<%BANNER%>

Identifying Suitable Areas for the Reestablishment of Pinus Elliottii Var. Densa on Previously Farmed Lands in the Hole-...

Permanent Link: http://ufdc.ufl.edu/UFE0024293/00001

Material Information

Title: Identifying Suitable Areas for the Reestablishment of Pinus Elliottii Var. Densa on Previously Farmed Lands in the Hole-in-the-Donut Restoration, Everglades National Park
Physical Description: 1 online resource (154 p.)
Language: english
Creator: Serra, Lauren
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: donut, everglades, mesocosm, pine, planting, reestablishment, restoration, soil
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Within Everglades National Park is the largest remaining tract of pine rockland in the United States, a subtropical vegetation community that has been classified as globally imperiled community. Aerial interpretation of 1940?s photos revealed that approximately 15% of the Hole-in-the-Donut, or HID, was comprised of pine rockland before the lands were rockplowed and farmed. Agricultural practices in the HID altered soil properties such that 2,670 ha of fallow fields became invaded by woody vegetation, primarily the non-native Schinus terebinthifolius. Following removal of the disturbed soil to limestone bedrock, Pinus elliottii var. densa (South Florida slash pine) regenerated adjacent to undisturbed pine rockland; however, reestablishment was limited by distance from seed source and elevation via hydrology. Therefore, the central hypothesis of this project was that South Florida slash pine germination and short-term survival was attributable to surface elevation, hydroperiod, and microsite soil differences; with hydroperiod being the primary controlling site characteristic in the Hole-in-the-Donut. In order to promote slash pine occurrence in areas far removed from seed source the objectives of this research were to 1) characterize the site with historical aerial photographic interpretation (1940?s) and current soil analyses, 2) test the effects of hydroperiod on South Florida slash pine using a mesocosm experiment modeled on field observations in order to determine suitable Pinus elliottii var. densa areas post-scraping for Schinus, and 3) monitor the growth and survival of planted and seeded Pinus elliottii var. densa in the HID by elevation treatment. Indirect relationships were discovered for soils, as depth, total carbon, and total nitrogen increased as elevation decreased, and pH increased with an increase in elevation. There were observable differences in hydroperiod over the 2007 and 2008 rainy seasons; however, the lowest elevation supported a hydroperiod typical of marl prairie wetland that was unsuitable for South Florida slash pine. A mesocosm that simulated the high, low, and midpoint of HID elevation ranges and hydroperiods for South Florida slash pine reestablishment revealed that partially flooded pines had higher growth and biomass after one growing season. Nitrogen and C:N tissue levels were similar to Pinus elliottii var. densa tissue samples in the Everglades, and this research demonstrated that this subspecies has apparently adapted to survive with lower phosphorus levels than Pinus elliottii var. elliottii. Suitable areas for slash pine in HID were identified as areas that were inundated for five weeks or less during the rainy season, while marginal areas were flooded for a total duration of five to ten weeks. Pines planted in the HID had significantly greater height and root collar diameter growth when compared to seeded pines of the same age. It appeared that the survival of South Florida slash pine across the higher elevation ranges was similar to the historical occurrence of Pinus elliottii var. densa that existed in the Hole-in-the-Donut prior to farming, despite lower elevations associated with the restoration process. Overall, site characterization in conjunction with pine growth and survival under flooded conditions will guide the restoration of South Florida slash pine in the Hole-in-the-Donut and throughout South 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.
Statement of Responsibility: by Lauren Serra.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Comerford, Nicholas B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024293:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024293/00001

Material Information

Title: Identifying Suitable Areas for the Reestablishment of Pinus Elliottii Var. Densa on Previously Farmed Lands in the Hole-in-the-Donut Restoration, Everglades National Park
Physical Description: 1 online resource (154 p.)
Language: english
Creator: Serra, Lauren
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: donut, everglades, mesocosm, pine, planting, reestablishment, restoration, soil
Soil and Water Science -- Dissertations, Academic -- UF
Genre: Soil and Water Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Within Everglades National Park is the largest remaining tract of pine rockland in the United States, a subtropical vegetation community that has been classified as globally imperiled community. Aerial interpretation of 1940?s photos revealed that approximately 15% of the Hole-in-the-Donut, or HID, was comprised of pine rockland before the lands were rockplowed and farmed. Agricultural practices in the HID altered soil properties such that 2,670 ha of fallow fields became invaded by woody vegetation, primarily the non-native Schinus terebinthifolius. Following removal of the disturbed soil to limestone bedrock, Pinus elliottii var. densa (South Florida slash pine) regenerated adjacent to undisturbed pine rockland; however, reestablishment was limited by distance from seed source and elevation via hydrology. Therefore, the central hypothesis of this project was that South Florida slash pine germination and short-term survival was attributable to surface elevation, hydroperiod, and microsite soil differences; with hydroperiod being the primary controlling site characteristic in the Hole-in-the-Donut. In order to promote slash pine occurrence in areas far removed from seed source the objectives of this research were to 1) characterize the site with historical aerial photographic interpretation (1940?s) and current soil analyses, 2) test the effects of hydroperiod on South Florida slash pine using a mesocosm experiment modeled on field observations in order to determine suitable Pinus elliottii var. densa areas post-scraping for Schinus, and 3) monitor the growth and survival of planted and seeded Pinus elliottii var. densa in the HID by elevation treatment. Indirect relationships were discovered for soils, as depth, total carbon, and total nitrogen increased as elevation decreased, and pH increased with an increase in elevation. There were observable differences in hydroperiod over the 2007 and 2008 rainy seasons; however, the lowest elevation supported a hydroperiod typical of marl prairie wetland that was unsuitable for South Florida slash pine. A mesocosm that simulated the high, low, and midpoint of HID elevation ranges and hydroperiods for South Florida slash pine reestablishment revealed that partially flooded pines had higher growth and biomass after one growing season. Nitrogen and C:N tissue levels were similar to Pinus elliottii var. densa tissue samples in the Everglades, and this research demonstrated that this subspecies has apparently adapted to survive with lower phosphorus levels than Pinus elliottii var. elliottii. Suitable areas for slash pine in HID were identified as areas that were inundated for five weeks or less during the rainy season, while marginal areas were flooded for a total duration of five to ten weeks. Pines planted in the HID had significantly greater height and root collar diameter growth when compared to seeded pines of the same age. It appeared that the survival of South Florida slash pine across the higher elevation ranges was similar to the historical occurrence of Pinus elliottii var. densa that existed in the Hole-in-the-Donut prior to farming, despite lower elevations associated with the restoration process. Overall, site characterization in conjunction with pine growth and survival under flooded conditions will guide the restoration of South Florida slash pine in the Hole-in-the-Donut and throughout South 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.
Statement of Responsibility: by Lauren Serra.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Comerford, Nicholas B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024293:00001


This item has the following downloads:


