treatments, mean herbaceous cover (46.5% and 46.1%, respectively) was significantly higher than th e control (22.1%). Following fire, however, 25

PAGE 26

herbaceous cover no longer differed significan tly among treatments (p = 0.302). Herbaceous cover decreased significantly following fire in all but the control and sulfometuron plots (Figure 2-4b). Much of this decrease can be attributed to the decrease in wiregrass crown size following fire (see below). Overall vegeta tion cover before and after fire is summarized in Figure 2-5. Wiregrass Cover Mean wiregrass cover four y ears after initial herbicide application differed significantly among treatments (p = 0.04) (Figure 2-6a). Wiregr ass cover in the sulfo + hexa treatment was three times greater than the cont rol (Figure 2-6). In all treatme nts except the control, wiregrass cover 4 years after treatment was higher than pre-tr eatment levels (Figure 2-6c). Following fire, wiregrass cover decreased in all treatments (F igure 2-6b). The overall decrease in wiregrass cover following fire is not attributed to wiregr ass mortality, but to the decrease in wiregrass crown size following fire. Wiregrass cover was su rveyed only four mont hs after the February 2007 fire, and it was observed that individual wiregrass clumps were alive and regenerating, but greatly reduced in size from pre-fire levels. A similar effect was documented after the herbicide treatment in 2002, with an init ial decrease in wiregrass cove r followed by a rebound over the next nine months (Figure 2-6c). Shrub Stem Density Four years after treatment, mean shrub stem density for Quercus Ilex and Serenoa varied significantly among treatments (p = 0.012), displaying a similar pa ttern to percent cover (Table 2-1). Mean stem density in the imazapyr treatment (20 stems/m2) was significantly lower than sulfometuron (34.7 stems/m2) and hexazinone (32.9 stems/m2). None of the treatments varied significantly from the control, in whic h we measured a mean stem density of 31.75 stems/m2. In the first nine months after herbicid e application, imazapyr was the only treatment that did not experience an increase in shrub stem density (Figure 2-7b). In the fourth year after 26

PAGE 27

application, this effect was no longer evident, as shrub stem density no longer varied significantly among treatments. Following the pr escribed fire in February 2007, mean shrub stem density increased again by around 50% in ev ery treatment (Figure 2-7a). These increases were significant in all treatment s but the control. Variation in Ericaceae stem density among treatments approached significance in June 2006 (p = .061) (Table 2-1), but this difference was moderated by the smaller control sample size an d uneven pre-treatment distributions. Ericaceae stem density did not change significantly in a ny of the treaments following fire (Table 2-1). Shrub height Mean shrub stem height for Quercus Ilex and Serenoa did not differ significantly among treatments in June 2006 (p = 0.2341). Stem height ranged from 14.6cm in the imazapyr treatment to 17.1cm in the hexazinone treatment (T able 2-1). Following the prescribed fire in February 2007, mean shrub height increased in the imazapyr, sulfo, and sulfo + hexa treatments (Table 2-1). Differences among treatments following fire were not significant (p = 0.40). Mean stem height of Ericaceae did not differ signif icantly among treatments in June 2006 (p = 0.304) or in June 2007 (p = 0.900), though height increases within some treatments following fire were significant (Table 2-1). Understory responses A total of 93 species were identified dur ing the June 2006 and June 2007 plant surveys (Appendix A). Weedy species were not prevalent, and we found numerous plant specimens belonging to functional groups identified by Walk er and Peet (1983) and Glitzenstein et al. (2003) as high-risk for being lost in shrub-invaded flatwoods, such as bunch grasses (Poaceae), sedges (Cyperaceae), sundews (Droser aceae), and basal rosette species. Overall species richness four years after he rbicide application varied significantly among treatments (p = 0.01), ranging from 7.4 species/m2 in the control treatment to 9.3 species/m2 in 27

PAGE 28

the sulfometuron treatment (Table 2-3). Herbace ous species richness, however, did not vary significantly among treatments (p = 0.498) (Table 23). Following fire, overall species richness no longer varied significantly among treatments (p = 0.15). Within treatments, responses to the fire varied (Table 2-3). Over all species richness increased signi ficantly in the control, but did not change significantly in the other treatments. Herbaceous species richness did not vary among treatments before or after fire (p = 0.498 and p = 0.114, respectively). Species Diversity Overall Shannon diversity (woody and herbaceous species combined) four years after herbicide application did not vary significantl y among treatments (p = 0.07) (Figure 2-8a, Table 2-3). However, herbaceous species diversity did vary signifi cantly among treatments (p = 0.036), with the highest diversity in the sulfometuron treatment a nd the lowest in the sulfo + hexa treatment (Table 2-3). Within treatments, re sponses to the prescribed fire varied (Figure 28b). Overall diversity increased following fire in the sulfo + hexa and hexazinone treatments, but did not change significantly in any of the other treatments (Figure 2-8b). Herbaceous diversity increased in the imazapyr, sulfo + hexa and hexazinone treatments following fire (Table 2-3). Initial gains in overall Shannon diversity during the first year after treatment (Ranasinghe 2003) were no longer evident after four years (Figure 2-8c). Species Evenness Mean species evenness (1/D) f our years after herbicide appl ication, calculated for woody and herbaceous species combined, did not vary significantly among treatments overall (p = 0.12) (Table 2-3). Herbaceous species evenness, how ever, did vary significantly among treatments (p = 0.014), with the highest evenness in the sulfom eturon treament and the lo west in the sulfo + hexa treatment (Table 2-3). Following fire, only the sulfo + hexa tr eatment experienced a significant change in overall evenness (from 3.12 to 4.54) (Table 2-3) Herbaceous species 28

PAGE 29

evenness increased in the imazapyr, sulfo + hexa and hexazinone treatments following fire (Table 2-3). Community Ordination Analysis of the plant community four years after treatment using NMDS and MRPP revealed only a slight relationship between treat ment and species composition (Fig 2-8), with a small chance-corrected within-gro up agreement (A) of 0.02 (p = 0.01) for the pre-fire survey quadrats. Indicator Species Analysis following the method of Dufrene and LeGendre (1997) revealed significant relationships between some species and treatments (Monte Carlo test of significance) (Table 2-2). Species of Special Interest Curtiss sandgrass ( Calamovilfa curtissii Scribn.), a florida panhandle endemic, was identified in one of the imazapyr treatment pl ots during both the June 2006 and June 2007 plant surveys. Prior to treatment, there was a very low density of legumes on our study site: out of 90 survey quadrats, only two legume specimens were identified, both of which were Mimosa quadrivalvis var. angustata Barneby (Ranasinghe, unpublished data). After four years, we found Desmodium lineatum DC, Desmodium strictum DC, and Tephrosia hispidula Pers. in addition to M. quadrivalvis, again at a low density: a total of eight legume specimens were identified out of 156 survey quadrats. In the post-fire survey, one D. lineatum four M. quadrivalvis, and one Baptisia lanceolata Elliott specimen were identified. These species were not significantly associated with any of the treatments. Fire Temperature The prescribed fire was cool, fast-moving, and patchy. Mean maximum fire temperature by treatment ranged from 252C to 306C at 30cm and 166C to 257C at 80cm (Table 2-4), and did not differ significantly among herbicide treatm ents at either height (p = 0.42 and p = 0.55, 29

PAGE 30

respectively). A significant negative relationship existed between percent Quercus cover and maximum fire temperature at 30cm (r2 = 0.06, p = 0.032) (Figure 2-9a), whereas a significant positive relationship existed between percent I. glabra cover and maximum fire temperature at 30cm (r2 = 0.07, p = 0.017) (Figure 2-9b). The ne gative relationship observed with Quercus is largely attributable to the fact that areas with high Quercus cover were often only partially burned or did not ignite at all, whereas the I. glabra tended to ignite readily and burn completely. Temperature was not significantly related to wiregrass cover (p = 0.73) or overall herbaceous cover (p = 0.28). Discussion Four years after initial herbic ide application, differential treatment effects on understory composition were still evident. Sulfometuron me thyl had a releasing e ffect on the two most common oak species, Q. minima and Q. elliottii indicating that this is not an appropriate herbicide for flatwoods with a high density of oa ks. However, in spite of having higher shrub cover and lower herbaceous cover than the othe r herbicide treatments, sulfometuron plots also had the highest fourth year herbaceous species ri chness, diversity, and evenness. One reason for this may have been that wiregrass cover in th e sulfometuron plots was so mewhat lower than the other treatments, leaving more space for other small herbaceous sp ecies. These gains in species richness, diversity, and evenness are therefore un likely to persist through future fire cycles because it will be more difficult to apply prescribed fire to this treatment. Indeed, sulfometuron was the only treatment in which species richne ss, diversity, and eve nness did not increase following the prescribed fire in 2007. Woody shrub cover, height, and stem density we re lowest in the imazapyr treatment, but the differences between imazapyr and the other treatments were not as pronounced as those observed in the first year after a pplication. These results suggest that the increases in shrub stem 30

PAGE 31

density evident in the first year (Figure 2-7b) were a short-term phenomenon, and subsequent dieback of sprouts occurred as these shrubs matured and grew taller. The responses of Ericaceae to our treatments were surprising, given that in other studies members of this family have been highly hexa zinone-resistant (Wilkins et al. 1993, Outcalt et al. 1999, Miller et al. 1999), and hexazinone is comm only used as a release treatment for blueberry crops. If anything, we expected to see higher Ericaceae cover in the hexazinone treatment, when in fact this treatment had the lowest Ericaceae cove r of all. This result was particularly notable given that the hexazinone plots had the highest percent cover of Ericaceous shrubs prior to treatment (Figure 2-1). The sensitivity of Eric aceous shrubs to hexazinone may have been a function of the timing of application and the poor ly drained, sandy soil; Velpar L is labeled for blueberry release only prior to budbreak in the spring and in conditions where there is no standing water (DuPont, 2007). Control of thes e species may be desirable for restoration scenarios in which the goal is to increase he rbaceous cover on a site dominated by Ericaceous shrubs (Outcalt et al. 1999). However, many of these shrubs are important wildlife food plants (Hay-Smith and Tanner 1994). If the goal is to avoid negative impacts to these populations, our results suggest that cauti on may be in order with regard to soil type, soil water, and timing of application. Both imazapyr and hexazinone significantly increased overall herbaceous cover relative to the control, and the sulfo + hexa treatment significantly increased wire grass cover relative to the control. Though the shrub control initially provided by the herbicide treatments was relatively short-live d, these results suggest th at the short-term shrub control allowed herbaceous species (and wiregrass in particular) to gain more of a foothold in the understory. 31

PAGE 32

The mean fire temperatures we recorded were low for a dormant-season fire in a mesic flatwoods: mean maximum fire temperatures for several other dormant season fires on flatwoods sites ranged from 383C to 588C (Dre wa et al. 2002), whereas our mean maximum temperatures were 200C at 80cm and 274C at 30cm. Across the study site, we made the general observation that I. glabra burned more readily and thoroughly than Quercus spp., which often burned only partially, and, in some highly oakdominated patches, failed to ignite at all. Since dead pine needles are a critical component of pyrogenic ity in longleaf pine systems (Williamson and Black 1981, Clewell 1989), it is like ly that the complete removal of overstory pines contributed to our inability to apply fire effectively to the site. In our study, stem counts of both Quercus spp. and I. glabra more than doubled in all treatments in the four months following fire. Because we did not have an unburned control, we cannot say to what degree the observed post-fire increases in shrub c over and density were greater than the increases expected from normal sp ring growth of these shrubs in the absence of overstory competition. However, it is safe to sa y that the prescribed fire was ineffective at controlling the shrubs, because it did not decrease shrub cover or even maintain it at pre-fire levels. This result can be attri buted to aggressive shrub regrowth due to the lack of competition from overstory pines, the loss of needlefall as a fine fuel supply, and the length of time since the previous fire and subsequent herb icide applications (four and five years). Shrub control, and by extension floral diversity, are maximized by fire re turn intervals of one to three years (Kirkman et al. 2001; Glitzenstein et al. 2003 ). If future fires are conduct ed on a shorter interval, better prescribed fire results may be obtained. The season of burn may also have contribu ted to the difficulty we had in applying prescribed fire effectively. It is generally be lieved that summer fires were probably the most 32

