<%BANNER%>

Seed Bank and Regeneration Ecology of an Annual Invasive Sedge (Scleria lacustris) in Florida Wetlands

Permanent Link: http://ufdc.ufl.edu/UFE0022036/00001

Material Information

Title: Seed Bank and Regeneration Ecology of an Annual Invasive Sedge (Scleria lacustris) in Florida Wetlands
Physical Description: 1 online resource (147 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: annual, bank, hydrology, invasive, plant, seed, seedling, wetland
Agronomy -- Dissertations, Academic -- UF
Genre: Agronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Our study supports a working hypothesis that the annual flood/dry cycle characterizing Florida's seasonal wetlands is the primary factor driving the incidence of the invasive annual sedge, Scleria lacustris, by selecting strategies integral to seed bank function and seedling regeneration. Particular strategies of seed bank persistence, seed bank survival, dormancy break and seedling emergence were identified and demonstrated as influenced by the hydrologic regime. Basic characteristics of the S. lacustris seed bank were evaluated in 2004 at two wetlands in south central Florida where soil was sampled twice yearly (before and after seedling emergence) and seedling emergence was monitored continuously. Seed extraction from the soil samples demonstrated the presence of up to 2,331 seeds m-2 with 88% viability in the top 9 cm of soil before springtime germination. Although seeds were concentrated in the upper 3 cm, seed depletion from germination increased with depth. Regardless of significant levels of seed bank depletion and a resulting seedling monoculture, a viable seed bank at the end of summer was indicative of a functionally persistent seed bank strategy. Persistence was considered to be a response of the innate dormancy, which was confirmed in 2005. Seed storage studies, conducted from 2004 to 2006 under field and controlled conditions, demonstrated that dormancy was influenced by the hydrologic environment of the seed bank which additionally functioned to affect viability and germination potential. Seed bank persistence served to provide for seedling populations for at least four years under annual, experimental flood/dry cycles. Under continuous inundation, seed bank viability was maintained in a presumably active metabolic state, for at least four years. On the other hand, a dry storage environment maintained viability through dormancy for the short term, but led to sooner and greater mortality rates. Fluctuating and intermediately moist conditions also tended to induce a higher state of physiological activity in the seed bank. In 2005 and 2006 seed bank evaluation and vegetation monitoring were conducted along three transects of the hydrologic gradient at a single depression marsh. These climactically opposing years supported experimental trials revealing seedling regeneration as restricted to well drained soils following surface water dry down. Furthermore, seedling regeneration, survival and adult productivity were significantly different along the gradient, as indicated by a cut off from the optimum conditions provided by the previously inundated portion of the marsh. The contributing influence of vegetation gaps that developed in the previous inundated regime was significant. Results support acceptance of the working hypotheses by indicating that both seed bank and regeneration strategies are highly promoted/selected for by the hydrologic regime of the seasonal marsh and may explain the invasive colonization and apparent advantages of this annual sedge in seasonal wetlands of Florida.
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 (Ph.D.)--University of Florida, 2008.
Local: Adviser: Fox, Alison M.
Local: Co-adviser: Langeland, Kenneth A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022036:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022036/00001

Material Information

Title: Seed Bank and Regeneration Ecology of an Annual Invasive Sedge (Scleria lacustris) in Florida Wetlands
Physical Description: 1 online resource (147 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: annual, bank, hydrology, invasive, plant, seed, seedling, wetland
Agronomy -- Dissertations, Academic -- UF
Genre: Agronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Our study supports a working hypothesis that the annual flood/dry cycle characterizing Florida's seasonal wetlands is the primary factor driving the incidence of the invasive annual sedge, Scleria lacustris, by selecting strategies integral to seed bank function and seedling regeneration. Particular strategies of seed bank persistence, seed bank survival, dormancy break and seedling emergence were identified and demonstrated as influenced by the hydrologic regime. Basic characteristics of the S. lacustris seed bank were evaluated in 2004 at two wetlands in south central Florida where soil was sampled twice yearly (before and after seedling emergence) and seedling emergence was monitored continuously. Seed extraction from the soil samples demonstrated the presence of up to 2,331 seeds m-2 with 88% viability in the top 9 cm of soil before springtime germination. Although seeds were concentrated in the upper 3 cm, seed depletion from germination increased with depth. Regardless of significant levels of seed bank depletion and a resulting seedling monoculture, a viable seed bank at the end of summer was indicative of a functionally persistent seed bank strategy. Persistence was considered to be a response of the innate dormancy, which was confirmed in 2005. Seed storage studies, conducted from 2004 to 2006 under field and controlled conditions, demonstrated that dormancy was influenced by the hydrologic environment of the seed bank which additionally functioned to affect viability and germination potential. Seed bank persistence served to provide for seedling populations for at least four years under annual, experimental flood/dry cycles. Under continuous inundation, seed bank viability was maintained in a presumably active metabolic state, for at least four years. On the other hand, a dry storage environment maintained viability through dormancy for the short term, but led to sooner and greater mortality rates. Fluctuating and intermediately moist conditions also tended to induce a higher state of physiological activity in the seed bank. In 2005 and 2006 seed bank evaluation and vegetation monitoring were conducted along three transects of the hydrologic gradient at a single depression marsh. These climactically opposing years supported experimental trials revealing seedling regeneration as restricted to well drained soils following surface water dry down. Furthermore, seedling regeneration, survival and adult productivity were significantly different along the gradient, as indicated by a cut off from the optimum conditions provided by the previously inundated portion of the marsh. The contributing influence of vegetation gaps that developed in the previous inundated regime was significant. Results support acceptance of the working hypotheses by indicating that both seed bank and regeneration strategies are highly promoted/selected for by the hydrologic regime of the seasonal marsh and may explain the invasive colonization and apparent advantages of this annual sedge in seasonal wetlands of Florida.
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 (Ph.D.)--University of Florida, 2008.
Local: Adviser: Fox, Alison M.
Local: Co-adviser: Langeland, Kenneth A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022036:00001


This item has the following downloads:


Full Text

PAGE 1

SEED BANK AND REGENERATION ECOL OGY OF AN ANNUAL INVASIVE SEDGE ( Scleria lacustris ) IN FLORIDA WETLANDS By COLETTE CLARE JACONO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Colette Clare Jacono

PAGE 3

3 To my Grandmother Agness Mungall Dell January 25, 1910 July 18, 2007 and Great Aunt May Mungall Clark February 12, 1923 February 9, 2008

PAGE 4

4 ACKNOWLEDGMENTS I am indebted to my advisor Alison Fox and to my committee members Kaoru Kitajima, Kenneth Langeland, Jerry Bennett, and Leo Nico, who each in their own way offered academic enlightenment, to my husband, Richard Smith, for acting as an unbiased sounding board in science, and to my friend Sherry Bostick, for ex pertise and patience in processing the electronic dissertation. I thank Karen Shepherd H upp, Richard Smith, Brent Bachelder, Susan Herbolsheimer, Myriah Richerson, Betsy Salisb ury, Lisa Huey, Sherry Bostick and Shawna Whaley and for their assistance, endurance, a nd cheerful companionship during travel, field and laboratory work. I thank George Yeargin, Shan e Ruessler, and Bob Lewis for their creativity and interest in engineering equipment and in building and outfitting th e greenhouse. I thank Howard Jelks and my other colleagues at the Ce nter for Aquatic Resources Studies for their encouragement and support. I thank Ken Snyder and the St. Johns Water Management District as well as the Disney Wilderness Preserve fo r providing logistic support. I gratefully acknowledge the Biological Resources Division of the United States Geological Survey, the Bureau of Invasive Plant Manage ment with the Florida Departme nt of Environmental Protection, the Agronomy Department at the University of Florida, and the Florida Exotic Pest Plant Council for their financial s upport in this study.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION..................................................................................................................15 Background..................................................................................................................... .15 Scleria lacustris ...............................................................................................................15 2 SEED BANK CHARACTERIZATION OF Scleria lacustris AN ANNUAL INVASIVE SEDGE IN FLORIDA WETLANDS.................................................................26 Introduction................................................................................................................... ..26 Materials and Methods....................................................................................................30 Field Sites and Experimental Design.......................................................................30 Seed Bank Sampling................................................................................................31 Seed Bank Viability Assessment..............................................................................32 Seedling Monitoring.................................................................................................33 Statistical Analysis...................................................................................................34 Results........................................................................................................................ .....34 Seed Bank Viability Assessment..............................................................................34 Seed Bank Depth......................................................................................................34 Seed Bank Density...................................................................................................35 Seed Bank Depletion................................................................................................35 Seed Bank Persistence..............................................................................................36 Seed Bank to Seedling Stage Transition..................................................................36 Discussion..................................................................................................................... ...37 Seed Bank Viability.................................................................................................37 Seed Bank Depth......................................................................................................37 Seed Bank Density...................................................................................................38 Seed Bank Depletion and Persistence......................................................................39 Seed Bank to Seedling Stage Transition..................................................................40 3 SEEDLING REGENERATION IN Scleria lacustris, AN INVASIVE ANNUAL WETLAND SEDGE...............................................................................................................53 Introduction................................................................................................................... ..53

PAGE 6

6 Materials and Methods....................................................................................................55 Year 1.......................................................................................................................55 Year 2.......................................................................................................................56 Statistical Analyses..................................................................................................58 Results........................................................................................................................ .....58 Discussion..................................................................................................................... ...60 4 INFLUENCE OF A SEASONAL HYDROL OGIC REGIME ON THE SEED BANK AND REGENERATION ECOLOGY OF Scleria lacustris AN INVASIVE WETLAND SEDGE...............................................................................................................67 Introduction................................................................................................................... ..67 Materials and Methods....................................................................................................70 Field Site and Experimental Design.........................................................................70 Field Sampling and Monitoring...............................................................................71 Hydrology............................................................................................................71 Seed bank dynamics............................................................................................72 Plant demography................................................................................................73 Statistical analyses...............................................................................................74 Experimental Tests of Seed Dormancy....................................................................75 Field Experimental Tests of Seed Bank Survival.....................................................76 Year 2004 seed source.........................................................................................76 Year 2005 seed source.........................................................................................78 Statistical analyses...............................................................................................79 Outdoor Laboratory Tests of Seed Bank Survival...................................................79 Results........................................................................................................................ .....80 Field Sampling and Monitoring...............................................................................80 Hydrology............................................................................................................80 Seed bank dynamics............................................................................................81 Plant demography................................................................................................82 Relationships between hydrologic and seed bank/plant variables......................83 Experimental Tests of Seed Dormancy....................................................................88 Field Experimental Tests of Seed Bank Survival.....................................................89 Year 2004 seed source.........................................................................................89 Year 2005 seed source.........................................................................................90 Outdoor Laboratory Tests of Seed Bank Survival...................................................92 Discussion..................................................................................................................... ...93 Seed Bank Temporal Dynamics...............................................................................93 Seed Bank Function and Plant Demogr aphics in Relation to Hydrology................94 Dormancy.................................................................................................................97 Seed Bank Survival..................................................................................................98 Conclusion.....................................................................................................................101 5 CONCLUSION AND SYNTHESIS....................................................................................129

PAGE 7

7 APPENDIX A TETRAZOLIUM STAINING PATTERNS.........................................................................131 B 2005 PLANT COMMUNITY INVENTORY......................................................................133 C HYDROLOGY AT 2004 FIELD SITES..............................................................................136 D ORGANIC CONTENT AND SOIL MOISTURE OF SUBSTRATE CORES AT VERO BEACH STUDY AREA, MARCH 2004.............................................................................139 LIST OF REFERENCES.............................................................................................................141 BIOGRAPHICAL SKETCH.......................................................................................................147

PAGE 8

8 LIST OF TABLES Table page 1-1 Locality data for Scleria lacustris in Florida. Data compiled and verified by C. Jacono......................................................................................................................... .......22 2-1 Parameters for a linear model predicting vi ability of intact seed as determined with tetrazolium assay.............................................................................................................. ..42 2-2 Viable number of Scleria lacustris seed at three soil depth ranges with preemergence and postemergence sampling at seasonal marsh sites........................................................43 2-3 Intact number of Scleria lacustris seed at three soil depth ranges with preemergence and postemergence sampling at seasonal marsh sites........................................................43 2-4 Percentage of total viable seed w ith soil depth at seasonal marsh sites.............................44 2-5 Density of the Scleria lacustris seed bank at seasonal marsh sites....................................45 2-6 The coefficient of variation (CV) for seed bank sampling time for each research site.....46 2-7 Proportion of persistence in the Scleria lacustris seed bank (0-9 cm depth) at seasonal marsh sites...........................................................................................................46 2-8 Proportion of persistence in the Scleria lacustris seed bank with soil depth.....................47 2-9 Data for mean viable seed extracted at preemergence sampling time and the total number of seedlings for each station, both extrapolated to m2..........................................48 3-1 Year 1 and Year 2 results of green house experiments on the recruitment of Scleria lacustris seed in different hydrologic treatments...............................................................62 4-1 Parameters for a linear model predicting vi ability of intact seed as determined with tetrazolium assay..............................................................................................................102 4-2 Density of Scleria lacustris seeds in the 2005 soil seed bank as estimated from the mean number of viable seed (m-2 1 std. dev.) extracted during preemergence and postemergence sampling times........................................................................................103 4-3 Density of Scleria lacustris seeds in the 2006 soil seed bank as estimated from the mean number of viable seed (m-2 1 std. dev.) extracted during preemergence and postemergence sampling times........................................................................................104 4-4 Coefficients of determination (r2) used to identify the strongest predictor for surface water, WavgSu (mean summer surface wate r) or WHi (annual surface water high) and for soil moisture, MavgSp (mean spring soil moisture) or MavgM (mean April soil moisture)................................................................................................................. ..105

PAGE 9

9 4-5 Density of Scleria lacustris seeds in the soil seed ba nk as categorized by annual surface standing water (water regime categ ory) transecting the gradient of the depression marsh..............................................................................................................106 4-6 Percent viability and persistence of Scleria lacustris seeds in the soil seed bank as categorized by annual surface standing water (w ater regime category) transecting the gradient of the depression marsh.....................................................................................107 4-7 The ANOVA summary of demographic data from a seasonal depression marsh categorized by annual surface standing water (water regime categories) transecting the hydrologic gradient....................................................................................................108 4-8 Viability and germination responses of 2005 fresh seed.................................................109 4-9 Seed fate of 2005 fresh seed after a germin ation trial of 21 wk (147 d) in an unheated greenhouse..................................................................................................................... ..109 4-10 Viability and germination results for 2004 seed after 5 months and 17 months storage at locations transecting the hydrologic gr adient of the depression marsh, and in ambient garage.................................................................................................................110 4-11 Viability, germination and seed death for 2005 seed after 5 months storage at locations transecting the hydrologic gradient of the marsh, and in ambient garage........111

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Distribution map of Scleria lacustris in North America....................................................21 2-1 Tetrazolium (1%) staining re sults for viable seed of Scleria lacustris ..............................49 2-2 Preemergence dispersion of intact seeds in replicate cores (0-9 cm depth) at A) Vero Beach and B) Kissimmee...................................................................................................50 2-3 Postemergence dispersion of intact seeds in replicate cores (0-9 cm depth) at A) Vero Beach and B) Kissimmee...................................................................................................51 2-4 Scatter plot of S lacustris seed bank density m-2 (preemergence) with the total number of seedlings m-2 that emerged at each station at the A) Vero Beach and B) Kissimmee study sites........................................................................................................52 3-1 Year 1 greenhouse experiment used to test the hydrologic conditions affecting regeneration of the annual sedge, Scleria lacustris ...........................................................63 3-2 Year 2 greenhouse experiment used to test the hydrologic conditions affecting regeneration of the annual sedge, Scleria lacustris ...........................................................64 3-3 Year 2 greenhouse experiment on the m echanism of regeneration, 39 days after initiation..................................................................................................................... ........65 3-4 Rate response of the treatments that de monstrated significant regeneration responses in Year 1 and Year 2 greenhouse experiments..................................................................66 4-1 The characteristic profile of a preemer gence soil core (15 cm diam. x 6 cm deep) extracted from the recently dewatered substrate of the Vero Beach depression marsh, April 2006..................................................................................................................... ...112 4-2 Scleria lacustris A) fruiting stems and B) shed seed collected Oct. 2005 from standing plants at Cell C, Blue Cypress Water Management Area, St. Johns Water Management District........................................................................................................112 4-3 March 2006 greenhouse germination tria ls at A) day 1, and B) day 21..........................113 4-4 Surface water fluctuation along T II, a 36m transect representing the hydroperiod of a depression marsh near Vero Beach, over 19 months, the duration of two consecutive growing seasons...........................................................................................113 4-5 Vero Beach marsh mi dway at Transect II........................................................................114

PAGE 11

11 4-6 Soil volumetric water content (represented as the mean of three readings for each station) for stations along the three tr ansects (TI, TII, and TIII) during the 2006 spring dry down period at the Vero Beach st udy site demonstrated the absence of a trend along the hydrologic gradie nt of the depression marsh..........................................115 4-7 Intact and viable seed bank for Scleria lacustris at sampling stations (0-7) along three permanent transects of the depression marsh..........................................................116 4-8 The 2006 preemergence seed bank (means for each station m) and seedlings produced at each station/m2.............................................................................................117 4-9 Adult biomass of the 2006 seedling c ohorts, quantified 18 September 2006, from Scleria lacustris at the Vero Beach depression marsh.....................................................118 4-10 Graphical summarization of Table 4-4 deta iling the greatest si gnificant effects of hydrology, surface water (WavgSu) and soil water (MavgM), on seed bank and plant variables at the Vero Be ach depression marsh, 2006......................................................119 4-11 Surface water depth measured at four loca tions along the marsh gradient of the Vero Beach study site...............................................................................................................120 4-12 Greenhouse germination rate response of 2004 seed source following 5 months storage........................................................................................................................ ......121 4-13 Greenhouse germination rate response of 2004 seed source following 17 months storage........................................................................................................................ ......122 4-14 Surface water depth measured along the gr adient of Transect I (TI) at the Vero Beach depression marsh...................................................................................................123 4-15 Greenhouse germination rate response of 2005 seed after five months storage..............124 4-16 Categorical format accounting for the fate of 2005 seed retrieve d after a five month overwintering storage to test seed bank survival.............................................................125 4-17 Experimental seed bank (2003 seed sour ce) stored under constant flooding at an outdoor laboratory remained viable for four years..........................................................126 4-18 Response over time in years of an experi mental seed bank (2003 seed source) stored under seasonally cyclic flood-dry conditions (early spring-summer dry down / later summer winter flooding) at an outdoor laboratory........................................................127 4-19 Logarithmic reduction of intact seed ( 2003 seed source) stored under experimentally induced seasonal hydrology (later summer-w inter flooding and spring-early summer dry down) at an outdoor laboratory.................................................................................128 A-1 Excised embryos from the seed of Scleria lacustris illustrate the various patterns observed with tetrazolium staining (1%) when assaye d for seed viability......................132

PAGE 12

12 C-1 Surface water fluctuation during the 2004 st udy period at the A) Vero Beach and B) Kissimmee seasonal marshes as measured with a meter stick at each station.................137 C-2 Soil moisture for the 2004 spring/early summer dry down period at the A) Vero Beach and B) Kissimmee study sites demons trated small differences among the 11 sampling stations with in monitoring dates......................................................................138 D-1 Gravimetric determination of percent mo isture and organic matter was made after drying and combustion of replicate soil cores.................................................................140

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SEED BANK AND REGENERATION ECOL OGY OF AN ANNUAL INVASIVE SEDGE ( Scleria lacustris ) IN FLORIDA SEASONAL WETLANDS By Colette C. Jacono May 2008 Chair: Alison M. Fox Major: Agronomy Our study supports a working hypothesis that th e annual flood/dry cycle characterizing Floridas seasonal wetlands is the primary factor driving the incidence of the invasive annual sedge, Scleria lacustris by selecting strategies integral to seed ba nk function and seedling regeneration. Particular strategies of seed ba nk persistence, seed bank survival, dormancy break and seedling emergence were identified and de monstrated as influenced by the hydrologic regime. Basic characteristics of the S lacustris seed bank were evaluate d in 2004 at two wetlands in south central Florida where soil was samp led twice yearly (bef ore and after seedling emergence) and seedling emergence was monitore d continuously. Seed extraction from the soil samples demonstrated the pr esence of up to 2,331 seeds m-2 with 88% viability in the top 9 cm of soil before springtime germination. Although seed s were concentrated in the upper 3 cm, seed depletion from germination increased with depth. Regardless of significant levels of seed bank depletion and a resulting seedling monoculture, a viable seed bank at th e end of summer was indicative of a functionally pe rsistent seed bank strategy. Persistence was considered to be a response of the innate dormancy, which was confirmed in 2005. Seed storage studies, conducted from 2004 to 2006 under field and controlled

PAGE 14

14 conditions, demonstrated that dormancy was in fluenced by the hydrologic environment of the seed bank which additionally functioned to affect viability and germination potential. Seed bank persistence served to provide for seedling popul ations for at least four years under annual, experimental flood/dry cycles. Under continuous inundation, seed bank viability was maintained in a presumably active metabolic state, for at le ast four years. On the other hand, a dry storage environment maintained viability through dormanc y for the short term, but led to sooner and greater mortality rates. Fluctu ating and intermediately moist c onditions also tended to induce a higher state of physiological activity in the seed bank. In 2005 and 2006 seed bank evaluation and ve getation monitoring were conducted along three transects of the hyd rologic gradient at a single depr ession marsh. These climactically opposing years supported experimental trials revealing seedling regene ration as restricted to well drained soils following surface water dry down. Furthermore, seedling regeneration, survival and adult productivity were significantly different along the gradient, as indicated by a cut off from the optimum conditions provided by the prev iously inundated portio n of the marsh. The contributing influence of vegetation gaps that developed in the previ ous inundated regime was significant. Results support ac ceptance of the working hypothese s by indicating that both seed bank and regeneration strategies are highly promoted/selected for by the hydrologic regime of the seasonal marsh and may explain the invasive colonization and apparent advantages of this annual sedge in seasonal wetlands of Florida.