Full Text

PAGE 1

IDENTIFYING SUITABLE AREAS FOR THE REESTABLISHMENT OF PINUS ELLIOTTII VAR. DENSA ON PREVIOUSLY FARMED LANDS IN THE HOLE-IN-THE-DONUT RESTORATION, EVERGLAD ES NATIONAL PARK By LAUREN ANN SERRA 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 2009 1

PAGE 2

2009 Lauren A. Serra 2

PAGE 3

To all those who supported and insp ired me these past five years 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank my major advisor, Nic holas Comerford, and committee members Eric Jokela, Yuncong Li, Craig Smith and James Snyd er, for your knowledge and support. I would also like to thank my HID coworkers and interns for assisting me with data collection, and the laboratory staff at TREC and UF Aja, Gui-Qin, Laura, and Da isy; you made sure I used the equipment correctly and walked aw ay with pragmatic analyses. I thank Steve Gilly at DOF for the germination advice, Nancy OHare for opening my eyes to this research, Joy Klein for caring so much about the pine rocklands, Guodong Liu for sharing your mesocosm ideas, and Jonathan Taylor for supporting my continuance on this jo urney. The dedication of time and resources from Everglades Research and Fire made this pr oject possible. I thank Mom, Dane, Nor, Fred, and Windy, for encouragement and balance. Mike Norland, this is for you. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................8LIST OF FIGURES .......................................................................................................................10ABSTRACT ...................................................................................................................... .............13CHAPTER 1 INTRODUCTION TO THE DISSERTATION .....................................................................15History of the Hole-in-the-Donut ...........................................................................................15The Restoration Project ..........................................................................................................18South Florida Slash Pine Reestablishment .............................................................................20Research Goals and Objectives ..............................................................................................23Research Experiments .......................................................................................................... ..24Aerial Photographic Interpretation and Site Characterization (Chapter 2) .....................24Hydroperiod for South Florida Slash Pine Sa pling Survival and Growth (Chapter 3) ...25Reestablishing and Monitoring of Pinus elliottii var. densa (Chapter 4) ........................26Implications for South Florida Slas h Pine in the Hole-in-the-Donut .....................................272 BEFORE AND AFTER RESTORATI ON: USING HISTORICAL AERIAL PHOTOGRAPHY AND CURRENT SITE CH ARACTERIZATION TO GUIDE THE REESTABLISHMENT OF PINUS ELLIO TTII VAR. DENSA IN THE HOLE-INTHE-DONUT ..................................................................................................................... ....32Introduction .................................................................................................................. ...........32Methods ..................................................................................................................................37Aerial Interpretation of 1940s Photos ............................................................................37Research Plot Establishment ...........................................................................................37Soil Analyses ...................................................................................................................38Soil depth ..................................................................................................................39Soil pH ......................................................................................................................39Total phosphorus ......................................................................................................39Total carbon and total nitrogen ................................................................................40Statistical analyses ....................................................................................................41Hydrological Measurements ............................................................................................41Results .....................................................................................................................................41Discussion .................................................................................................................... ...........44Conclusion .................................................................................................................... ..........49 5

PAGE 6

3 THE EFFECTS OF FLOODING TREATMENT ON PINUS ELLIOTTII VAR. DENSA SAPLINGS USING FIELD-SIMUL ATED HYDROPERIOD DEPTHS AND RECOMMENDATIONS FOR RESTORATION ..................................................................58Introduction .................................................................................................................. ...........58Materials and Methods ...........................................................................................................62Study Site .........................................................................................................................62Mesocosms ..................................................................................................................... .63In-Situ Measurements and Growth ..................................................................................64Biomass ....................................................................................................................... ....64Tissue Analyses ...............................................................................................................65Chlorophyll a and b ..................................................................................................65Total-nitrogen, total-carbon, total-phosphorus, and iron .........................................66Hydroperiod and Pine Suitability Maps ..........................................................................67Statistics .................................................................................................................... .......67Results .....................................................................................................................................68Tree Growth and Survival ...............................................................................................68Biomass Measurements and Calculations .......................................................................69Chemical Analyses ..........................................................................................................70Discussion .................................................................................................................... ...........71Conclusion .................................................................................................................... ..........794 THE REESTABLISHMENT AND MONI TORING OF SOUTH FLORIDA SLASH PINE SEEDLINGS IN A RESTORED AGRICULTURAL AREA WITHIN EVERGLADES NATIONAL PARK .....................................................................................97Introduction .................................................................................................................. ...........97Materials & Methods ........................................................................................................... .102Seed Collection, Harvest, and Utilization .....................................................................102Viability Study ...............................................................................................................103Study Site .......................................................................................................................103Direct Seeding ...............................................................................................................104Seed scatter one ......................................................................................................104Seed scatter two ......................................................................................................105Comparing seeding events .....................................................................................105Pine Planting ..................................................................................................................105Water Levels ..................................................................................................................106Statistics .................................................................................................................... .....106Results ...................................................................................................................................108Greenhouse Viability Study ..........................................................................................108Germination in the Field ................................................................................................108Survival ...................................................................................................................... ....109Direct Seeding Comparisons .........................................................................................109Planting verses Direct Seeding ......................................................................................110Planting Covariate Effects on Survival and Growth .....................................................111Hydrology and Survival ................................................................................................111 6

PAGE 7

Discussion .................................................................................................................... .........112Conclusion .................................................................................................................... ........1185 SUMMARY ..................................................................................................................... .....132Reestablishment of Sout h Florida Slash Pine .......................................................................132Research Summary ........................................................................................................132Recommendations .........................................................................................................135Reflections and Refinements ................................................................................................136Future Considerations ...........................................................................................................139Research Investigations .................................................................................................139Pine Rockland Restoration ............................................................................................140APPENDIX: LOCATION OF RE SEARCH PLOTS IN HID .....................................................145Soil Sampling Coordinates ...................................................................................................145South Florida Slash Pine Research Plot Coordinates ...........................................................146LIST OF REFERENCES .............................................................................................................147BIOGRAPHICAL SKETCH .......................................................................................................154 7

PAGE 8

LIST OF TABLES Table page 2-1 Hole-in-the-Donut land cover in 1940, as interpreted using historic aerial photography. ......................................................................................................................502-2 True elevation ranges for pine plots within the HID re storation area, as eight-cm increments every ten centimeters, across five elevation ranges. .......................................502-3 Percent of exposed limestone bedrock and average soil depth, with and without limestone rock measurements, as associated with each elevation range. ..........................502-4 C:N and N:P ratios according to elevation. Means without letters or followed by the same letters are not significant for alpha at 0.05. ..............................................................503-1 Root collar diameter, he ight, and growth measuremen ts by treatment by block for pine seedlings before and after flooding treatments ..........................................................803-2 Covariate effects of initial root collar diameter or height on dependent variables of final root collar diameter or height measurements, with flooding as a fixed factor. .........803-3 Dry weight biomass for Pinus elliottii var. densa root, stem, and needles by flooding treatment. .................................................................................................................... .......813-4 Covariate effects of initial heights of S outh Florida slash pine saplings on dependent dry biomass variables, with flooding as a fixed factor. .....................................................813-5 Covariate effects of initial root collar di ameters of South Florida slash pine saplings on dependent dry biomass variables, with a fi xed factor of flooding as a fixed factor .....823-6 Carbon, nitrogen, and C:N in Pinus elliottii var. densa tissue following 20 weeks of flooding treatment. ........................................................................................................... ..823-7 Chlorophyll A and B concentrations (mg/g), and chlorophyll A:B after flooding treatment for Pinus elliottii var. densa tissue.....................................................................823-8 Hydroperiod data for Station DO-1 in the Hole-in-the-Donut, Everglades National Park. ......................................................................................................................... ..........834-1 Germination of South Florida slash pine during the initial greenhouse viability study conducted in January 2007. ..............................................................................................1194-2 Number of germinants and germinative capacity between elevations and among direct seeding events. .......................................................................................................1194-3 Two-way ANOVA of germinative capacity for direct seed scatters 1 and 2 by elevation treatment, with percent germination as the dependent variable.. .....................119 8