PAGE 33

prevalent selective force prior to European colonization of the S outheast, and are more effective at controlling shrubs and maintaining herbaceous plant diversity in the understory (Glitzenstein et al 1995, Streng et al. 1993). Others have reported shrub resprouti ng responses to dormantseason fire in flatwoods similar to those we obs erved. On mesic and wet sites in Florida and Lousiana, shrub stem densities we re greater than pre-treatment levels after dormant-season fire, and continued to increase with repeated dormant season fires, whereas long-term growing season fires maintained stem densities at the same le vel (Drewa et al. 2002). In another recent study, it was observed that stem densities of root-crown bearing shrubs ( Quercus and Ilex spp.) after repeated dormant-season fires were seven times gr eater than those subjec ted to growing-season fires (Drewa et al. 2006). In one case, stem densities of root-cro wn bearing shrubs were 50 times greater in South Carolina savannas managed w ith annual dormant season fires than in plots managed with annual growing-season fires over a 30 year period (Waldrop et al 1992). Some of the resprouting ability of Quercus can be attributed to a past hi story of fire suppression, during which underground stems are likely to have grown extensively (Drewa et al. 2006). Given that our site was previously a slash pine plantation, periods of fire suppression are likely to have occurred in the past during seed ling establishment and may have c ontributed to the recalcitrance of the dominant woody shrubs. However, results indicating that dormant-se ason fire increases shr ub cover in flatwoods are not universal. In one case, a long-term dorma nt season prescribed fire regime significantly reduced shrub cover on an I. glabra -dominated flatwoods while increasing herbaceous species cover and diversity (Brockway and Lewis 1997), sugge sting that dormant-season fire is a viable option for flatwoods management when growing-s eason fire cannot be use d. Hiers et al. (2000) offer experimental evidence, based on patterns of legume reproduction, that variable fire regimes 33

PAGE 34

will maintain a broader suite of native species than either growing-season or dormant-season fire alone, and cite studies suggesti ng that significant variation in the frequency and season of lightning may have occurred during the 7000 years BP that longleaf pine communities have been established in the sout hern coastal plain. Though dormant season fire may be effectiv e and even desirable in some cases, our results suggest that dormant season fire should be avoided when species of low flammability are abundant, there is a lack of fine fuels, and there are no overstory pines to compete with shrubs for light and soil resources. Ilex glabra is known to be one of the most flammable flatwoods species, while oaks are among the least flammable (Behm et al. 2004), and this may be another part of the reason for the succe ss of dormant-season fire on an I. glabra-dominated site reported by Brockway and Lewis (1997). In our study, though stem density of both Quercus and Ilex increased following fire, only Quercus increased in mean percent cover across treatments; mean I. glabra cover remained roughly the same or decrease d. It may be that in the coming growing seasons, some of the in itial flush of basal I. glabra sprouts will die back, and this species will not show an overall cove r increase. Since Quercus increased in cover as well as stem density, it seems more likely that the fire-related increase will have long-term effects. In summary, only positive herbicide impact s on wiregrass and herbaceous cover were observed in this study, and understory species diversity, richness, evenness, and community composition were largely unaffected. The imazapyr treatment, which had the most dramatic shrub control effects in the first year after application, still had the lowest shrub levels and the highest herbaceous cover after four years, indica ting that this was the mo st successful treatment overall. The poor vegetation control offered by sulfometuron, which was largely attributable to the resistance of Q. elliottii and Q. minima indicates that it is not an appropriate herbicide for 34

PAGE 35

use in an oak-dominated flatwoods community. However, the sulfometuron + hexazinone tank mix showed promise as a flatwoods restor ation treatment, offeri ng better control over Quercus spp. and significantly increasing wiregrass cove r relative to the control. Though hexazinone plots had significantly higher overall herbaceous cove r than the control, our results suggest that this herbicide, at the rate app lied, did not provide much control over the dominant shrubs in the long term. Because we applied a ve ry low rate of hexazinone, it is difficult to make any absolute comparisons between hexazinone and the other he rbicides based on the re sults of this study. However, given the success others have had w ith hexazinone (Brockway et al 1998, Brockway and Outcalt 2000, Provencher et al 2001), it seems lik ely that we would have seen greater shrub control and attendant increases in herb aceous cover had we used a higher rate. As of June 2007, it was too early to tell wh ether the herbicide-related increases in wiregrass and overall herbaceous cover will be maintained through future fire cycles. The Florida Division of Forestry normally burns this site on a three-year fire interval, and results from more frequent burning may be more successful than those we observed five years after the inital herbicide application a nd four years after the second a pplication. Because the entire overstory was harvested, it may be several fire cycles before the new longleaf pine canopy provides sufficient needlefall to sustain the level of prescribed fire necessary for shrub control. In the meantime, this limitation is likely to be exacerbated by the Florida Division of Forestrys inability to use growing-season fire on this site due to its proximity to a highway and several beach housing communities. If the trends of increasing shrub cover, height, and density continue, the herbicide-related gains in wiregra ss and overall herbacous cover may be lost. The aggressive post-fire resprouting of shrubs observed in this study underscores the utility of herbicide as a supplemental rest oration tool, especially in fl atwoods ecosystems where undesired 35

PAGE 36

changes in understory community composition can be rapid and diffi cult to reverse. It also underscores the need for effective prescribed fire in order to maintain the early gains in wiregrass and herbaceous cover conferre d by herbicide application. 36

PAGE 37

Table 2-1. Mean percent cover, st em density, and height of shrubs 4 years after treatment (YAT) (June 2006) and following prescribed fire (June 2007). Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Treatment ANOVA/ ANCOVA p = 0.215 18.7 13.6 22.9 13.5 19.9 p = 0.215 18.7 13.6 22.9 13.5 19.9 18.7 13.6 22.9 13.5 19.9 22.5 24.3* 34.0* 21.5 26.6 22.5 24.3* 34.0* 21.5 26.6 21.3 5.1 4.4 5.8 4.0 p = 0.060 21.3 5.1 4.4 5.8 4.0 21.3 5.1 4.4 5.8 4.0 p = 0.060 2.8 2.3 5.6 3.6 3.3 2.8 2.3 5.6 3.6 3.3 p = 0.174 31.8ab20.8b34.7a25.1ab32.9a31.8ab20.8b34.7a25.1ab32.9ap = 0.012 56.3 47.8* 64.9* 50.9* 59.1* 56.3 47.8* 64.9* 50.9* 59.1* p = 0.331 p = 0.097 % Cover Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Density (stems/m2) Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Height (cm) Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire % Cover Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire % Cover Q, S, I 2006 4YAT PostFire Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Density (stems/m2) Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Density (stems/m2) Q, S, I 2006 4YAT PostFire Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Height (cm) Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Stem Height (cm) Q, S, I 2006 4YAT PostFire Q, S, I 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire Ericaceae 2006 4YAT PostFire 16.1 14.6 16.9 15.9 17.1 p = 0.234 16.1 14.6 16.9 15.9 17.1 16.1 14.6 16.9 15.9 17.1 p = 0.234 21.1 19.7* 22.3* 21.6* 19.7 21.1 19.7* 22.3* 21.6* 19.7 21.1 19.7* 22.3* 21.6* 19.7 42.9 8.4 17.2 16.5 13.0 p = 0.061 42.9 8.4 17.2 16.5 13.0 42.9 8.4 17.2 16.5 13.0 p = 0.061 10.3 8.9 25.0 12.8 9.6 p = 0.108 10.3 8.9 25.0 12.8 9.6 10.3 8.9 25.0 12.8 9.6 p = 0.108 p = 0.400 11.6 12.3 12.2 12.0 11.9 p = 0.304 11.6 12.3 12.2 12.0 11.9 11.6 12.3 12.2 12.0 11.9 p = 0.304 12.5 12.1 15.2* 17.0* 16.1* p = 0.900 12.5 12.1 15.2* 17.0* 16.1* 12.5 12.1 15.2* 17.0* 16.1* p = 0.900 *Letters denote significant differences between means in the same column at = 0.05 based on Tukey-Kramer HSD test. (*) denot es significant post-fire change within treatment (Tukey-Kramer HSD). P-values are for randomized block ANOVA / ANCOVA within columns (significant at = 0.05). Q, S, I stands for Quercus spp., S. repens and I. glabra 37

PAGE 38

Table 2-2. Indicator species analysis following the method of Dufrene and LeGendre (1997). IV from randomized groups Asclepias cineria Pteridium aquilinum Lachnocaulon auceps Lachnanthes caroliniana Quercus elliottii Species Quercus minima Smilax pumila Treatment Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Control Observed Indicator Value (IV) 15.5 20.3 12.2 16.7 19.1 29.7 12.5 Mean 5.4 11.7 5 4.1 12.6 21.7 3.7 S. Dev 2.91 3.4 2.86 2.5 3.27 3.56 2.39 p* 0.010 0.020 0.025 0.010 0.040 0.031 0.029 IV from randomized groups IV from randomized groups Asclepias cineria Pteridium aquilinum Lachnocaulon auceps Lachnanthes caroliniana Quercus elliottii Species Quercus minima Smilax pumila Treatment Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Control Observed Indicator Value (IV) 15.5 20.3 12.2 16.7 19.1 29.7 12.5 Mean 5.4 11.7 5 4.1 12.6 21.7 3.7 S. Dev 2.91 3.4 2.86 2.5 3.27 3.56 2.39 p* 0.010 0.020 0.025 0.010 0.040 0.031 0.029 Asclepias cineria Pteridium aquilinum Lachnocaulon auceps Lachnanthes caroliniana Quercus elliottii Species Quercus minima Smilax pumila Asclepias cineria Pteridium aquilinum Lachnocaulon auceps Lachnanthes caroliniana Quercus elliottii Species Quercus minima Smilax pumila Treatment Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Control Treatment Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Control Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Sulfo + Hexa Sulfometuron Imazapyr Hexazinone Sulfometuron Sulfometuron Control Observed Indicator Value (IV) 15.5 20.3 12.2 16.7 19.1 29.7 12.5 Observed Indicator Value (IV) 15.5 20.3 12.2 16.7 19.1 29.7 12.5 15.5 20.3 12.2 16.7 19.1 29.7 15.5 20.3 12.2 16.7 19.1 29.7 12.5 Mean 5.4 11.7 5 4.1 12.6 21.7 3.7 Mean 5.4 11.7 5 4.1 12.6 21.7 3.7 5.4 11.7 5 4.1 12.6 21.7 5.4 11.7 5 4.1 12.6 21.7 3.7 S. Dev 2.91 3.4 2.86 2.5 3.27 3.56 2.39 S. Dev 2.91 3.4 2.86 2.5 3.27 3.56 2.39 2.91 3.4 2.86 2.5 3.27 3.56 2.91 3.4 2.86 2.5 3.27 3.56 2.39 p* 0.010 0.020 0.025 0.010 0.040 0.031 0.029 p* 0.010 0.020 0.025 0.010 0.040 0.031 0.029 0.010 0.020 0.025 0.010 0.040 0.031 0.010 0.020 0.025 0.010 0.040 0.031 0.029 Monte Carlo test of significance, =0.05 38