PAGE 15

15 CHAPTER 1 INTRODUCTION Background The wetlands of peninsular Florida are characterized by hydroperiod, the depth and duration of surface water and soil saturation. A nnual fluctuations in hydroperiod are a direct response to rainfall patterns. In seasonal wetl ands, this is typically portrayed as dry surface conditions during the winter dry season and surface inundation with the advent of summer rains. In their natural state, the seasona l wetlands of Florida are part of vast ecosystems composed, like a tapestry, of complex marsh communities inters persed with sloughs, ponds and wet prairies. Their ecological function relies on the conservati on of the natural hydrologic regime (Miller et al. 1998; Toth 1993; Ki tchens et al. 2001). In 1999 and 2000, La Nina condition s resulted in lower than average rainfall and into 2001 Florida experienced severe drought conditions that exacerbated sp ring dry down, the recession of surface water, in many area wetlands (Steinman et al. 2002). At the same time, records of a new plant were increasing. Plants were identified as the introduced species, Scleria lacustris and associations were soon suspected between its occurrence and wetlands ex periencing effects of severe hydrologic drought. Regardless, little was possible in the way of scientific predictions on its ecology beyond recognition that the new species was annual in habit and thereby remained intrinsically dependent on seed for regeneration. Results of this study indica te that this single trait may contribute greatly to its apparent co mpetitive advantage and invasiveness in Florida wetlands. Scleria lacustris Scleria (Cyperaceae) is a genus primarily of tr opical and warm-temperate climates that stands apart from other genera in the sedge family for its unisexual flowers and hard, boney

PAGE 16

16 achenes (Kessler 1987). Scleria lacustris C. Wright, given the vernacular name Wrights Nutrush, is distinguished from native North American species by its wide leaves (1.5 to 2.5 cm), culms with terminal panicles that can exceed 2 m lengths, and slightly three-sided, nearly ovoid to ellipsoid achenes (Jacono 2001). Achenes (called nutlets, but, here after referred to as seeds) measure 2.95 to 3.95 mm long and 2.3 to 2.5 mm in diameter (Jacono unpublished) which are considered very large (Leck and Sch tz 2005). They are shining and green when fresh and turn marbled gray or tan to white upon drying. The smooth surface of the seed lacks specialized structures for dispersal. Neve rtheless, seeds float for severa l days before sinking (Jacono 2001) and are likely candidates for both primary and secondary dispersal by water (hydrochory), a common dispersal mode in wetland Cyperaceae (Leck and Sh tz 2005). Airboats and other contaminated equipment are capable of transpor ting seed across independ ent drainage regions. A hardened pad at the base of the seed, the hypogynium, is an important taxonomic and functional character throughout the genus Scleria (Core 1936; Kessler 1987). In S lacustris the hypogynium is obscurely triangular in shape and, becau se it is not completely sclerified, shrivels when the seed matures to initiate seed sh ed. During seedling emergence the hypogynium is pushed off for radicle and shoot emergence and, in light of the hardness of the achenes, is probably the place where water is im bibed. Seedlings are large, have plicate leaves, and can be recognized in the field by sheathing leaf bases an d roots tinged a dark purplish red. The seed can usually be found remaining attached to the seedli ng, at the junction of the primary root with the base of the stem, when pulled from the ground. Much of the plant surface including the three sided culms, the leaf margin, lower midrib, rachis and bracts, is harsh and cutt ing to the touch. This is due to minute prickles that develop as outgrowths of silica laden epidermal cells, a wi despread feature in th e Cyperaceae (Metcalfe

PAGE 17

17 1971). The size of the plants and extent of thei r prickles warrant wear ing gloves and full length clothing to protect the skin during field work. The native distribution of Scleria lacustris is Trans-Atlantic, extending to distant, yet scattered areas of Africa and the Neotropics. In Africa a smatte ring of records exist from the west coast, extending from Sierra Leone, to Gabon and Zaire, from the center of the continent in the Central African Republic south to Zambia and Botswana, and from Madagascar (Nelmes 1955; Hennessy 1985; Fairey 1972; Cook 2004). Plan ts were described as rare or locally distributed along river floodplains and at isolated marshes. The association with aquatic environments was emphasized by populations noted as growing in deep water. S lacustris was originally described as endemic to Cuba (C ore 1936). However, recent revision of the Neotropical members of the genus (Camelbeck e and Goetghebeur in prep.) has confirmed specimens from Costa Rica, Jamaica, Guyana, Suriname, French Guiana, Brazil and Paraguay. Nevertheless, within this widespread distribution S. lacustris is similarly regarded as a very rare species (K. Camelbecke, pers. comm.). Perhaps due to its rarity, S. lacustris has not been studied outsid e of taxonomic treatise and floristic inventories, and many basic life history traits are no t well understood. For example, it remains uncertain if plants in Africa might perennate (Nelmes 1955; Fairey 1972; Hennessy 1985) or if juveniles might grow submerged (Cook 2004). Hennessy (1985) suggested that the annual habit in S lacustris could serve well as a drought esca pe mechanism in tropical regions that experience seasonal drought. Florida specimens of Scleria lacustris were first collected in 1988, at a marsh within the headwaters of the St. Johns Rive r. Populations must have been developing at that time in a broad arc across the south central peninsula, because within the en suing years, plants were found

PAGE 18

18 in independent drainages far to the southwest and westward of th at initial local. The current distribution extends to more than twenty natu ral areas and seven counties within four major drainage regions of the state (F igure 1; Table 1). Many of the areas affected are hydrologically linked to expansive conservation wetland ecosys tems. Despite its widespread distribution, S. lacustris does not occur throughout Flor idas natural areas. Locality data indicate a particular affinity with herbaceous, seasonally inundated pl ant communities such as basin, floodplain, and depression marshes, as well as wet prairie. All experience surface flooding approximately 200 250 d/y (FNAI and FDNR 1990). Where found in Florida, S lacustris typically grows to dominate the local plant community. It occurs in high densities and is capable of reaching heig hts above 2 m. Under flooded conditions of late summer it changes from an upright growth habit to lodge, root at the nodes and sprawl across what natura lly exists as sparsely vegetate d and open water habitat. In this manner, the native community composition a nd structure become severely altered. In 2005, S lacustris was listed as a Category II i nvasive exotic, following consid eration of its increase in abundance and frequency in natura l areas (FLEPPC 2005). It has b ecome a species of concern to land managers and is suspected to obscure and di splace prey for sight feeding birds, especially the Florida snail kite ( Rostramus sociabilis ), an endangered species which commonly inhabits the expansive regions infested. In wetland plant communities, hydrologic fluctu ation is the key environmental factor selecting for life history strategies, species su rvival and population dynamics (van der Valk 1981; Keddy and Reznicek 1982; Seabloom et al. 2001). The cyclical patterns defined by the presence or absence of surface flooding directly affects seed regeneration and influences ecosystem processes, such as decomposition and nutrient availability which indirectly affect

PAGE 19

19 seedling growth and recruitment (Smith and Kadlec 1983; Gerritsen and Greening 1989). A premise basic to this study is that the annual hydrologic fluctuati ons that similarly characterize Florida marshes may be the key f actor governing the incidence of S lacustris It is hypothesized that seasonal water periodicity is the driving fo rce behind the function of the seed bank and the mechanisms of seedling regeneration in S lacustris The objectives are th reefold in regard to S lacustris : 1) to define the characteri stics that contribute to the dynamics of its seed bank, 2) to understand its mechanisms and demographic patterns of seedling regeneration, and 3) to identify the specific hydrologic conditions that at once both promote and restrict its incidence. To meet these objectives, seed bank analys is, demographic monitoring, and hydrologic measurement was completed at two wetlands in south central Florida between 2004 and 2006. In 2004 data was collected at Disney Wilderness Pr eserve, Kissimmee, Fla. at a floodplain marsh (KS) and at a depression mars h west of Vero Beach, Fla. (VB). In 2005 and 2006 data was collected entirely at the VB s ite. Chapter 2 details primary characteristics relevant to the S lacustris soil seed bank. Chapter 3 delineates th e mechanism for seedling regeneration in S lacustris Chapters 2 is based on 2004 data collected within the infe station zone at the two study sites, thus allowing comparison a nd contrast between the differen t habitats. Chapter 3 involves an experimental greenhouse study for investigati ng the influence of hydrology as a mechanism in seedling regeneration. The results from Chapters 2 and 3 provide a foundation for understanding the relationship between seed ba nk dynamics and population inciden ce, themes which are further addressed in Chapter 4. Chapter 4 provides co rrelative evidence from seed bank, plant and environmental variables registered along the hydr ologic gradient of the Vero Beach depression marsh in 2005 and 2006. Field work was compleme nted with experiment al studies to better substantiate seed bank dynamics. Seed surviv al experiments were assessed using greenhouse

PAGE 20

20 germination tests complemented with tetrazoli um viability assays. Chapter 5 provides a synthesis of the research chapters and a conclusion to the work as a whole.

PAGE 21

21 Figure 1-1. Distribution map of Scleria lacustris in North America. Map prepared by C. Jacono according to data in Table 1-1.

PAGE 22

22 Table 1-1. Locality data for Scleria lacustris in Florida. Data compiled and verified by C. Jacono. County Hydrologic unit Locale Year Voucher specimen Reporter of occurrence Latitude Longitude Brevard Upper St. Johns Jane Green Swamp 1988 Hall s n (FLAS) 28 02' N 80 48' W Brevard Upper St. Johns Jane Green Swamp 2000 Harrison #259 (FLAS) 28 01 22 N 074 48 38 W Hendry Caloosahatchee Okaloacoochee Slough State Forest 2002 McCollom s n (FLAS) 26 34 35.6 N 81 22 35.1 W Hendry Caloosahatchee Okaloacoochee Slough State Forest 2003 D. Lambers, FDOACS 26 29 25.9 N 81 14 52.5 W Indian River Upper St. Johns Blue Cypress Conservation Marsh, ca. 600 m E of N Blue Cypress Lake 2001 Jacono #204 (FLAS) 27 41.869 N 80 41.659 W Indian River Upper St. Johns Blue Cypress Conservation Marsh 2001 Nichols s n (FLAS) 27 41.869 N 80 41.659 W Indian River Upper St. Johns Blue Cypress Conservation Marsh, W of Lake Miami Ranch 2001 R. Bennetts USGS, BRD 27 41 34.5 N 80 41 40.92 W Indian River Vero Beach Blue Cypress Water Management Area, Jewish Federation 2001 C. Jacono, G. Nichols 27 40 41.8 N 80 36 10.4 W Indian River Vero Beach Blue Cypress Water Management Area, Ansin E 2001 Jacono #344 (NY) 27 41 30 N 80 36 20.7 W Indian River Vero Beach Blue Cypress Water Management Area, Ansin E Cell 3 2001 Jacono #358 (fcsc) 27 40 40.1 N 80 35 13.1 W

PAGE 23

23 County Hydrologic unit Locale Year Voucher specimen Reporter of occurrence Latitude Longitude Indian River Vero Beach Blue Cypress Water Management Area, Ansin E, Cell 3 2001 C. Jacono, R. Bennetts 27 40 51.3 N 80 34 54.7 W Indian River Vero Beach Blue Cypress Water Management Area, Ansin E. 2001 C. Jacono, G. Nichols 27 40 57.4 N 80 36 18.8 W Indian River Vero Beach Blue Cypress Water Management Area 2001 C. Jacono, G. Nichols 27 40 45.1 N 80 36 18.0 W Indian River Upper St. Johns Fort Drum Conservation Area, S Ft. Drum Creek 2001 K. Snyder, SJRWMD 27 34 12.7 N 80 45 50.1 W Indian River Upper St. Johns Fort Drum Conservation Area 2002 Read, SJRWMD 27 34 09 N 80 45 43 W Indian River Upper St. Johns Fort Drum Conservation Area 2006 K. Snyder, SJRWMD Indian River Vero Beach Florida Conservation Commission, cypress heads and seasonal marsh ponds 2003 T. Towles, C. Jacono 27 34 8.76 N 80 23 30.96 W Indian River Upper St. Johns Blue Cypress Conservation Marsh 2003 C. Jacono, K. Snyder 27 40 25.9 N 80 41 20.7 W Lee Big Cypress Ft. Myers, ca. 10 mi SE, herbaceous wetland 1989 Cox s n (FLAS) Lee Big Cypress Ft. Meyers, Griffen Dr., 0.5 mi N of Gateway Blvd., wet prairie 1990 Roessler s n (FTG) Lee Big Cypress Ft..Meyers, Commerce Lakes Dr., marsh pond 2001 Vandiver s n (fcsc) 26 34 33.06 N 81 44 33.48 W

PAGE 24

24 County Hydrologic unit Locale Year Voucher specimen Reporter of occurrence Latitude Longitude Lee Big Cypress Ft. Meyers, SW Florida Intern. Airport, Mitigation F, wet prairie and cypress dome depression marsh 2002 Jacono #411 (fcsc) 26 33 26.8 N 81 44 21.7 W Lee Big Cypress Ft. Meyers, SW Florida Intern. Airport, Mitigation F 2002 Jacono #410 (fcsc) 26 33 30.2 N 81 44 26.8 W Lee Big Cypress Ft. Meyers, Fuel Farm Rd, marsh pond; seedlings 2002 Jacono #409 (fcsc) 26 32 55.4 N 81 44 31.1 W Lee Big Cypress Ft. Meyers, Fuel Farm Rd, marsh pond; juveniles 2002 Jacono #475 (fcsc) Lee Big Cypress Wild Turkey Strand Preserve, cypress domes and strands 2004 J. Key, FGCU Lee Big Cypress Wild Turkey Strand Preserve, grazed cypress domes and disturbed wetlands 2005 K. Bradley, Inst. Reg. Conserv. Okeechobee Kissimmee Eagle Island Rd, ca.5 mi W of US 441 on SR 724 2000 Lane s n (FLAS) 27 29' N 80 55' W Osceola Kissimmee East Lake Tohopekaliga 2001 E. Harris, FDEP 28 19 35.52 N 81 17 1.26 W Osceola Kissimmee Lake Kissimmee, SW corner, near Hwy 60 and river 2001 E. Harris, FDEP 27 50 29.4 N 81 13 55.14 W Osceola Kissimmee Lake Kissimmee, E littoral, S of Overstreet Landing 2001 E. Harris, FDEP 27 55 44.46 N 81 13 41.46 W

PAGE 25

25 County Hydrologic unit Locale Year Voucher specimen Reporter of occurrence Latitude Longitude Osceola Kissimmee Gardner-Cobb Marsh, trail C-35, E of Long Hammock 2001 Jacono #359 (FLAS) 28 00 18.1 N 84 16 47.9 W Osceola Kissimmee East Lake Tohopekaliga 2002 C. Jacono, E. Harris 28 19 33.5 N 81 16 22.0 W Osceola Kissimmee Lake Hatchineha proper, lakeside S of cypress band 2002 E. Harris, FDEP 28 02 10.0 N 81 25 8.5 W Osceola Kissimmee Lake Tohopekaliga 2002 E. Harris, FDEP 28 11 6.66 N 81 20 48.66 W Osceola Kissimmee Lake Cypress 2002 E. Harris, FDEP 28 05 17.22 N 81 20 32.76 W Polk Kissimmee London Creek, Tract A 1999 MacGregor s n (USF) Polk Kissimmee Disney Wilderness Preserve, N side Lake Hatchineha, HU12 2001 Jacono #266 (FLAS) Polk Kissimmee W of Dead River, Johnson Island, McKinney Tract 2000 Hansen #12,894 (USF) Polk Kissimmee Disney Wilderness Preserve, HU10 water structure and "super marsh" 2002 Jacono #456 (fcsc) 28 02 33.8 N 81 24 37.4 W Polk Kissimmee Lower Reedy Creek, between Cypress Lake and Lake Hatchineha 2000 E. Harris, FDEP 28 4.098 N 81 21.559 W

PAGE 26

26 CHAPTER 2 SEED BANK CHARACTERIZATION OF Scleria lacustris AN ANNUAL INVASIVE SEDGE IN FLORIDA WETLANDS Introduction Seed banks are an integral component of vegetation dynamics (Simpson et al. 1989; Chambers and MacMahon 1994). Their particular importance in the perseverance of annual plants has been predicted with evolutionary models (Brown and Venable 1986) and demonstrated in natural wetlands (Leck and Gr aveline 1979; Parker and Leck 1985; Baldwin et al. 2001). Relative to habitat restoration and community management the seed bank present in natural areas is typically assesse d for its ability to contribute to floristic diversity (Keddy and Reznicek 1982; van der Valk and Rosburg 1997; Lee 1994; Wetzel et al. 2001; Smith et al. 2002; Leck and Leck 2005). These types of assessm ents are usually based on seedling emergence techniques which attempt to quantify the number of species and their frequency in the soil seed bank community. Intrinsic to this method, however, are problems associated with the inability to detect seed dormancy and to miss specific germinat ion requirements for particular species, which will result in underestimation of the total seed bank (Baskin and Baskin 1989; M. Leck, personal communication; van der Valk and Rosburg 1997). Seed extraction techniques may be used to increase accuracy, especially when attempting to quantify the distribution of a species on a broad spatial scale or when the species at question has large or easily identifiable seeds (Gross 1990). A downside to the extraction method is the ab sence of information on biological activity, however, when extraction is supplemented with tetrazolium assay (AOSA 2000), direct estimates on both seed bank number and viability can be achieved. This paper investigates the seed bank of an annual herbaceous species for its functional significance in the invasion ecology of natural area wetlands. Insi ght into understa nding the role the seed bank plays in regeneration ecology may be gained by investigatin g the basic characters

PAGE 27

27 that contribute to seed bank dynamics. Relative to the soil seed bank, such characters include seed longevity, density, availability (as indicated by burial depth in the substrate) and transition to the seedling stage. Thompson and Grime (1979) proposed two main seed bank strategies pertaining to seed longevity in surface soils: transi ent seed banks in which none of the seed survives beyond one year following dispersal, and persistent seed ba nks, which maintain a component that survives for more than a single year. Although revised through the years (Csontos and Tamas 2003) this simplest of classifications remains useful fo r surface soils under natural field conditions. Seed bank strategy is typically independent of the environment. However, associations can be made between seed bank strategy and ha bitat, particularly when the habitat is characterized according to disturbance patterns in the soil and vegetation. Compared to more stable environments (e.g., woodlands), which tend to support transient seed banks, more disturbed environments (e.g., farmlands) tend to select for persistence in the seed bank (Thompson et al. 1998 ). Wetlands typically rank midway on this generalized disturbance continuum of habitats. Spatial predictability in the pattern of disturbance may carry further influence in selecting for seed bank strategy. For example, transient seed banks can be expected to have adapted to exploit gaps resulting fr om seasonally predictable disturbance in the vegetation, while persistent seed banks likel y adapted to provide for regeneration where disturbance was less predictable in time and/ or space (Thompson and Grime 1979; Baskin et al. 1993; Abernethy and Willby 1999). Wetland hydrology can select for seed bank strate gy in its function as a natural disturbance factor. Such an influence was ascribed by L eck and Brock (2000) when comparing seed bank characteristics between wetlands of contrasting water regime. Under the predictable ebb and

PAGE 28

28 flow governing a coastal, freshwater tidal ma rsh (New Jersey, U.S.A) the seed bank was dominated by short lived, transien t species that regenerated in spri ng time. This scenario was in contrast with the unpredictable water fluctua tions recognized for a high elevation, temporary wetland (New South Wales, Australia) where long lived species formed a persistent seed bank. Because most wetlands generally fall between these two extremes offer environments for species with both persistent and transient seed banks ( Fenner and Thompson 2005). In south central Florida, the s easonality portrayed in natural area wetlands is characterized by annual wet/dry fluctuations in hydrology, wh ich although variable in space, are typically timed for draw down of surface water in spring and the return of surface flooding with summer rains. Such annual cycles are natural disturba nce factors that, although se vere, are essential in the role they play in the dynami cs of the biotic community (DeAngelis 1994; Smith and Kadlec 1983). Because hydrologic fluctuation occurs on a seasonally predictable ba sis, expectations for seed bank strategy might lend to preference for shor t lived, transient species rather than for those that accumulate a persist seed bank in Florida wetlands Although investigation directed at seed bank strategy has not been explor ed in the region, recent studies have indicated that native seed banks tend not to persist after the natural hydrol ogy has been disrupted (Lee 1994; van der Valk and Rosburg 1997; Wetzel et al. 2001; Smith et al. 2002). Seed size is commonly associated with burial depth, which in turn reflects the availability and longevity of the seed bank. Small seeded species tend to be more deeply buried, less available to contribute to the seedling stage, and persistent in strategy; meanwhile species with transient seed banks are typically larger seeded and tend to rema in higher up in the soil profile (Thompson and Grime 1979; Leck 1989; Moles et al. 2000; Leck and Brock 2000). In Florida, the seeds of S lacustris (2.95 to 3.95 mm long by 2.3 to 2.5 mm wide; Jacono, unpublished) are

PAGE 29

29 especially large, not only fo r an herbaceous annual, but fo r any wetland sedge (Leck and Sch tz 2005). The size (density) of the seed bank is an im portant factor for maintaining species in habitats that experience waterlevel fluctuation (Keddy and Rezn icek 1982). Abundant and often diverse seed banks have been demonstrated fo r wetlands of varying types (van der Valk and Davis 1976; Leck and Gravelin e 1979) although lower densitie s have been more commonly demonstrated for wetland sedges (Leck 1989). Of the temperate sedges reviewed by Leck and Schtz (2005), 40% were characteri zed as having transient seed banks. In turn, species with transient seed banks were found to occur in relati vely low densities, generally less than 500 seeds m-2. Studies relating persisten ce strategy with seed bank dens ity for the warmer climate members of the family are generally not availabl e, although in northern Florida, the seed bank of Rhyncospora inundata a native sedge, was clos e to 1000 viable seeds/m-2 at a deep water marsh once subjected to drought (G erritsen and Greening 1989). Commonly omitted from seed bank studies is in quiry into the seed bank to seedling stage transition, simply put, the proportion of the seed bank that contributes to a regeneration event. While research typically evaluates the relatio nship between seed bank species composition to that of the standing vegetation (Wet zel et al. 2001; Leck 2003) litt le data are available that may be used to infer relationships between the number of seeds of a particular species in a natural wetland seed bank and its acting cont ribution to regeneration. In a broader aspect, wetland seed banks have been shown to be equally repres ented by both annual and pe rennial species (Leck 2003) while the annual component of the wetland seed bank was [found to be] the key to yearly regeneration of the annual-dominated vegetation (Parker and Leck 1985).

PAGE 30

30 To understand the role the seed bank plays in the invasion ecology of S lacustris the hypotheses were presented that the s eed bank would: 1) be buried at a shallow depth; 2) occur in relatively low densities; 3) be transient in strate gy, and 4) transition prop ortionally to seedlings. To test these hypotheses, wetlands with previ ously known infestations were experimentally sampled twice yearly to evaluate the depth, density, depletion, pe rsistence and stage transition of the S lacustris seed bank. Materials and Methods Field Sites and Exp erimental Design Two different wetland habitats located in se gregate drainage regi ons of south central Florida were selected for comparative research, a seasonal depression marsh west of Vero Beach, Indian River County (VB), and a seasonal fl oodplain marsh on the north side of Lake Hatchineha, Disney Wilderness Preserve, in Polk County (KS). VB was defined by a shallow, seasonal pond bordered by cypress strands and located within a wet prairie. KS was a broad, low marsh with an agricultural and grazing history. Sandy soils at VB averaged 8% organic matter and comparably heavy soils at KS averaged 29% organic matter (gravimetric determination after combustion; MM2 User Manual Version 2, 2000, Delta-T Devices and The Macaulay Land Use Research Institute). Dense populations of S lacustris had been vouchered at VB since the previous year (2003) and at KS for five years (since 1999). In early March 2004, eleven 50 cm x 50 cm PV C grids were placed randomly within the zone of the previous population. The zone was i ndicated by decomposing st alk and leaf litter on the substrate where the grids were to serve as sampling stations. Surface water depth and soil moisture were measured monthly for the dura tion of the study period (Feb.-Oct.) and twice monthly during the period of seedling emergence (Appendix C).