PAGE 9

4-4 Survival data of South Florida slash pine for direct seed ing scatters one and two at 12 months, as well as scatter one and pl anted pines at 21 months of age. ...........................1204-5 Kruskal-Wallis test results for survival rates of seeded and planted South Florida slash pines following exposure to one season of inundation. ..........................................1204-6 Two-way ANOVA of height and root collar at 12 months, with elevation and timing of seed scatter as fixed factors, and height or root collar as the dependent variable .......1214-7 Two-way ANOVA for height and root collar diameter of scattered and planted South Florida slash pines at 18 months, as well as growth over 9 months.. ..............................1214-8 Covariate effects of initial height or r oot collar diameter by elevation treatment fixed factors, with survival as the dependent variable ..............................................................1224-9 Covariate effects of initial height or r oot collar diameter by elevation treatment fixed factors. Dependent variables were fina l height or root collar and growth ......................1225-1 Recommendations for the reestablishment of South Florida slash pine on rocky substrate. ..........................................................................................................................143A-1 GPS coordinates, in UTMs, for the firs t set of research plots where soil samples were collected for analyses ..............................................................................................145A-2 GPS coordinates, in UTMs, for all Sout h Florida slash pine re search plots in the Hole-in-the-Donut where pine was seeded and planted ..................................................146 9

PAGE 10

LIST OF FIGURES Figure page 1-1 The Hole-in-the-Donut restor ation project in Everglades National Park. Located in South Florida, portions of the Hole-in-the-Donut were farmed until 1975. ......................281-2 The Hole-in-the-Donut as visible from space. Note the restoration area in the center of the image, while the bright green areas remain to be restored ......................................291-3 Hole-in-the-Donut Restoration site s by year, described as ResYr. ...................................301-4 Post-restoration elev ation data for the Hole-in-the-Donut. Elevation grids given in meters and derived from photogra mmetric data by 0.1 m range. ......................................301-5 Location of the Pinus elliottii var. densa study site in the 2001 scraped area of the Hole-in-the-Donut. ............................................................................................................ .312-1 Aerial interpretation of the Hole-inthe-Donut by vegetation cover circa 1940. Approximately 15% of the Hole-in-the-Donut was dominated by pine rockland. ............512-2 Plot locations, for the first set of research plots. Note the majority of research plots in the highest four elevations were located in pine rockland during 1940. ...........................522-3 Soil depth as a function of elevation ra nge, excluding exposed bedrock. Logarithmic means were back-transformed to arithmetic units. ............................................................532-4 Soil pH verses elevation range for the firs t set of pine research plots. Note that periphyton pH was analyzed at the 63 to 71 cm elevation. ...............................................542-5 Distribution of total-phosphorus in soils as related to elevation ranges for pine research plots.. ...................................................................................................................552-6 Total-carbon and total-nitrog en concentrations as a func tion of elevation, with error bars at one standard error above the mean and an alpha = 0.05. .......................................562-7 Comparison of 2007 and 2008 rainy season wate r levels in the 63 to 71 cm elevation, using data from the first set of research plots where pine was seeded and planted. ..........573-1 Rainy season average weekly water le vels for 2007 by elevation range, for plots seeded with Pinus elliottii var. densa .. ..............................................................................843-2 Total daily rainfall at the IFS hydrolog ical station rain ga uge for 20 weeks of mesocosm flooding treatment dur ing the 2008 rainy season. ............................................853-3 Initial and final height measurements of Pinus elliottii var. densa saplings before and after flooding treatment.. .................................................................................................... 86 10

PAGE 11

3-4 Initial root collar diameter and final r oot collar measurements by flooding treatment for Pinus elliottii var. densa saplings. ................................................................................873-5 Height growth and root collar diameter growth after 20 weeks of flooding treatment for Pinus elliottii var. densa saplings .. ...............................................................................883-6 Repeated measures over 20 weeks for pe rcent dead needles by flooding treatment for Pinus elliottii var. densa saplings. Logarithmic data were back-transformed. .................893-7 Bi-weekly repeated measures of chloro sis estimates for various flooding treatments in a mesocosm study of Pinus elliottii var. densa saplings.. ..............................................903-8 Root:shoot ratios of dried Pinus elliottii var. densa saplings according to flooding treatment. .................................................................................................................... .......913-9 Logarithmic root mass as a function of logarithmic shoot mass (in grams) by flooding treatment with logarithmic regression values. .....................................................923-10 Total tissue phosphorus of South Florida slash pine needles by mesocosm flooding treatment. Means were back-transformed from logarithmic to arithmetic units. .............933-11 Iron tissue measurements ve rses flooding treatment for Pinus elliottii var. densa Means are shown as back-transformed logarithmic to arithmetic values ..........................943-12 Hydroperiod map for the Hole-in-the-D onut, based upon number of weeks flooded in one year. Data were averaged for 20 07 and 2008 flooding duration by elevation. ..........953-13 Pine suitability map for the Hole-inthe-Donut, as based upon number of weeks flooded and elevation data. Areas are identifi ed as suitable, marginal, or unsuitable. .....964-1 Locations of the two sets of 20 research plots by elevation range in the Hole-in-theDonut study area. There were four plots es tablished across five el evation treatments. .1234-2 An example of the x-y coordinate grid used to map pine seedlings in all study plots. This figure shows the direct seeding results after three months in a 73-81 cm plot ........1244-3 Cumulative number of germinated South Fl orida slash pine seedlings and survival in the greenhouse over nine months.. ...................................................................................1244-4 Survival of pine plantings by elevat ion following one year in the ground. In December 2008, pines were approximately 21 months old. ............................................1254-5 A comparison of height betw een elevations and direct seed scatters 1 and 2 at twelve months after seeding of Pinus elliottii var. densa .. ..........................................................1264-6 A comparison of root collar diameter betw een elevations and direct seed scatters 1 and 2 twelve months after seeding of Pinus elliottii var. densa ......................................127 11