PAGE 39

Table 2-3. Species Richness (species/m2), Diversity (Shannon Index), and Evenness (Simpsons Index, 1/D) 4 years after treatment (YAT) (June 2006) and following fire (June 2007). Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Treatment ANOVA/ ANCOVA p = 0.014 Species Richness (spp/m2) Overall 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire Species Richness (spp/m2) Overall 2006 4YAT PostFire Overall 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire p = 0.498 p = 0.151 PostFire Shannon Diversity Overall 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire Shannon Diversity Overall 2006 4YAT PostFire Herbaceous 2006 4YAT Shannon Diversity Overall 2006 4YAT PostFire Overall 2006 4YAT PostFire Herbaceous 2006 4YAT Herbaceous 2006 4YAT p = 0.066 p = 0.131 p = 0.114 Evenness (1/D) Overall 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire Evenness (1/D) Overall 2006 4YAT PostFire Overall 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire Herbaceous 2006 4YAT PostFire p = 0.116 1.34 1.48 1.57 1.34 1.35 1.34 1.48 1.57 1.34 1.35 1.10ab1.01ab1.17a0.85b0.88b1.10ab1.01ab1.17a0.85b0.88b7.36c8.97ab9.25a8.03bc8.22bc7.36c8.97ab9.25a8.03bc8.22bc4.58 5.75 5.41 5.13 5.33 4.58 5.75 5.41 5.13 5.33 3.54ab3.51ab4.15a3.12b3.01b3.54ab3.51ab4.15a3.12b3.01b3.02ab2.45ab3.03a2.11b2.13b3.02ab2.45ab3.03a2.11b2.13b9.75* 9.0 8.19 9.0 8.86 9.75* 9.0 8.19 9.0 8.86 6.83 5.97 5.05 6.14 5.94 6.83 5.97 5.05 6.14 5.94 1.67 1.70 1.51 1.69* 1.65* 1.67 1.70 1.51 1.69* 1.65* 1.59 1.40* 1.30 1.40* 1.45* 1.59 1.40* 1.30 1.40* 1.45* 4.24 4.43 3.80 4.54* 4.07 4.24 4.43 3.80 4.54* 4.07 p = 0.249 p = 0.036 p = 0.293 p = 0.014 p = 0.533 4.36 3.67* 3.53 3.79* 3.95* 4.36 3.67* 3.53 3.79* 3.95* *Letters denote significant differences among means in the same column at = 0.05 based on Tukey-Kramer HSD test. (*) denotes sign ificant post-fire change within treatment (Tukey-Kramer HSD). P-values are for randomized block ANOVA / ANCOVA within columns (significant at = 0.05). 39

PAGE 40

Table 2-4. Mean maximum fire temperature (C) by treatment. Treatment 30cm80cm Mean Fire Temp. ( C ) Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone ANOVA 276 (57) 275 (32) 306 (29) 260 (26) 252 (30) 181 (36) 190 (27) 197 (21) 166 (22) 257 (30) 276 (57) 275 (32) 306 (29) 260 (26) 252 (30) 276 (57) 275 (32) 306 (29) 260 (26) 252 (30) 181 (36) 190 (27) 197 (21) 166 (22) 257 (30) 181 (36) 190 (27) 197 (21) 166 (22) 257 (30) p = 0.420p =0.546 P-values for randomized block ANOVA are significant at = 0.05. Standard error in parentheses. 40

PAGE 41

0 5 10 15 20 25 30 35 ControlImazapyrSulfoSulfo + HexaHexazinone TreatmentMean Ericaceae Cover Pre-Harvest 4 yrs post-treatmentFigure 2-1. Pre-harvest (June 2001) and June 2006 percent cover of Eric aceae by treatment. Pre-harvest Ericaceae cove r was not uniform among treatment plots (ANOVA, p = 0.009). 4 years post-treatment, variance in Ericaceae cover by treatment approached significance (ANCOVA, p = 0.060). 41

PAGE 42

0 10 20 30 40 50 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean % Shrub Cover Ericaceae Serenoa Quercus Ilex 0 5 10 15 20 25 30 35 40 45 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean % Shrub Cover Ericaceae Serenoa Quercus IlexJune 2006 (4 years after treatment) June 2007 (post-fire)a) b) Figure 2-2. a) Mean percent shrub cover f our years after trea tment (June 2006) for Quercus spp., I. glabra, S. repens and Ericaceous shrubs (Vaccinium spp. K. hirsuta Gaylussacia spp., and L. lucida). b) Mean percent shrub cover in J une 2007, four mont hs after fire. 42

PAGE 43

0 5 10 15 20 25 30 35 40 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean % Cover Pre-Fire Post-FireNon-Ericaceaea) b) 0 5 10 15 20 25 30 35 ControlImazapyrSulfoSulfo + HexaHexa TreatmentMean % Cover Pre-Fire Post-FireEricaceaeFigure 2-3. Mean percent shrub cover before (June 2006) and after fire (June 2007) for a) Quercus spp., I. glabra, and S. repens and b) Ericaceae. Differences between and within treatments are not significant on either graph. 43

PAGE 44

0 10 20 30 40 50 60 ControlImazapyrSulfoSulfo+HexaHexazinone TreatmentMean % Herbaceous Cover Pre-Fire Post-Firec a bc abc ab* *Figure 2-4. Mean % herbaceous cover (including both forbs and graminoids) 4 years after treatment (Pre-fire) and 5 year s after treatment (Post-fire). Pre-fire means not sharing the same letter are significantly different (Tukey-Kramer HSD, = .05). (*) indicates significant pre-fire / post-fire differences within treatments (Tukey-Kramer HSD, =.05). 44

PAGE 45

0 10 20 30 40 50 60 70 80 ControlImazapyrSulfoSulfo + HexaHexazinoneMean % Cover Herbaceous Shrubs Ericaceous Shrubs 0 10 20 30 40 50 60 70 80 ControlImazapyrSulfoSulfo+HexaHexazinone TreatmentMean % Cover 0 10 20 30 40 50 60 70 80 ControlImazapyrSulfoSulfo + HexaHexazinoneMean % Cover Herbaceous Shrubs Ericaceous Shrubs 0 10 20 30 40 50 60 70 80 ControlImazapyrSulfoSulfo+HexaHexazinone TreatmentMean % Covera) b)June 2006 (4 years after treatment) June 2007 (Following Feb. 2007 fire)Figure 2-5. Overall vegetation cove r in a) June 2006 (4 years afte r treatment) and b) June 2007 (4 months after fire). Cover is separated into herbaceous, shrub, and Ericaceous shrub components. 45

PAGE 46

0 5 10 15 20 25 30 35 40 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean % Cover Wiregrass Pre-Fire Post-Firea) b) 0 5 10 15 20 25 30 35 40 Pre-harvest3 Mos 9 Mos 4 yrsPost-fire TimeMean % cover Wiregrass Control Imazapyr Sulfo Sulfo+Hexa Hexab b a ab ab* *Figure 2-6. a) Mean % cover of wiregrass four years after initial treatment (pre-fire) and five years after treatment (post-fire). Pre-fi re means not sharing the same letter are significantly different (Hsus MCB, = 0.05). (*) indicates a significant prefire/post-fire differen ce (Tukey-Kramer HSD, = 0.05). b) Mean % wiregrass cover within treatments over the course of the 5year experiment. First-year results from Ranasinghe (2003 and unpublished data), and Ranasinghe et al. (2005). 46

PAGE 47

0 10 20 30 40 50 60 70 80 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean Shrub Stem Density Pre-Fire Post-Fire* a) 0 20 40 60 80 100 120 Pre-harvest3 Mos9 Mos 4 yrsPost-fire TreatmentMean Shrub Stem Density Control Imazapyr Sulfo Sulfo+Hexa Hexab) 0 20 40 60 80 100 120 Pre-harvest3 Mos9 Mos 4 yrsPost-fire TreatmentMean Shrub Stem Density Control Imazapyr Sulfo Sulfo+Hexa Hexab)ab b a ab aFigure 2-7. a) Mean shrub stem density (stems/m2) before (June 2006) and after (June 2007) fire for Quercus spp., I. glabra, and S. repens Pre-fire means not sharing the same letter are significantly different. (*) indicates a significant change following fire (TukeyKramer HSD, = 0.05). b) Changes in shrub stem density within each treatment over the course of the 5-year experiment. Fi rst-year results from Ranasinghe (2003 and unpublished data), and Rana singhe et al. (2005). 47

PAGE 48

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 Pre-treat3 Mos9 Mos4 yrsPost-Fire TimeMean Shannon Diversity Control Imazapyr Sulfo Sulfo+Hexa Hexa 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 ControlImazapyrSulfoSulfo+HexaHexa TreatmentMean Shannon Diversity Pre-Fire Post-Fire* *a) b)Figure 2-8. a) Mean species diversity (Shannon Index) by tr eatment four years after treatment (Pre-fire) and five years after treatment (Post-fire). Pre-fire differences among treatments are not significant (ANOVA, p = 0.13). (*) indicates a significant prefire/post-fire differen ce (Tukey-Kramer HSD, = 0.05). b) Changes in species richness within each treatment over the cour se of the 5-year experiment. First-year results from Ranasinghe (2003 and unpublishe d data), and Ranasi nghe et al. (2005). 48

PAGE 49

Control Imazapyr Sulfo+ Hexa Sulfo Hexazinone Control Imazapyr Sulfo+ Hexa Sulfo Hexazinone Axis 1 Axis 2Figure 2-9. NMDS ordination of plant communities Each point represents one plant survey quadrat. Chance-corre cted within-group ag reement (A) = 0.02 ( = 0.05, p = 0.01). 49

PAGE 50

Figure 2-10. a) relationship between mean Quercus cover and mean fire temperature (C) (r2 = 0.06, p = 0.03) in treatment subplots, and b) relationship between mean Ilex cover and mean fire temperature (C) (r2 = 0.13, p = 0.002) in treatment subplots. 0 100 200 300 400 500 600Fire Temperature at 30cm 0 10 20 30 40 50 Quercus % Cover 0 100 200 300 400 500 600Fire Temperature at 30cm 0 5 10 15 20 25 30 Ilex % Cover 50

PAGE 51

CHAPTER 3 HERBICIDE AND SUSTAINABLE L ONGLEAF PINE SILVICULTURE Introduction Perhaps more than any other forested ecosy stem in the United States, the longleaf pine ( Pinus palustris Mill.) savanna lends its elf to simultaneous management for timber and biodiversity. The understory communities asso ciated with longleaf pine are dominated by herbaceous plants adapted to frequent ground layer disturbance in the form of fire, and regular overstory disturbance by lightning, windthrow, an d hurricanes (Platt and Rathbun 1993, Platt et al. 2002). In longleaf pine systems with a freque nt fire regime and an intact seedbank, most native herbaceous species respond positively to the increased light levels in newly-opened canopy gaps (Keddy et al. 2006, Platt et al. 2006), and it has been hypothe sized that many of these species are such weak competitors for light that they could be considered fugitive or ephemeral (Keddy et al. 2006). In one case, remova l of the overstory resulted in a 20% increase in the number of understory species in gaps within five years, and surveys of older gaps indicate that similar increases have been maintained ove r several decades (Platt et al. 2006). Wiregrass ( Aristida stricta Michx.), the dominant bunchgrass in most longleaf pine systems and a keystone species with regard to pyrogenic ity, is also released in response to overstory removal (McGuire 2001). Since the renowned plant biodiversity of th e longleaf pine ecosystem resides largely in the herbaceous component (Walker and Peet 1983, Kirkman et al. 2001), the removal of large overstory trees for timber harvest has the potential to benefit the ecosystem. However, there are several important caveat s to the analogy between silviculture and natural disturbance in longleaf pine systems (Mitchell et al. 2006). One such caveat is that the removal of overstory pines across a large area also releases shrubs (Brockway et al. 1998, Kush et al. 1999, McGuire et al 2001, Kirkman et al. 2007, Pecot et al 2007), which, in the absence of 51