PAGE 31

31 Seed Bank Sampling The soil seed bank was sampled twice, first in early spring and again in mid to late summer, to evaluate seed bank characteristics. From these data, persistence was estimated on the premise that densities that vary within a season offer evidence for persistence (Leck and Simpson 1987; M. Leck, pers. comm.). Sampling events were timed to quantify 1) seed depth and density in the soil profile before seedli ng emergence in early spring and 2) seed depletion after seedling emergence had ended for the year. Sampling events were thus termed preemergence and postemergence. Preemergence sampling was con ducted in March as the substrates became dewatered at both sites (Appendix C). Postemergence sampling was completed in August at VB. At KS, sampling was started in August, but due to delay from hurrican es, was not completed until September. A portable corer was used that had been engi neered from lightweight aluminum (designed and built by George Yeargin, Alach ua, Fla.) which was capable of sampling inundated substrates as well as terrestrial soils rooted with herb aceous vegetation. Under flooded conditions, the hydrostatic pressure pushed a plunger up into the corer tube allowing water to fill the tube above the core sample The plunger was seated before the core was extracted by means of a push rod and the core was drawn. The water was released by either decanting from the top of the tube or by unseating the plunger with the pus h rod to allowing flow out of the core cavity and removal of the core from the sampling device. Four replicate soil cores were collected at ea ch sampling station outside of each of the four corners of the sampling grids. Core samples were 15 cm in diameter and approximately 12 cm in depth. Each soil core sample was released into a cradle created from a 15 cm PVC pipe that had been halved lengthwise. The cradle was notched at 3, 6, and 9 cm to guide slicing of the soil

PAGE 32

32 profile into three horizontal layers. Each layer was placed into a one gallon plastic bag that was sealed and stored in the shade unt il transported to the laboratory. Postemergence sampling was initiated after se edling emergence had stopped, or in early July at VB and early August at KS. Sampling procedures followed those of preemergence, except that the soil cores were collected from outside of the midpoint of each of the four sides of the sampling grid, rather than from outside of each of the four corners. In this manner there was no overlap of destructively sampled substrates. Six stations had remained un-sampled at KS in August before hurricanes blocked access to the study site. An attempt was made in September to recover seed banks from the rema ining stations. By that time however, matured seeds of the season appeared to have augmented the seed bank, thus seriously compromising the late postemergence sampling. Therefore, analysis of postemergence seed banks for KS entailed only the five stations sampled in August. Once removed from the study sites soil samples we re transported, usuall y within 18 hrs, to the U.S. Geological Survey (USGS) facility, Gain esville, Florida, where floor drains and holding ponds offer facilities conducive to i nvasive species containment. S eeds were counted directly by using the extraction method as follows. For each sample, which consisted of a 3cm depth slice of soil core, large organic debris was remove d by hand and mineral soil was washed with running well water through a set of metal sieves. Mesh openings in the top sieve were larger than 4 mm while the lower sieve, havi ng 1.7 mm wide openings, retained seeds of S lacustris and several other species. Seeds of S lacustris were visually detected and extracted from the lower sieve by hand held forceps. Seed Bank Viability Assessment Seeds extracted from the soil samples were sorted as spent or intact. Spent seeds consisted of a white, hollow shell with a hole where the hypogynium was missi ng, indicating that the seed

PAGE 33

33 had germinated at some time in the past. Conve rsely, intact seeds were entire in morphology and proved to be either firm or soft when probed w ith forceps. The viability of intact seeds was assessed by embryo staining as follows. Seeds were held individually in a vise (custom made by George Yeargin, Alachua, Fla.) and bisected longitudinally with a ha nd held single edge blade. Each half was incubated with a 1% solution of 2,3,5-triphenyl tetrazolium chloride in 50mM Tris-HCl, pH 6.8, under darkness for 18 to 24 hours, at room temp (~27 C). Embryo staining was distinguished using both halves of each s eed and the aid of 25-40X magnification (Appendix A). Nijalingappa (1986) was c onsulted for identification of embryo morphology. Viability was evaluated according to accepted gu idelines with particular consideration made to the note that some minimal unstained areas in the scutellar region would not affect viability of the seed (AOSA 2000). Either all or a subset of the intact seeds we re assayed with tetrazolium. At VB, 86% of preemergence and 100% of postemergence seeds we re assayed. At KS 69% of preemergence and 74% of postemergence samples were assayed. The subsets equally represented the sampling core and depth layer from which they were extracted. The viability determined for each subset was factored with the remaining seeds for that set. Finally, the resulting values for viable seed were extrapolated from the original 15 cm diameter sampling core to m-2 before analysis and presentation. Seedling Monitoring Seedling regeneration was m onitored within the 0.25 m2 sampling stations throughout most of the year. Seedlings of S lacustris emerged at the study sites between March and July when they were identified at th e 1.5 to 2 leaf stage. New s eedlings were tagged with vinyl coated telephone wire, which afte r encircling the seedling, was fasten ed with the tail end, into the

PAGE 34

34 substrate. The total number of seedlings produced at each sampling stati on was extrapolated to the number of seedling m-2 before use in analysis. Statistical Analysis Paired t -tests were conducted on preemergence a nd postemergence samples to detect seed bank depletion. One-way analysis of variance (A NOVA) was used to compare research sites. Two-way ANOVA was used on arcsin-sq.rt transformed data to investigate seed bank persistence with depth. All data (except where indicated) were based on values for viable seed as determined by tetrazolium assay. Duncans multiple range tests were used to evaluate differences. All statistical analyses were perfor med with procedures in SAS (SAS Institute, Inc. 2002-2003). Results Seed Bank Viability Assessment Tetrazolium assay indicated that intact seed was a good predictor of viable seed (Figure 21) when overall viability of the seed bank was high Regression revealed a strong relationship between intact seed and the subset determined to be viable at preemergence sampling time (Table 2-1). When viability was high (at preem ergence), the slope was close or equal to one, indicating that most seeds that appeared intact were viable. Overall, seed bank viability was found to decrease at the postemergence sampling tim e at both sites. As viability decreased (postemergence), intact seed was less predictive of y. Seed Bank Depth The S lacustris seed bank was not deeply buried. Significantly more viable seed was extracted from the upper, 0-3 cm layer than from the deeper profiles. No significant differences were found in seed bank number between the 3-6 cm and 6-9 cm soil depths. This spatial pattern of shallowly buried seed was apparent at both sites, and was consistent during both sampling

PAGE 35

35 times (Table 2-2). The fraction of viable seed pa ralleled that found for in tact seed at all three soil depths, indicating that viable seed was spatia lly distributed in the same manner as that of intact seed (Table 2-3). At both sites, and at both sampling times, a relatively greater proportion of seed was concentrated in the upper 0-3 cm layer at Kissim mee than it was at Vero Beach (Table 2-4). Seed Bank Density Seed bank density is reported for each stat ion according to preor postemergence sampling (Table 2-5). The patterns in dispersion for seed ba nk soil core data illustrated that the lower seed bank densities at Vero Beach were coupled with lower variation among the four replicate sample cores per station, while the higher seed bank dens ities at Kissimmee were coupled with greater variation among the cores sampled pe r station (Figures 2-2 and 2-3). No correlation was found between seed bank density and surface water or soil moisture (Appendix D) on a per station basis thr oughout the study site (data not shown). Significantly greater seed bank densities were found at Kissimmee than at Vero Beach for both sampling times (Preemergence F =28.81; P<0.0001; Postemergence F =22.45, P<0.0001). The coefficient of variation (CV) was calculated to provide a measure of va riability for the four cores representing each station (Table 2-5). ANOVA on the CV for each station indicated that the degree of station to station heterogeneity was not different be tween the sites (Table 2-6). Thus, while overall seed bank heterogeneity wa s similar between sites, the real source of heterogeneity probably originated from variati on among the four replicate cores at each station, as illustrated in Figures 2-2 and 2-3. Seed Bank Depletion Seasonal depletion of the seed bank was detected using paired T -tests of preemergence and postemergence densities. T -test probability values indicated that a significant depletion in the

PAGE 36

36 seed occurred at both sites following seedling emergence while the effect of depletion was stronger, or less variable at Vero Beach than at Kissimmee (Table 2-5). Seed Bank Persistenc e Persistence of the seed bank is represented for each station in Table 27. When the stations were averaged at each site and seed bank persistence compared between Vero Beach and Kissimmee, no significant differenc es were distinguished (Table 2-7). Seed bank persistence after seedling emergence was greater in the surfac e layers and declined significantly with each depth layer (Table 2-8). No relationship was found between seed bank persistence and seed bank density, soil moisture, or surface water on a per station basi s throughout either study sites (data not shown). Seed Bank to Seedling Stage Transition There was no significant relationship between seed bank density and seedling numbers (VB r2 = 0.033, p=0.593; KS r2 = 0.354; n=11; Figure 2-4). Likewise, there was no effect of seed bank depletion (preemergence-postemergence/ preemergence) on seedling numbers (VB r2 = 0.303, p=0.079, n=11; KS r2 = 0.099; p=0.684, n=4). Thus, seedling regeneration was not representative of the num ber of viable seeds present in the soil either before or after seedling emergence. No relationship was found between seedling re generation and soil moisture at Kissimmee (r2 = 0.038, p=0.565, n=11), however at Vero Beach the number of seedlings significantly decreased as soil moisture increased on a per station basis thr oughout the study site (r2 = 0.448, p=0.024, n=11). The proportion of the seed bank that was deplet ed after seedling emergence appeared to be greater at Vero Beach compared to Kissimmee. Notwithstanding apparent deviation in the data (likely compounded by comparison of the two samp ling times) approximatel y 70% of the viable

PAGE 37

37 seed bank appeared to have been depleted at Vero Beach, compared to only about 50% of that at Kissimmee (calculated from Table 2-5). The differing depletion in the seed banks however, could not be explained by the transition to seedlings observed at each site. This became apparent when ANOVA was conducted on seed bank to seedling transition probab ilities. Results indica ted that the transition probability, the contribution of the seed bank to emerged seedlings m-2 was not significantly different between sites (Table 2-9). In summary, though pree mergence seed bank density and total seedlings produced were significantly greater at Kissimmee than at Vero Beach, the seed to seedling transition probabilities were equal at the two sites. Discussion Seed Bank Viability Tetrazolium assay indicated that the S lacustris seed bank occurred at an especially high viability level (see Gross 1990) and was sensitive in detecting narrow declines of viability in the postemergence seed bank. Overall, it supported the ascertainment of seed bank persistence by demonstrating that the seed bank rema ining after germination was viable. Seed Bank Depth This study supports earlier findings that large seeds occur at shallow depths. The shallow seed bank found with S lacustris may be attributed, however, to fact ors other than just seed size, such as disturbance regime and substrate comp osition that are characteristic to the seasonal wetland habitat. While hydrologic dry down ranks as a catastr ophic disturbance fact or in the aquatic environment, it has little if any role in disturbing the soil s eed bank. When inundated wetlands dry down, the substrate remains rela tively intact and seeds, whethe r present in the water or on the sediment surface, are not subject to mechanical forces of burial. Typically, some degree of

PAGE 38

38 burial is necessary for initial roots to take hold and seedlings to establish. Thus deposition in water has accounted for the shallo w burial of large seed s that would not be readily incorporated without active force (Leck and Schtz 2005). The surface substrate at both sites was a coarse ly textured organic layer that typically comprised the entire upper 0-3 cm and was com posed of detritus, the decaying algae, plants, insects and vertebrates that had settled from the water column. This layer trapped the large seeds of S lacustris which after deposition with dry down were incapable of penetrating the underlying mineral soil (Appendix D). The tendency for de eper burial at Vero Beach could indicate a greater permeability of it s light, sandy under layers compared to the heavy, lower substrates at Kissimmee. Seed Bank Density Sample heterogeneity is intrinsic with seed banks and is mainly due to the irregular distribution and clustering of seeds in the soil (Wetzel et al. 2001). Indeed, especially high degrees of variability in seed bank density are ch aracteristically reported for sedge species (Leck and Leck 2005). In this study, th e variation in seed di stribution demonstrated across the infested area was comparable at both sites, while the true source of variation appeared to stem from the heterogeneity existing within each individual sampli ng station. A larger number of smaller sized core samples, as detailed by Bigwood and I nouye (1988) might have resulted in smaller deviations and greater precision in estimates at the individual stations. On the other hand, the greater edge effect produced by a greater number of smaller cores could have been detrimental by contributing to greater mixing of su rface seed within the deeper profile. The seed bank densities determined for S. lacustris at either site surpass those reported as usual for individual sedge species, <500 seeds m-2 (Leck and Schtz 2005). Large seed banks and shallow burial depth may in itself be expected to carry ecological significance in the rapid

PAGE 39

39 colonization of wetlands with natural disturba nce regimes, such as found by Leck and Leck (2005) for numerous species in a newly constr ucted tidal freshwater marsh, by Nicholson and Keddy (1983) for a lake shoreline, and by Aber nathy and Willby (1999) for a diversity of annuals in the fluctuating we tlands of a riverine floodplain. In this study the density of S lacustris in sandy soils was as high as 2,136 1286 seeds m-2 and in heavy soils as high as 3,323 1,378 seeds m-2. Compared to similar habitats in Florida, these values exceeded those fo r individual native species of Rhyncospora Eriocaulon and Xyris each nearing 1000 seeds m-2 in naturally fluctuating mars hes of the Okefenokee Swamp (Gerritsen and Greening 1989) and rivaled those of Juncus effuses at 4000 seeds m-2 in former marshes of altered hydrology of the Kissimmee Ri ver floodplain (Wetzel et al. 2001). When compared to the entire community, multiple specie s seed bank in existing wetlands of central and southern Florida, the densities for S lacustris remain high. For example, in existing wetlands on the Kissimmee River community seed banks range from 179 to 857 seeds m-2 (Wetzel et al. 2001) and in the Everglades, disturbed, Typha dominated wetlands hosted near 3500 seeds m-2 and the more natural, Cladium dominated wetlands neared total multi-species seed banks of 1300 seeds m-2 (van der Valk and Rosburg 1997). Seed Bank Depletion and Persistence A significant seasonal reduction in seed bank density was demonstrated after spring germination both within stations at each site and for all stations averaged between the two sites. Such findings follow Leck and Simpson (1987) where measurable seed bank reductions were reported with annual species at a freshwater tidal marsh. While depletion in this study occurred at all soil depths, the greatest degree of depletion was recorded at Vero Beach from the deepest, 6-9 cm layer. This was unexpected in that as shallow seed banks readily contribute to regeneration, seed bank depleti on should be expected to decrea se with depth. Recognizing the

PAGE 40

40 need for caution in drawing conclusions from small numbers of seed, it is noted that the large seeds and robust seedlings of S lacustris appeared to find deep bur ial of little hindrance to regeneration in the lighter subs trate. Future studies could be used to clarify the limits to regeneration by controlli ng burial depth in differing substrates of seasonally dewatered habitats. Despite the level of seasonal depletion recorded in the seed bank, the presence of viable seed remaining at postemergence sampling (July/A ug.) ascertains a persiste nt seed bank strategy for S lacustris Persistence precludes th e requirement for annual renewal of the seed bank, providing insurance for a succeeding population w ithout augmentation from seed rain. The higher seed bank density at Kissimmee over Vero Beach could have been influenced by two factors, the build up over a longer period of time, at least four additional years, and the greater probability of persistence in heavier soil types. Si milarly, the greater reduction in viability of the postemergence seed bank at Kissimmee could indi cate an older seed bank, such as might be expected of building up over multiple years of infe station. In such a case, a more diverse age structure with proportionally older seeds could be expected to account for the greater loss in viability at postemergence sampling. Seed Bank to Seedling Stage Transition While depletion was greater at Vero Beach ( 70%) than at Kissimmee (50%), the seed to seedling transition probability was the same (25% and 24%, respectively). Therefore, seedling regeneration might only account for less than half of the depleted seed bank. Of the unaccounted for losses, predation and disease may have had so me influence, although, th e pervasive factor of spatial heterogeneity could be most likely. No relationship was found between the num ber of viable seed comprising the preemergence seed bank and the final number of s eedlings produced at the tw o sites. Lack of a relationship between the seed bank and vegeta tion is not unusual, as has been reported

PAGE 41

41 previously in wetlands (Leck and Simpson 1987; Leck 2003). In this study, eith er the affect of seed bank heterogeneity extends beyond such pred ictions, or microsite va riables, directly or indirectly, may be dominating the pattern of seedling regeneration at the wetland habitats. Persistent seed banks have been shown to be especially important to annual species to assure their recurrence in vari able and unpredictable environmen ts, ranging from deserts (Pake and Venable 1996) to vernal pools ( Zedler 1990). Recent development of the persistent seed bank of an invasive annual plan t at seasonal wetlands may indicat e the current unpredictability of these habitats in Florida. The ecological conseq uences of a large and persistent seed bank as found in this study for S lacustris can be significant in the seasonal wetland by providing a continuous source for colonization and thus a se rious impediment to ecosystem integrity and restoration management.

PAGE 42

42 Table 2-1. Parameters for a linear model predicti ng viability of intact seed as determined with tetrazolium assay. The regression was forced to go through the origin (viable seed = intercept + slope intact seed). Site Sampling time # Cores ( n ) Slope r2 Vero Beach Preemergence 44 0.893 0.994 Postemergence 44 0.733 0.925 Kissimmee Preemergence 44 0.903 0.997 Postemergence 20 0.728 0.975

PAGE 43

43Table 2-2. Viable number of Scleria lacustris seed at three soil depth ranges with pr eemergence and postemergence sampling at seasonal marsh sites. Values are estimate s of the mean number of viable seed m-2 ( 1 std. dev.) per station according to extrapolations of data from 15 cm diameter cores. Vero Beach Kissimmee n 0-3 cm 3-6 cm 6-9 cm F n 0-3 cm 3-6 cm 6-9 cm F Preemergence 44 794 539A 147 256B 17 54B 63.79*** 442038 1440A 256 456B20 39B 70.36*** Postemergence 44 235 200A 46 86B 1 8B 42.63*** 201043 1063A 49 133B0 0B 18.09*** F values are the result of ANOVA to compare seed number with depth. *P<0.05; **P<0.01;***P <0.001. Within a site and row means with the same letter are not significantly diffe rent (alpha = 0.05) by Duncans multiple range test. Table 2-3. Intact number of Scleria lacustris seed at three soil depth ranges with preemergence and postemergence sampling at seasonal marsh sites. Values are estimate s of the mean number of intact seed m-2 ( 1 std. dev.) per station according to extrapolations of data from 15 cm diameter cores. Vero Beach Kissimmee n 0-3 cm 3-6 cm 6-9 cm F n 0-3 cm 3-6 cm 6-9 cm F Preemergence 44 895 587A 173 287B 18 60B 67.00*** 442272 1573A 281 481B21 42B 73.89*** Postemergence 44 342 222A 60 103B 1 8B 72.96*** 201626 1702A 72 189B 0 0B 17.24*** F values are the result of ANOVA to compare seed number with depth. *P<0.05; **P<0.01;***P <0.001. Within a site and row means with the same letter are not significantly diffe rent (alpha = 0.05) by Duncans multiple range test.

PAGE 44

44 Table 2-4. Percentage of total viable seed with soil depth at seasonal marsh sites. Vero Beach Kissimmee n 0-3 cm 3-6 cm6-9 cm n 0-3 cm 3-6 cm 6-9 cm Preemergence 44 83 15 2 44 88 11 1 Postemergence 44 84 15 1 20 96 4 0

PAGE 45

45 Table 2-5. Density of the Scleria lacustris seed bank at seasonal marsh sites. Estimates for the mean number of viable seed m-2 ( 1 std. dev.) are extrapolations from 15 cm diameter soil cores (0-9 cm depth) at preemergence and postemergence sampling times. Probability results for paired T tests indicate depletion (difference of the mean preemergence minus the mean postemergence per station) based on 11 stations at Vero Beach and 5 stations at Kissimmee. Coefficient of varia tion (CV) = standard deviation/mean per station. Station Preemergence Postemergence Pre-Post T -value CV Pre CV Post Vero Beach 1 1047 439 622 360 .42 .58 2 764 547 183 116 .72 .63 3 1103 589 311 286 .53 .92 4 1344 187 580 370 .14 .64 5 651 499 240 169 .76 .70 6 1230 427 268 203 .35 .76 7 2136 1286 325 162 .60 .50 8 566 541 311 56 .96 .18 9 665 213 71 107 .32 1.51 10 665 56 46 .53 .82 11 368 134 141 149 .37 1.06 Mean 958 494 283 183 675 438 5.11** Kissimmee 1 2646 2496 664 765 .94 1.15 2 1340 1967 4 7 1.46 2.00 3 2018 867 1157 700 .43 .60 4 3323 1378 1184 410 .41 .35 5 1551 936 2454 1330 .60 .54 Mean 2175 814 1092 899 1083 12221.98 6 2761 726 .26 7 2118 143 .68 8 1819 1467 .81 9 2235 1662 .74 10 2914 1928 .66 11 2857 1923 .67 Mean 2175 814 *P<0.05; **P<0.01;***P<0.001

PAGE 46

46 Table 2-6. The coefficient of va riation (CV) for seed bank sampling time for each research site. Values give the mean CV ( 1 std. dev.). Site CV Preemergence n F CV Postemergence n F Vero Beach 0.517 0.233 11 2.30 0.754 0.338 11 0.5 Kissimmee 0.698 0.319 11 0.929 0.669 5 Table 2-7. Proportion of persistence in the Scleria lacustris seed bank (0-9 cm depth) at seasonal marsh sites. The fraction persistence was calculated as the mean postemergence divided by the mean preemergence number of seeds for each station. For a one-way ANOVA comparison between sites, F = 1.43, Pr>F 0.236. Results are based on 11 stations at Vero Beach and 5 stations at Kissimmee. Vero Beach Kissimmee Station Persistence n Persistence n 1 0.61 43 0.33 0.20 2 0.46 45 0.00 0.00 3 0.26 13 0.66 0.39 4 0.45 13 0.43 0.9 5 0.44 21 0.90 0.19 6 0.20 7 7 0.16 5 8 0.72 34 9 0.11 15 10 0.10 10 11 0.36 43 Mean 0.35 31 11 0.46 38 5

PAGE 47

47 Table 2-8. Proportion of persistence in the Scleria lacustris seed bank with soil depth. The fraction persistence was calculated from the mean postemergence divided by the mean preemergence number of seeds for the two sites totaled and then arcsinsq.rt transformed. F value is for two-wa y ANOVA comparisons among soil depths. Depth (cm) Persistence n F Pr> F 0-3 0.604A 64 61.08 P<0.0001 3-6 0.185B 64 6-9 0.000C 64 Means within a column with different letters are significan tly different (alpha=0.05) by a Duncans multiple range test.

PAGE 48

48Table 2-9. Data for mean viable seed ex tracted at preemergence sampling time and the total number of seed lings for each statio n, both extrapolated to m2. The transition probability for the preemergence seed bank turnover to seedlings reflects the percentage of the seed bank to produce se edlings as calculated from th e total number of seedlings m-2 divided by the preemergence seed bank m-2. The final row provides the ANOVA summary for the m ean ( 1 std. dev.) preemergence seed bank density at the two sites n=44; F 108.72***; for the m ean number of seedlings at the two sites: n=44; F 41.24***; and for the mean transition probability at the two sites; n=44; F 0.17. Vero Beach Kissimmee Station Preemergence seed bank m-2 Seedlings m-2 Seed to seedling transition probability (%) Preemergence seed bank m-2 Seedlings m-2 Seed to seedling transition probability (%) 1 1047 40 4 2646 444 17 2 764 56 7 1344 444 33 3 1103 612 55 2023 552 27 4 1344 452 34 3339 1020 31 5 651 120 18 1556 224 14 6 1231 364 30 2759 1036 38 7 2136 108 5 2660 476 18 8 566 32 6 1203 248 21 9 665 460 69 2235 832 37 10 679 204 30 2914 472 16 11 368 92 25 2858 432 15 (Range) median (32) 120 (4) 25 (224-1036) 472(14) 21 Mean 959 4762 230 196B 25 20 2288 6981 556 273A 24 8 F values are comparisons between sites: *P<0.05; **P<0.01;***P< 0.001. Means within a row with th e same number, with the same number, letter, or with the same symbol (), are not significantly different (alpha 0.05) by a Duncan s multiple range test.

PAGE 49

49 A B Figure 2-1. Tetrazolium (1%) staini ng results for viable seed of Scleria lacustris A) The longitudinal section of an intact seed illustrates the embryo, basally located and stained pinkish red, and the endosperm, remaining unstained and white (36X magnification). B) An excised embryo is characterized by the apical cotyledon which is in the shape of a circular disk t opped with a cone (60X magnification).

PAGE 50

50 A B 0 10 20 30 40 50 60 70 80 90 100 110 01234567891011 Station# Intact Seeds / Core 0 10 20 30 40 50 60 70 80 90 100 110 01234567891011 Station# Intact Seeds / Core Figure 2-2. Preemergence dispersion of int act seeds in replicate cores (0-9 cm dept h) at A) Vero Beac h and B) Kissimmee.

PAGE 51

51 A B 0 15 30 45 60 75 90 105 120 01234567891011 Station# Intact Seeds / Core 0 15 30 45 60 75 90 105 120 01234567891011 Station# Intact Seeds / Core Figure 2-3. Postemergence dispersion of int act seeds in replicate cores (0-9 cm depth) at A) Vero Beach and B) Kissimmee.

PAGE 52

52 A B 0 100 200 300 400 500 600 70005001000150020002500 Seed Bank Densit y Seedlings 0 200 400 600 800 1000 1200 01000200030004000 Seed Bank DensitySeedlings Figure 2-4. Scatter plot of S lacustris seed bank density m-2 (preemergence) with the total number of seedlings m-2 that emerged at each station at the A) Vero Beach and B) Kissimmee study sites.