PAGE 12

4-7 Pinus elliottii var. densa height and root collar diam eter for scattered and planted pines by elevation at 18 months of age. ...........................................................................1284-8 Growth rates of planted and seeded S outh Florida slash pines by elevation following nine months. Height growth and root collar diameter growth are depicted. ...................1294-9 A comparison of water depth and total live seedlings in the 63 to 71 cm elevation for the first direct seeding of Pinus elliottii var. densa during the 2007 rainy season. .........1304-10 A comparison of water depth and total live seedlings in the 63 to 71 cm elevation for the second direct seeding of Pinus elliottii var. densa during the 2008 rainy season. .....1304-11 A comparison of water depth and total liv e seedlings in the 63 to 71 cm elevation range for planted Pinus elliottii var. densa seedlings during the 2008 rainy season. ......1315-1 A comparison of current suitable and marg inal habitat for South Florida slash pine with historical pine rockland occu rrence in the Hole-in-the-Donut. ...............................144 12

PAGE 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 IDENTIFYING SUITABLE AREAS FOR THE REESTABLISHMENT OF PINUS ELLIOTTII VAR. DENSA ON PREVIOUSLY FARMED LANDS IN THE HOLE-IN-THE-DONUT RESTORATION, EVERGL ADES NATIONAL PARK By Lauren A. Serra August 2009 Chair: Nicholas B. Comerford Major: Soil and Water Science Within Everglades National Park is the larg est remaining tract of pine rockland in the United States, a subtropical vegetation community that has been classified as globally imperiled community. Aerial interpretati on of 1940s photos revealed that approximately 15% of the Holein-the-Donut, or HID, was comprised of pine ro ckland before the lands were rockplowed and farmed. Agricultural practices in the HID altered so il properties such that 2,670 ha of fallow fields became invaded by woody vegetation, primarily the non-native Schinus terebinthifolius Following removal of the disturbe d soil to limestone bedrock, Pinus elliottii var. densa (South Florida slash pine) regenerated adjacent to undisturbed pine rockland; however, reestablishment was limited by distance from seed source and elev ation via hydrology. Therefore, the central hypothesis of this project was that South Florida slash pine germination and short-term survival was attributable to surface elevation, hydroperiod, and microsite soil differences; with hydroperiod being the primary controlling site charac teristic in the Hole-in-the-Donut. In order to promote slash pine occurrence in areas far re moved from seed source the objectives of this research were to 1) characterize the site with historical aerial photographic interpretation (1940s) and current soil analyses 2) test the effects of hydrope riod on South Florida slash pine 13

PAGE 14

14 using a mesocosm experiment modeled on field ob servations in order to determine suitable Pinus elliottii var. densa areas post-scraping for Schinus, and 3) monitor the growth and survival of planted and seeded Pinus elliottii var. densa in the HID by elevation treatment. Indirect relationships were di scovered for soils, as depth, total carbon, and to tal nitrogen increased as elevation decreased, and pH increased with an increase in elevation. There were observable differences in hydroperiod over the 2007 and 2008 rainy seasons; however, the lowest elevation supported a hydrop eriod typical of marl prairie wetland that was unsuitable for South Florida slash pine. A mesocosm that simulated the high, low, and midpoint of HID elevation ranges and hydroperiods for South Florid a slash pine reestablishment revealed that partially flooded pines had higher growth and bi omass after one growing season. Nitrogen and C:N tissue levels were similar to Pinus elliottii var. densa tissue samples in the Everglades, and this research demonstrated that this subspecies has apparently adapted to survive with lower phosphorus levels than Pinus elliottii var. elliottii Suitable areas for slash pine in HID were identified as areas that were inundated for five weeks or less during the rainy season, while marginal areas were flooded for a total duration of five to ten weeks. Pines planted in the HID had significantly greater height a nd root collar diameter growth when compared to seeded pines of the same age. It appeared that the surviv al of South Florida slas h pine across the higher elevation ranges was similar to the historical occurrence of Pinus elliottii var. densa that existed in the Hole-in-the-Donut prior to farming, despite lower elevations associated with the restoration process. Overall, site characterizat ion in conjunction with pine growth and survival under flooded conditions will guide the restoration of South Florida slash pine in the Hole-in-theDonut and throughout South Florida.

PAGE 15

CHAPTER 1 INTRODUCTION TO THE DISSERTATION History of the Hole-in-the-Donut Among the wilderness of Everglades National Park is a 2,670 hectare wetland restoration project known as the Hole-in-the-Donut (HID), where the National Park Service is currently restoring previously farmed land by removing i nvasive woody vegetation and disturbed soils (Figure 1-1). Encompassing about 8,900 ha, farmla nd owners insisted in 1947, when Everglades National Park was established, that the entire Ho le-in-the-Donut footprint be excluded from the park boundary (Cornwell and Atkins 1975). Fa rming began during late 1916, and though only 3,640 ha were actually farmed, portions of the HID remained as agricultural areas until 1975. Old Ingraham Highway reached Royal Palm Hammock in Paradise Key during 1915 and Flamingo in 1922, which allowed farmers to more extensively utilize the Hole-in-the-Donut (Cornwell and Atkins 1975). The Florida Federa tion of Womens Clubs, in part responsible for enacting Royal Palm State Park in 1916, rented the Hole -in-the-Donut lands to tomato farmers. By the 1930s, areas along the Old Ingraham Highw ay and the western edge of the Hole-in-theDonut had been farmed (Loope and Dunevitz 1981). Aerial photography from the 1940s revealed irregular farms confined to the deeper marl fingerglades that ran between the higher elevated pine rocklands, as well larger farmed areas in the western HID and the southeastern HID, located in proximity to S outh Taylor Slough (Krauss 1987). With the invention of the rock-plow in the 1950s, farmers were able to expand operations into larger and more regularly shaped agricultura l areas, like those of conventional farms. In order to farm these areas, the rockplow broke up the underlying limestone bedrock and intermixed it with the shallow soils of marl prairies and pine rocklands, thereby artificially creating a more workable soil depth (Orth and Conover 1975). With these newly farmed areas 15

PAGE 16

was the abandonment of deeper marl areas wi th longer hydroperiods, in favor of shorter hydroperiods and more successful crop production. Over the sixty years that these lands were in and out of production, tomatoes, squash, cucumbers, guava, bananas, and even oranges were grown in the Hole-in-the-Donut. Farming ceased by 1975, once all the lands had been purchased throughout the 1970s as a series of acquisitions by Everglades National Park. During the 1973 to 1974 growing season, tomatoes had a total production value of $6,472,500 for 698 ha, while squash was valued at $1,210,300 for 283 ha (Cornwell and Atkins 1975). Given that fields were rotated or abandoned at various stages throughout HIDs farming hi story, agricultural fields were also either rockplowed or nonrockplowed. As the lands were incorporat ed into the park from 1973 to 1975, Loope and Dunevitz (1981) estimated that 1700 ha of ro ckplowed land was left fallow. Rockplowing history, length of time the area was farmed, and fert ilizer or nutrient additions may have played a role in the vegetation communities that developed in the abandoned farmlands. This research is currently under investigation. The loss of pine rockland over time was due in part to the invention of the rock plow during the 1950s. Though the rockplow provided fa rmers with more soil, it also created soils that were aerated, higher, drier, and more nutrient-enriched than the soils that were naturally found in the poorly aerated, oligotrophic Evergl ades ecosystem (Ewel et al. 1982, Dalrymple et al. 1993). There was a higher presence of my corrhizae in the farmed, shorter hydroperiodcreated soils, an indication that anaerobic soils had become aerobic, given a lack of mycorrhizal presence under flooded conditions (Mikola 1967, Meador 1977). Meador (1977) also described higher bulk densities and soil vol umes on farmed sites due to compaction and disking by farm equipment, as well as increased bulk density related to the additi on of limestone into rockplowed 16