PAGE 52

adequate control, eventually leads to the exclusion of herbaceous plants (Gilliam and Platt 1999, Glitzenstein et al. 2003). The release of shrubs following oversto ry harvest is likely to be more extensive on a site with a past history of fire suppression, due to the preponderance of underground hardwood rootstocks established during periods wit hout fire (Olson and Platt 1995; Drewa et al. 2006). Shrub invasion excludes herbaceous specie s not only through competition for light and resources, but through the alteration of fire regimes away from the high-frequency fire intervals perpetuated by wi regrass and pine needles. Hardwoods produce litter of lower flammability than pines or wiregrass, which can l ead to longer fire return intervals in gaps, ultimately exerting a negative feedback on the systems pyrogenicity and perpetuating a hardwood midstory (Williamson and Black 1981, Rebe rtus et al. 1993, Jacq main et al. 1999). When a large portion of the overstory is removed, th is negative feedback cycle is likely to be exacerbated by the subsequent lack of needlefall and the potential loss of wiregrass due to harvest-related soil disturbance. These are the two fine fuel inputs essential to maintaining a frequent fire return interval in a longleaf pine ecosystem, and w ithout them the keystone species in the system will not regenerate (Williamson and Black 1981, Clewell 1989). The ability of longleaf pine seedlings to surv ive and emerge from the grass stage will also be impacted if there are not adequate fine fuels to carry fire through the system and keep hardwoods suppressed. Longleaf seedlings are hi ghly intolerant of competition and will not initiate height growth if an excessive amount of competing vegetation is present (Boyer 1963, Haywood 2000, Ramsey et al. 2003). One approach to these problems, which has re ceived increasing attention in recent years, is to develop an uneven-aged silvicultural system that replicates natural smaller-scale patterns of disturbance and succession so as not to disrupt th e overall continuity of fine fuels and understory 52

PAGE 53

dynamics (Palik et al. 2002, Mitchell et al. 2006, Brockway et al. 2006). It has long been observed that longleaf pine seed lings regenerate only where di sturbances have created canopy openings (Boyer 1963). Several rece nt studies have looked at patte rns of regeneration in natural gaps as well as gaps created by groupor single-tree selection. Naturally-regenerating longleaf pine seedlings on a xeric sandhill were clustered near the center of gaps but were absent around the outer edge of each gap, an observation whic h led Brockway and Outcalt (1998) to propose the existence of a 12-16m seedling exclusion zone resulting from 1) higher fire intensity due to high litter accumulation at the bases of trees ar ound the gap edge, and 2) competition with the root systems of adult pines around the gap edge. Some studies of longl eaf pine regeneration have supported this hypothesis. As basal area in a mature longl eaf pine stand increased, firerelated seedling mortality also increased and seedling size decr eased (Grace and Platt 1995), while higher light intensity and N availability in gap centers were correlated with seedling survival and growth (Palik et al. 1997). Though subsequent research has not found evidence for a seedling exclusion zone per se, suppressed growth with increasing proximity to adults has been shown in all studies. Container-g rown seedlings planted in gaps exhibited maximum growth in response to high light intensity at the center of gaps as small as .1-ha, but no strong response to nutrient availability (McGuire et al. 2001). Others have reported that seedling survival was actually higher toward the outer edges of gaps, though the seedlings able to survive in the center were significantly larger than t hose closer to the edges (Gagnon et al. 2003, Pecot et al. 2007). Results from a trenching experiment suggest that this tradeoff is due to underground facilitation by adult longleaf pines at gap edge s vs. greater light availability at the gap center (Pecot et al. 2007). Trenching also revealed that hardwoods were immediately released when belowground competition from mature canopy trees was remove d (Pecot et al. 2007). The upshot of these 53

PAGE 54

recent studies is that small group selection a nd single-tree selecti on are recommended as silvicultural treatments for l ongleaf pine systems in which bot h biodiversity conservation and timber yield are desired. These methods enable new seedlings to establish at acceptably high survival rates without a major disruption in the supply of fine fuels or excessive hardwood release. But what of a system in which there are no longleaf pines in the overstory, and the desired outcome is a comple te conversion of the overstory from slash pine (Pinus elliottii Engelm.) to longleaf pine? The same issues with hardwood control and fine fuel supply must be considered. It has recently been shown that a gr adual replacement approach is also an effective method for making the conversion from slash to longleaf. The single-tree selection method was used to create plots with variable basal area densities, into whic h longleaf pine seedlings were introduced in conjunction with various mechanical and herbicidal hardwood control treatments. High rates of seedling survival were achieved at all densities, though rates of grass stage emergence were low and were not expected to in crease until the next harvest cycle (Kirkman et al. 2007). The second approach to slash-longleaf convers ion, and the one employed in this study, is to remove the entire slash pine canopy, replant with longleaf pine seedlin gs, and use mechanical and herbicidal treatments to control competing ve getation. The natural analog of this type of clearcut is a very larg e-scale disturbance, such as a hurri cane (Mitchell et al. 2006). Since both longleaf pine seedlings and native herbaceous sp ecies respond positively to full sunlight, this method, too, has the potential to restore a long leaf pine canopy and a diverse, herbaceous understory. Some of the potential benefits of this approach, as compared to single-tree and group-selection methods, are: 1) less logging dama ge to the site over time because only one 54

PAGE 55

harvest is required, 2) lower management costs, 3) lower personnel skill level and inventory information needs, 4) higher short-term timber output, and 5) better equipment access and maneuverability for effective site preparation in an area with severe understory competition, such as a saw palmetto-dominated flatwoods (Palik et al. 2002, Brockway et al. 2006). If desired, subsequent management of the new longleaf pi ne stand can follow a single-tree or groupselection prescription in order to perpetuate the structural and functional diversity of the system over the long-term. The biggest challenges facing this method will be 1) to achieve control of competing shrubs without negatively impacting native understory vegetation, 2) to maintain or enhance wiregrass cover, and 3) to successf ully apply fire to the system in spite of the loss of overstory fine fuel inputs. The studies described above indicate that supplemental hardwood control in the form of herbicide or mechanical treatment will almost always be necessary after clearcutting a pine overstory, because of the rele ase of hardwoods and the lack of fine fuels for prescribed fire. Even in a groupor single-tree selection syst em, additional hardwood control may be necessary if pyrogenicity is already comp romised due to a past histor y of fire suppression and an entrenched shrub component. In the single-tr ee selection study described above, the herbicide treatment was the only one in wh ich hardwood stem density did not increase over the course of the study (Kirkman et al. 2007). Following overstory harvest, longleaf pine pl antings must be carefully managed in order to achieve survival and growth rates adequate for stand replacement. A reasonable range of survival rates for containerized pl antings is 60-75%, but in order to achieve this range of survival rates, site preparation is necessa ry before seedlings are planted (Brockway et al. 2006). Grasses and other herbaceous species are the most seriou s short-term competitors for newly established 55

PAGE 56

longleaf seedlings, while woody species control is im portant for the development of fine fuels to perpetuate the ecosystem in the longer-term (Haywood 2005). On a frequently-burned site dominated by herbaceous species, prescribed fire may provide adequate site preparation, but on sites with heavy woody plant competition, mechan ical or herbicide treatments are necessary (Brockway et al. 2006). Intens ive mechanical site preparatio n techniques such as bedding, disking, and harrowing are effective for longleaf pine establishment (Johnson and Gjerstad 2006), but are very detrimental to wiregrass (Cle well 1989) and are therefor e not options on sites where understory restoration is a goal. Ther efore, herbicide, which removes competing vegetation without disturbi ng the soil, has potential utility as a component of a sustainable forest management strategy. In order to gain a better unders tanding of the role of herbicid e in sustainable longleaf pine silviculture, we compared the effects of five post-planting, banded herbicide treatments on the growth and survival of containeri zed longleaf pine seedlings in a coastal flatwoods in the Florida panhandle: 1) Sulfometuron methyl (Oust, E.I. du Pont de Ne mours and Company, Wilmington, Del.), 2) Imazapyr (Arsenal, BASF Corporation, Research Triangle Park, NC), 3) Hexazinone (Velpar L, E.I. du Pont de Nemours and Co mpany, Wilmington, Del.), 4) A tank mix of Oust and Velpar L, and 5) a no-herbicide control. These herbicide treatments were chosen because they are commonly used for loblolly ( Pinus taeda L.) and slash pine release in the Southeast, but their effects on longleaf pine growth and flatwoods community composition are not well studied. A recent survey by Shepard et al. (2004) reported that the forestry herbicide most commonly used in the Southeast was Arsenal, followed by tank mixes of Arsenal + Oust and Oust + Velpar. These tank mixes are used to achieve broader spectrum control, but as a result they may also have more negative implications for native plant communities (Keyser 56

PAGE 57

and Ford 2006). Sulfometuron methyl and hexazinone are effective as ove r-the-top pine release applications on longleaf pi ne seedlings (Haywood 2000, Ramsey and Jose 2004, Haywood 2005), but effects of a mixed applic ation on longleaf seedlings or on the understory of a natural flatwoods system have not been studied. Imazapyr mixed with triclopyr was effective as a preplanting application to release longleaf pine seedlings on a poor ly-drained site (Knapp et al. 2006), but information on imazapyr as an over-thetop release treatment for longleaf is also limited. We examined the longer-term effects of these herbicide treatments on longleaf pine seedlings five growing seasons after initial application (see Ranasinghe 2003 for first year results). We then conducted a dormant-season pres cribed fire on all treatment plots and assessed post-fire survival at the beginning of the sixth growing season. Methods Study Area The study was conducted at the Point Washi ngton State Forest, which is located on the gulf coast of Walton County, Florida (30 20 N, 86 04 W). The 20-acre study site is part of a larger mesic flatwoods/wet flat woods matrix, dominated by saw palmetto ( Serenoa repens W. Bartram), gallberry ( Ilex glabra A. Gray), rhizomatous oaks ( Quercus minima Small and Q. elliottii Wilbur), blueberries ( Vaccinium myrsinites Lam. and V. darrowii Camp.), huckleberries ( Gaylussacia dumosa Torr. and A. Gray and G. frondosa L. Torr and A. Gray ex. Torr), bluestem grasses ( Andropogon virginicus var. glaucus (L.) Heck and Andropogon virginicus L.) and wiregrass ( Aristida stricta ). The majority of the site can be classified as mesic, but due to subtle topographi cal variation across the site some areas are at the wetter end of the mesic flatwoods spectrum and some are at the drier end. For th is reason, we used a randomized complete block design, with three bl ocks laid out in mesic-wet areas and three 57