PAGE 53

53 CHAPTER 3 SEEDLING REGENERATION IN Scleria lacustris, AN INVASIVE ANNUAL WETLAND SEDGE Introduction The mechanisms of seeding regeneration are i nherent traits, particular to a species, yet strongly influenced by the envir onmental variables defining a hab itat. Although seed banks play a critical role in wetland re-v egetation, acting as a safeguard fo r annual plants (Baldwin et al. 2001) and providing the insurance for community change, hydrologic regime ranks as the principle environmental variable selecting for life history trai ts and defining the presence or absence and abundance of species (van der Valk 1981; Seabloom et al. 2001). Indeed, the composition of wetland vegetation has been shown to correlate more strongly with the water regime than with the seed bank composition (Nicol et al. 2003). Thus, similar rationale may be attributed to studies that have found the presence and abundance of a species in the vegetation to greatly outweigh its importance in the wetland seed bank (Parker and Leck 1985). Based on the simplest rendition for wetland hydrologic regimes, van der Valk (1981) proposed two major groups of speci es: those that recruit when ther e is no standi ng water (Type I) and those that recruit under floode d conditions (Type II). Type I species are primarily emergent plants and include many short lived annuals. Type II species tend to be plants with submersed or floating habits, although they also include some emergent species. Thus simplified, this successional model has been upheld by field inves tigation and supported by experimental trials that controlled surface and so il water levels. For example, shallow, continuous flooding prevented the establishment of all but one out of 20 wetland species representing a diverse assemblage of functional types from eastern On tario (Weiher et al. 1996 ). Soil seed banks samples from wetlands in Florida responded in a similar fashion, although the seed banks were natural and of unknown species composition. On the other hand, Cook (2004) expected

PAGE 54

54 germination and recruitment for Scleria lacustris to occur under flooded conditions in its native range of southern Africa. The dry down of surface water from flooded we tlands and lake shorelines exposes the substrate and may often result in a rapid pulse of seedling regeneration from the existing seed bank (van der Valk and Davis 1978; Smith and Kadlec 1985; Leck 1996; Leck 1989). The timing and magnitude of such dry down events are critical in seedling regeneration, as has been shown at an arid playa wetland (Haukos and Sm ith 2001) and at a temperate piedmont lake (DeBerry and Perry 2005). In the Okefenokee Swamp of northern Florida, the springtime dry down of a deep water marsh, both unexpected and extreme in magnitude, re sulted in a massive cohort of seedlings of Rhyncospora inundata (Cyperaceae) from a persistent seed bank (Gerritsen and Greening 1989). Sedge seedlings can be among the most common occurring in wetlands (Leck 1989). Most wetland sedge seedlings favor light over dark and require large fluctuations in alternating temperatures for germination (Leck and Schtz 2005), requirements which are typically provided for in substrates that dry down during spri ng or summer months (Baskin et al. 2002). The wetlands of south central Florida are typica lly seasonal, as characterized by an annual pattern of water level fluctuation. The wint er dry season leads to surface water draw down during spring, and substrates are reflooded with su mmer rains. In such seasonal habitats the introduced sedge, Scleria lacustris has been spreading since 1988. Comparative investigation at two different wetland sites demonstrated that an nual populations recur from a persistent seed bank (Chapter 2). The native wetland sedge, Rhyncospora inundata although less extensive in occurrence, can often be found in association with S. lacustris infestations (Jacono, voucher collections). It is predicted that S lacustris like R inundata at Okefenokee Swamp, survives

PAGE 55

55 flooded conditions through the persistent seed ba nk and emerges to exploit the opportunities for establishment created by dry down at the seasonal marsh. The objective of this study wa s to determine the mechanism of seedling regeneration for the invasive annual, S lacustris by testing the hypothe sis that hydrologic dry down of the local environment functions to control regeneration. Greenhouse experiments were used to elucidate the microenvironments potentially available during a typical spring season at a seasonal marsh in Florida. Materials and Methods Year 1 Seed source: In Oct. 2004 fruiting stalks were clipped from sta nding plants of S lacustris in a seasonal depression marsh at the Florida Fish and Wildlife Conservation Commission field station, Vero Beach, Indian River County, Florida, USA (27.146 N and 80.516 W). Seeds that fell, on shaking, cleanly away from th e infructescence were immediately contained in aluminum screen packages and stored to a substr ate depth of approximately 3 cm at the center of the ponded area of the marsh. Surface water was 28 cm at the time of storage. Flooding is believed to have been maintained over the wi nter months, even though surface water had been measured as low as 2 cm (Figure 4-10). Wh en seeds were retrieve d in March 2005, surface flooding had risen to 23 cm. Seed s appearing full and intact were randomly counted into 4 replicates of 25 and assessed fo r viability with tetrazolium (see tetrazolium assay detailed in Materials and Methods, Chapter 2). Experimental design: Four greenhouse treatments were designated according to likely conditions during spring time at a seasonal ma rsh: flooded to 2.5 cm, flooded to 5.5 cm, saturated and drained. The th ree hydrologic treatments were established by placing 10 cm diameter plastic pots into three containers of differing, selected heights and filling the containers

PAGE 56

56 with water. The height of the container main tained the desired hydrologic conditions above the media surface for the flood treatments prescribe d, and 3 cm below the media surface, for the saturated treatment. The final, drained treatmen t consisted of a pot that was not placed in a water-holding container and as su ch drained freely from the bottom. All treatments were seated on the wire surface of a greenhouse bench and ma intained with overhead mister irrigation. Before pots were designated into treatments each was lined with 2 layers of newspaper and a single layer of washed pebbles before filling with Fafard Superfine Germination Mix. Five consecutive overhead waterings were appl ied, until drain-through, in order to leach the small amount of fertilizer incorporated with the mix. On compaction with watering, additional mix was added to achieve equal surface levels, 2 cm below the top edge of the pot. Whole and intact seeds were randomly counted into 16 repli cates of 25 seeds and were planted, 25 to a pot, at a depth of 1 cm. The surface was smoothed to level. Finally, 20 ml of play sand (purchased sterilized, strained and screened) was sprinkled evenly over the surface. The weight of the sand kept the underlying media in place, and though de vised for flooded conditions, was applied to all treatments. Thus prepared the pots were, dependi ng on the treatment, placed in containers or left free standing and established in the greenhouse on 2 Mar. 2005. The misters were automatically timed to run 2 times a day for 30 minutes under a one minute on, two minute off regime. The free standing pots, comprising the drained treatm ent, were rotated among each other two times per week. The experiment was maintained for 96 days, until 6 June, after which un-germinated seeds were extracted and assayed for viability with tetrazolium. Year 2 Seed source: In Oct. 2005 seed stalks were clipped from plants at the Blue Cypress Water Management Area, Cell C, St. Johns Water Management District (27 40 40.1 N and 80 35 13.1 W). The material was transported to the Ga inesville laboratory and within 24 hr the stalks

PAGE 57

57 were gently shaken. Seeds that fell on hand shaking were examined with 40X magnification. Those appearing full and intact we re selected and sewn into pack ages of black nylon screen. The seeds were then set aside for 20 days at the po ly-roofed outdoor laboratory at USGS, Gainesville. Contrary to procedures followed in Year 1, the seeds were not returned to the depression marsh for over wintering, but were stored under simu lated marsh conditions at the outdoor lab. The seed packages were buried to a 3 cm depth within 2 large plastic pots (27 cm diameter x 24 cm depth) that had been filled with a substrate of play sand augmented with 8% peat. The pots were set in a plastic storage bin that was itself set within a concrete tank at the outdoor laboratory. The bin was finally flooded with 10 cm of pond water. Flooding was maintained over winter, from late October until seed retrieval in early March. Sub samples were assayed for viability with tetrazolium, as in Year 1 with the excep tion that re-selection fo r intact seeds was not necessary, as the seed state had not changed over the storage period. Experimental design: The experimental design differed in several manners from that of the previous year: 1) the deeper flooding (5.5 cm) treatment was replaced with a treatment of extreme drought, 2) the individual pots compri sing the treatments for saturation and flooding were placed each in their own sm all container, and 3) each trea tment consisted of 25 replicate pots containing 8 seeds. The flooded, saturated and drained treatments were arranged in a random block design on the greenhouse bench. The drought treatment was pl aced adjacent to, but not interspersed with the other treatments. The irrigation stand over th is treatment was capped and a Plexiglas screen erected to block overspray from other stand pi pes in the greenhouse. The pots representing the drought treatment were watered until drain-thro ugh once per week, and were rotated on the

PAGE 58

58 bench two times per week. The experiment was initiated 13 March and maintained for 96 days (17 June). The un-germinated seeds were not extracted, thus precluding viability assessment. Statistical Analyses Analysis of variance (ANOVA) was used to co mpare the four treatments independently for each experimental year. Tukeys Studentized Range (HSD) tests were used to determine differences between the treatm ent means for seedling regenera tion, seedling establishment and seed viability. All statistical analyses were pe rformed with procedures in SAS (SAS Institute, Inc. 2002-2003). Results Seedling emergence and establishment were the cr iteria used to interpret the influence of the experimental hydrologic environment on recru itment. In the Year 1 greenhouse experiment seedling recruitment was significantly higher in the drained substrate than in either the saturated or flooded treatments (Table 3-1). By the end of the study, all of the 246 seedlings that had emerged in the drained treatment remained green and healthy and remained standing (established). Meanwhile the few seedlings th at had developed under the flooded treatments (4 seedlings under 2.5 cm inundation; 2 seedlings u nder 5.5 cm inundation) were dead. Finally, in the saturated treatment, 8 of the 10 seedlings that had emerged were still standing, and thus established, although severely chlorotic w ith necrotic leaf tips (Figure 3-1b). Year 2 greenhouse results repe ated the Year 1 trend in that seedling emergence and establishment were significantly higher in the drained than in the saturated and flooded treatments. Seeds did not germin ate, however, in pots treated w ith saturation or flooding (Table 3-1; Figure 3-2). Germination and seedling emergence occurred in the drought treatment as well as the drained, although in signifi cantly lower frequencies (Table 3-1). By the end of the study, all of the seedlings remained in both the drought and the moist well drained environments. It

PAGE 59

59 should be noted, however, that seedlings produced in the drained treatments of both years were uniformly vigorous in color, turgor, and upright habit (Figure 3-3a), whil e the seedlings of the drought treatment were dull and somewhat with ered (Figure 3-3b), as typified by drought induced stress. The rate of seeding emergence that occurred in the drained treatment varied between the two experimental years. In Y ear 1 the time for 50% of the mean cumulative number of seeds to emerge was only 11 days, which compared to 19 days during Year 2 (F igure 3-4). Maximum germination, however, was reached at about the sa me time (39 days) during each year. In Year 2, seedling emergence was comparable in rate fo r the drought and drained treatments. The time for 50% of the mean cumulative number of s eeds to emerge was 19 and 20 days, respectively (Figure 3-4), even though the final maximum ge rmination was significantly greater in the drained over the droughty pots. Tetrazolium assay indicated that the viability of the Year 1 and Year 2 seed sources differed considerably at the time of experiment al set up (Year 1 71% 15%; Year 2 58% 15 %; Table 3-1). Nevertheless, the variability accompanying the mean percentage values should not be discounted as it indicates a broad overlap in the data. There was no measured effect on actual viab ility, however, seedling emergence revealed with certainty that seed viability was as high following the Gainesville tank storage as it was following the Vero Beach field storage (Table 3-1). The number of seed that had not germinated by the end of the Year 1 experiment differed significantly between the draine d and the three wet treatments (Table 3-1). All of the ungerminated seed remaining in the drained trea tment tested as dead. In the saturated and

PAGE 60

60 flooded treatments, however, viabil ity of the ungerminated seed in creased significantly with the level of flooding. Discussion The results of this study were conclusive in finding that the hydrologic environment functions to control s eedling regeneration in S lacustris That drained and droughty substrates strongly favored regeneration over satura ted and flooded environments ranks S. lacustris as a Type I species that, like the drought adapted R inundata at Okefenokee Swamp (Gerritsen and Greening 1989) recruits only when there is no standing water. Fu rthermore, the intolerance of this species for germination under presumably anoxic conditions is reflected. The minimum water level to significantly limit seedling emerge nce was saturation, not surface flooding. Such results lead to interpretation that the magnitude of an actual dry down event may be critical in the regeneration ecology, and thus, invasion of this speci es in seasonal wetlands. It is apparent that complete removal of surface wate r and adequate draining of the substrate must occur before seedling emergence and establishment may occur. The large seeds of S. lacustris are located primarily in the surface (0-3 cm) layer where detritus and organic material are deposited with the sedimentation that occurs with dry down. Aside from trapping the seeds, the organic layer on the sediment surface (Chapter 2), once exposed with dry down, offered a moist and warm seedbed favorable for germination, root anchorage and seedling establishment. Even if a small proportion of seedlings were capable of em erging in waterlogged environments, their establishment would likely fail. Over wintering storage of seed, however, unde r the more moderate and relatively warmer conditions of the flooded field, over those of an above ground tank at a higher latitude, may account for a germination rate that was 12 days earlier for the Year 1 over the Year 2 drained

PAGE 61

61 treatment, indicating the field stored seeds were at a potentially more active state than tank stored seeds in early March. Results from this study indicate that th e influence of the hy drologic environment on seedling recruitment was stronger th an a potential influence of the s eed source. This was further demonstrated when seedling emergence was lower in the drought than in th e drained treatment of Year 2, both which originated from the same seed source. The reduced germination could indicate the prolongation of dor mancy, or the induction of a secondary dormancy, a syndrome that has been previously reported under droughty germination conditions. While regeneration was suppressed under saturate d and inundated substrates, results from this study indicate that seed longevity of the unge rminated seed was prom oted with increasingly inundation levels. That seeds remaining ungerm inated under the deepest level of surface flooding maintained significantly higher viabil ity indicates a possibl e trend between the presumably anoxic environments of saturated or flooded substrates and the longevity of ungerminated seed. This finding provides insight th at could have importance in the persistence of this, and potentially other annua l, wetland species and lends addi tional support to the prediction that the typical hydrologic e nvironment of the Florida seas onal marsh functions as the mechanism for regeneration of S lacustris

PAGE 62

62Table 3-1. Year 1 and Year 2 results of greenhouse experi ments on the recruitment of Scleria lacustris seed in different hydrologic treatments. Values are presented in mean percent ( 1 st d. dev). Ungerminated seed re maining intact = # remaining divided by # planted*100). Viability ungerminated seed (# live divided by # remaining*100). Year 1 d.f. =60, Year 2 d.f .=96. Ungerminated seed was not extracted in Year 2 and therefore not assayed for viability. Treatment Seed viability at setup Seedlings emerged Seedlings established Ungerminated seed Viability of ungerminated seed YEAR 1 F F F F 71% 15 Drained 61 12A 319*** 61 12A 355*** 59 15B 17*** 0C 157*** Saturated 3 4B 2 3B 84 9A 59 9B Flooded 2.5 cm 1 1B 0B 80 7A 58 14B Flooded 5.5 cm 1 1B 0B 81 10A 70 11A YEAR 2 58% 15 Drought 48 18B 178*** 48 18B 178*** Drained 76 21A 76 21A Saturated 0C 0C Flooded 2.5 cm 0C 0C F values are the result of ANOVA to co mpare treatments during each year: *P<0.0 5; **P<0.01; ***P<0.001. For each year, means within a column with the same letter are not signif icantly different (alpha 0.05) by a Tukeys range test.

PAGE 63

63A B Figure 3-1. Year 1 greenhouse experiment used to test the h ydrologic conditions affecting rege neration of the annual sedge, Scleria lacustris A) The two flooded, the saturated and the drained treatments were placed under overhead misters on the greenhouse bench. B) Seedlings estab lished under the saturated treatment.

PAGE 64

64 Figure 3-2. Year 2 greenhouse experiment used to test the h ydrologic conditions affecting rege neration of the annual sedge, Scleria lacustris The flooded, saturated and drained treatments we re randomized under overhead misters on the greenhouse bench. The drought treatment (foreground) was placed adjacent to the others, yet under a capped mister head and behind a Plexiglas screen (removed for photogr aph) to shield from overspray.

PAGE 65

65A B Figure 3-3. Year 2 greenhouse experiment on the mechanism of re generation, 39 days after initiation. A) The second row from the bottom demonstrates, from left to right, fl ooded, drained, and saturated treatments. B) An indi vidual pot from the drought treatment.

PAGE 66

66 0 10 20 30 40 50 60 70 80 90 100 071418212528323639424656646875Time in DaysMean Percent Germination Drained Yr 1 Drained Yr 2 Drought Yr 2 Figure 3-4. Rate response of the treatments that demonstrated significant regeneration responses in Year 1 and Year 2 greenhouse experiments. Pots in the draine d treatment received twice daily overhead irrigation while pots in the drought treatment received a single irrigation once per week.

PAGE 67

67 CHAPTER 4 INFLUENCE OF A SEASONAL HYDROL OGIC REGIME ON THE SEED BANK AND REGENERATION ECOLOGY OF Scleria lacustris AN INVASIVE WETLAND SEDGE Introduction The ecological consequences of persistent seed banks can be significant in wetland communities. Seed banks interact with the ever fluctuating environmenta l variable of hydrologic regime to contribute to species composition and vegetation struct ure in wetlands (van der Valk 1981; Keddy and Reznicek 1986; Baldwin et al. 2001). In seasona l wetlands of south central Florida the annual hydrologic regime is charact erized by a summer rainy season marked by the rapid development of surface flooding followed by a winter dry season during which surface water levels slowly fall, usually to reach dr y down in spring (DeAngelis 1994). Despite the general acceptance of dry down as a natural disturbance import ant to the maintenance of herbaceous communities, the functi onal role of seed banks in response to hydrologic fluctuation is only beginning to be underst ood in Florida marshes (van de r Valk and Rosburg 1997; Lee 1994; Wetzel et al. 2001; Smith et al. 2002). The seed banks of habitats that are subject to water level fluctuation are typically characterized by dormancy. Most seeds of aqua tic and wetland plants, including wetland sedges have a strong initial physiological dorman cy, a poorly understood condition of blocked germination caused by a mechanism in the embryo that inhibits radicl e emergence (Leck and Schtz 2005; Baskin and Baskin 1998). The seeds of wetland species may be released from dormancy during flooding. For example, important annual species (Cyperaceae, Lythraceae) of seasonal mudflats in Tennessee and Kentucky, whic h are dormant at matur ity in autumn, were released from dormancy during winter flooding and proceeded with regeneration on summer draw down (Baskin et al. 2000; Baskin et al 2002). Although flooding may not always be required for seeds to come out of dormancy, th e effects of even rela tively short periods of

PAGE 68

68 inundation ( 2 mo) can result in significantly higher percentages of germination than experienced with non flooded seeds (Baskin et al. 2002). After having been released from dormancy while they were flooded, many wetland plants germinate whenever the hydrologic condition is suitable, irrespective of the time during the growing season. Other wetland seeds are govern ed by seasonal dormancy and will germinate only during a particular season, pot entially reentering dormancy (sec ondary dormancy) if they do not germinate at the appropr iate time (Baskin et al. 1993; Baskin et al. 1996). The effects of the hydrologic environment on in situ seed bank survival have not been well studied within naturally fluctu ating wetland habitats. Bekker et al. (1998) made considerable efforts to simulate the hydrologic conditions in soil mesocosms that would typify wet grasslands in the Netherlands. Finding greater survival in se veral species of wet gras sland taxa subjected to constant water levels it was inferred that the seeds of wetland habitats have adaptations that allow for metabolic activity under saturated, low oxygen conditions. Results from more routine laboratory and greenhouse experiments support the prediction that while the seeds of nearly all wetland spec ies tolerate some leve l of inundation, most are highly intolerant of desiccati on (Baskin and Baskin 1998; Leck and Brock 2000). In temperate, humid climates the majority of persistent seeds occur naturally in an imbibed state while in the seed bank (Leck 1989). Even where wetland seed banks were only intermittently wetted, or imbibed, they survived longer than when they were dry (Priestly 1986). Dry laboratory storage of temperate Carex (Cyperaceae) species resulted in reduc ed viability, by as much as 95% (Leck and Schtz 2005). Contrastingly, in the arid climate of eastern Aust ralia, tolerance to desiccation is a prerequisite for many wetland species in or der to survive years of dry periods (Leck and Brock 2000).

PAGE 69

69 In natural community habitats, gaps exposed by small scale disturbances serve as sites of regeneration and recruitment (Fenner and Thomps on 2005). In seasonally de watered habitats of North America, including riparian mudflats of the humid eastern states, playas and vernal pools of the arid west, and, presumably, marshes in south central Florida, hydrologic dry down may likewise act as a disturbance mechanism as it e xposes small to potentially expansive areas of sparsely vegetated substrate that favor seed ba nk regeneration. Zedler (1990) identified the annual habit as an important life history strategy in vernal pool plants, to help explain their mechanisms for coping with the extreme hydrologic variation characteristic of the seasonal pool habitat He suggested that temporal variation in rainfall leads to recurrent gaps which favor the recruitment of annua l (~80%) over perennial species. Ad ditionally, he predicted the annual habit to be more capable of surviving extende d drought through minimal biomass, namely as seeds in the soil seed bank. Similarly, Hennessy (1985) sugges ted that the annual habit in Scleria lacustris (Cyperaceae) could serve well as a dr ought escape mechanism in its indigenous tropical range of south ce ntral Africa which experi ences seasonal drought. Scleria lacustris was introduced to and has been spr eading in south central Florida since 1988. It appears to demonstrate a high degree of habitat specificity by oc curring in wetlands that have experienced significant fluctuation in hydrologic regime and seasonal dry down. Results presented in Chapter 2 provided evidence for a persistent seed bank in S lacustris that is extremely viable, high in density, and shallowl y buried. Results presented in Chapter 3 demonstrated that flooded seed banks, having been inundated over wintering in the marsh, germinated to a high degree in moist and dry soil during spring, but did not germinate under saturated or inundated conditions. While increasingly populations in Florida seem to be linked to extended drought and resulting spring dry dow n, exceedingly reduced populations have been

PAGE 70

70 observed to coincide with consecutive years of wetland desiccation (Ed Harri s, pers. comm.). In this manner, the incidence of S. lacustris could be restricted at the two critical life stages the seed bank and the seedling. Evidence from Chapte r 3 leads to the predicti on to be explored in this chapter that surface flooding af ter mature seeds have shed in autumn, is important for both dormancy break and survival while an opposing spri ng dry down is essential for regeneration of that seed bank. The hypothesis to be tested in this chapte r is that the characteristic hydrological regime for seasonal marshes in Florida serves as the selective factor dr iving the incidence of S lacustris primarily through the functiona l role of its seed bank. It is predicted 1) that mature seed is dormant, 2) that inunda tion is strongly implicated in s eed dormancy break, and that 3) seed bank dynamics, seedling regeneration, seedling survival and adult productivity will vary, while being both promoted and restricted, along the hydrologic gradient of the marsh. For testing these hypotheses, field investigatio ns aimed at elucidating seed bank survival and dynamics were conducted over several annual cycles at a s easonal depression marsh in south central Florida. To interpret field results, in situ and ex situ seed storage treatments were applied for later evaluation with greenhouse regeneration. Materials and Methods Field Site and Experimental Design Field studies were conducted w ithin a 61 ha tract of seasonal wetlands at the Florida Fish and Wildlife Conservation Commi ssion field station, Vero Beach, Indian River County, Florida, USA (27.146 N and 80 23.516 W). The habita t was defined by wet prairie interspersed with depression marshes and cypress strands. Co nservation marshes that had been repeatedly infested with S. lacustris were located nearby. Nevertheless, ditching had isolated the study tract from adjacent properties; hence its water re gime was completely dependent on immediate rainfall and the groundwater tabl e. Annual hydrologic fluctuati ons have been historically

PAGE 71

71 predictable as portrayed by dry down in early spring, surface flooding with summer rains with the maintenance of flooding, more or less thr ough the winter months (T. Towles, personal communication). Between 2003 and 2005, S lacustris had been found scattered throughout prairie type wetlands and locally abundant in adjacent, but deeper, depression marshes. The experiment was started in Feb. 2005, when three transects were established along the elevation gradient at the larges t of the depression marshes at the tract. Two transects were oriented west to east, and the third was oriented north to south. All three transects began at the highest point of the bank rimming the marsh and stopped at the center depth, thus providing linear lengths ranging betw een 42 m and 48 m. Data were colle cted at sampling stations (0.5 m x 0.5 m2 PVC grids) placed every 6 m along each transect. This design allowed for the placement of 8, 7, and 8 stations, respectively, along the Transects I, II and III during 2005 (2005 TI, TII, TIII). The second year of field research was in itiated in Feb. 2006, and followed the same experimental design except that three new transects were placed just 10m north of the previous two west to east transects, and 10 m west of the previous north to south transect. The irregular topography of the depression marsh resulted in transect lengths of 48, 36 and 42 m lengths for TI, TII and TIII, and thus allowed for the pl acement of 8, 6, and 7 stations, respectively, along 2006 TI, TII, TIII. Field Sampling and Monitoring Hydrology Surface water level and soil moisture values were measured monthly from Feb. through Nov. and twice monthly, as possible, during the period of seedling regeneration. Surface water was measured above the substrate with a mete r stick. Soil moisture (volumetric soil water content) was measured with a portable, handheld Thetaprobe sensor (D elta-T Devices and The

PAGE 72

72 Macaulay Land Use Research Institute). The sens or outputs DC voltage in response to changes in the apparent dielectric cons tant, thereby reflecting volumetric water content within the soil pores. Voltage output data was converted to wa ter content (soil moisture) using soil specific calibration coefficients that had been determined for the soil type of the study site (the equation for the polynomial relationship was used following the MM2 User Manual Version 2, 2000). Calibration parameters were ao = 1.47 and a1 = 6.76 which were more characteristic of generalized coefficients typical for organic rather than mineral soils. Seed bank dynamics The seed bank was sampled twice per year to ev aluate persistence and depletion in the seed bank over time (Leck and Simpson 1987). Sampling time was set to quantify 1) seed density before seedling emergence in early spring and 2) seed bank depletion after seedling emergence had ended. Sampling events were term ed preemergence and postemergence. In 2005 preemergence sampling was initiate d in March however rainfall and the prolongation of surface flooding precluded comple tion until July. Postemergence sampling was delayed until September and as deep surface wate r continued to prove difficult and inefficient, the number of stations target ed was reduced from 23 to 11. In 2006, prevailingly dry weather contributed to complete dry down of the depression marsh by 2 Apr. Preemergence seed bank sampling was started in late Mar. and completed in early Apr. and post germination sampling was accomplishment in June. The soil seed bank was sampled by drawing soil cores and extracting seed as detailed in Chapter 2. Preemergence cores were collected in f our replicates at each station. Each replicate was drawn at the outside 4 corners of the 0.25 m2 sampling grid, before the emergence of S. lacustris seedlings. Results from the Chapter 2 s eed bank characterization demonstrated that most seeds were located in the upper 6 cm, regardle ss of sampling time. On that basis, all cores

PAGE 73

73 were collected as a single layer, 0-6 cm belo w the surface (Figure 4-1). Postemergence sampling commenced when seedling emergence stopped. Sampling procedures followed those of preemergence, with the exception that the core s were drawn adjacent to the outside of the midpoint of each of the four sides of the sampling grid, rather than from outside of each of the four corners. In this manner there was no overlap of destructively sampled substrate. Soil seed bank samples were transported to the USGS laboratory, Ga inesville, Florida, usually within 18 hrs. Subsequent wet sievi ng and seed extraction was conducted within one week following the procedure described in Chapter 2. All seeds, regardless of the species and condition, were garnered from the sieves. Magnifi cation (25X) was used to distinguish seeds of S. lacustris from those of S reticularis Michx., (netted nutrush) a na tive, annual congener of prairie type wetlands (Wunderlin and Hansen 2003). Partial seeds we re identified for both species which consisted of hollow shells (the pericarp) punctuated by an opening at the basal end, where the hypogynium had fallen away. As su ch, these shells were identified as the remainders of seeds that had germinated in the past and were thus designated as spent. Intact seeds of S lacustris were immersed in distilled water at room temperature (1-3 days) before testing viability with tetrazolium (TZ) as detailed in Chapter 2, Materials and Methods. All seed bank values were extrapolated to m-2 from the original data befo re analysis and presentation. Plant demography The emergence and survival of S lacustris seedlings were monitored biweekly during the peak springtime germination period and monthly through Oct. to determine regeneration patterns at the study site. New seedlings were identified at the 1.5 2 leaf stage and then tagged by encircling with vinyl coated telephone wire. Cohor ts were defined as each set of newly emerged seedlings, between the 1.5-2 leaf stage, found at each field visit. All seedling values were extrapolated to 1 m2 from the 0.25 m2 sampling station prior to analysis and presentation.