PAGE 17

sites. Increased calcium carbona te concentrations reacted with the phosphorus fertilizers, providing an unnaturally high phosph orus supply available for plant growth (Orth and Conover 1975). Over time, much of the farmed HID was both rockplowed and fertilized, such that when the farm fields were abandoned in th e 1970s or earlier, Brazilian pepper ( Schinus terebinthifolius) readily invaded many of the disturbed, fallow lands (Meador 1977, Dalrymple et al. 2003). Krauss (1987) summarized old field succession in the Hole-in-the-Donut with study plots that focused on a variety of hydroperiods, substr ates (or soils), age (since abandonment), and vegetation compositions. A study plot that was formerly marl prairie, rockplowed and farmed from 1960 and 1973 was covered with a canopy of 17.5% Baccharis halimifolia five years after farming ceased. Another plot that was formerly pine rockland, rockplowed, and farmed after 1952 until 1965, contained a canopy dominated by Schinus terebinthifolius over twenty years following abandonment. There was also a research plot located in the southeastern HID that was only farmed from the late 1930s until 1940, ye t was composed of a well-developed canopy with 80% Schinus terebinthifolius cover and an understo ry dominated with 95% Ardisia elliptica on the ridges between the furrows nearly 40 years la ter. Lastly, a study plot that was located in the western HID, farmed from 1940 to 1955, and non-rockplowed, resembled the surrounding mesic prairie with the presence of Cladium jamaicense, Muhlenbergia capillaris and Rhynchospora microcarpa. Prior to farming, the HID consisted of shor t hydroperiod graminoid wetlands and mesic to hydric pine rocklands (USFWS 1998a, USFWS 1 998b, Dalrymple et al. 2003). The subsequent invasion of these farmed lands with Schinus terebinthifolius led Everglades National Park toward early efforts to deal with Brazilian pepper a nd restore the native vegetation communities that 17

PAGE 18

occurred in HID prior to farming. Attempts included the planting and seeding of pine and hardwoods, transplanting native grasses and sedges, mowing, di sking, bulldozing, controlled burning, herbicides, and partial substrate removal. The Reforestation Proj ect failed to restore Pinus elliottii var. densa in the Hole-in-theDonut during the early 1970s because those pines that survived rainy season flooding were eventually outcompeted by invasive woody vege tation on the disturbed soils (Bancroft 1973). Fire was unsuccessful at limiting the reinvasion of Schinus terebinthifolius on a site that was bulldozed and burned (Krauss 1987). Although triclopyr and he xazinone were effective at killing Schinus terebinthifolius this treatment caused native ve getation to suffer, it was costly, and did not stop the reinvasion of Brazilian pepper over time (Ewe l et al. 1982). In 1989, a pilot study to restore marl prairie found that partial substrate removal had a higher canopy cover of Schinus terebinthifolius and lacked the diversity of herbaceo us species that were found under complete substrate removal (Dalrymple et al. 2003) Therefore, it was determined that complete substrate removal of rockplowed soils inhibited the reinvasion of Brazilian pepper, and yet, facilitated the restoration of a wetland ecosystem dominated by native plants. The Restoration Project Mechanically clearing the non-native Brazilian pepper and scraping away the disturbed soils to limestone bedrock restored conditi ons conducive to colonization and dominance by native wetland vegetation (Figure 1-2). This improved Everglades ecosystem structure and function which resulted in a massive mitigation effort to remove the monoculture of Schinus terebinthifolius and scrape the soil to bedrock. The lim estone rock and minimal soil layer left behind created an unfavorable environment for Schinus terebinthifolius and thereby allowed the reestablishment of native plants (Dalrymple et al. 2003). The fi rst restoration effort was the 18

PAGE 19

1989 test site, after which larger scale land clear ing began in 1997 and continued almost annually through 2005 (Figure 1-3). In order to restore lands invaded with Schinus terebinthifolius large-scale land clearing efforts were necessary (personal observation). First, a hydroax cut through the Brazilian pepper forest and the plant material was further mulche d into smaller pieces. Bulldozers then piled the vegetation, along with the disturbed soil into windrows. Front-end loaders moved this material to dumptrucks, which hauled the debris to an on-site soil disposal mound, approximately 30 acres in size. Lastly, a final scrape of the land was required, such that graters and sweepers removed any excess soil from the limestone bedrock. This resulted in the exposure of oolitic limestone bedrock and soil pockets that promoted native plant colonization. Following restoration, communitie s of short to medium hydroperiod prairie developed within the first year (OHare and Dalrymple 2005). It was known that restor ed areas in the Holein-the-Donut that were historically marl pr airie were functioning as a wetland, and though slightly different in marl prairi e plant structure, the site becam e habitat for wading bird species, as well as bear and the Florida panther. While the dominant plant composition was not statistically similar to the surr ounding native marl prairie, the ma jority of plant species found in restored sites were natives and layers of periphyton had developed in restored areas (OHare and Dalrymple 2005). In addition, this area was also reclaimed by marl prairie wetland species as a result of the surrounding native we tland vegetation and hydrol ogical sheet flow of restored sites. Restoration of pine rockland was much slower than marl prairie, and posed more of a restoration challenge. The Hole-in-the-Donut restoration did not recr eate the microtopography that increases pine rockland plant diversity, there was limited recruitment of Pinus elliottii var. densa (South Florida slash pine) and there was a lack of character istic species like saw palmetto 19