PAGE 58

blocks laid out in mesic-dry areas. A pre-trea tment vegetation survey identified wetter areas by the presence of indicator species such as sedge s (Cyperaceae). The soils of the study site are sandy, siliceous, thermic, aeric alaquods of the Leon series (Spodosols consisting of deep, poorly to very poorly drained soils derived from marine parent material (Ranasinghe 2003)). Yearly weather patterns during the study period were va riable. Total annual precipitation during the study period was 103cm in 2002, 200cm in 2003, 135cm in 2004, 146cm in 2005, and 106cm in 2006 (NOAA 2007). Rainfall in the months following the prescribed fire was very low, with totals of .5cm in March 2007, 4cm in April 2007, and .3cm in May 2007 (NOAA 2007). Prior to August 2001, the 20-acre study site was a planted slash pine stand with a mean stand age of about 26 years, which had been burned on a 3-year rotation by the Division of Forestry since its acquisition of Point Washington State Forest in 1992. A pre-harvest, baseline data survey was conducted in the areas slated to become study blocks in June 2001. In August 2001, the entire overstory was harves ted and the site was prepared for the study with a single, light roller chopping followed by a prescribed burn in October 2001. In December 2001, six study blocks, each divided into five 36.6m x 24.4m treatment plots, were established according to the aforementioned randomized complete block design. Each treatment plot was hand-planted with 100 containerized longleaf pine seedlings laid out in 10 rows, with 3.1m between rows and 1.8m between the seedlings in each row. Treatments Within each block, each 100-seedling plot was randomly selected to receive one of five herbicide treatments in March of 2002: 1) Sulfometuron methyl (Oust) at 0.26 kg a.i./ha. This is a moderate application rate for sulfometuron methyl; other studies have us ed 0.16 kg a.i./ha (Lauer and Glover, 1998) and 0.21 kg ai/ha (Shiver and Mart in 2002, Keyser and Ford 2006), while the manufacturer58

PAGE 59

recommended rate for herbaceous weed control in longleaf pine stands is 0.10 to 0.42 kg a.i./ha. Sulfometuron methyl is a selective herbicide primarily effective for controlling herbaceous species (Lauer and Glover 1998), and mixing with hexazinon e is recommended if broader spectrum control is desired (DuPont 2007). 2) Hexazinone: (Velpar L) at 0.56 kg a.i./ha. Recommended application rates for hexazinone range from 2.0 to 6.7 kg a.i./ha (E.I. Du Pont De Nemours), while rates successfully used to control hardwoods in sandhills ecosystems have ranged from 1.1 kg a.i./ha to 2.4 kg a.i /ha (Brockway et al. 1998, Provencher et al. 2001) The hexazinone application rate in our study was therefore very low. This rate was chosen because hexazinone had been deemed effective at the aforementioned rates, and we were interested in whether an even lower application rate would provide vegeta tion control. 3) Sulfometuron + Hexazinone mix (0.26 kg a.i./ha Sulfometuron methyl + 0.56 kg a.i./ha Hexazinone). This is a common tank mix used thr oughout the Southeast (Shepard et al. 2004) with the potential to control a broad sp ectrum of woody and herbaceous plants. 4) Imazapyr (Arsenal) at 0.21 kg a.i./ha. For a restoration project, this was a moderate application rate for imazapyr. Signifi cant hardwood control and habitat improvements have been obtained using imazapyr at 0.42 kg a. i./ha (Jones and Chamberlain 2004), and at all rates from 0.08 kg a.i./ha to 0.24 kg a.i..ha (t hough higher rates did not offer significant improvements over lower rates) (K eyser and Ford 2006). Arsenal is not labeled for use over longleaf pine at the recommended rates of 0.32 to 0.42 kg a.i./ha until after the end of the second growing season (BASF 2007), and has caused stunted growth in loblolly pi ne seedlings (Barber 1991). We chose the rate of 0.21 kg a.i./ha in order to determine whether a lower-thanrecommended application rate would prove less injurious to pine seedlings. 59

PAGE 60

The reasons for using a banded post-plan ting herbicide application on unshielded seedlings were both operational and environmenta l. Prior to harvest, the study site already contained many desirable understory species (including wiregrass) and only a few weedy species, none of which were exotic (Ranasinghe 2003). The banded application allowed us to pinpoint the delivery of herbicide to an area directly around planted seedlings, while native vegetation remained intact in the strips between rows. This application method also lowered the cost of herbicide treatment relative to a broadcast application. The herbicides were applied in a 1.2m ba nd over the seedling rows using a backpack sprayer. Within each 10-row treatment plot, three rows received a second herbicide application in March of 2003, three rows received a second app lication in April of 2003, and four rows were left with only the first-year treatment. The results from these subplots were pooled with the first year data for most parameters because of a l ack of significant differences between one and twoyear applications. Data Collection and Analysis In November 2006, the root collar diameter (RCD) and height to the top of bud were measured on every seedling. Following a prescribed fire in February of 2007 (Chapter 2), postfire seedling mortality was assess ed in June 2006. Survival rate was calculated as a percentage of each 10-tree, 56m2 treatment row. Rate of grass-stage release was calculated as a percentage of surviving seedlings in each treatment row with a bud height greater than 12cm, following Haywood (2000). Height and RCD co mparisons were made separately for seedlings in the grass stage (GS) and out of the grass stage (OOGS). The study addressed only the main effects of herbicide treatment, and tests of these eff ects were not dependent on the assumption of no treatment x block interaction. Block effects were therefore tr eated as random effects in a univariate ANOVA model with two independent variables: treat ment with Block&Random as 60

PAGE 61

a covariate. Data were log-transformed where necessary to meet the assumptions of ANOVA. Significant differences between treatments were separated with the Tukey-Kramer HSD test. Post-fire survival was analyzed with ANCOVA, us ing pre-fire survival as a covariate. All parameters were analyzed in JMP IN version 5 (SAS Institute, Inc.). Results Seedling Survival Five growing seasons after treatment (Novemb er 2006), survival rate varied significantly among herbicide treatments ( = 0.05, p < 0.0001) (Figure 3-1). The highest survival rates occurred in the hexazinone and control treatme nts (80% and 78.5%, respectively) (Figure 3-1). Survival rates were significantly lower in th e sulfometuron + hexazinone (sulfo + hexa) and imazapyr treatments (64.2% and 63.2%, respectively). It should be noted that the mortality in the imazapyr and sulfo + hexa treatments o ccurred shortly after herbicide application (Ranasinghe 2003), and there was little subsequent mort ality in any of the treatments. None of the treatment plots experienced significant mortality following fire (Figure 3-2). However, there were some slight differences in fire-related mortal ity among the different treatments, and as a result, surv ival rate no longer varied signifi cantly by treatment following fire (p = 0.238). In the imazapyr treatment, the surv ival rate dropped by only .7% following fire, as compared to decreases of 3% to 3.5% in the other treatments. Percent Out of Grass Stage Five growing seasons after herbicide application (November 2006), the mean grass-stage release rate varied significantly among treatments (p < 0.0001). The highest level of grass-stage release after four years was obs erved in the imazapyr treatment, though both sulfo + hexa and hexazinone also significantly ra ised the release rate relative to the control (Figure 3-3). 61

PAGE 62

Growth: Out of Grass Stage Seedlings The mean RCD, height, and SVI of seedlin gs out of the grass stage (OOGS) varied significantly among treatments (p = 0.0013, p < 0.0001, and p < 0.0001, respectively). In all parameters, imazapyr showed the greatest improveme nt relative to the co ntrol, though sulfo + hexa also significantly improved height relative to the contro l (Table 3-1, Figure 3-4, Figure 35). RCD of OOGS seedlings varied less among treat ments than height or SVI, which is to be expected due to the preferential investment in height growth characteri stic of longleaf pine seedlings in the bolting stage. Growth: Grass Stage Seedlings Root collar diameter (RCD), height, and SVI of grass-stage seedlings varied significantly among treatments (p < 0.0001, p = 0.003, and p = 0.0015, respectively) (Table 3-1, Figure 3-5), with the smallest seedlings in the sulfometuron treatment. Among s eedlings that were still in the grass stage after five growing seasons, none of the herbicide treatments showed significant improvements relative to the control. This pa ttern has been observed since the first growing season after treatment (R anasinghe 2003), though it was unclear whether the lower seedling size was due to seedling injury fr om sulfometuron or the poor sh rub control evident in the sulfometuron treatment since the fi rst growing season (Chapter 2). Discussion Five growing seasons after treatment, imazapyr plots had a grass stag e release rate more than twice that of the control, and the mean stem volume index of OOGS seedlings in the imazapyr treatment was also more than double that in the control. Although direct application of imazapyr to recently-planted seedlings caused significant mortality in the first year after planting, the overall post-fire surviv al rate of 62.5% in this treatmen t is still within the acceptable range of 60-75% recommended for adequate over story restocking (Brockway et al. 2006). 62

PAGE 63

Seedlings in this treatment experienced a fire-r elated mortality rate of only .06%, whereas fire mortality was around 3% in all of the other treatments. As a resu lt, the differences in survival rate between imazapyr and the other treatments were no longer significant following fire. Though the imazapyr treatment provided the be st overall pine re lease results, both hexazinone and the sulfometuron + hexazinone ta nk mix also showed promise as over-the-top longleaf pine release treatments for flatwoods ecosystems. Significant improvements in the grass-stage release rate were ach ieved with both of these treatments, and the OOGS seedlings in plots treated with the tank mix were significantly ta ller than those in the control. An absolute comparison between the effectiveness of thes e two formulations versus imazapyr is not warranted because we used a very low rate of hexazinone. However, based on success others have had using hexazinone for longleaf pi ne release (Haywood 2000, Ramsey and Jose 2004, Haywood 2005), we can expect that at a high er rateboth alone and in the tank mixthis herbicide would have compared more favorably with imazapyr. Contrary to the results previously observed with longleaf pine seedlings planted in an old field (Ramsey and Jose 2004), sulfometuron alone did not provide any pine release benefits over the control. This was most likely due to the higher degree of woody compe tition on this site, against which sulfometuron was significantly less effective than the other treatments (Chapter 2). For grass stage seedlings, variation among treatments was much less pronounced. Grass stage seedlings in the sulfometuron treatment were significantly smaller than those in the control, but none of the other treatments di ffered significantly from the cont rol in any of the grass stage parameters. A possible explanation for this patte rn is that as vegetati on regrew in the years following application, any seedlings not released relatively early in the process were suppressed and remained in the grass stage. As differen ces in understory vegeta tion between treatments 63

PAGE 64

became less pronounced over the years, differences in biomass accumulation rates between suppressed seedlings probably al so decreased. In contrast, th e earlier in the process a pine seedling was released, the better its chances ma y have been of gaining a competitive advantage for light, water, and soil resources. The early gains made by seedlings in the imazapyr plots (Ranasinghe 2003) apparently translated into an increasing competitive advantage over the years, leading to much larger seedlings in these plot s even as vegetation differences between plots became less pronounced. In a scenario such as this, where the dual goals of silviculture and ecosystem restoration dictate light site prepara tion and banded herbicide a pplication in spite of the heavy flatwoods competition, early release fr om the grass stage may be critical to the ultimate success of the planted pines. The site preparation burn conducted followi ng overstory harvest in 2001 consumed the last of the accumulated slash pine needles, and, though none of the herbicide treatments decreased wiregrass cover, distri bution of wiregrass across the site was patchy (Chapter 2). Five years after initial herbicide treatment, we had di fficulty applying a prescribed fire of sufficient intensity to control shrubs a phenomenon which was undoubtedly exacerbated by length of time since the application of herbicide and the need to apply fire during the dormant season (Chapter 2). The lack of fine fuels was also probably a major factor limiting our ability to apply fire effectively, given the key ro le played by dead pine needles in the pyrogenicity of a longleaf pine system (Williamson and Black 1981, Noss 1989). Our results support the theory proposed by Kirkman et al. (2007) that cl earcutting, even when the goal is a total canopy conversion from slash pine to longleaf pine, may ultimately hi nder the success of a restoration project. The overstory and understory effects of imazapyr, hexazinone, and the sulfometuron + hexazinone tank mix lined up well with the restora tion goals for this site. These treatments, 64