PAGE 74

74 Percent cover was visually estimated within e ach sampling station in April of both years. Frequency was based on a m odified rating scale of Bra un-Blanquet (Mueller-Dombois and Ellenberg 1974) as follows: 1 = 0-4.9%; 2 = 5-9.9%; 3=10-24.9%; 4=25-49.9%; 5=50-74.9%; 6 = 75-100%. Statistical analyses Temporal differences between pre and poste mergence seed bank numbers were analyzed with paired T -tests on the means of soil core data from each station or from each transect. Relationships between the hydrologic variable s for surface water and soil moisture and the biological variables for seed ba nk and plant data were analyzed using Spearman correlation. Surface water monitoring data was calculated to relate the mean summer surface water (WavgSu; mid Jun. through Oct. 2005 and 2006) a nd the annual surface wa ter high (WHi; mid Jun. 2005 and Jul. 2006). In the same manner, soil volumetric water content was used to calculate the mean soil moisture during the peri od of seedling emergence (MavgSp; 2 Apr. 21 Jun. 2006), the mean soil moisture for the mont h of highest seedling emergence (MavgM; 2 Apr. 29 Apr. 2006). As such, these four hydrol ogic variables were exp ected to describe or influence biological responses. Spearman correla tion coefficients were calculated to quantify relationships between the four hydrologic variable s and the seed bank or plant demographic data on a per station basis for each sampling time. Resu lting coefficients of determination were used to identify the strongest independent pr edictors among the hydrologic variables. To detail the source of the effect of the hydr ologic predictor as a pplied to the shallow hydrologic gradient of the marsh, data were reorde red for analysis of variance (ANOVA). Each sampling station was categorized according to its collective surf ace standing water for the year, which necessarily excluded the typical spring dry down period. Such a categorization resulted in the transformation of 23 stations into 4 water regime categorie s as follows: A stations that

PAGE 75

75 always lacked surface inundation, C stations that were cyclically both inundated and dry, IS stations that were conti nuously inundated and shallow in wate r depth, and ID stations that were continuously inundated and deeper in wa ter depth. ANOVA was then applied to evaluate the response of seed bank and plant variables according to th e water regime category. All statistical analyses were performed with pro cedures in SAS (SAS Institute, Inc. 2002-2003) using an alpha level of 0.05, unless otherwise indicated. Experimental Tests of Seed Dormancy Fruiting stems of S lacustris were clipped Oct. 2005 from st anding plants in Cell C of the Blue Cypress Water Management Area, St. Johns Water Management District. This section of restoration marsh was located approximately 5 m iles from the Vero Beach study site and hosted S lacustris growing with the na tive community types Panicum hemitomum and Eleocharis cellulosa Stems were transported to the USGS la boratory and mature seed was selected the following day. Seed that had fallen cleanly away during transport or after gentle shaking (Figure 4-2) were further discerned with 40X magnification. Seeds that a ppeared intact and filled, which bore a hypogynium that was smooth and free of s ubtending bracts, and which felt hard when probed with forceps, were identified as mature Mature seed was stor ed under cover outdoors for ten days before testing viability, and for one month before setting up growth room and greenhouse experiments used to in vestigate initial seed dormancy. Viability: The 2005 seed was first tetrazolium assaye d to estimate viability of fresh seed before use in growth chamber and greenhouse germ ination trials. Eight sets of 25 seeds were tested. Growth chamber germination experiment: Incubation was conducted in the light and in the dark (wrapped in foil) with in a growth chamber programme d for alternating 30C day/20C

PAGE 76

76 night environments. Each treatment consisted of 8 plates each with 25 fresh seeds. Seeds were plated on 40 g of washed, sterilized sand. A volume of distilled water (~10ml) was added to bring the sand to near 100% hydration without pr oducing a film of standing water. The set of plates that were not foil wrapped were monitored at 5 day intervals for germination. The germination criterion was designate d as radical emergence. The experiment ended after 3 weeks, germination was assessed in each treatment a nd tetrazolium assay was conducted on 10 seeds from each plate (40%). Greenhouse germination experiment: Fresh seed was sown into seedling flats and kept in an unheated greenhouse in Gainesville. Fl ats had been filled w ith Fafard Superfine Germination Mix and watered until run-through five times to leach the small amount of fertilizer in the mix. Two flats, 48 cells each, were planted with 2 seeds/cell at a depth of 1 cm. Overhead water was applied sparingly and only after th e surface had dried. Seedling emergence was monitored, and flats were rotated in bench space, twice a week. High and low temperatures were monitored with a Fisher min/max thermomete r daily and weekly until the experiment was ended after 21 weeks. Cells containing ungerminated seeds were sieved to evaluate seed fate. Field Experimental Tests of Seed Bank Survival Year 2004 seed source Field storage: Mature fruiting stalks we re clipped from standing pl ants at the Vero Beach study site in Oct. 2004 and gently sh aken to collect seed that fell cl eanly from the infructescence. Seed was contained in aluminum screen packages and buried to an approximate depth of 3 cm at four locations along the gradient of the marsh. Tent stak es were driven through two corners of each package for fastening to the substrate. Th e storage locations were identified as center, midway, and edge of the ponded area of the depression marsh. The fourth package was buried uphill adjacent to the cabbage palm ( Sabal palmetto ) tree line that marked the highest point of a

PAGE 77

77 discontinuous bank rimming the marsh. Because th e floors of airboats were observed to carry seeds of S lacustris when leaving an infested wetland, an additional storage location was maintained under ambient conditions at the USGS garage to evaluate viability after extended airboat storage. Viability and greenhouse germination: After 5 months (March 2005) and 17 months (March 2006) seed was retrieved to test its viab ility and ability to germinate. Necessary seed was retrieved from the storag e locations while in the field and the remaining seed was immediately reburied in the same location. March 2005: Before germination trials, a subset of the retrieved seed was tested to determine viability using tetrazolium assay with 4 replicates of 25 seeds each. Germination tests of the retrie ved seed were conducted in a heated greenhouse employing 6 replicates of 25 seeds from each stor age location. Seeds were planted in 4-inch plastic pots at a depth of 1 cm, 25 seeds to a pot. Pots contained Fafard Superfine Germination Mix that had been leached of fertilizer. March 2006: A subset of the retrieved seed was tested with tetrazolium to determine viability before germination in 2 sets of 25 seeds from each storage location. Greenhouse germination tria ls were conducted in a heated greenhouse employing seedling flats (50 cells /flat) that containing media de scribed previously. Each cell received 2 seeds to a depth of 1cm providing 2 fl ats with 100 seeds each to be tested for each treatment (Figure 4-3A). Flats contained Fafa rd Superfine Germination Mix that had been leached of fertilizer and seeds were planted at a depth of 1 cm. In both years greenhouse trials were maintained for 49 days during which seedli ngs were tallied upon emer gence (Figure 4-3B). Overhead misters were automatically timed to irrigate for 30 minutes twice daily and were adjusted over the experimental period to provide for a consistently moist, well drained media.

PAGE 78

78 To reduce potential experimental error from dissimilar microenvironments along the bench, pots and flats were rotated two times weekly. Year 2005 seed source In 2005 seed was collected seve ral miles north of the Vero Beach study site at the Blue Cypress Water Management Area, Cell C, St. J ohns Water Management District due to the absence of S lacustris at the depression marsh. This was th e same mature seed source that in Oct. 2005 was used in growth room and greenhouse trials to investigate innate dormancy. Fresh seed was counted into lots of 200, machine sewn with nylon thread into packages of black nylon screen, and protected outdoors fo r 20 days before placing at the study site. Six duplicate packages of seeds were contained within a layer of aluminum screen and placed at five locations along Transect I of the study site. The locations considered field stor age treatments, were placed along the transect at 1) Treeline Surface, 10 m above the high water mark of the 2004 wet season, adjacent to cabbage palm and level wi th the soil surface (vegetation having been clipped); 2) Treeline Buried, adjacent to cabbage palm and buried 3 cm below the soil surface; 3) TIStation 0, the edge of th e ponded area of the marsh; 4) TI-S tation 1: the shallow zone of the ponded area; and 5) TI-Station 7, the deepest zone of the ponded area. Six packages were also maintained in ambient conditions at the US GS garage. After 5 months (March 2006) seed was retrieved for viability a nd greenhouse emergence trials. Viability and greenhouse germination: After 5 months (March 2006) seed was retrieved to test its viability and ability to germinate. Before germination trials, 25 seeds from each storage location were assayed in two sets usi ng tetrazolium to determine viability. Greenhouse germination trials were then conducted in a he ated greenhouse employing seedling flats (50 cells /flat) that containing media desc ribed previously. Each cell recei ved 2 seeds to a depth of 1cm

PAGE 79

79 providing 2 flats with 100 seeds each to be tested for each treatment. Pots or flats were rotated and overhead misters were operated as previously described. Flats were maintained for 49 days during which seedling germination was tallied upon emergence and attainment of a height of 4 cm. At the end of the experiment, media from cel ls containing ungerminated seed were sieved to recover the intact seed remaining. Ungerminated seed that could be recovered was tetrazolium assayed to designate seed fate as either dead or dormant following the germination event. Statistical analyses Rate and completeness of germination were plotted according to mean percentage over time. Seed viability and germination res ponses following storage were compared among treatments with ANOVA. Seed fate following gr eenhouse germination trials was analyzed with Chi-square. Outdoor Laboratory Tests of Seed Bank Survival Mature seed was collected from plants at the Vero Beach study site in Nov. 2003 in the manner described previously for experimental te sts of seed dormancy. The seed was counted into sets of 300 and packaged into short segments of nylon stocking before subjecting to constant or seasonal flooding. Constant flooding: Seven packages of seed were placed in a heavy plastic bin and maintained under flooding with 30 cm of pond water outdoors at the Gaines ville laboratory for 3 and 4 years. In March 2005 and again, in Ma rch 2006, 50 seeds were retrieved for tetrazolium viability assay and 200 seeds were retrieved for gr eenhouse germination trials in seedling flats as described earlier for field stored seed. Seed fate following greenhouse germination was analyzed with Chi-square. Seasonal flooding: Seven nylon stocking packages were buried to a depth of 3 cm in two 20 cm plastic pots containing play sand mixed to 8% with Canadian peat. The pots were placed

PAGE 80

80 side by side in a heavy plastic bin and stored outdoors at the Gainesville la boratory. Pots were flooded to 15 cm with pond water until March of the following year when overhead watering was stopped and pots were allowed to draw dow n naturally. During the dry down period, pots were provided full drainage and watered daily to several times weekly as required to maintain moist, yet well drained substrate. Seedlings we re counted and removed on emergence. After seedling emergence had ended for the s eason, Aug. 2004, 2005, and 2006, the nylon packages were exhumed. The fate of remaining seeds were determined and counted as germinated, intact, or decayed. Intact seeds were repackaged in new nylon stockings and reburied. Pots were flooded as before and maintained as such until March of the following year. The seasonally flooded/drawn down conditions empl oyed were intended to model t hose typically experienced at depression marshes in south central Florida. Seed fate following annual dry down was analyzed with Chi-square. Results Field Sampling and Monitoring Hydrology The surface hydrologic conditions experienced at the depression marsh contrasted sharply between the two study years. In 2005, early spring rainfall events contributed to a continuation of surface flooding. Although a degree of dry down occurred along a narrow zone of the highest elevation (in early March) the vast majority of the study area rema ined flooded into the following year (Figure 4-4). In 2006, however, dr y climactic conditions prevailed and the marsh experienced a rapid drying event in early spring th at resulted in the removal of 0.5 to 25 cm of surface water within little over one month. The marsh remained dry until mid June, when surface flooding returned with summer rainfall. Short periods of dry down were experienced

PAGE 81

81 intermittently in Aug. and Oct. as well, although so il water content remained at or near saturation (data not shown). The plant communities appearing at the marsh paralleled the hydrologic conditions in that they contrasted sharply between the two study years (Figure 4-5). In 2005, the marsh was dominated by a taxonomically diverse and stru cturally open community of native, mainly perennial species (Appendix B). In 2006, however, the marsh was dominated by the introduced, annual species S lacustris It grew overtopping in height and later lodged horizontally in a manner that occluded a visual assessment of the native plants as was conducted in 2005. Repeated measurements of soil volumetric c ontent were made during the spring dry period of 2006 (Figure 4-6). Values of 1.0 indicated standing water; however, the volumetric response was not linear and values of 0.5 to 0.6 and above constituted essentially saturated conditions. Field monitoring consistently demo nstrated that seedling regenera tion did not occur in flooded or saturated substrates. Nevertheless, soil volumetric water content demonstrated only small numerical differences between the stations and overall, the absen ce of a trend along the naturally shallow hydrologic gradient of the depression marsh. Seed bank dynamics Tetrazolium assay of seeds extracted from the so il cores demonstrated, as determined from combining samples across the site, that overall, intact seed was a good pr edictor of the viable seed bank. A strong correlation wa s revealed between intact seed from the 0-6 cm profile and the resulting subset assayed by tetr azolium as viable (Table 4-1). Seed bank densities from the preemergen ce and postemergence soil sampling were reported as the mean viable seed m-2 ( 1 std. dev.) for 2005 (Table 4-2) and 2006 (Table 4-3). Zero data values at the uppermost stations of the gradients (TI stati on 0 and TII station 1)

PAGE 82

82 demonstrated the absence of a seed bank for S lacustris at higher elevations of the depression marsh. Paired t -tests were conducted to detect differences between preemergence and postemergence seed bank density. Results indicate d that the number of seeds in the soil was not significantly different between th e two sampling periods in 2005 (TI: n=4, mean diff. -38 45, t =-1.72, p=0.183; TII: n=4, mean diff. 215 272; t =1.58, p=0.211; TIII: n=4, mean diff. -211 132; t =0; p=1.0). In 2006, significant differe nces were found between preemergence and postemergence densities at two of the three transects (TI: n=8, mean diff. 79 125, t =3.9, p=0.0003; TII: n=6, mean diff. 135 319; t =2.3, p=0.027). Heterogeneity among the sample cores may have weakened the statistical power at the third transect (TII I: n=7, mean diff. 74 258, t =1.7, p=0.097) where overall seed bank density al so tended to be higher (Table 4-3). Heterogeneity arose from the four cores compri sing each station. Heterogeneity in seed bank samples is linked to the irregular distribution, or patchiness of s eeds in the soil (Wetzel et al. 2001), and is evidenced in this st udy throughout seed bank means data. A considerable difference in the early spri ng seed bank density was apparent between study years when seed bank density was found to be high in 2005 and much reduced in 2006. This effect was directly related to a reduction in the density of intact seed at preemergence sampling, as was illustrated by simultaneously plotting variables for both total intact seed and the subset testing as viable for each year (Figure 4-7). Plant demography The year 2005 was marked by the absence of S lacustris in the above ground vegetation at the depression marsh, despite the abundance of viable seed that was demonstrated with seed bank sampling. The functional importance of the seed bank was revealed in 2006 when S

PAGE 83

83 lacustris seedling regeneration was notably abundant, de spite the fact that the viable seed bank had been considerably reduced from that measured in 2005. The number of seedlings corresponded closely to the values determined for intact and viable seed (Figure 4-8), validating the accuracy of the field sampling and TZ assessment methods. Figure 4-8 is also important for demonstrating the contribution of the 2006 seed bank to the S lacustris standing population. In 2006, seedling regeneration began early Apr il, concurrent with dry down at the depression marsh. Five cohorts of seedlings we re tagged between 2 April and 21 June (Figure 49). The earliest cohort was restricted to the upper two or three stations of each transect while the remaining stations were still inundated. Cohorts emerging late r in April were important in populating the mid to lower regions of the tran sects and, as illustrated, were overall more important in contributing to biomass of the adult population. The termination of seedling emergence coincided with the return of surface wa ter flooding in late June (Figure 4-4). No seedlings emerged in the Aug. and Oct. dry down events. Relationships between hydrologic and seed bank/plant variables The hydrologic variables for surface water (WavgSu and WHi) produced 12 significant correlations with the seed bank and plant variab les while variables for soil moisture (MavgSp and MavgM) produced 10 and 9 sign ificant correlations, re spectively (Table 44). The strongest variables, WavgSu and MavgSp, demons trated little, if any overlap, thus were more effective as predictors when used together. Relationships were more comprehensive across the board when the two variables, which never coexisted in the fi eld situation, were both utilized. For example, soil moisture often demonstrated a relationship when one was absent with surface water, and vice versa. No correlation was found between surface water (WavgSu) and soil moisture (MavgSp) (r2 = 0.02, p = 0.53), substantiating the need to treat the two as i ndependent predictive variables.

PAGE 84

84 Fewer correlations occurred w ith hydrologic predictors for 2005 than for 2006. The direct effects of surface water and soil moisture on seed bank and plant variables are represented for 2006 in Figure 4-10. Interest ingly, all the variables for S lacustris were positively correlated with the hydrologic predic tors while the variables for native components (seedlings of the native congener S reticularis and cover of associated plants) were negatively correlated with the hydrologic predictors. Despite the significant correla tions often found between hydrol ogic and plant variables, resulting coefficients using seed bank vari ables often indicated the absence of a linear relationship and plotted results commonly demonstrated a clus tering of the data points. It is possible that the shallow hydrologi c gradient at the marsh might not have been strong enough to produce a measurable effect when the available data were treated with conventional correlation statistics. Subsequently, the categorization of each station according to the annual surface water regime lent for ANOVA to evaluate the response of seed bank and plant variables according to the water regime category. The ANOVA results for seed bank density data are presented in Table 4-5. A viable S lacustris seed bank was present at all four water regi me categories, although density and viability were somewhat reduced in the always dry (A) regime. In 2005, although significant differences were not indicated for the density of intact or of viable seed across the categories, seed bank density was markedly lower where the substrate was always dry (A), than where the substrate was cyclically (C) or shallowly inundated (IS). Somewhat lower seed bank densities were also found where the seed bank was deeply inundated (ID ). In 2006, this pattern was repeated at preemergence sampling time, when seed bank densiti es were higher at the mid-zone of the marsh (C and IS water regimes). These differences, however, were not apparent at postemergence

PAGE 85

85 sampling, several months later. The density of viable seed at preemergence sampling in 2006 was, as during 2005, significantly lower at the u pper, always dry (A) ca tegory, but again, these statistical differences were not apparent at po stemergence sampling. Sample heterogeneity, as indicated by the high standard deviations, coul d have influenced the absence of statistical significance with these results. Ironically, the strongest statisti cal difference across the water regimes was with spent seed when significantly mo re spent seed was found at the IS category at both sampling times in 2005. This variable likely carried the least biolog ical significance and its accumulation at a particular zone was probably th e result of the transient deposition of light weight, hollow particles under i nundated field conditions rather than as cast offs from germination that year. Indeed, during 2006, the y ear of germination, zonation of spent seed was not evident. In reviewing the seed bank data as presented in Table 4-5 the overall reduction in seed bank density between the year 2005 and 2006, as observed earlier w ith the independent transect data, is apparent. Categorization by wate r regime indicated that intact seed was more reduced at the A and C categories (64% and 66%) than at the IS and ID categories (34% and 28%) in 2006. ANOVA results for seed bank viability and pe rsistence are presented as mean percent (Table 4-6). The 2005 preemergence viability was equally ve ry high at the A, IS and ID water regime categories and significantly lower at ca tegory C. At postemergence sampling, however, viability was equal across the categories. In fact viability did not differ statistically across the water regimes, even though the means data were reduced at the always dry (A) and cyclic (C) regimes. In 2006, preemergence viability was signif icantly lower at the A th an at the C, IS and ID regime and at 3.8% was much reduced from 95% viability recorded one year earlier at the always dry water regime. 2006 postemergence sa mpling somewhat contradicted these values,

PAGE 86

86 nevertheless a trend of reduced s eed bank viability over time at the always dry (A) regime was statistically evident. Persiste nce in the seed bank (calculate d as the mean postemergence seed bank divided by the mean preemergence seed bank) did not differ signifi cantly across the water regime categories for either study year. The effect of sampling heterogeneity, however, was likely inherent to this calculati on with means which were actually associated with wide standard deviations. Nevertheless, it should be noted that mean seed bank persistence of intact seed in 2006 was markedly less at most of the water regi me categories from values determined in 2005. ANOVA results for plant variables are presen ted in Table 4-7. In 2005 regeneration of S. lacustris was not detected at the depression marsh; therefore adults a nd seed production was absent as well (Table 4-7). Seedlings of the native species S reticulari s, which were initially mistaken for S lacustris provided an unexpected seedli ng variable. Seedlings of S reticulari s were significantly more abundant in the always dr y (A) than in the cyclic (C) and inundated (IS and ID) water regimes (alpha = 0.05; p< 0.0001), in fact, seedlings of the native Scleria were completely absent in the IS and ID categor ies where deep flooding remained constant. Simultaneously, total cover of associated native vegetation was significantly higher in the A and C regimes than it was down the gradient at the IS and ID zones (a lpha = 0.05; p<0.0001). In 2006, S lacustris reappeared from seedlings at the depression marsh (Table 4-7). For the entire germination period, si gnificantly fewer seedlings emer ged at the A than at the C regime, however there was no difference among seed ling values in the A and IS or ID or among the C and IS or ID water regimes. This distri bution is reminiscent of that found for seed bank density. Present as well were high standard deviations, presumably related to seed bank patchiness. Cohorts, denoted as the mean number of seedling cohorts produced at each water regime, were without question, si gnificantly reduced at the dry (A ) than at the wetter water