PAGE 20

( Serenoa repens) In fact, the pine rockland community is typically dominated by an open canopy of Pinus elliottii var. densa with a diverse understory of West Indian woody species such as Tetrazygia bicolor and Myrsine guianensis as well as herbaceous species endemic to South Florida (Snyder et al. 1990). Ma ny of these endemics are prot ected species, and the pine rockland community provides habitat for endangered species like the Florida panther. It is important to protect pine rocklands because of the original 74,000 ha of pine rockland habitat that once existed, the largest remaining tract of pine rockland is only 8,000 ha and found within the Everglades National Park. Currently less than two percent of Miami Rock Ridge pine rockland remains outside of the park, which has rendered the subtropical species associations that comprise the unique pine rockland ecosystem to be listed as a gl obally imperiled habitat (Snyder et al. 1990, OHare and Dalrymple 2006). South Florida Slash Pine Reestablishment Since Pinus elliottii var. densa is a canopy dominant of the pine rockland ecosystem, it was the first pine rockland species to be anthr opogenically reintroduced to the HID following complete substrate removal. In this research, Pinus elliottii var. densa was distinguished as a different subspecies than Pinus elliottii var. elliottii found to the north. There are differences in the autecological characteristics of both slash pine varieties, and there is debate whether these differences should be recognized at a taxonomic level. Little and Dorman (1952) described Pinus elliottii var. elliottii with needle fascicles of two and three, and typical stemmed seedlings, while Pinus elliottii var. densa predominantly had needle fascicles of two, and seedlings that exhibited a grass stage and thicke r tap root. Additiona lly, the wood of South Florida slash pine was heavier and had a noticeably th icker summerwood. The cones of Pinus elliottii var. densa were about 20% smaller than Pinus elliottii var. elliottii, with average survival of large seed plantings from these cones at 75.3% one year following planting (Langdon 1958). 20

PAGE 21

Squillace (1966) found clinal variation exis ted throughout the slash pine range, with genetic traits such as germinability, total he ight, and thickness of hypoderm showing a northsouth trend in variation. Today, Pinus elliottii var. elliottii is grown for pulp and saw-timber while South Florida slash pine is protected for th e niche it serves in the pine rockland ecosystem. Therefore, it is vital to preserve and restore Pinus elliottii var densa Once this dominant overstory tree is established, it is then possible to further deve lop a suite of understory plant species that are of the pine rockland community association with time. The reestablishment of South Florida slash pine was one of the revegetation concerns of Everglades National Park. Yet, little was know n about the autecology of South Florida slash pine and the conditions that promote its germin ation and survival on restored lands. South Florida slash pine in Miami Dade County and the Florida Keys grows in shallow soils atop oolitic limestone rock outcrops; however, the extent that South Fl orida slash pine can recolonize in restored areas like the HI D is unknown. This is because th e HID restoration leaves behind minimal soil in an area removed from seed s ource with altered hydrological conditions. The degree to which South Florida slas h pine can be reestablished in areas that were formerly pine rockland prior to farming and land clearing was the topic of this research. When an area of HID adjacent to pine ro ckland was cleared to limestone bedrock and restored in 1998, there was natural pine recruitmen t and survival the following year. Within four years of restoration there were 3,013 seedlings gr eater than ten cm in height on the 12 ha of restored land known as 98 North (OHare and Dalrymple 2006). It was determined that recruitment from seed dispersal was limited to a distance of 25 meters from the stand of mature South Florida slash pine trees Growth and survival of Pinus elliottii var. densa in the 98 North area of restored HID were also correlated with elevation; therefore, hydroperiod. Elevations 21

PAGE 22

above 107 cm had 50% or less mortality, while 81% mortality occurred at 105 cm and 100% seedling mortality occurred at 90 cm. This led to the idea that perhaps some areas of the HID could be restored to pineland, rather than prairie. Though current elevations were lower than pr e-farming elevations, there were restored areas where the ground surface elevation was high enough to promote a shorter hydroperiod and survival of slash pine. Due to drainage of the Everglades ecosystem by the Central and South Florida canal system, lower elevations may actua lly have the same hydrology today that existed prior to drainage. High accuracy elevation data (+/-0.1 m) existed for 1,538 ha of restored sites (Figure 1-4). The highest elevations had either natural recruitment of Pinus elliottii var. densa or elevations favorable for pine re storation. Favorable areas had the potential to support South Florida slash pine, yet were rem oved from pine seed source. Though South Florida slash pine had recolonized areas of the HID, there were unanswered questions that made this research innovative. The duration of flooding S outh Florida slash pine seedlings could withstand, and hence, how low of an elevation that the species could tolerate and survive within the HID was critical to determ ining where South Florida slash pine could be reintroduced. It was also not clear the degree to which scatte red and planted seedlings could survive the hydrological and edaphi c conditions of restor ed lands, or the extent to which the potential South Florida slash pine occurrence compared to the species former range in the HID prior to farming. Elevations suitable for slash pine existed in restored areas other than 98 North and there was even pine rockland understory plant recruitmen t in these vacant pine areas. However, there were less than twenty recorded seedlings on these si tes, probably due to dist ance from seed trees. Artificially seeding areas could accelerate the recruitm ent process in areas capable of supporting 22

PAGE 23

pine, which would otherwise take decades under na tural conditions. Therefore, all elevations capable of supporting pine seedlings within the Hole-in-the-Donut were delineated and tested for pine recruitment by directly seeding Pinus elliottii var. densa over a series of plots at different elevation on restored areas. Elevation was related to the local hydr operiod, such that a mesocosm study of flooding treatment on South Fl orida slash pine and pine planting field observations tested the effects of hydroperiod on the survival of South Florida slash pine in the Hole-in-the-Donut. This was used to aid in the development of a pine restoration plan. Research Goals and Objectives The long-term goal of the National Park Serv ice was to determine whether the restored areas in the Hole-in-the-Donut were capable of supporting pine rockla nd species. The overall objective of this project was to determine the c onditions that optimized South Florida slash pine germination and short-term survival in the scra ped area within the HID. The approach included the use of both a historical perspective of the lo cation of natural pine stands, as well as current investigations which tested South Florida slas h pine germination, surv ival, and growth of Pinus elliottii var. densa seedlings in the HID. The central hypot hesis of this proj ect was that South Florida slash pine germination and short-term su rvival were attributable to surface elevation, hydroperiod, and microsite soil differences; with hydroperiod being the primary controlling site characteristic. This concept flowed from observations that So uth Florida slash pine was naturally found at higher elevations where the hydroperiod was the s hortest, soil was present, and where fires had promoted germination and survival. Specifically, the objectives of this research were to (1) document the distribution of pine rockland befo re farming as based upon interpretation of 1940s aerial photography, as well as to char acterize the HID South Florida sl ash pine research site with current soil and hydrological data as a function of elevation, (2 ) simulate the effects of HID 23

PAGE 24

hydroperiod data on South Florida slash pine sapli ngs in a mesocosm study that tested the effects of flooding depth and duration on pi ne seedling growth and survival and (3) test the effects of hydroperiod on slash pine reestablishment in the field by direct seeding and planting viable seeds across five elevation ranges in HID areas rem oved from a pine seed source. Overall, the following three chapters identified suitable areas for South Florida slash pine within the HID. Research Experiments Aerial Photographic Interpretation a nd Site Characterization (Chapter 2) Interpretation of historical aerial photographs, current elev ation survey, soil analyses, and hydrology were used to evaluate a nd characterize potential areas for pine restoration in the Holein-the-Donut. A digital evaluation of the 1940 s aerial photographs was used to delineate the historical extent of pine rock land and other vegetation communities prior to farming. These data were combined with current elevation survey, soil analyses, and hydroperiod durations, in order to characterize the site a nd define suitable areas for Pinus elliottii var. densa reestablishment. The working hypothesis was that hydroperiod, as cl osely related to elevation, was the controlling factor for South Florida slash pine reestablishment. In order to test this hypothesis, 2 m2 research plots were estab lished across five elevation treatments on a 32 ha study site located in th e 2001 scraped area (Fig ure 1-5). Among these plots, soil physical and chemical properties were tested and analyzed as a function of elevation, and hence hydroperiod. Weekly water level pres ence and depths were also recorded at all inundated plots for the 2007 and 2008 rainy seasons. The edaphic and hydrologic characteristics of the site in relation to Sout h Florida slash pine seeding and planting results in the HID were used to further evaluate the HIDs current So uth Florida slash pine range. Additionally, hydroperiod data from the 2007 field data were app lied in a mesocosm study to specifically test 24