PAGE 65

especially imazapyr, released pines while also maintaining or improving wiregrass cover and overall cover of native herbaceous species. While tank mixes have been widely studied in silvicultural applications, they have understandably been avoided as habitat treatments thus far, out of concern that a broad spectrum mix will nega tively impact native species (Keyser and Ford 2006). However, our results indicate that this tank mix not only released pines, but had positive impacts on the understory, which might have been even more significant if a higher level of hexazinone had been included in th e mix (Chapter 2). An important caveat to this conclusion is that the use of a banded applica tion method may have been a key to the positive result. On pine plantations in the Virginia piedmont, broad cast application of an imazapyr + sulfometuron methyl tank mix significantly lowered herbaceous speci es diversity relative to the control and to imazapyr-only treatments, and banded application wa s suggested as an alternative for sites with wildlife habitat concerns (Keyser and Ford 2006). Recent studies of sustainable silviculture in longleaf pine systems have cited the potential negative impacts of herbicide as a justification for the use of gr oup or single-tree selection, since these harvesting methods should enhance the effec tiveness of prescribed fire as a management alternative to herbicide (Mitchell et al. 2006, Peco t et al. 2007, Kirkman et al. 2007). However, we observed only beneficial impacts on understo ry species following chemical treatment, indicating that herbicide can in fact play a positive supporting role in conservation-oriented longleaf pine silviculture. Based on the result s of this study, we recommend banded application of imazapyr as an ecological re storation tool for flatwoods sites with aggressive understory shrubs. 65

PAGE 66

Table 3-1. Root collar diameter (mm), height (mm), and stem volume index (basal area2 x height) of grass stage and OOG S seedlings five years after initial herbicide treatment. Root Collar Diameter (mm)Height (mm) Stem Volume Index Out of Grass Out of Grass Out of Grass Stage 25.4a23.6ab22.3b25.0a25.1a25.4a23.6ab22.3b25.0a25.1a Co *Letters denote significant differences between means in the same column at = 0.05 based on Tukey-Kramer HSD test P-values are for randomized block ANOVA within columns (significant at = 0.05). ntrol Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Control Imazapyr Sulfometuron Sulfo + Hexa Hexazinone Treatment ANOVA Grass Stage 39.5abc41.8a38.6c39.4bc40.6ab39.5abc41.8a38.6c39.4bc40.6abp < 0.0001 Stage Grass Stage Stage Grass Stage 63.1ab70.2a62.2b71.0a65.4ab63.1ab70.2a62.2b71.0a65.4ab492ab467ab388b519a499a 492ab467ab388b519a499a492ab467ab388b519a499a389.1c663.6a490.8bc513.1b484.1bc389.1c663.6a490.8bc513.1b484.1bcp < 0.0001 p = 0.0015 6332b14,542a9437b10,042b6332b6332b14,542a9437b10,042b6332bp < 0.0001 p = 0.003 p = 0.0013 66

PAGE 67

20% 30% 40% 50% 60% 70% 80% 90% ControlImazapyrSulfoSulfo + HexaHexa Treatment% Surviving (trees/row )a b a b a Figure 3-1. Mean survival rate by treatment after five growing seasons, calculated as # surving / # planted in each treatment row. Means not sharing the same letter are significantly different (Tukey-Kramer HSD). 67

PAGE 68

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% ControlImazapyrSulfoSulfo+HexaHexa Treatment% Surviving (trees/row) Pre-Fire Post-Fire Figure 3-2. Percent seedling survival (trees/row), before and after fire. None of the treatments experienced significant mortality du e to fire (Tukey-Kramer HSD, = 0.05). 68

PAGE 69

0% 10% 20% 30% 40% 50% 60% 70% 80% ControlImazapyrSulfoSulfo + HexaHexa Treatment% out of grass stagec a c b ab Figure 3-3. Mean grass-stage releas e rate by treatment after five gr owing seasons, calculated as # released / # surviving in each treatment row. Means not sharing the same letter are significantly different (Tukey-Kramer HSD). 69

PAGE 70

100 200 300 400 500 600 700 800 ControlImazapyrSulfoSulfo + HexaHexazinone TreatmentMean height (mm) c a bc b bc Figure 3-4. Mean height (mm) of out of grass stage seedlings by treatment. Means not sharing the same letter are significantly different (Tukey-Kramer HSD, = 0.05) 70

PAGE 71

100 150 200 250 300 350 400 450 500 550 600 ControlImazapyrSulfoSulfo + HexaHexazinone TreatmentMean SVI 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 ControlImazapyrSulfoSulfo + HexaHexazinone TreatmentMean SVIab b ab a aa)b a b b bb)Grass Stage Out of Grass Stage Figure 3-5. Mean stem-volume index (SVI) of a) grass stage seedlings and b) OOGS seedlings by treatment. Means not sharing the same letter are significantly different (TukeyKramer HSD, = 0.05). 71

PAGE 72

CHAPTER 4 SUMMARY AND CONCLUSIONS One of the biggest challenges to succe ssful restoration of the longleaf pine ( Pinus palustris Mill.) ecosystem is the persis tence of shrubs in the unde rstory, which suppress longleaf pine seedlings as well as native herbaceous plants. Herbicide can be used as a supplement to fire in order to enhance shrub control, but must be studied carefully because of the potential for negative impacts on native plants. Information is lacking about the effects of herbicide on natural longleaf pine flatwoods communities. We used a banded a pplication of three herbicides and one tank mix as shrub cont rol treatments following harves t of a slash pine stand and replanting with containerized longleaf pine seed lings in a mesic-wet flatwoods. The herbicides tested were Arsenal (imazapyr), Oust (sulfometuron methyl), Velpar L (hexazinone), and an Oust + Velpar L tank mix. Imazapyr herbicide significantly improved longleaf pine seedling growth, due to its effectiveness at controlling competing shrubs during the first growing seasons after planting. Though these shrub control effects were shortlived, the short term shrub control allowed herbaceous species to gain more of a foothold in the understor y, and significant increases in herbaceous cover relative to the c ontrol were recorded in imazapyr -treated plots four years after initial treatment. In a longleaf pine silviculture project with both timber and conservation goals, this was a very desirable result, because of the high level of biodive rsity in the herbacous component of longleaf pine ecosystems (Walker and Peet 1983, Kirkman et al. 2001). No negative impacts on understory species richness, diversity, evenness, or community composition resulted from the application of this herbic ide. Though over-the-top application of imazapyr caused significant longle af pine seedling mortality, the mean survival rate of seedlings in imazapyr-treated plots was 62.5% after five gr owing seasons and 61.8% after a fifth-year 72

PAGE 73

prescribed fire, which is stil l within the acceptable range of 60-75% recommended for adequate overstory restocking (Brockway et al. 2006). We therefore recommend imazapyr for use in sustainable longleaf pine silvic ulture, as it enhanced both long leaf pine seedling growth and herbaceous cover in the understory. Hexazinone, both alone and in a tank mix with sulfometuron, also showed promise as both a longleaf pine rele ase treatment and as an understory restoration tool. Four years after treatment, hexazinone plots had significantly higher herbaceous cover than the control, and plots treated with a sulfometuron + hexazinone tank mix had significantly higher wiregrass cover than the control. Significant improvements in the gras s stage release rate were achieved with both of these treatments, and released seedlings in plots treated with the tank mix were significantly taller than those in the control. An absolute comparison between these two formulations versus imazapyr is not warranted because we used a ve ry low rate of hexazinone. However, based on success others have had using hexazinone for longleaf pine release (Haywood 2000, Ramsey and Jose 2004, Haywood 2005) and rest oration (Brockway et al. 1998, Brockway and Outcalt 2000, Provencher et al. 2001) we can expe ct that at a higher rateboth alone and in the tank mixthis herbicide would have compared more favorably with imazapyr. Sulfometuron methyl did not improve pine rel ease relative to the cont rol, and instead had a releasing effect on the tw o most common oak species ( Quercus minima and Quercus elliottii ), indicating that this is not an appropriate herbicide for flatwoods with a high density of oaks. However, in spite of having higher shrub cove r and lower herbaceous cover than the other herbicide treatments, sulfometuron plots also had the highest herbaceous species richness, diversity, and evenness four years af ter initial treatment. One reas on for this may have been that wiregrass cover in the sulfom eturon plots was somewhat lowe r than the other herbicide 73

PAGE 74

treatments, leaving more space for other small herbaceous species. These gains in species richness, diversity, and evenness are therefore un likely to persist through future fire cycles because it will be more difficult to apply prescribed fire to this treatment. In support of this prediction, we observed that su lfometuron was the only treatment in which species richness, diversity, and evenness did not increase following the prescr ibed fire in 2007. The long-term sustainability of any longleaf pine ecosystem restoration project depends on the managers ability to apply prescribed fire with enough frequency a nd intensity to control shrubs. The herbicides applied in this study pr ovided short-term shr ub control and had positive impacts on longleaf pine growth, wiregrass cove r, and overall herbaceous cover. However, complete removal of the slash pine overstory at the start of the pr oject resulted in a lack of fine fuels and presumably also encouraged aggressi ve regrowth of shrubs due to the loss of competition from mature pines (Brockway et al. 1998, Kush et al. 1999, McGuire et al. 2001, Kirkman et al. 2007, Pecot et al. 2007). These hindrances were most likely exacerbated by the length of time since herbicide a pplication (five years since first application and four years since second application), as well as the use of dormant -season fire due to the sites proximity to a highway and residential areas (Waldrop et al. 1992, Drewa et al. 2002, Drewa et al. 2006). Though three of the four herbicid es we tested were successful at releasing longleaf pine seedlings, the herbicide-related gains in wiregr ass and overall herbaceous cover may be lost in the coming years due to the aggressive resprouti ng of shrubs. We therefore make the following recommendation to those managing flatwoods si tes for both timber and biodiversity: when possible, some overstory pines shou ld be retained following harvest to serve as a source of fine fuels. Banded application of imazapyr is r ecommended in order to improve longleaf pine 74

PAGE 75

seedling growth and enhance herbaceous cover in the understory, espe cially on sites where growing-season fire is not an option. 75