PAGE 87

87 regimes. Yet the C regime produced significantly more seedling cohorts th an did the ID regime, at the lower end of the trans ects. The earliest 2 April c ohort produced significantly more seedlings at the C than at the other three wate r regime categories (alpha = 0.1; p<0.01). While the A regime had remained without flooding, th e IS and ID categories were still undergoing drawdown at that time. The 17 April cohort, impor tant for its substantial contribution to total population biomass, did not dem onstrate significant differences in seedling numbers among the water regimes, even though the mean value was markedly reduced at the A category. Of all 13 plant response variable s tested with water regime in ANOVA, that of seedling survival produced the highest level of statis tical significance when comparing water regimes across the gradient of the marsh. Mean seedli ng survival at the prev iously inundated water regimes (IS 88% and ID 86%), was significantly higher than at the C category and at the A category, which were likewise different from each other (alpha = 0.5; p<0.001). Regardless of the poorer survival of seedlings at the A and C categories, adult density remained undifferentiated across the water regimes, which like the number of seedlings in the 17 April cohort, remained tied to seed bank heterogeneity. Adult biomass (dry weight) and fecundity further supported the si gnificant trend to promote S lacustris populations towards the lower portion of the hydrologic gradient, as was demonstrated earlier with seedling survival. Tota l adult biomass was significantly greater in the previously inundated water regimes (IS and ID) th an in the always dry (A) regime, while the cyclic (C) regime ranged intermediate among th e categories (alpha = 0.05; p<0.05). The mean values for developed seed were significantly higher in the IS and ID categories than in the A category, where no seed was produced, while the C category ranged intermediate. Failed seed followed the pattern of developed seed across the categories indicating that water regime did not

PAGE 88

88 select for the development of failed seed over that of developed seed. While S lacustris adult biomass and seed production were greater in the IS and ID regimes, associated native cover was significantly reduced in these previously inunda ted parts of the marsh. For example, in 2006, seedlings of S reticulari s were significantly more abundant in the cyclic (C) water regime than in the IS and ID categories. Additionally, the associated native plant cover was significantly higher in the always dry regime (75-100% cover) than in the lowe r regimes that had experienced some level of surface flooding, whether intermittently or continuously. Experimental Tests of Seed Dormancy The viability of fresh seed, as determined w ith tetrazolium, was above 60% (Table 4-8). Growth chamber incubation, whethe r in the light or dark, did not result in germination, even though there was no loss in viability after this tr eatment. Germination only occurred with fresh seed after over wintering for 12 wk in the unhe ated greenhouse (Table 4-8). During the 2 wks preceding germination (late Jan. to early Feb.) daily mean high temperatures in the greenhouse had risen to 30C while nightly temperatures remained low at a mean 11C. Only two weeks after the first seedling had emer ged, 50% germination was reached (within 14 wk of planting). Germination quickly tapered off by 15 wk, altho ugh it continued at a very slow rate until the experiment was ended. At the final 21 wk, me an germination (60.4%) was near the initial viability estimated for fresh seed (63.5%). Seedling survival was 100%. The total fate of greenhouse cultured fresh s eed is accounted for in Table 4-9. Dormant, viable seeds comprised <1% of the starting mate rial. A considerable portion, 8.2%, had aborted or failed in germination, while the majority of the starting seed that ha d not germinated into seedlings (30%) was dead by the end of the 147d greenhouse study.

PAGE 89

89 Field Experimental Tests of Seed Bank Survival Year 2004 seed source Hydrologic conditions: Surface water at the seed storag e locations are illustrated from burial in Oct. 2004 to dry down in Apr. 2006 (Figure 4-11). Although monitoring data was lacking, Nov. 2004 through Jan 2005 reportedly rema ined inundated at the deep and midway storage locations; the marsh edge experienced gradual water withdraw al, and the treeline remained dry (personal communication Tim Towles FGFFC). The edge, mid and deep locations experienced little hydrologic cha nge through 2005 due to the unusually wet climactic year and in 2006 the 2004 midway location was eliminated becau se water depth mirrored that of the deep marsh. During the more typical climactic year of 2006, greater change was experienced from tree line to deep mars h along the gradient. Viability and germination: Tetrazolium assay after a five month over-wintering period indicated that there was equally high viability in 2004 seed that had been stored at the garage and at the marsh proper and significan tly lower viability in treeline buried seed (Table 4-10). Greenhouse germination trials revealed however, that treeline buried seed germinated equally well as seed stored at the marsh proper while the garage stored s eed was significantly less capable of germinating. These results imply that all of the field conditions served to provide for dormancy release and germination of 2004 s eed after only five months storage. The time for 50% of the mean cumulative number of seeds to germinate was 10-11 days for seed stored at the deep, midw ay and treeline locations and 14 days at seed stored at the edge of the ponded marsh, meanwhile the garage st ored seed never achieved more than 5% germination at 49 days when the study ended (F igure 4-12). Maximum germination for the deep and midway stored seed was reached after 36 days (5 wk), and for the edge and treeline seed, after 43 days (6 wk).

PAGE 90

90 Tetrazolium assay after 17 months of storage indicated that th e viability of field stored seed had declined little if any fr om that determined at 5 months, compared to the viability of garage stored seed which had declined by nearly one half (Table 4-10). Germination of seed that had been stored 17 months in the garage was signif icantly lower than at any field stored location. Meanwhile in the field, the deep marsh location, which had remained flooded over the 17 months period, produced significantly greater germinati on. The small increase in germination that garage stored seed experienced from 5 to 17 mont hs was probably due to a slow release of the initial dormancy state over time, since long term storage at room temperature has the capacity, to some extent, to trigger dormancy release (Baskin and Baskin 1998). The time for 50% of the mean cumulative number of seeds to germinate was 15 days for the deep location and 17 days for the edge, while by the end of the experiment (49d) the treeline and garage stored seed never achieved more than 38% and 15% ge rmination, respectively (Figure 4-13). Maximum germination was sooner after 17 months than af ter 5 months of field storage, meanwhile a decline in germination was evident with the edge a nd treeline stored seed (Figure 4-15). Seed stored at the deep and edge marsh locations reached maximum germination at 3.4 weeks (24d); seed from the treeline took 6. 4 weeks (45d) and garage stored seed required 5.4 weeks (38d) before reaching maximum germination. Year 2005 seed source Hydrologic conditions: Surface water depth at the seed storage locations are illustrated from establishment in Oct. 2005 to marsh dry dow n in Apr. 2006 (Figure 4-14). At the deep and shallow marsh locations surface water was mainta ined over the length of the storage period, the marsh edge experienced flooding for the first three months, while the treeline remained continuously dry.

PAGE 91

91 Viability and germination: After the 5 month over-wintering period, viability and germination percentages for the 2005 seed mirrore d those for the 2004 seed after 17 months of storage. Viability by tetrazolium assay was sim ilar among the field and garage stored seed, yet actual germination was significantly higher in seed that had been stored deep in the flooded marsh (Table 4-11). The shallow marsh and edge of marsh seed comprised a second group with a significantly lower germination response. A third group, comprised of seed stored on the surface of the treeline and in the garage, demons trated poor germination which was significantly lower than that from the other storage locations. The time for 50% of the mean cumulative number of seeds to germinate was 18 days for the deep stored seed, 25 days for seeds buried at the treeline and 35 days for shallow marsh buried seed (Figure 4-15). Maximum germinati on was reached in 30 days for all of the buried treatments, whether at the treeline or in the marsh, however, the surface treeline and garage stored seed maintained very low germina tion rates through the experiment (49 days). Chi-square analysis among the field storage locations indicated that seed fate was highly dependent on storage location ( 2=308.6; p <0.0001; df 10, alpha 0.05; sample size=1200 (Figure 4-16). Seed fate following storag e at the treeline surf ace was similar to that of garage stored seed most seed died after 5 months, and of that remaining, dormancy was favored over germination. Seeds were more capable of germination after stor age at the more moderate field environments (buried at the tr eeline or at the edge of th e ponded marsh) than where dry environments were more constant. However, even seed stored at the bu ried treeline and marsh edge experienced a significant reduction in survival as indic aed by their significantly higher proportions of dead seed (Table 4-11). The mo st favorable experimental location for promoting both survival and germination was the consta ntly flooded environment of the deep marsh.

PAGE 92

92 Outdoor Laboratory Tests of Seed Bank Survival Constant flooding: The 2003 experimental seed bank stored under constant flooding in an outdoor bin remained viable for four years. By the fourth year, inundated seeds recruited successfully at substantial amounts while the do rmant seed component was gone (Figure 4-17). Chi-square analysis conducted on seed fate after three and four years of flooding demonstrated that seed fate was statistically dependent on time in years ( 2=35.3; p <0.0001; df 2, alpha 0.05; sample size=400). Partial sloughing of the external, hard seed coat was observed after the third year of inundated storage. By the fourth year, all rema ins of the initial seed coat were gone and the seeds were soft. Although this did not significantly reduce the viability of seeds, it could account for the simultaneous loss of the dormant seed component after four years. Seasonal flooding: Fate of the 2003 experimental s eed bank exposed to three years of seasonally cyclic flood-dry conditions is illust rated in Figure 4-18. E ach April, soon after surface dry down, strong pulses of germination a nd seedling emergence occurred. The pulse was followed by low levels of regeneration until flooding was imposed each July. Chi-square analysis for seed fate among the years indicated that seed fate was strongly dependent on time in fluctuating storage (( 2=989.0; p <0.0001; df 12, alpha 0.05; sample size=1200. The reduction of intact seed with time follo wed a first order decay of logarithmic decline (Figure 4-19). Notwithstanding the annual losses by germination and decay, the ar tificial seed bank had not reached depletion by the end of the third, final year of the study. The remaining seed, in Aug. 2006, was 23% viable and dormant, much reduced from the 60% measured after one year of study.

PAGE 93

93 Discussion Seed Bank Temporal Dynamics When the depression marsh remained almost entirely flooded through the year, soil sampling demonstrated little evid ence of change in seed bank density. This was supported by demographic findings which were underlined by the absence of S lacustris seedling emergence at the site. Despite the lack of seedlings, sampling confirmed th e existence of a considerably large and viable seed bank with high levels of persistence. When, hydrologic seasonality returned in the following year, as marked by spring dry down and summ er flooding, a significant decline was measured in seed bank density over the growing season. The loss in the seed bank was in turn explained by the number of seedlin gs that germinated and emerged. An analogous regeneration scenario was portray ed with experimental seed ba nks in outdoor pots subjected to seasonal flooding and where seedlings regenera ted annually upon spring dry down from a single, persistent seed bank. Only after four years of flood/dry seasonal cycles did the seed bank become nearly exhausted from depletion and decay. Notwithstanding the measured seed bank d ecline and the regeneration event of 2006, a persistent, viable seed bank remained at the de pression marsh two years af ter the last input of seed rain. Experimental studies confirmed that the seed bank was age structured, in that not all the seeds from a single source germinated the year following their production. Therefore, irrespective of the succe ss of the 2006 standing population, the persisting seed bank will provide a source for future populations, pending dry down of the habitat. The reduction detected in the S lacustris seed bank between years, the time between late summer 2005 and early spring 2006, may be explaine d by making inference to results from the experimental tests for seed bank survival. Based on the short history of in festation at the site, it can be assumed that the majority of the seed bank accounted for in 2005 originated from the

PAGE 94

94 2004 seed rain. Therefore, the seed collected in 2004 for burial in packages along the gradient of the marsh was the same age and comparable in storage environment as the natural seed bank. Experimental results demonstrated a 54% mort ality rate for seed buried at the treeline, meanwhile field sampling detected a 64% reduction in seed bank de nsity at the A water regime. The A regime and the treeline storage site were similar in that their substrates remained well drained and were never flooded. At the othe r extreme of the hydrol ogic gradient, a 28% mortality rate was demonstrated for seed stored at the deep marsh site, while field sampling detected a 28% reduction in the seed bank density at the deeply inundated (ID) regime. Thus it is probable that death and decay accounted for the substantially greater loss of the seed bank at the higher than at the lower elevations along th e depression marsh gradient. This account is more plausible than possible theories of abor ted germination or undetected seedling regeneration, especially because there was no support, from storage or greenhouse trials, for regeneration under flooding, nor during seasons other than the sp ring time. The possibility of relocation of the seed bank should also not be discounted, however a transect-based sampling design failed to reveal a zonal redistribution of seed between th e two years. Although disp ersal in water may be expected to produce a well mixed seed bank, outside studies have found that the seed banks of specific wetland zones reflect the dominant vegeta tion of those zones implication that wetland seeds generally stay in place (Leck an d Graveline 1979; Smith and Kadlec 1983). Seed Bank Function and Plant Demogr aphics in Relation to Hydrology The posit that response va riables would vary with the hydrologic gradient was demonstrated more strongly by plant than by seed bank variables. Correlation along the sampling transects inferred an optimum range for the biological responses as a reflection of the topography of the seasonal pond which is not a linear gradation or dec line but more a step function. Analysis based on the annual su rface water regime clarified the groupings by

PAGE 95

95 demonstrating relationships between hydrology and biological responses along the marsh gradient. Relationships between the seed bank data (density, viability, pe rsistence) and the hydrologic environment were commonly hinted at by means data, however, analysis was encumbered by heterogeneity sourced from within the seed bank distribution. Density of the viable seed bank was lowest at the highest elevations that were never flooded, more concentrated along the cyclically and shallo wly inundated zones and slightly lower again at the deeply inundated center of the marsh, however, heteroge neity prevented statistical distinctions among the water regime categories. The blueprint for seed bank density distribution lingered into the early seedling life stage in that the values for total number of seedlings were similar in distribution to that of seed bank density. Thus it may be inferred that at dry down, seed bank activity may regulate regeneration more than the available hydrologic environment. This could be a function of a time period when the narro w hydrologic requirement for regeneration is equally available across the marsh (i.e. drained substrate). Neve rtheless, soon after regeneration, the association between plant response and seed bank array vanished as st rong relationships were demonstrated between the later life stage va riables (seedling survival, vegetative vigor, fecundity) and the water regime categories representative of the hydrologic gradient. Winter flooding followed by spring dry down re sulted in a depression marsh defined by hydrologic zones. The lower elevation zones, which having been previously flooded, defined the ponded region of the marsh, promoted S lacustris through seedling survival, vegetative vigor, seed production, and, as demonstrated more clearly in experimental trials, seed bank persistence. Thus the magnitude of invasive populations was linked primarily to the previously inundated portions of the marsh. The fact that greater biomass and seed production were demonstrated

PAGE 96

96 from the relatively fewer stem numbers populat ing the ponded area prov ides further evidence that the influence of the hydrologi c regime dominated over that of the contribution of the active seed bank. A similar effect has been re ported in a perennial wetland Cyperaceae, Eleocharis where zonation was explained in part by seedli ng recruitment, not by the seed bank, and it was concluded that seedling survival may be more impor tant than seed bank pers istence in explaining zonation patterns (Be ll and Clarke 2004). The significant effect of inundation on na tive plant cover, and simultaneously, on the inverse of cover or bare substrate, was also de monstrated along the gradient of the marsh. In early 2005 the inverse of native pl ant cover was equivalent to ope n water and in early 2006, prior to S lacustris regeneration, the inverse of native plant cover was equivalent to an essentially bare, dewatered substrate. While native plan ts more vigorously inhabited the upper elevation zones of the depression marsh, the nonnative S. lacustris demonstrated greater fitness in the uncolonized expanse of the previ ously ponded area of the marsh. In the summer months of 2006, as the dry marsh transitioned into an aquatic environment, Scleria lacustris grew with and acclimated to the changing habitat. Pre-flowerin g plants adjusted to increasing surface water levels by increasing aerenchyma tissue in their lo wer stems and by developing adventitious roots at stem nodes as they became inundated. Equippe d with these morphologica l alterations plants were capable of modify their upr ight growth habit and to spra wl horizontally across the water surface. In this manner Scleria lacustris demonstrates adaptive traits that confer fitness in the rapid and extreme water level fluc tuations that are typical for a seasonal marsh in south central Florida. Similar features and habits were f ound to confer success to invasive populations of Phalaris arundinaceae in the Great Lakes region of No rth America (Detenbeck et al. 1999).

PAGE 97

97 In 2005, while S. lacustris remained absent from the flooded marsh, the native congener S. reticularis was frequent along the higher elev ation. Seedling regeneration of S. reticularis occurred earlier in the springtime and in soil s that were too high in water content for S lacustris to recruit. In the drought year of 2006, Scleria reticularis became infrequent and remained confined to the upper, densely vegetated zones. This pattern resulted in little overlap in distribution of the congeners along th e depression gradient. Meanwhile S. lacustris aggressively colonized the lower water regimes, which were essentially vacant of vegetation on dry down. Outside studies have used experimental met hods to find out why some congeners are better colonizers and more invasive, as was casually observed in this study. For example, among three species of Centaurea (Asteraceae), the most invasive congener excelled from its stronger response to canopy gaps and its longer grow ing period (Gerlach and Rice 2003). Similar characteristics appear to se parate the two species of Scleria at the depression marsh. Significantly shorter in stature, the native S reticularis (~30 cm ht) could not tolerate summer water levels in the lower gap regi ons most effectively colonized by S. lacustris. Future studies could incorporate experimental desi gn to directly contrast this apparent association between the two species and the wetland water regime. Dormancy The high viability, yet delayed germination of fresh seed indicated the involvement of an innate dormancy mechanism. Typically, a stro ng primary dormancy in fresh seeds will prevent germination immediately after maturation and se ed shed and will promote development of the seed bank. The ecological significance of primary dormancy is that it serves as a mechanism to prevent germination at inopportune times (Fenner and Thompson 2005). In species within the Cyperaceae, large seeded species tended to bear thicker, harder seed coats and to be more dormant and for some wetland species of Carex, weakening of the seed coat

PAGE 98

98 was required for two to three years before seed s could germinate (Schtz 2002). However, these reports were for mainly for temperate taxa and in this study, only severa l months (over winter) were needed before initial dormancy was weaken ed and germination proceeded. This indicates that seeds of S. lacustris are probably not physically dorma nt, but were controlled by a physiological dormancy. The use of unheated greenhouse germination te sts (phenology studies of Baskin and Baskin 1998), as employed in this study, help in understanding when seeds will germinate, especially the ambient temperat ure is monitored. During late Jan. and early Feb., the large temperature ranges experienced between day and night in the days preceding seedling emergence, were probably important in initiating the germination of over wintered seed. Acid and heat treatments (data not s hown) were also ineffective for breaking the ini tial dormancy in fresh seeds of S lacustris and further support the assumption th at initial dormancy was probably not physical, such as from a hard seed coat, but physiological. Findings from this study might have indicated that the viabil ity percentages determined by tetrazolium assay on fresh seed could be used as pr edictors for germination in the field. Indeed, viability results appeared to be duplicated in the unheated greenhouse. However, as supported by field experiments of stored seed, direct predic tions may not made between initial viability and germination because varying degrees of dor mancy release will result depending on the hydrologic environment of the seed bank. Seed Bank Survival It is unclear why the results for viability and germination of 2005 seed stored for 5 months were closer to those for 2004 seed stored for 17, rather than for 5 months. Because the trend was apparent with the tetrazolium viability assay, which was conducted before greenhouse germination trials, variable gr eenhouse conditions across the years may be ruled out. Handling

PAGE 99

99 methods of the fresh seed, which in 2004 involve d immediate field storage, but for 2005 seed required several weeks away from the field, may be more likely as having had an influence. Misinterpretation of tetrazolium assay could be the cause for the inaccurately low viability estimates of 2004 seeds stored at the treeline. Se ed metabolic rate appears to influence staining intensity, for example as seen in this study, freshly shed (dormant) seed produced weaker staining intensity than did seed that had been stored at somewhat hydra ting conditions (data not shown) while seed extracted in the springtime from preemergence soil cores produced the most vivid embryo staining results. Tetrazolium viability assay was not ideal for predicting the survival of experimental seed banks stored in the field; however, when combined with germination trials, it was very effective for detecting the presence and understanding the influence of seed dormancy. Dormant seed states were indicated by germination percentages th at fell short of the TZ viability ratings, such as was seen with garage stored seed. Extr eme desiccation over wint er, as experienced under garage storage, was incapable of inducing seed ou t of its initial dormant state. Naturally dry conditions on the ground surface at the treeline, while prompting a small degree of germination, similarly maintained a significant dormant func tion in the seed bank. While viability tests indicated that seeds can tolera te high levels of desiccation by prolonging dormancy for a matter of months, the ability to germ inate following extreme, long term desiccation will drop markedly beyond the first growing season, or year. As suggested by Hennessy (1985), dormancy may serve as a mechanism for S. lacustris to escape drought over seasonal dr y periods in Africa. The same feature, however, may not function to mainta in the viability a seed bank to after extended years of wetland desiccation. Th ese indications corroborate observ ations by resource managers

PAGE 100

100 that invasive populations of S. lacustris declined after consecutive ye ars of continuous climactic drought and marsh dry down (Ed Harris, pers. comm.). Conversely, mechanisms that resulted in seed hy dration over the winter months be it burial in drained soil, fluctuating or constant flooding, served to induce dormancy break and result in significantly greater degrees of ge rmination and seedling emergence The most effective condition for dormancy break and germination was deep flooding. Thus, similar to that found with seeds of temperate wetland spec ies (Baskin et al. 2002) seeds of S lacustris come out of dormancy with hydration, and that degrees of flooding can result in significantly higher percentages of germination ove r that of non flooded seeds. Long term (3 year) constant flooding positivel y influenced seed bank survival while 3 years of seasonal wet/dry envir onments depleted the seed bank through a series of spring germination events that would have lead to colonization of a marsh under dry down and rejuvenation of the seed bank through annual s eed rain. The outdoor la boratory experiment supported this finding which was as basic to the soil seed bank as it was to experimental seed bank storage. Under continuously flooded condi tions, seeds maintained high viability, after three years the recruitment levels were equivalent to those of fr esh non-dormant seed (68%). At the same time, the component of dormant seed wa s very small, indicating, as was found in field storage experiments, that flooded seeds may be maintained at an active, nondormant metabolic level. In summary, three functionally distinct sets of experimental seed bank environments were identified along the hydrologic gradient of the de pression marsh; dry, intermediate (moist or cyclically flooded) and constan tly flooded. These environments promoted poor, moderate and excellent survival of S lacustris seed, respectively, over time. Wh ile seeds were highly tolerant

PAGE 101

101 of constant flooding they were intolerant of drought, which ranged experimentally from the extreme desiccation of the garage to the naturally dry ambient at the surface of the treeline above the marsh. The intermediate or cyclic marsh environment promoted a moderate level of both germination and survival by maintaining a le vel of prolonged dormancy, a seed state that contributes to seed bank persis tence over time. While the cons tantly flooded conditions (>20cm of standing surface water) maintained seeds at th e highest nondormant state, as demonstrated by higher regeneration rates and greater germinati on responses with greenhouse trials, constant flooding was still effective at maintaining a porti on of viable seed in the dormant condition, and thus provided the best overall conditions fo r ensuring both a seasonal standing population and seed bank persistence over time. Conclusion This study provides quantitative evidence fo r understanding how the hydrologic regime induces shifts in the plant community of a seasonal marsh during opposing years of flood and drought by directly affecting the presence or ab sence of a seed bank based invasive species, S lacustris Extended hydroperiods of surface water flooding prevented seedling regeneration and prolonged seed bank viability. Extended sp ring drought and ensuing dry down induced regeneration and recruitment in the newly exposed substrates of the previously flooded marsh. Short, intermittent dry downs of late summer and early autumn did not induce regeneration, perhaps because soil water content remained at or near saturation, a condition which preclude S. lacustris regeneration in springtime, and/or becaus e of a prolonged dormancy mechanism in the persisting seed bank. In light of the high rates of seed bank pers istence, seedling recruitment and seed production demonstrated in this study, a rela tively small seed bank will only be promoted by the fluctuating hydrologic regime of seasonal marshes in Florid a, thus ensuring the species a permanent place in the flora.