PAGE 25

South Florida slash pine growth and survival on saplings that were exposed to fully flooded, partially flooded, and non-flooded conditions. Hydroperiod for South Florida Slash Pine Sapling Survival and Growth (Chapter 3) Survival and growth of South Florida slas h pine under simulated HID flooded conditions served as an indicator for slash pine survival in the field. Since elevation was correlated with hydroperiod, a controlled mesocosm study was used to represent high, mid, and low elevations currently available for pine restoration in the HID, by testing the effects of flooding on Pinus elliottii var. densa. The working hypothesis was that hydr operiod was the controlling factor for South Florida slash pine survival and growt h, as changing the hydroperiod will alter pine survival and growth in the mesocosms. The impact of hydroperiod on growth and survival under the controlled conditions of the mesocosms will th en be used to determine the extent to which hydroperiod controls growth and survival in the fi eld. In order to evalua te this, three mesocosms were constructed and a randomized complete block design was used to test two flooding treatments and a control of no flooding. After simulating a rainy season of flooding in the Holein-the-Donut, South Florida slash pine growth rates, biomass, and tissue analyses were examined. The 2007 and 2008 flooding data from HID res earch plots in conjunction with local hydrological station data were used to constr uct a hydroperiod durati on map for Hole-in-theDonut restored areas as related to the curre nt elevation map. H ydrology and soils data, mesocosm pine sapling results, and Pinus elliottii var. densa seedling field survival data were used to create a map of suitable, marginal and unsuitable areas for South Florida slash pine in restored HID. Marginal areas were defined as the transition between pineland and marsh, where growth could be restricted in response to flooding pressure s. The mesocosm study flooding 25

PAGE 26

response was applied to Pinus elliottii var. densa in the field situation, where survival, growth characteristics, and age provided recommendations for South Florida slash pine reestablishment. Reestablishing and Monitoring of Pinus elliottii var. densa (Chapter 4) South Florida slash pine sa pling response to mesocosm flooding treatment was further evaluated in the field, where South Florida sl ash pine was directly seeded and planted by elevation treatment in order to evaluate germina tion and short-term survival in the HID. The working hypothesis was that pine survival was attributable to elevation, as related to hydroperiod. Seeds were collected, germinated a nd grown in the greenhouse, and direct seed scatters and plantings were evalua ted in the field for eighteen months. The first direct seeding event was in the original 20 research plots, where the upper four elevation treatments were located where pine rockland historically occurred and the lowest elevation treatment was located in historic marl prairie. A second direct seeding was completed in 20 new research plots during December of 2007, while eight containerized pines were plan ted within each of the 20 original 2 m2 plots. Initial and subsequent measurements were made for surv ival, growth (height and root collar diameter), and health, including initial measurements as c ovariates. Pine reestablishment methodology was evaluated, such that the best method of establis hment will be used to guide the National Park Service in the reforestation of restored lands, i. e. seeding or planting. Elevations in the HID where South Florida slash pine would be reestabl ished with the greatest chance of success were identified, such that edaphic and hydrologic f actors will promote pine reestablishment. An effective management plan for the reestablishment of South Florida slash pi ne in the Hole-in-theDonut restoration area was then formulated according to the research results. 26

PAGE 27

27 Implications for South Florida Slas h Pine in the Hole-in-the-Donut While drastic restoration efforts like the Hole-in-the-Donut restoration make it nearly impossible to return the area to exactly what it was prior to agriculture and restoration, it is possible to recreate an ecosystem high in eco system function and composed of native plant structure. Through this research, it was feasib le to reintroduce and in crease the abundance of South Florida slash pine in suitable areas where the species will persist over time. The following dissertation describes the efforts in volved in the reestablishment of Pinus elliottii var. densa including 1) an assessment of hi storical slash pine occurrence and current survival, as well as edaphic and hydrologic indicators, 2 a defined hydroperiod for Sout h Florida slash pine as an inherent silvical characteristic for reestablishment of the species, and 3) a methodology for South Florida slash pine reestablishment. Through this research, recommendations for the reestablishment of Pinus elliottii var. densa were made not only for Everglades National Park, but with applicability fo r South Florida rocklands.

PAGE 28

Figure 1-1. The Hole-in-the-Donut restoration project in Everglades National Park. Located in South Florida, portions of the Hole -in-the-Donut were farmed until 1975. 28

PAGE 29

Figure 1-2. The Hole-in-the-Donut as visible from space. Note th e restoration area composed of native plants in the center of the image, while the bright green areas are full of Schinus terebinthifolius and remain to be restored. Pine rockland borders the Holein-the-Donut to the north, while marl prairie is found to the south. 29

PAGE 30

Figure 1-3. Hole-in-the-Donut Rest oration sites by year, described as ResYr. Image courtesy of OHare and Dalrymple (2005). Meters Figure 1-4. Post-restoration eleva tion data for the Hole-in-the-Donut. Elevation grids given in meters and derived from photogr ammetric data by 0.1 m range. 30

PAGE 31

31 Figure 1-5. Location of the Pinus elliottii var. densa study site in the 2001 scraped area of the Hole-in-the-Donut.

PAGE 32

CHAPTER 2 BEFORE AND AFTER RESTORATION: USI NG HISTORICAL AERIAL PHOTOGRAPHY AND CURRENT SITE CHARACTERIZATION TO GUIDE THE REESTABLISHMENT OF PINUS ELLIOTTII VAR. DENSA IN THE HOLE-IN-THE-DONUT Introduction Suitability for South Florida slash pine ( Pinus elliottii var. densa ) reestablishment in the Hole-in-the-Donut (HID) re quires an understanding of the land cover that existed prior to largescale agricultural development, as well as the conditions that are found on-site today. Potentially, land that was pine rockland prior to agriculture an d subsequent restoration could support South Florida slash pine given optimal hydr ological conditions, or elevations, that exist in the restored HID. The drainage of the Ever glades ecosystem by the Central and South Florida canal system, in conjunction with the lower elevations and increased hydroperiod found in HID restored areas may actually create current hydrol ogical conditions that are similar to those that existed prior to drainage. Although Dalrymple et al. (2003) stated that the HID cons isted of short hydroperiod graminoid wetlands and mesic pine rockland, only in ferences were made as to where pine was found and there were no digital interpretation records. Ther e were also hardwood hammocks scattered amongst the HID landscape. These hammo cks remained relative ly dry throughout the rainy season and would have been domina ted by Caribbean broadleaved deciduous and evergreen tree species, such as gumbo limbo ( Bursera simaruba) and pigeon plum ( Coccoloba diversifolia) (Snyder et al. 1990, Ross et al. 2001). Unlike hardwood hammock, pine rockland is a fire dependent community, characterized as an ecosystem with a unique assemblage of upl and plant species growi ng on shallow soils atop jagged oolitic limestone rock out crops (Snyder et al. 1990). Pine rocklands are dominated by an open canopy of Pinus elliottii var. densa with an understory of w oody species of West Indian 32