PAGE 76

APPENDIX SPECIES LIST SpTable A-1. Species lis t. ecies name Common name Family Life Form Andropogon virginicus var. glaucus Chalky Bluestem PoaceaeGrass Andropogon virginicus var. virginicus Broomsedge BluestemPoaceaeGrass Aristida spiciformis Bottlebrush ThreeawnPoaceaeGrass Aristida stricta Wiregrass PoaceaeGrass Asclepias cinerea Carolina Milkweed ApocynaceaeForb Balduina uniflora Oneflower HoneycombheadAsteraceaeForb Baptisia lanceolata Gopherweed FabaceaeLegume Calamovilfa curtissii Curtiss's Sandgrass PoaceaeGrass Carphephorus odoratissimus Deer's Tongue AsteraceaeForb Chrysopsis gossypina subsp. cruseana Cruise's Goldenaster AsteraceaeForb Chrysopsis mariana Maryland GoldenasterAsteraceaeForb Cladoniaceae Deer Moss CladoniaceaeMoss Ctenium aromaticum Toothache Grass PoaceaeGrass Cynanchum angustifolium Gulf Coast SwallowwortApocynaceaeForb Cyrilla racemiflora Swamp Titi CyrillaceaeShrub Desmodium lineatum Sand Ticktrefoil FabaceaeLegume Desmodium strictum Pinebarren TicktrefoilFabaceaeLegume Dichanthelium aciculare Needleleaf WitchgrassPoaceaeGrass Dichanthelium ensifolium Cypress Witchgrass PoaceaeGrass Dichanthelium ovale Eggleaf Witchgrass PoaceaeGrass Dichanthelium erectifolium Erectleaf Witchgrass PoaceaeGrass Dichanthelium strigosum Roughhair WitchgrassPoaceaeGrass Drosera capillaris Pink Sundew DroseraceaeCarnivorous Eryngium yuccifolium Button RattlesnakemasterApiaceaeForb Eupatorium capillifolium Dog Fennel AsteraceaeForb Eupatorium compostifolium Yankeeweed AsteraceaeForb Eupatorium mohrii Mohr's ThoroughwortAsteraceaeForb Eupatorium pilosum Rough Boneset AsteraceaeForb Eupatorium rotundifolium Roundleaf ThoroughwortAsteraceaeForb Euphorbia inundata Florida Pineland SpurgeEuphorbiaceaeForb Eurybia eryngiifolia Thistleleaf Aster AsteraceaeForb Euthamia graminifolia Flattop Goldenrod AsteraceaeForb Gaylussacia dumosa Dwarf Huckleberry EricaceaeShrub Gaylussacia frondosa Dangleberry EricaceaeShrub Gelsemium sempervirens Yellow Jessamine GelsemiaceaeVine Gratiola hispida Rough Hedgehyssop VeronicaceaeForb Houstonia procumbens Roundleaf Bluet RubiaceaeForb Hypericum brachyphyllum Coastalplain St. John's WortClusiaceaeShrub Hypericum hypericoides St. Andrew's Cross ClusiaceaeShrub Hypoxis sessilus Yellow Star Grass HypoxidaceaeMonocot Nomenclature follows Wunderlin (2004). Species ar e divided into the following life forms: Grass = member of the Poaceae family, Sedge = member of the Cyperaceae family, Monocot = other monocot, Legume = member of the Fabaceae family, Carnivorous = carnivorous forb, Forb = other forb, Fern = member of the order Filicophyta, Shrub = woody un derstory species, Tree = woody overstory species. 76

PAGE 77

Table A-1. Continued. Species name Common name Family Life Form Ilex coriacea Giant Gallberry AquifoliaceaeShrub Ilex glabra Gallberry AquifoliaceaeShrub Ilex opaca American Holly AquifoliaceaeShrub Kalmia hirsuta Hairy Wicky EricaceaeShrub Lachnanthes caroliniana Carolina Redroot HaemodoraceaeMonocot Lachnocaulon auceps Whitehead Bog ButtonEriocaulaceaeMonocot Liatris gracilis Slender Blazing Star AsteraceaeForb Liatris tenuifolia Shortleaf Blazing StarAsteraceaeForb Lycania michauxii Gopher Apple ChrysobalanaceaeForb Lyonia lucida Fetterbush EricaceaeShrub Mimosa quadrivalvis Sensitive Vine FabaceaeLegume Mitreola sessilifolia Swamp Hornpod LoganiaceaeForb Muhlenbergia capillaris Cutover Muhly PoaceaeGrass Opuntia humifusa Prickly Pear CactaceaeForb Orchidaceae Orchid sp. OrchidaceaeMonocot Osmunda cinnamomea Cinnamon Fern OsmundaceaeFern Photinia pyrifolia Red Chokeberry RosaceaeShrub Pinus elliottii Slash Pine PinaceaeTree Piteopsis graminifolia Grassleaf GoldenasterAsteraceaeForb Polygala lutea Orange Milkwort PolygalaceaeForb Polygala nana Candyroot PolygalaceaeForb Polygonella polygama October Weed PolygonaceaeForb Pteridium aquilinum Bracken Fern DennstaedtiaceaeFern Pterocaulon pychnostachium Blackroot AsteraceaeForb Quercus elliottii Runner Oak FagaceaeShrub Quercus incana Bluejack Oak FagaceaeTree Quercus minima Dwarf Live Oak FagaceaeShrub Rhexia alifanus Savannah Meadow-BeautyMelastomataceaeForb Rhexia petiolata Fringed Meadow-BeautyMelastomataceaeForb Rhynchospora filifolia Threadleaf BeaksedgeCyperaceaeSedge Rhynchospora plumosa Plumed Beaksedge CyperaceaeSedge Sabatia brevifolia Shortleaf RosegentianGentianaceaeForb Scleria ciliata Fringed Nutrush CyperaceaeSedge Serenoa repens Saw Palmetto AracaceaeShrub Seriocarpus tortifolius Whitetop Aster AsteraceaeForb Smilax auriculata Earleaf Greenbrier SmilacaceaeMonocot Smilax glauca Wild Sarsparilla SmilacaceaeMonocot Smilax laurifolia Laurel-leaf GreenbrierSmilacaceaeMonocot Smilax pumila Sarsparilla Vine SmilacaceaeMonocot Solidago odora Sweet Goldenrod AsteraceaeForb Sporobolus junceus Pineywoods DropseedPoaceaeGrass Stillingia sylvatica Queen's Pleasure EuphorbiaceaeForb Stylisma patens Coastalplain DawnflowerConvolvulaceaeForb Stylisma villosa Hairy Dawnflower ConvolvulaceaeForb Symphiotrichum adnatum Scaleleaf Aster AsteraceaeForb Tephrosia hispidula Sprawling Hoarypea FabaceaeLegume Tragia urens Wavyleaf Noseburn EuphorbiaceaeForb Vaccinium darrowii Darrow's Blueberry EricaceaeShrub Vaccinium elliottii Elliott's Blueberry EricaceaeShrub Vaccinium myrsinites Shiny Blueberry EricaceaeShrub Vernonia angustifolia Tall Ironweed AsteraceaeForb Vitis rotundifolia Muscadine Grape VitaceaeVine Xyris caroliniana Carolina Yelloweyed GrassXyridaceaeMonocot 77

PAGE 78

LIST OF REFERENCES BASF Corporation. 2007. Arsenal AC label. http://www.cdms.net/LDat/ld542007.pdf BASF Agricultural Products, Research Triangle Park, NC. Behm, A.L., Duryea, M.L., Long, A.J., and W.C. Zipperer. 2004. Flammability of native understory species in pine flatwood and hardwood hammock ecosystems and implications for the wildland-urban interface. Inte rnational Journal of Wildland Fire 13:355-365. Boyd, R.S., Freeman, J.D., Miller, J.H., and M.B. Edwards. 1995. Forest herbicide influences on floristic diversity seven years after broadcast pine release treatments in central Georgia, U.S.A. New Forests 10: 17-37. Boyer, W.D. 1963. Development of longleaf pi ne seedlings under parent trees. USDA Forest Service, Southern Forest Experiment Stati on, Research Paper SO-4, New Orleans, LA. Brockway, D.G. and C.E. Lewis. 1997. Long-term effects of dormant season prescribed fire on plant community diversity, structure, and productivity in a longleaf pine-wiregrass ecosystem. Forest Ecology and Management 96: 167-183. Brockway, D.G. and K.W. Outca lt. 1998. Gap-phase regenerati on in longleaf pine wiregrass ecosystems. Forest Ecology and Management 106: 125-139. Brockway, D.G., Outcalt, K.W., and R.N. Wilkin s. 1998. Restoring longleaf pine wiregrass ecosystems: plant cover, diversity, and bioma ss following low-rate hexazinone application on Florida sandhills. Forest Ecology and Management 103: 159-175. Brockway, D.G. and K.W. Outcalt. 2000. Rest oring longleaf pine wi regrass ecosystems: hexazinone application enhances effects of pres cribed fire. Forest Ecology and Management 137: 121-138 Brockway, D.G., Outcalt, K.W., and W.D. Boye r. 2006. Pages 95-133 in Jose, S., Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration. Springer Science + Bu siness Media, LLC., New York. Clewell, A.F. 1989. Natura l history of wiregrass ( Aristida stricta Michx., Graminae). Natural Areas Journal 9: 234-245. Croker, T.C. Jr. 1975. Seedbed preparation aids natural regeneration of longleaf pine. USDA Forest Service Research Paper. Southe rn Forest Experiment Station No. SO-112. Daubenmire, R.F. 1959. Canopy coverage method of vegetation analysis. Northwest Scientist 33: 43-64. Drewa, P.B., Platt, W.J., and E.B. Moser. 2002. Fire effects on resprouting of shrubs in headwaters of southeastern long leaf pine savannas. Ecology 83(3): 755-767. 78

PAGE 79

Drewa, P.B., Thaxton, J.M., and W.J. Platt. 2006 Responses of root-crown bearing shrubs to differences in fire regimes in Pinus palustris (longleaf pine) savannas: Exploring oldgrowth questions in second growth system s. Applied Vegetation Science 9: 27-36. Dufrene, M. & P. Legendre. 1997. Species asse mblages and indicator species: the need for a flexible asymmetrical appr oach. Ecological Monographs 67: 345-366. DuPontTM. 2007. Velpar L Label. http://www.dupont.com/ag/vm/products/velparl.html E. I. du Pont de Nemours and Company,Wilmington, DE. DuPontTM. 2007. Oust XP Label. http://www.dupont.com/ag/us/prodi nfo/prodsearch/information/H65144.pdf E. I. du Pont de Nemours and Company,Wilmington, DE. Frost, C.C. 2006. History and Future of the L ongleaf Pine Ecosystem. Pages 9-47 in Jose, S., Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration. Springer Science + Business Media, LLC., New York. Gagnon, J.L., Jokela, E.J., Moser, W.K., and D.A. Huber. 2003 Dynamics of artificial regeneration in gaps within a longleaf pine flatwoods ecosystem. Forest Ecology and Management 172: 133-144. Gilliam, F.S. and W.J. Platt. 1999. Effects of long-term fire exclusion on tree species composition and stand structure in an old-growth Pinus palustris forest. Plant Ecology 140: 15-26 Glitzenstein, J.S., Streng, D.R ., and D.D. Wade. 2003. Fire frequency effects on longleaf pine ( Pinus palustris Mill.) vegetation in South Carolina a nd northeast Florida, USA. Natural Areas Journal 23(1): 22-37. Grace, S.L., and W.J. Platt. 1995. Effects of a dult tree density and fire on the demography off pregrass stage juvenile longleaf pine ( Pinus palustris Mill.) Journal of Ecology 83: 75-86. Hay-Smith, L. and G.W. Tanner. 1994. Responses of wildlife food plants to a forest herbicide. University of Florida Cooperative Extension Fact Sheet WRS-9. Haywood, J.D. 2000. Mulch and hexazinone herbicid e shorten the time longleaf pine seedlings are in the grass stage and increase height growth. New Forests 19: 279-290. Haywood, J.D. 2005. Effects of herbaceous and woody plant control on Pinus palustris growth and foliar nutrients through six growing seasons. Forest Ecology and Management 214: 384397. Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., Orze ll, S.L., E.L. Bridges. 1993. Longleaf pine vegetation of the Southern Atlantic and Eastern Gulf Coast regions: a preliminary classification. Proceedings of the Tall Timbers Fire Ecology Conference 18: 83-104 79