PAGE 102

102 Table 4-1. Parameters for a linear model predicti ng viability of intact seed as determined with tetrazolium assay. The regression was forced to go through the origin (viable seed = intercept + slope intact seed). Year Sample time # Cores ( n ) Slope r2 2005 Preemergence 92 0.856 0.990 2005 Postemergence 48 0.906 0.998 2006 Preemergence 84 0.768 0.912 2006 Postemergence 84 0.765 0.995

PAGE 103

103Table 4-2. Density of Scleria lacustris seeds in the 2005 soil seed bank as estimated from the mean number of viable seed (m-2 1 std. dev.) extracted during preemergence and postemergence sa mpling times. Values are extrapolations from 15 cm diameter soil cores coll ected at each station ( n =4) running from the highest point ri mming the marsh (Station 0) to the lowest center of the marsh (Station 7). Transect I Transect II Transect III Station Preemergence Postemergence Preemergence Postemergence Preemergence Postemergence 0 0 0 0 0 212 117 71 54 1 127 54 198 108 0 0 0 0 368 255 467 198 2 269 125 1471 757 1174 672 340 273 3 637 246 1358 231 863 117 4 170 146 170 113 1217 582 651 312 538 383 453 212 5 354 192 1061 432 325 228 6 170 160 255 164 127 28 778 1011 7 57 65 141 135 127 85 354 371 481 247

PAGE 104

104Table 4-3. Density of Scleria lacustris seeds in the 2006 soil seed bank as estimated from the mean number of viable seed (m-2 1 std. dev.) extracted during preemergence and postemergence sa mpling times. Values are extrapolations from 15 cm diameter soil cores coll ected at each station ( n =4) running from the highest point ri mming the marsh (Station 0) to the lowest center of the marsh (Station 7). Transect I Transect II Transect III Station Preemergence Postemergence Preemergence Postemergence Preemergence Postemergence 0 0 0 0 0 1 14 28 0 0 0 0 0 0 71 85 127 97 2 99 85 28 33 439 354 184 223 85 73 57 65 3 71 85 28 33 99 54 170 139 283 191 226 201 4 127 128 14 28 354 162 85 73 368 435 241 71 5 0 0 14 28 113 104 170 226 566 327 622 588 6 113 122 0 0 57 65 0 0 198 247 28 57 7 99 54 28 33 99 54 42 54

PAGE 105

105 Table 4-4. Coefficients of determination (r2) used to identify the strongest predictor for surface water, WavgSu (mean summer surface wate r) or WHi (annual surface water high) and for soil moisture, MavgSp (mean spring soil moisture) or MavgM (mean April soil moisture). Spearman correlation anal ysis was conducted on a per station basis between hydrologic variable s and biological data; 2005: n =23; 2006: n =21. *P < 0.05; **P < 0.01; ***P < 0.001. Surface water Soil moisture Yr WavgSu WHi MavgSp MavgM Scleria lacustris Intact seed, preemergence 05 -0.099 -0.099 0.333 0.137 06 -0.024 -0.016 0.584** 0.548** Intact seed, postemergence 05 0.164 0.121 0.521 -0.353 06 -0.041 -0.033 0.597** 0.525 Spent seed, preemergence 05 -0.017 -0.001 0.179 0.139 06 0.280 0.280 0.400 0.371 Spent seed, postemergence 05 0.338 0.296 0.521 0.362 06 0.254 0.269 0.531* 0.497* Viable seed, preemergence 05 -0.061 -0.068 0.275 0.137 06 0.272 0.289 0.425 0.390 Viable seed, postemergence 05 0.234 0.182 0.521 -0.353 06 0.272 -0.044 0.556** 0.483* %Viability, preemergence 05 0.563** 0.524* 0.467 0.0 06 0.515* 0.536* 0.100 0.087 % Viability, postemergence 05 0.686* 0.686* 0.637 -0.353 06 0.357 0.350 0.400 0.411 % Persistence, intact seed 05 0.669* 0.622* 0.405 -0.353 06 0.169 0.170 0.556** 0.508* % Persistence, viable seed 05 0.607* 0.580* 0.405 -0.353 06 -0.106 -0.091 0.488* 0.417 # Seedlings 06 0.138 0.145 0.741*** 0.692*** # Seedling cohorts 06 0.104 0.113 0.601** 0.566** % Seedling survival 06 0.553** 0.532* -0.100 -0.078 # Adults 06 0.399 0.398 0.582** 0.540* Biomass, adult 06 0.522* 0.522* 0.525* 0.543* # Seed, developed 06 0.463* 0.467* 0.370 0.345 Plant associates # Scleria reticularis seedlings 05 -0.466* -0.467* 0.093 1.0*** 06 -0.743*** -0.755***-0.167 -0.151 Native cover 05 -0.847*** -0.857***-0.284 0.561 06 -0.646** -0.668***-0.149 -0.162

PAGE 106

106Table 4-5. Density of Scleria lacustris seeds in the soil seed bank as categorized by annua l surface standing water (water regime category) transecting the gradient of the depression marsh. Values are the mean (m-2 1 std. dev.) extrapolated from 15 cm diameter soil cores. 2005: Preemergence A, n = 4; C, n = 24; IS, n = 28; ID, n = 36; df = 88; Postemergence A, n = 4; C, n = 20; IS, n = 4; ID, n = 20; df = 44. 2006: A, n = 12; C, n = 8; IS, n = 32; ID, n = 32; df = 80. Water regime category A C IS ID F 2005 Intact seed, preemergence 226 122A518 747AB 853 537A 399 506AB 3.77* Intact seed, postemergence 99 85A 407 572A 523 300A 339 307A 1.06 Spent seed, preemergence 71 141B433 539B 1360 1091A 309 329B 14.2*** Spent seed, postemergence 42 54B 167 217B 1754 1047A 240 289B 23.2*** Viable seed, preemergence 212 117B384 601AB 743 484A 369 477AB 3.75* Viable seed, postemergence 71 54A 368 530A 452 211A 314 278A 0.76 2006 Intact seed, preemergence 146 283B347 304A 279 289AB 115 136B 3.62* Intact seed, postemergence 57 113A156 183A 230 363A 42 86A 3.6 Spent seed, preemergence 1169 506A1527 302A 1308 546A 1316 501A 0.8 Spent seed, postemergence 47 170C112 668AB 1476 1734A 493 438BC 1.37 Viable seed, preemergence 23 57C 261 302A 231 263AB 101 117BC 4.75** Viable seed, postemergence 42 80A 120 166A 175 279A 32 84A 3.42 F values are the results of ANOVA to compare water regime categories for each seed bank variable: *P<0.05; **P<0.01; ***P<0.001. Means within a row with the same letter are not si gnificantly different (alpha = 0.05) by a Duncans multiple rang e test.

PAGE 107

107Table 4-6. Percent viabili ty and persistence of Scleria lacustris seeds in the soil seed bank as ca tegorized by annual surface standing water (water regime category) transecting the gradient of the depression marsh. Valu es are the mean percent ( 1 std. dev.) extrapolated from 15cm diameter soil cores. 2005: Viability preemergence A, n = 4; C, n = 24; IS, n = 28; ID, n = 36; df = 88. Viability postemergence A, n = 4; C, n = 20; IS, n = 4; ID, n = 20; df = 44. Persistence A, n = 4; C, n = 20; IS, n = 4; ID, n = 20; df = 44. 2006: A, n = 12; C, n = 8; IS, n = 32; ID, n = 32; df = 80. Water regime category A C IS ID F 2005 Viability, preemergence 95 10A 49 38B 85 20A 85 30A 9.07*** Viability, postemergence 58 42A 54 46A 89 13A 91 22A 4.15 Persistence, intact seed 96 138A 86 128A 143 119A 116 134A 0.31 Persistence, viable seed 69 90A 94 115A 129 91A 119 130A 0.33 2006 Viability, preemergence 4 9B 67 31A 69 37A 25 42A 7.42*** Viability, postemergence 20 37B 51 45AB 61 41A 26 42B 5.07** Persistence, intact seed 13 27A 77 99A 88 126A 55 150A 1.11 Persistence, viable seed 25 62A 51 71A 73 98A 44 147A 0.64 F values are the results of ANOVA to compare water regime categories for each seed bank variable: *P<0.05; **P<0.01; ***P<0.001. Means within a row with the same letter are not si gnificantly different (alpha = 0.05) by a Duncans multiple rang e test.

PAGE 108

108Table 4-7. The ANOVA summary of demographic data from a seasonal depression marsh ca tegorized by annual surface standing water (water regime categories) transecting the hydr ologic gradient. Values are the mean (m-2 1 std. dev.) extrapolated from the 0.5 m2 sampling grid. 2005: A, n = 1; C, n = 6; IS, n = 7; ID, n = 9; df = 19. 2006: A, n = 3; C, n = 2; IS, n = 8; ID, n = 8; df =17. Water regime category A C IS ID F 2005 Scleria lacustris # Seedlings 0 0 0 0 # Adults 0 0 0 0 Plant associates # Scleria reticularis seedlings 76 0A 3 8B 0 0B 0 0B 103.0*** Native cover (rating) 5 0A 4 1A 2 1B 1 0B 14.1*** 2006 Scleria lacustris # Seedlings 31 46B 242 3A 196 172AB 146 90AB 1.59 # Seedling cohorts 1 1C 4 0A 3 1AB 3 1B 6.43** # Seedlings 2 April cohort 24 41B 136 28A 1 1B 0 0B 43.0*** # Seedlings 17 April cohort 7 6A 70 31A 159 151A 99 53A 1.73 % Seedling survival 11 19C 43 2B 88 8A 86 10A 48.0*** # Adults 9 16A 104 5A 166 146A 125 74A 1.63 Biomass adults (g) 14 25B 351 117AB556 377A 703 186A 4.98* Biomass 2 April cohort (g) 14 25B 211 114A 2 5B 0 0B 31.0*** Biomass 17 April cohort (g) 0 0C 91 2BC 475 364AB 499 137A 4.21* # Seed, developed 0 0B 576 814AB1133 735A 1408 270A 5.52** # Seed, aborted 0 0B 616 226AB930 560A 1296 368A 6.83* Plant associates # Scleria reticularis seedlings 48 25B 136 28A 22 54B 3 7B 7.34* Native cover (rating) 6 1A 3 0B 3 2B 2 1B 6.42** F values are the results of ANOVA to compare water regime categories for each seed bank variable: *P<0.05; **P<0.01; ***P<0.001. Means within a row with the same letter are not si gnificantly different (alpha = 0.05) by a Duncans multiple rang e test.

PAGE 109

109 Table 4-8. Viability and germination responses of 2005 fresh seed. Mean percent viability ( std. dev.) as determined with tetrazolium is given for control seed (time 0, Oct. 2005), and after growth chamber incubation of th e seed at 3 wk. Germination, in mean percent ( std. dev.), was determined according to seedling emergence in an unheated greenhouse. Viability in mean percent ( std. dev.) was determined on seed remaining ungerminated in the greenhouse at 21 wk Treatment Sample size Time Mean percent germination Mean percent viability Control 200 0 0 63.5 8.5 Growth chamber light 200 3 wk 0 62.5 20.5 Growth chamber dark 200 3 wk 0 68.7 18.0 Greenhouse (unheated) 192 3 wk 0 9 wk 0 12 wk 1.0 1.4 15 wk 52.1 8.8 18 wk 59.9 5.1 21 wk 60.4 4.4 0.5 0.7 Table 4-9. Seed fate of 2005 fresh seed after a ge rmination trial of 21 wk (147 d) in an unheated greenhouse. Seed that had not germinated by 21 wk was extracted and assayed as dormant (viable) or dead. Seed fate is given in mean percent ( std. dev.) from a starting sample of 192 seeds. Seed fate Mean percent Germinated 60.4 4.2 Dormant (intact; viable) 0.5 0.7 Aborted (germination failed) 8.2 1.4 Dead (intact; dead) 4.2 0.7 Decayed (not found) 26.6 0.7

PAGE 110

110Table 4-10. Viability and germination resu lts for 2004 seed after 5 months and 17 mont hs storage at locations transecting the hydrologic gradient of the depression marsh, and in ambient ga rage. Values are the mean pe rcent std. dev. for each of the seed storage locations. Seed viability was determined befo re germination trials using te trazolium assay on 4 sets of 25 seeds retrieved from each location. Germination was determin ed according to the response of 6 sets of 25 seeds during a 49 d greenhouse trial. Seed storage location Garage Treeline Edge Midway Deep F 5 Months Viability 72 5.6A 48 7.3B 65 8.8AB 76 7.3A 71 15.4A 5.32** Germination 5 1.0B 71 0.8A 67 2.5A 69 1.2A 73 2.6A 112.68*** 17 Months Viability 34 8.5C 46 14.1BC 80 0A 72 0AB 13.73* Germination 15 4.2C 38 4.2B 55 3.5AB 68 8.5A 34.59** F values are the result of ANOVA to compare seed storage locations for each variable: *P<0.05; **P<0.01; ***P<0.001. Means within a row with the same letter are not significantly di fferent (alpha = 0.05) by a T ukeys Studentized range test.

PAGE 111

111Table 4-11. Viability, germination and seed death for 2005 seed after 5 months storag e at locations transecting the hydrologic gradient of the marsh, and in ambient garage. Values are the mean percent std. dev. for each of the seed storage locations. Seed viability was determined before germination trials using tetrazolium assay on 2 sets of 25 seeds retrieved from each location. Germination was determined according to th e response of 2 sets of 200 seeds during a 49 d greenhouse trial. Seed death was assessed using tetrazolium assay of ungerminated seed after completion of the germination trial. Seed storage location Garage Treeline surface Treeline buried Edge (TISta0) Shallow (TI-Sta1) Deep (TI-Sta7) F Viability 42 2.8A 62 14.1A 56 16.9A 34 8.5A 48 17.0A 58 2.8A 1.47 Germination 1 0.0C 13 2.1C 50 3.5B 47 9.9B 50 3.5B 74 8.4A 43.89*** Death 72 3.5A 62 3.5A 38 0.7B 39 5.7B 32 0.7BC 20 4.9C 56.50*** F values are the result of ANOVA to compare seed storage locations for each variable: *P<0.05; **P<0.01; ***P<0.001. Means within a row with the same letter are not significantly di fferent (alpha = 0.05) by a T ukeys Studentized range test.

PAGE 112

112 Figure 4-1. The characteristic profile of a preemerge nce soil core (15 cm diam. x 6 cm deep) extracted from the recently dewatered substrate of the Vero Beach depression marsh, April 2006. An organic surface layer extende d to ~3 cm depth and was distinctly underlain with mineral sand. A B Figure 4-2. Scleria lacustris A) fruiting stems and B) shed seed collected Oct. 2005 from standing plants at Cell C, Blue Cypress Water Management Area, St. Johns Water Management District. Mature seed deviated in color from green to mottled grey to white, while a shining exterior coat pervaded.

PAGE 113

113 A B Figure 4-3. March 2006 greenhouse germinati on trials at A) day 1, and B) day 21. 0 5 10 15 20 25 30 35 40Mar Apr May Jun Jul Aug Sept Oct Dec Jan Feb Mar Apr May Jun Jul Aug Sept Oct 20052006 Surface Water (cm) Sta. 1 Sta. 2 Sta. 3 Sta. 4 Sta. 5 Sta. 6 Figure 4-4. Surface water fluctuation along T II, a 36m transect representing the hydroperiod of a depression marsh near Vero Beach, over 19 months, the duration of two consecutive growing seasons.

PAGE 114

114 A B Figure 4-5. Vero Beach marsh midway at Tran sect II. A) In August 2005 was defined by a diverse native plant community. B) In September 2006 was dominated by Scleria lacustris

PAGE 115

115 0 0.1 0.2 0.3 0.4 0.5 0.6 012345671234561234567 T1TIITIII Mean Soil Volumetric Water Conte n 2-Apr 17-Apr 29-Apr 14-May 21-Jun Figure 4-6. Soil volumetric water content (re presented as the mean of three readings for each station) for stations along the three transects (TI, TII, and TIII) during the 2006 spring dry down pe riod at the Vero Beach study site demonstrated the absence of a trend along the hydrologic gradient of the depression marsh. Dry Dry Dry Wet Wet Wet

PAGE 116

116 0 20 40 60 80 100 120 140 01234567123456701234567 TITIITIII Seed Density 2005 Intact 2005 Viable 2006 Intact 2006 Viable Figure 4-7. Intact and viable seed bank for Scleria lacustris at sampling stations (0-7) along three permanent tr ansects of the depression marsh. Values represent pr eemergence densities before the 2005 and 2006 growing seasons as the sum of the four replicate 15 cm diameter soil cores sampled at each station (total area 707 cm2). Dry Dry Dry Wet Wet Wet

PAGE 117

117 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700012345671234561234567 Transect ITransect IITransect III Density / Square Meter Intact Seed Viable Seed Seedlings Figure 4-8. The 2006 preemergence seed bank (means for each station m) and seedlings produced at each station/m2. Values are extrapolations of data from 15 cm diameter soil cores and 0.25 m2 monitoring grids. Dry Dry Dry Wet Wet Wet TI TII TIII

PAGE 118

118 0 100 200 300 400 500 600 700 800 900 1000012345671234561234567 TITIITIII Dry weight g / sq meter 2-Apr 17-Apr 29-Apr 14-May 21-Jun Figure 4-9. Adult biomass of the 2006 seedli ng cohorts, quantified 18 September 2006, from Scleria lacustris at the Vero Beach depression marsh. Values are extra polations of data from the 0.25 m2 monitoring station that have been scaled to 1.0 m2. Dry Dry Dry Wet Wet Wet

PAGE 119

119 Figure 4-10. Graphical summarization of Table 4-4 detailing the grea test significant effects of hydrology, surface water (Wavg Su) and soil water (MavgM), on seed bank and plant variables at the Vero Beach depression marsh, 2006. Coefficients are given for each; arrows designa te the direction of effect. Native congener seedlings WavgSu % Viability Seedling survival Associated native cover Seed production MavgM Biomass Seed bank % persistence viable Seed bank density intact Seed bank density viable Seed bank density spent Seedlings Adults Seedling cohorts 0.68* -0.74*** 0.46* 0.55** -0.64** 0.52* 0.52* 0.48** 0.59*** 0.55*** 0.74*** 0.58** 0.53* 0.60**

PAGE 120

120 0 5 10 15 20 25 30 35 40 OctNovDecJanFebMarAprMayJunJulAugSeptOctNovDecJan FebMarApr 200420052006 Surface Water (cm) Tree Line Marsh Edge Marsh Midway Marsh Deep Figure 4-11. Surface water depth measured at f our locations along the marsh gradient of the Vero Beach study site. Locations denote storage sites for the 2004 Scleria lacustris seed source. Seeds were retrieved from sc reen packages for germination trials in March 2005 (5 months storage) and in March 2006 (17 months storage).

PAGE 121

121 0 10 20 30 40 50 60 70 80 90 100 05122229364349 Times in DaysMean Percent Germination Deep Midway Edge Treeline Garage Figure 4-12. Greenhouse germination rate resp onse of 2004 seed source following 5 months storage. Storage locations were the US GS garage and four locations along the gradient of the depression marsh describe d as: the upper treeline bordering the marsh, the marsh edge, midway to the center of the marsh and at the center point where surface water was at its deepest. Values are for mean percent ( std. dev.).

PAGE 122

122 0 10 20 30 40 50 60 70 80 90 100 036101317202427313238414548 Time in DaysMean Percent Germination Deep Edge Treeline Garage Figure 4-13. Greenhouse germination rate resp onse of 2004 seed source following 17 months storage. Storage treatments were the gara ge and four locations along the gradient of the depression marsh: the upper treeline bordering the marsh, the marsh edge, midway to the center of the marsh and at the center point where surface water was at its deepest. Values are for the mean percent ( std. dev.).

PAGE 123

123 0 5 10 15 20 25 30 35 40 OctNovDecJanFebMarApr 20052006 Surface Water Depth (cm) Tree Line Surface Tree Line Buried TI Sta 0 (Edge) TI Sta 1 (Shallow) TI Sta 7 (Deep) Figure 4-14. Surface water depth measured along the gradient of Trans ect I (TI) at the Vero Beach depression marsh. The five locations indicate where the 2005 seed source of Scleria lacustris was stored by burial in screen packages. Field storage was maintained until March 2006 (5 months). Values are the mean percent ( std. dev.).

PAGE 124

124 0 10 20 30 40 50 60 70 80 90 100 047111418212528323539424649 Time in DaysMean Percent Germinatio n Deep Shallow Edge Treeline Buried Treeline Surface Gara g e Figure 4-15. Greenhouse germinatio n rate response of 2005 seed after five months storage. Storage treatments were garage and five locations along the gradient Transect I (TI) at the Vero Beach depression marsh: th e upper treeline bordering the marsh, on the surface (where vegetation had been clipped), and buried to 3cm, TI Station 0 (marsh edge), TI Station 1 (shallow marsh), and TI Station 7 (d eep marsh). Values are the mean percent ( std. dev.).

PAGE 125

125 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110Garage Treeline Surface Treeline Buried Edge (TI Sta 0) Shallow (TI Sta 1) Deep (TI Sta 7) 5 Months Seed Fate (Mean Percent) Dead Dormant Germinated Figure 4-16. Categorical format accounting for the fate of 2005 seed retrie ved after a five month overwintering storage to test seed bank su rvival. Values for seed fate represent responses after greenhouse germination tria ls and viability testing of ungerminated seed ( n =200). Seed fate was as follows: germinated, did not germinate due to dormancy, or did not germinate because it was dead. Seed fate is presented in mean percentage ( std. dev.).

PAGE 126

126 0 10 20 30 40 50 60 70 80 90 100 3 Years4 Years 20062007 Seed Fate (Mean Percent) Dead Dormant Germinated Figure 4-17. Experimental seed bank (2003 seed source) stored under c onstant flooding at an outdoor laboratory remained viable for four years. Seed fate was evaluated with replicate greenhouse culture ( n =200) following the third and fourth years of experimental storage. Categorical values represented in stack bars reflect the mean percentage ( std. dev.) of seed that ha d germinated, remained dormant or died.

PAGE 127

127 0 10 20 30 40 50 60 70 80 90 100 2003200420052006Seed Fate (mean %) Past Decayed Past Emerged Decayed Germinated Intact Figure 4-18. Response over time in years of an experimental seed bank (2003 seed source) stored under seasonally cyclic flood-dry c onditions (early spring-summer dry down / later summer winter flooding) at an outdoor laboratory. Categorical values for seed fate reflect the mean percentage of seed th at had germinated, remained intact, or had decayed ( n =300) for each year of study.

PAGE 128

128 1 10 10020032004 Year 1 2005 Year 2 2006 Year 3 Intact Seed (mean %) Figure 4-19. Logarithmic reduction of int act seed (2003 seed source) stored under experimentally induced seasonal hydro logy (later summer-winter flooding and spring-early summer dry down) at an outdoor laboratory.