PAGE 80

Hiers, J.K., Wyatt, R., and R.J. Mitchell. 2000. The effects of fire regime on legume reproduction in longleaf pine savannas: is a season selective? Oecologia 125: 521-530. Hurst, G.A. and W.E. Palmer. 1987. Vegetati on responses to imazapyr for pine release. Proceedings of the Southern Weed Science Society 40: 247. Hobbs, R.J., Currall, J.E.P., and C.H. Gimingham. 1984. The use of thermocolor pyrometers in the study of heath fire be havior. Journal of Ecology 72: 241-250. Iverson, L.R., Yaussy, D.A., Rebbeck, J., Hutchi nson, T.F., Long, R.P., and A.M. Prasad. 2004. A comparison of thermocouples and temperatur e paints to monitor spatial and temporal characteristics of landscape-scale prescribed fi res. International Journal of Wildland Fire. 13(3): 311-322. Jacqmain, E.I., Jones, R.H., and R.J. Mitchell. 1999. Influences of frequent cool-season burning across a soil moisture gradient on oak commun ity structure in longleaf pine ecosystems. American Midland Naturalist 141: 85-100. Jones, J.D., and M.J. Chamberlain. 2004. Efficacy of herbicides and fire to improve vegetative conditions for northern bobwhites in mature pine forests. Wildlife Society Bulletin 32(4): 1077-1084. Johnson, R. and D.J. Gjerstad. 2006. Restoring the overstory of longleaf pine ecosystems. Pages 271-295 in Jose, S., Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restor ation. Springer Science + Business Media, LLC., New York. Keyser, P.D., and V.L. Ford. 2006. Wildlife habitat and herbicides: an evaluation of a widely applied tank mix. Southern Journal of Applied Forestry 30(1): 46-51. Keddy, P.A., Smith, L., Campbell, D.R., Clark, M., and G. Montz. 2006. Patterns of herbaceous plant diversity in a Louisiana pine savanna. Applied Vegetation Science 9: 17-26. Kirkman, L.K., Mitchell, R.J., Helton, R.C., a nd M.B. Drew. 2001. Productivity and species richness across and environmental gradient in a fire-dependent ecosystem. American Journal of Botany 88(11): 2119-2128. Kirkman, L.K., Mitchell, R.J., Kaiser, M.J., Peco t, S.D., and K.L. Coffey. 2007. The perpetual forest: using undesirable species to bridge restoration. Journal of Applied Ecology 44: 604614. Knapp, B.O., Wang, G.G., Walker, J.L., and S. Cohen. 2006. Effects of site preparation treatments on early growth and surv ival of planted longleaf pine ( Pinus palustris Mill.) seedlings in North Carolina. Forest Ecology and Management 226: 122-128. 80

PAGE 81

Kush, J.S., R.S. Meldahl, and W.D. Boyer. 1999. Understory plant co mmunity response after 23 years of hardwood control treatm ents in natural longleaf pine (Pinus palustris ) forests. Canadian Journal of Forest Research 29: 1047-1054. Lauer, D.K. and G.R. Glover. 1998. Early pine response to control of herbaceous and shrub vegetation in flatwoods. Southern Journal of Applied Forestry 15(4): 201-208. Leach, M.K. and T.J. Givnish. 1996. Ecologica l determinants of species loss in remnant prairies. Science 273: 1555-1558. Litt, A.R., Herring, B.J. and L. Provencher. 2001. Herbicide effects on ground-layer vegetation in southern pinelands, USA: a review. Natural Areas Journal 21(2): 177-188. McNease. L.L. and G.A. Hurst. 1991. Preh arvest treatment with imazapyr for seed tree regeneration and habitat improvement. Proceed ings of the Southern Weed Science Society 44: 265. Mehlman, D.W. 1992. Effects of fire on plant community composition of North Florida second growth pineland. Bulletin of the Torrey Botanical Club 119: 376-383. McCune, B. and Grace, J. 2002. Analysis of Ecological Communities. MJM Software Design. Gleneden Beach, OR. McGuire, J.P., Mitchell, R.J., Moser, E.B., Pecot, S.D., Gjerstad, D.H., and C.W. Hedman. 2001. Gaps in a gappy forest: plant resources, longleaf pine regeneration, and understory response to tree removal in longleaf pine savann as. Canadian Journal of Forest Research 31: 765-778. Miller, J.H., R.S. Boyd, and M.B. Edwards. 1999. Floristic diversity, stand structure, and composition 11 years after herbicide site preparat ion. Canadian Journal of Forest Research 29: 1073-1083. Mitchell, R.J., Hiers, J.K., OBrien, J.J., Jack, S. B., and R.T. Engstrom. 2006. Silviculture that sustains: the nexus between silviculture, fre quent prescribed fire, and conservation of biodiversity in longleaf pine forests of the sout heastern United States. Canadian Journal of Forest Research 36: 2724-2736. Neary, D.G., J.E. Smith, B.F. Swindel, and K.V. Miller. 1991. Effects of forestry herbicides on plant species diversity. Proceedings of the Southern Weed Science Society 44: 266-272. NOAA Satellite and Information Service: National Climatic Data Center. 2007. Panama City, FL Station 5N. http://www4.ncdc.noaa.gov/cgiwin/ wwcgi.dll?wwDI~StnSrch~StnID~20004593 Noss, R.F. 1989. Longleaf pine and wiregras s: keystone components of an endangered ecosystem. Natural Areas Journal 9: 211-213. 81

PAGE 82

Noss, R.F., LaRoc, III, E.T., and J.M. Scott. 1995. Endangered ecosystems of the United States: a preliminary assessment of loss and degrada tion. USDI National Biological Service, Washington, D.C., Biological Report No. 28. Olson, M.S. and W.J. Platt. 1995. Effects of ha bitat and growing season fires on resprouting of shrubs in longleaf pine savannas. Vegetatio 119: 101-118. Outcalt, K.W., Williams, M.E., and O. Onokpise. 1999. Restoring Aristida stricta to Pinus palustris ecosystems on the Atlantic Coastal Plain, U.S.A. Restoration Ecology 7(3): 262270. Peet, R.K., Wentworth, T.R., and P.S. White 1998. Method for recording vegetation composition and structure. Castanea 63: 262-274. Palik, B.J., Mitchell, R.J., Houseal, G., and N. Pederson. 1997. Effects of canopy structure on resource availability and seed ling responses in a longleaf pine ecosystem. Canadian Journal of Forest Research 27: 1458-1464. Palik, B.J., Mitchell, R.J., J.K. Hiers. 2002. M odeling silviculture after natural disturbance to sustain biodiversity in the longleaf pine ( Pinus palustris ) ecosystem: balancing complexity and implementation. Forest Ecology and Management 155: 347-356. Pecot, S.D., Mitchell, R.J., Palik, B.J., Moser, E.B., and J.K. Hiers. 2007. Competitive responses of seedlings and understory plants in longleaf pine w oodlands: separating canopy influences above and below ground. Canadian Journal of Forest Research 37: 634-648. Peet, R.K. 2006. Ecological clas sification of longleaf pine woodl ands. Pages 51-93 in Jose, S., Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration. Springer Science + Business Media, LLC., New York. Peet, R.K. and D.J. Allard. 1993. Longleaf pine-dominated vegetation of the southern Atlantic and eastern Gulf Coast region, USA. Pro ceedings of the Tall Timbers Fire Ecology Conference 18: 45-81. Platt, W.J. and S.L. Rathbun. 1993. Dynamics of an old-growth longleaf pine population. Proceedings of the Tall Timb ers Fire Ecology Conference 18: 275-297. Platt, W.J., Beckage, B., Doren, R.F. & H.H. Slater. 2002. Interactions of large-scale disturbances: prior fire regimes and hurrica ne mortality of savanna pines. Ecology 83: 15661572. Platt, W.J., Carr, S.M., Reilly, M., and J. Fahr 2006. Pine savanna overstorey influences on ground-cover biodiversity. A pplied Vegetation Science 9: 37-50. 82

PAGE 83

Provencher, L., Herring, B.J., Gordon, D.R., R ogers, H.L., Galley, K.E.M., Tanner, G.W., Hardesty, J.L., and L.A. Brennan. 2001. Effects of hardwood reduction techniques on longleaf pine sandhill vegetation in nort hwest Florida. Restoration Ecology 9: 13-27. Ramsey, C.L., S. Jose, B.J. Brecke, and S. Merritt. 2003. Growth response of longleaf pine ( Pinus palustris Mill.) seedlings to fertilization and herbaceous weed control in an old field in southern USA. Fore st Ecology and Management 172(2-3): 281-289. Ramsey, C.L., and S. Jose. 2004. Growth, surviv al, and physiological eff ects of hexazinone and sulfometuron methyl applied overtop of longleaf pine seed lings. Southern Journal of Applied Forestry 28(1): 48-54. Ranasinghe, S. 2003. Role of herbicides in longl eaf pine flatwoods restoration: pine growth, understory vegetation response, and fate of applie d herbicides. Masters thesis. University of Florida, Gainesville. Ranasinghe, S., Jose, S., and C. Ramsey. 2005. Understory species dynamics following herbicidal release treatments on a longleaf pine site in the Southern U.S.A. In New Research on Forest Ecosystems, A.R. Burk, ed. Nova Science Publishers, Inc. Rebertus, A.J., Williamson, G.B., and W.J. Platt. 1993. Impacts of temporal variation in fire regime on savanna oaks and pines. Pro ceedings of the Tall Timbers Fire Ecology Conference 18: 215-225. Shepard, J.P., Creighton, J., and H. Duzan. 2004. Forestry herbicides in the United States: an overview. Wildlife Society Bulletin 32(4): 1020-1027. Varner III, J.M., and J.S. Kush. 2004. Remnant ol d-growth longleaf pine (pinus palustris Mill.) savannas and forests of the Southeastern USA; status and threats. Natural Areas Journal 24: 141-149. Waldrop, T.A., White, D.L., and S.M. Jones. 1992. Fire regimes for pine-grassland communities in the southeastern United States. Forest Ecology and Management 47: 195210. Walker, J., and R.K. Peet. 1983. Composition and species diversity of pine-wiregrass savannas of the Green Swamp, Nort h Carolina. Vegetatio 55: 163-179. Walker, J. and A.M. Silletti. 2006. Restoring the Ground Layer of Longleaf Pine Ecosystems. Pages 297-333 in Jose, S., Jokela, E., and Miller, D.L., editors. The Longleaf Pine Ecosystem: Ecology, Silviculture, and Restoration. Springer Science + Business Media, LLC., New York. Welch, J.R., Miller, K.V., Palmer, W.E., and T.B. Harrington. 2004. Response of understory vegetation important to the northern bobwh ite following imazapyr and mechanical treatments. Wildlife Society Bulletin 32(4): 1071-1076. 83

PAGE 84

Wilkins, R.N., W.R. Marion, D.G. Neary, and G. W. Tanner. 1993. Vascular plant community dynamics following hexazinone site preparati on in the lower Coastal Plain. Canadian Journal of Forest Research 23: 2216-2229 Williamson, G.B. and E.M. Black. 1981. High te mperature of forest fires under pines as a selective advantage over oaks. Nature 293: 643-644. 84

PAGE 85

85 BIOGRAPHICAL SKETCH Johanna Freeman is a native of western New Yo rk state. She received a B.S. in natural resources from Cornell University in 2001, where her studies focused on animal behavior and environmental policy. She worked as a land-use planner at the New York City Department of Parks and Recreation for three years prior to en tering the masters program in interdisciplinary ecology at the University of Florida.