PAGE 129

129 CHAPTER 5 CONCLUSION AND SYNTHESIS This study confirmed the hypothesis that seasona l fluctuation in surface water drives seed bank dynamics and seedling recruitment in the invasive annual plant, Scleria lacustris In Chapter 2, I identified a seed bank for S lacustris and discovered that it occurred in relatively high numbers in the surface sediments of marshe s in south central Florida. Of functional importance was finding that, even after transiti oning into a standing population, a substantial seed bank remained to persist. The development of a persistent seed banks insure a source of invasion in Florida wetlands. In Chapter 3, I used experimental method to confirm the hypothesis that hydrologic dry down of the marsh environment func tioned to induce regeneration in S lacustris Additionally I found that while dry down lead to recruitment, co nstant flooding resulted in the maintenance of viability at high levels in the ungerminated seed. The ecological consequences of this finding heightens the threat of seed bank persistence to ecosyste m integrity under the current working principles of using hydrologic dry down in restoration management. In Chapter 4, quantitative evidence from seed bank, plant and hydrologic variables were used to clarify the relationship between seed bank dynamics and population incidence their relationship with the fluctuati ng hydrologic regime of the seasona l marsh. One application of this study is that it allows for predicting the fundamental dynamics (presence or absence) of an invasive species in tempor ally fluctuating habitats. Two underlying variables identified as highly influential in this study were 1) within year spatial heterogeneity of the seed bank and 2) year to year variability in the hydrologic environment of the seasonal wetland habitat. While these factors portray inherent ecosystem properties, they provide obstacl es in the ability to conduct e ffectively detailed demographic

PAGE 130

130 models (i.e. transition matrices ) for seed bank dynamics, seedling recruitment, and survival. Future studies may benefit from the use of pe rmanently placed instru mentation to capture subtleties in the environment which would allo w longer term monitoring and a more complete analysis of seed bank longev ity and population dynamics. Investigations conducted in other wetland habita ts have demonstrated that environmental fluctuations will enhance and sometimes maintain species diversity (e.g., Keddy and Reznicek 1982). Floridas wetlands are dive rse and unique but highly vulnerab le to the impacts of habitat alteration, primarily through human disruption of na tural hydrologic patterns. In the way that water level fluctuation is essential for the long term survival of species (Keddy and Reznicek 1982) or for healthy marsh systems, in this study I found seasonal hydrologic fluctuation working to enhancing the recruitment of a single invasive species at near monoculture levels. While Floridas wetlands are uniquely diverse, they are highly vulnerable to the impacts of habitat alteration, primarily thr ough human disruption of their na tural hydrologic fluctuations, but also through climactic change. In Florida, this impact can be expected to advance if hydroperiods continue to be shortened from th e combined effect that climactic drought and societal withdrawal will have on lowering of the groundwater ta ble. Anticipated extremes in the frequency and intensity of rainfall and drought ev ents, such as has been predicted with global climate change (IPPC 2007), can be expected to further select for the increasing incidence of S. lacustris in Florida.

PAGE 131

131 APPENDIX A TETRAZOLIUM STAINING PATTERNS

PAGE 132

132A B C D Figure A-1. Excised embryos from the seed of Scleria lacustris illustrate the various patterns obs erved with tetrazolium staining (1%) when assayed for seed viability. Embryos are interpreted as dead A) when only the apex stains red (30X magnification) and B) when only the base stains red ( 30X magnification). Embryos are interprete d as viable C) when only a small portion of the cotyledon, or scutellarlike region, remains unstained (50X), and D) when the entire embryo stains red (50X magnification). Seeds from the embryos pi ctured were collected Nov. 2003 at Vero Beach and stored under inundation at an outdoor laboratory, in Gainesville, until March 2006.

PAGE 133

133 APPENDIX B 2005 PLANT COMMUNITY INVENTORY

PAGE 134

134 The plant community of the Vero Beach depression mars h was inventoried 18 August 2005. Species were identified as occurring within th e lateral zone of the respective station under which they are listed and not only within each 0.25 m2 station. The plant species distributions demonstrate the natural community vegetation r unning the hydrologic gradient of the study site from the highest (0) to the lowest (7) point. Species ar e listed in the approximate order of more to less common. Nomenclature follows Wunder lin and Hansen (2003). Identification was determined from field specimens using Godf rey and Wooten (1979, 1981). Voucher specimens were deposited at the herbarium of the Fl orida Museum of Natural History (FLAS). Transect I Station 0 Transect III Station 0 Panicum repens Paspalum notatum Xyris jupicai Andropogon virginicus virginicus Rhyncospora inundata Taxodium ascendens Lachnanthes caroliniana Eupatorium sp. Hypericum fasiculatum Urena lobata Sabal palmetto Scleria reticulata Rhyncospora inundata Panicum hemitomum Erigeron sp. D Persea palustris Myrica cerifera Pinus elliottii Baccharis halmifolia Hyptis alata Transect I Station 1 Transect II Station 1 Transect III Station 1 Rhyncospora inundata Panicum repens Panicum repens Fuirena scirpoidea Scleria reticulata Rhyncospora inundata Xyris jupicai Fuirena scirpoi dea Lachnanthes caroliniana Lachnanthes caroliniana Diodia vi rginiana Solidago leavenworthii Centella asiatica Ipomoea sp. Xyris jupicai Pluchea rosea Andropogon virginicus Transect I Station 2 Transect II Station 2 Transect III Station 2 Xyris jupicai Panicum repens Panicum repens Rhyncospora inundata Centella asiatica Xyris jupacai Lachnanthes caroliniana Sagittaria lancifolia Rhyncospora inundata Fuirena scirpoidea Diodia virg iniana Lachnanthes caroliniana Sagittaria lancifolia Rhyncospora inundata

PAGE 135

135 Centella asiatica Lachnanthes caroliniana Panicum hemitomum Hydrocotyle umbellata Transect I Station 3 Transect II Station 3 Transect III Station 3 Xyris jupicai Centella asiatica Panicum hemitomum Fuirena scirpoidea Rhyncospor a inundata Fuirena scirpoidea Rhyncospora inundata Sagittaria lancifolia Rhyncospora inundata Lachnanthes caroliniana Fuirena scirpoidea Centella asiatica Panicum hemitomum Lachnanthes caroliniana Rhyncospora tracyi Sagittaria lancifolia Rhyncospora tracyi Lachnanthes caroliniana Centella asiatica Utricularia foliosa Rhyncospora tracyi Panicum repens Transect I Station 4 Transect II Station 4 Transect III Station 4 Fuirena scirpoidea Rhyncospora tracyi Panicum hemitomum Rhyncospora tracyi Fuirena sc irpoidea Stillingia aquatica Xyris jupicai Stillingia aquatica Sagittaria lancifolia Lachnanthes caroliniana Bacopa caroliniana Rhyncospora tracyi Sagittaria lancifolia Justicia americana Lachnanthes caroliniana Bacopa caroliniana Lachnanthes caroliniana Fuirena scirpoidea Stillingia aquatica Utricularia foliosa Centella asiatica Pluchea roseae Transect I Station 5 Transect II Station 5 Transect III Station 5 Fuirena scirpoidea Bacopa caroliniana Panicum hemitomum Rhyncospora tracyi Fuirena scirpoidea Cyperus haspans Bacopa caroliniana Rhyncospora tracyi Lachnanthes caroliniana Sagittaria lancifolia Sagittaria lancifolia Fuirena scirpoidea Stillingia aquatica Bacopa caroliniana Sagittaria lancifolia Stillingia aquatica Bacopa caroliniana Rhyncospora inundata Ludwigia sp. Transect I Station 6 Transect II Station 6 Transect III Station 6 Bacopa caroliniana Rhyncospora tracyi Nymphaea odorata Sagittaria lancifolia Stillingia aquatica Stillingia aquatica Stillingia aquatica Sagittaria lancifolia Sagittaria lancifolia Rhyncospora tracyi Justicia americana Rhyncospora tracyi Bacopa caroliniana Fuirena scirpoidea Transect I Station 7 Transect II Station 7 Transect III Station 7 Rhyncospora tracyi Stillingi a aquatica Stillingia aquatica Sagittaria lancifolia Fuirena scirpoidea Sagittaria lancifolia Bacopa caroliniana Justicia americana Rhyncospora tracyi Justicia americana Rhyncospora filifolia Fuirena scirpoidea Stillingia aquatica Justicia americana Bacopa caroliniana Utricularia foliosa

PAGE 136

136 APPENDIX C HYDROLOGY AT 2004 FIELD SITES

PAGE 137

137 A 0 5 10 15 20 25 30 3523-Feb 14-Mar 3-Apr 23-Apr 13-May 2-Jun 22-Jun 12-Jul 1-Aug 21-Aug 10-Sep 30-Sep 20-Oct Surface Water Depth (cm) Sta 1 Sta 2 Sta 3 Sta 4 Sta 5 Sta 6 Sta 7 Sta 8 Sta 9 Sta 10 Sta 11 B 0 5 10 15 20 25 30 3523-Feb 14-Mar 3-Apr 23-Apr 13-May 2-Jun 22-Jun 12-Jul 1-Aug 21-Aug 10-Sep 30-Sep Surface Water Depth (cm) Sta 1 Sta 2 Sta 3 Sta 4 Sta 5 Sta 6 Sta 7 Sta 8 Sta 9 Sta 10 Sta 11 Figure C-1. Surface water fluctuation during the 2004 study period at the A) Vero Beach and B) Kissimmee seasonal marshes as measured with a meter stick at each station. Relatively small differences were demonstrated among each of the 11 stations. Maximum depth at Kissimmee ranged betw een 52-63 cm on 20 Sept. (off scale).

PAGE 138

138 A 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1234567891011 StationSoil Volumetric Water Content 29-Apr 25-May 14-Jun 28-Jun 27-JulB 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1234567891011 StationSoil Volumetric Water Content 17-May 21-Jun 27-Jul Figure C-2. Soil moisture for the 2004 spring/early summer dry down period at the A) Vero Be ach and B) Kissimmee study sites demonstrated small differences among the 11 sampling stations within monitoring dates. Soil moisture (volumetric soil water content) was measured with a portable Thetaprobe se nsor (Delta-T Devices and The Macaulay Land Use Research Institute). The volumetric response wa s not linear; thereby values of 0.5 to 0.6 and above are near saturation, 1.0.

PAGE 139

139 APPENDIX D ORGANIC CONTENT AND SOIL MOISTURE OF SUBSTRATE CORES AT VERO BEACH STUDY AREA, MARCH 2004

PAGE 140

140 0 10 20 30 40 50 60 70 80 903cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm 3cm 6cm 9cm Station 1Station 2Station 3Station 4Station 5Station 6Station 7Station 8Station 9Station 10 % Soil Moisture % Organic Matter Figure D-1. Gravimetric determination of pe rcent moisture and organic matter was made after drying and com bustion of replicate soil cores.

PAGE 141

141 LIST OF REFERENCES Abernethy, V.J. and N.J. Willby. 1999. Changes along a disturbance gradient in the density and composition of propagule banks in floodplain aquatic habitats. Plant Ecology 140:177190. AOSA (Association of Official Seed Analysts). 2000. Tetrazolium Testing Handbook. J. Peters (ed.). Contribution No. 29. Association of Official Seed Analysts, NM, USA. Baldwin, A.H., M.S. Egnotovich, and E. Clarke 2001. Hydrologic change and vegetation of tidal freshwater marshes: field, greenhous e, and seed bank experiments. Wetlands 21:519-31. Baskin, C.C. and J.M. Baskin. 1998. S eeds, Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press, San Diego, CA, USA. Baskin, C.C., J.M. Baskin, and E.W. Chester. 1993. Seed germinati on ecophysiology of four summer annual mudflat species of C yperaceae. Aquatic Botany 45:41-52. Baskin, C.C., J.M. Baskin, and E.W. Chester. 2000. Effect of flooding on the annual dormancy cycle and on germination of seeds of the summer annual Schoenoplectus purshianus (Cyperaceae). Aqua tic Botany 67:109-16. Baskin, C.C., J.M. Baskin, and E.W. Chester. 2002. Effects of flooding and temperature on dormancy break in seeds of the summer annual mudflat species Ammannia coccinea and Rotala ramosior (Lythraceae). Wetlands 22:661-68. Baskin, C.C., E.W. Chester, and J.M. Bask in. 1996. Effect of flooding on annual dormancy cycles in buried seeds of two wetland Carex species. Wetlands 16:84-88. Baskin, J.M. and C.C. Baskin. 1989. Physiol ogy of dormancy and germ ination in relation to seed bank ecology. In M.A. Leck, V.T. Parker, and R.L. Simpson (eds.) Ecology of Soil Seed Banks. Academic Press, San Diego, CA, USA. Bekker, R.M., M.J.M. Oomes, and J.P. Bakker. 1998. The impact of groundwater level on soil seed bank survival. Seed Science Research 8:399-404. Bell, D M. and P.J. Clarke. 2004. Seed-bank dynamics of Eleocharis : Can spatial and temporal variability explain habitat segregation? Australian Journal of Botany 52: 119-31. Bigwood, D.W. and D.W. Inouye. 1988. Spatial pa ttern analysis of se ed banks: an improved method and optimized sampling. Ecology 69:497-507. Brown, J.S. and D.L. Venable. 1986. Evolutiona ry ecology of seed-bank annuals in temporally varying environments. The American Naturalist 127:31-47.

PAGE 142

142 Chambers, J.C. and J.A. MacMahon. 1994. A day in the life of a seed: m ovements and fates of seeds and their implications for natural and managed systems. Annual Review of Ecology and Systematics 25:263-92. Cook, C.D.K. 2004. Aquatic and Wetland Plants of Southern Africa. Backhuys Publishers, Leiden, The Netherlands. Core, E.L. 1936. The American species of Scleria Brittonia 2:1-105. Csontos, P. and J. Tamas. 2003. Comparisons of soil seed bank classification systems. Seed Science Research 13:101-111. DeAngelis, D.L. 1994. Synthesis: spatial and temporal characteristics of the environment. In S.M. Davis and J.C. Ogden (eds.) Everglad es: The Ecosystem and Its Restoration. St. Lucie Press, Boca Raton, FL, USA. DeBerry, D.A. and J.E. Perry. 2005. A draw down flora in Virginia. Castanea 70:276-86. Detenbeck, N.E., S.M. Galtowitsch, and J. Atkinson. 1999. Evaluati ng perturbations and developing restoration strategi es for inland wetlands in the Great Lakes Basin. Wetlands 19:789-820. Fairey, J.E., III. 1972. The genus Scleria in North America. Disse rtation, West Virginia University, Morgantown, WV, USA. Fenner, M. and K. Thompson. 2005. The Ecology of Seeds. University Press, Cambridge, England. FLEPPC (Florida Exotic Pest Plant Council). 2005. 2005 Inva sive Species List. Wildland Weeds 8:3-16. FNAI and FDNR (Florida Natural Areas Inve ntory and Florida Department of Natural Resources). 1990. Guide to the Natural Comm unities of Florida. Tallahassee, FL, USA. Gerlach, J.D. and K.J. Rice. 2003. Testing life history correlates of invasiveness using congeneric plant species. Ec ological Applications 13:167-179. Gerritsen, J. and H.S. Greening. 1989. Marsh se ed banks of the Okefenokee Swamp: effects of hydrologic regime and nutrients. Ecology 70:7550-763. Godfrey, R.K. and J.W. Woote n. 1979. Aquatic and Wetland Pl ants of Southeastern United States: Monocotyledons. University of Georgia Press, Athens, Georgia, USA. Godfrey, R.K. and J.W. Woote n. 1981. Aquatic and Wetland Pl ants of Southeastern United States: Dicotyledons. University of Georgia Press, Athens, Georgia, USA.

PAGE 143

143 Gross, K.L. 1990. A comparison of methods for estim ating seed numbers in the soil. Journal of Ecology 78:1079-93. Haukos, DA. and L.M. Smith. 2001. Temporal em ergence patterns of seedlings from playa wetlands. Wetlands 21:274-80. Hennessy, E.F.F. 1985. The genus Scleria in southern Africa. Bothalia 15:505-30. Intergovernmental Panel on Climate Change (IPPC). 2007. Synthesis Report of the IPCC Fourth Assessment Report (AR4). Jacono, C.C. 2001. Scleria lacustris (Cyperaceae), an a quatic and wetland se dge introduced to Florida. 2001. Sida 19:1163. James, C.S., S.J. Capon, M.G. White, S.C. Rayburg and M.C. Thomas. 2007. Spatial variability of the soil seed bank in a heterogeneou s ephemeral wetland system in semi-arid Australia. Plant Ecology 190:205-17. Kalisz, S. and M.A. Peek. 1992. Demography of an age-structured annu al: resampled projection matrices, elasticity analyses, and se ed bank effects. Ecology 73:1082-93. Keddy, P.A. and A.A. Reznicek. 1982. The role of seed banks in the persistence of Ontarios coastal plain flora. American Journal of Botany 69:13. Keddy, P.A. and A.A. Reznicek. 1986. Great Lake s vegetation dynamics: the role of fluctuating water levels on buried seeds. Jour nal of Great Lakes Research 12:25-36. Kessler, J.W. 1987. A treatment of Scleria (Cyperaceae) for North Am erica north of Mexico. Sida 12:391-407. Kitchens, W.M., R.E. Bennetts, and D.L. DeAn gelis. 2001. Linkages between the Snail Kite population and wetland dynamics in a highly fr agmented South Florida hydroscape. In J.W. Porter and K.G. Porter (eds.) Linkages Between Ec osystems: The South Florida Hydroscape. St. Lucie Press, Delray Beach, FL, USA. Leck, M.A. 1989. Wetland seed banks. In M.A. Leck, V.T. Parker, and R.L Simpson (eds.) Ecology of Soil Seed Banks. Academic Press, San Diego, CA, USA. Leck, M.A. 1996. Germination of macrophytes from a Delaware Ri ver tidal freshwater wetland. Bulletin of the Torrey Botanical Club 123:48-67. Leck, M.A. 2003. Seed-bank and vegetation deve lopment in a created tidal freshwater wetland on the Delaware River, Trenton, Ne w Jersey, USA. Wetlands 23:310-43. Leck, M.A and M.A. Brock. 2000. Ecological and evolutionary trends in wetlands evidence from seeds and seed banks. Plant Species Biology 15:97-112.

PAGE 144

144 Leck, M.A. and K.J. Graveline. 1979. The seed bank of a freshwater tidal marsh. American Journal of Botany 66:1006-15. Leck, M.A. and C.F. Leck. 2005. Vascular plants of a Delaware River ti dal freshwater wetland and adjacent terrestrial areas : seed bank and vegetation comparisons of reference and constructed marshes and annotated species list. J. Torrey Botanic Society Journal 132:323-34. Leck, M.A. and W. Sch tz. 2005. Regeneration of Cyperaceae, with particular reference to seed ecology and seed banks. Perspectives in Plant Ecology, Evolution and Systematics 7:95133. Leck, M.A. and R.L. Simpson. 1987. Seed bank of a freshwater tidal wetland: turnover and relationship to vegetation change. Am erican Journal of Botany 74: 360-70. Lee, M.A. 1994. Seed Banks of Marsh and Re storation Sites. Upper St. Johns River Basin, Technical Memorandum No. 4. Department of Water Resources, St. Johns River Water Management District, Palatka, FL, USA. Metcalfe, C.R. 1971. Anatomy of the Monocotyl edons. V. Cyperaceae. Oxford University Press, London. Miller, S.J., M.A. Lee, and E.F. Lowe. 1998. The Upper St. Johns River Basin Project: merging flood control with aquatic ec osystem restoration and preservation. In K.G. Wadsworth (ed.): Transactions of the 63rd North American Wildlife and Natural Resources Conference, March 20-25 1998; Orlando, FL, USA. Moles, A.T., D.W. Hodson, and C.J. Webb. 2000. Seed size and shape a nd persistence in the soil in the New Zealand flora. Oikos 89:541-45. Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and Methods of Vegetation Ecology, John Wiley and Sons, New York, NY, USA. Nelmes, E. 1955. Notes on Cyperaceae, XXXVIII. Kew Bulletin 10:415-53. Nicholson, A. and P.a. Keddy. 1983. The depth prof ile of a shoreline seed bank in Matchedash Lake, Ontario. Canadian Journal of Botany 61:3293-3296. Nicol, J.M., G.G. Ganf, and G.A. Pelton. 2003. S eed banks of a southern Australian wetland: the influence of water regime on the fina l floristic composition. Plant Ecology 168:191205. Nijalingappa, B.H.M. 1986. Embryology of Scleria foliosa (Cyperaceae). Plant Systematics and Evolution 152:219-30.

PAGE 145

145 Pake, C.E. and D.L. Venable. 1996. Seed banks in desert annuals: implications for persistence and coexistence in variable environments. Ecology 77:1427-35. Parker, V.T. and M.A. Leck. 1985. Relationships of seed banks to plant distribution patterns in a freshwater tidal wetland. Amer ican Journal of Botany 72:161-174. Priestly,D.A. 1986. Seed Aging; Implications fo r Seed Storage and Persistence in the Soil. Cornell University Press, Ithaca, NY, USA. SAS Institute, Inc. 2002-2003. SAS 9.1, SAS Institute, Inc., Cary, NC, USA. Schtz, W. 2002. Ecology of seed dormancy and germination in sedges ( Carex ). Perspectives in Plant Ecology, Evolution and Systematics 3:67-89. Seabloom, E.W., K.A. Maloney and A.G. van der Valk. 2001. C onstraints on the establishment of plants along a fluctuating waterdepth gradient. Ec ology 82(8):2216-32. Simpson, R.L., M.A. Leck, and V.T. Parker. 1989. Seed banks: general concepts and methodological issues. In M.A. Leck, V.T. Parker, and R.L Simpson (eds.) Ecology of Soil Seed Banks. Academic Press, San Diego, CA, USA. Smith, L.M. and J.A. Kadlec. 1983. Seed banks and their role duri ng drawdown of a North American marsh. Journal of Applied Ecology 20:673-84. Smith, L.M. and J.A. Kadlec. 1985. The effects of disturbance on marsh seed banks. Canadian Journal of Botany 63:2133-37. Smith, S.M., P.V McCormick, J.A. Leeds, and P. B. Garrett. 2002. Constraints of seed bank species composition and water de pth for restoring vegetation in the Florida everglades, U.S.A. Society for Ecological Restoration 10:138-145. Steinman, A., K. Havens, and L. Hornung. 2002. The managed recession of Lake Okeechobee, Florida: integrating scien ce and natural resource management. Conservation Ecology 6:17. Toth, L.A. 1993. The ecological basis of the Kissim mee River restoration pl an. Florida Scientist 56:25-51. Thompson, K., J.P. Bakker, R.M. Bekker, and J. G. Hodgson. 1998. Ecological correlates of seed persistence in soil in the NW Eur opean flora. Journal of Ecology 86:163-9. Thompson, K. and J.P. Grime. 1979. Seasonal vari ation in the seed banks of herbaceous species in ten contrasting habitats. Journal of Ecology 67:893-921. van der Valk, A.G. 1981. Succession in wetla nds: a Gleasonian approach. Ecology 62:688-96.

PAGE 146

146 van der Valk, A.G. and C.B. Davis. 1976. The seed banks of prairie glacial marshes. Canadian Journal of Botany 54:1832-38. van der Valk, A.G. and C.B. Davis. 1978. The ro le of seed bank in the vegetation dynamics of prairie glacial marshes. Ecology 59:322-35. van der Valk, A.G. and T.R. Rosburg. 1997. Seed bank composition along a phosphorus gradient in the northern Florid a everglades. Wetlands 17:228-36. Weiher, E., I.C. Wisheu, P.A. Keddy, and D.R.J. Moore. 1996. Establishment, persistence and management implications of experimental wetland plant communities. Wetlands 16:20819. Wetzel, P.R., A.G. van der Valk, and L.A. Tot h. 2001. Restoration of wetland vegetation on the Kissimmee River floodplain: pot ential role of seed banks. Wetlands 21:189-198. Wunderlin, R.P. and B.F. Hansen. 2003. Guide to the Vascular Plants of Florida. 2nd Ed. University Press of Florida, Gainesville, FL, USA. Zedler, P.H. 1990. Life histories of vernal pool vascular plants. In D.H. Ikeda and R.A. Schlising (eds.) Vernal pool pl ants-their habitat a nd biology. Studies from the Herbarium No. 8. California State Un iversity, Chico, CA, USA.

PAGE 147

147 BIOGRAPHICAL SKETCH I received a B.S. with honors in Plant Science at the University of Delaware in 1984. My professors W. Mitchell and H. Pill were tw o of many exemplary mentors who offered unique opportunities in undergraduate resear ch and encouraged my passion for the plant sciences. With Peace Corps, Honduras I was stationed with an AID sponsored vegetable coop and developed field trials on sustainable production in cucumber for disease and insect management. On return home I worked on maize breeding with commercial seed companies and later with the University of Delaware where I helped devel op applications of plan t tissue culture for disease resistance. I was offered an assistantship in Plant Pathology at the Univers ity of Florida and completed a there M.S. with a thesis on antigenic and mol ecular characterization of cucumoviruses under W. Zettler in 1989. On completing my degree I enjo yed working with molecular transformation and tissue culture for viral resistance in citrus in Horticultural Science the University of Florida. In the ensuing years I worked with serological applications in immunology until joining the U.S. Geological Survey in 1996. At the Geological Survey I managed a national database for tracking the distribution of nonnative aquatic plants, whic h employed GIS, plant taxonomy and internet media. In 2002 our Center Direct or R. Hall offered me the opportuni ty to enter a student career program to pursue a doctoral degree while em ployed with the Geological Survey. This dissertation is a produc t of that program.