PAGE 33

origin and herbaceous plants endemic to Sout h Florida. Species include willow bustic ( Bumelia salicifolia) locust berry ( Byrsonima lucida) and rough velvetseed ( Guettarda scabra) Among the limestone rock outcrops are solution holes that contain wetland plants like pond apple ( Annona glabra) South Florida slash pine can be found in upl and and wetland conditions, such that this species is associated with mesic and hydric pine flatwood communities (USFWS 1998a, USFWS 1998b). In South Florida, the mesic pine rock lands have non-hydric soils, with a midstory dominated by species such as saw palmetto ( Serenoa repens ) and wax myrtle (Myrica cerifera ) (USFWS 1998a). On-the-contrary, hydric pine rocklands have marl soils and an active periphyton mat during the rainy season (USFWS 1998b). Additionally, South Florida slash pine in hydric pine rocklands are associated with a wetland understory, includi ng species like cypress ( Taxodium spp.), dahoon holly ( Ilex cassine ), and sawgrass ( Cladium jamaicense ). There are about 15 acres of pine rockland remnants in the southern HID that exist as both mesic and hydric pine rockland following mechanical removal of Schinus terebinthifolius in the understory. This is an indication that hydric pine is expected in the HID after restoration, especially in areas where mesic pine rockland tran sitions to marl prairie. There is typically an ecotone between the pine rockland a nd marl prairie wetland, where upland transitions to marsh. The freshwater marl prairie ecosystem has both a shorter hydroperiod and lacks the peat f ound in the freshwater slough th at is known as the River of Grass. In fact, the accumulation of peat is inhi bited by the formation of calcitic marl by both the exposed limestone bedrock and the algal periphyt on community, as well as a shorter hydroperiod (Davis et al. 2005). This fine layer of calcium carbonate supports sawgrass (Cladium 33

PAGE 34

jamaicense) muhly grass ( Muhlenbergia capillaris var. filipes), and spikerush (Eleocharis cellulosa) as dominant marsh species (Armentano et al. 2006). The marl prairie has a hydroperiod that ranges from about four to six months, as defined by the plant species tolerances that dominate the marsh. Sawgrass is found in areas with a five to nine month hydroperiod, muhly has a two to four month hydroperiod (or as much as six months), and spikerush is found in areas that flood for six to nine months of the year (Armentano et al. 2006). Soils of the marl prairies that existed in HID were classi fied as Biscayne marl if the limestone bedrock had a depth upwards of 51 cm or Perrine marl if the limestone bedrock occurred at a depth of 51-102 cm (USDA 1996). The southeastern portion of the HID lies in proximity to Taylor Slough and would have had soils typical of the Perrine Series; that is a course-silty, carbonati c Typic Fluvaquent (USDA 1996). This silt loam is poorly drained to very poorly drained, with an A-horizon less than 28 cm, and a Cg horizon of an angular blocky stru cture atop the soft, porous, oolitic limestone. There would have been no need to rockplow th ese soils, as they were deep enough to farm. Biscayne soils are similar; how ever, the hard oolitic limestone is described around 38 cm, and therefore, it is likely these area s were rockplowed. This was pr obably typical of most HID marl prairie areas. On the slightly higher elevated pine ro ckland ridges were the very shallow, nonrockplowed soils that were classified as Rockdale in 1958 Soil Survey (USDA 1958), which closely resembles the Cardsound Series descript ion today (USDA 1996). There is only an Ahorizon of 0 to 10 cm that forms an abrupt irregular boundary above the hard limestone bedrock (USDA 1996). This Lithic Udorthent is a well dr ained silty clay loam, w ith friable weak fine and granular structure, with approximately 12% gravel. 34

PAGE 35

Once rockplowed, pine rockland soils are of th e Krome soil series, a calcareous gravelly loam with an Ap layer atop the oolitic limest one bedrock. Krome soils are anthropogenic, and are typically rockplowed or s carified annually in order to gr ow crops like tomatoes (USDA 1996). Krome soils are shallow (0 to 18 cm), carbonatic, moderately well drained, and yet, moderately permeable. They are extremely friable, of weak granular stru cture, and contain 35% to 60% limestone fragments. Typically, roc kplowing increases soil volume, leading to a decrease in bulk density and an increase in th e soil surface elevation (Meador 1977). Once low in nutrients and poorly drained, Krome soils are aerobic and contain higher nutrients, like phosphorus (Doren and Whiteaker 1990). As a result of fertilization, the Hole-in-the-Donut fallow farm fields of marl soil contained up to 500% more total phosphorus 150% greater available phos phorus, and 12% greater nitrate when compared to adjacent, unfarmed pa rk land in 1975 (Orth and Conover 1975). High nutrient, well-aerated soils were favorable to invasion by exotic plants (Loope and Dunevitz 1981). Attempts to restore pine to the HID we re made during the early 1970s; however, since the extent of pine rockland was not known, pines we re planted in areas that were historically marsh, as well as pine rockland (Bancroft 1973) Pines planted in marsh did not survive inundation, while those planted in former pine rockland were soon overgrown with Schinus terebinthifolius since the plantings were done on disturbed soils (Meador 1977). OHare and Dalrymple (2006) found that the growth and survival of South Florida slash pine on HID restored lands where disturbed soil was scra ped off were correlat ed with elevation, given that elevation was closely tied with hydroperiod. The Evergl ades region is very flat and only a few centimeters change in elevation, a nd hence hydrology, can influence where vegetation communities occur (Mitsch and Gosselink 2000). Th ese slight changes in elevation can mean 35

PAGE 36

the difference between a freshwater marsh, a pineland, or hammock. The elevation of the limestone bedrock and the limited soil atop th e limestone define the character of these ecosystems. Combined, they have an overwhelm ing influence on the nature and distribution of vegetation. While it has been shown that natural recruitment was co rrelated with distance and elevation, this researched was unique in that it focused on testing the hydrological limit to which South Florida slash pine could be reestablished in the HID. Areas devoid of pine, yet suitable in elevation, were seeded and plan ted with pine. Then, survival was monitored over the rainy season in order to redefine the South Florida slash pine range. The soil properties of the upland to wetland transition, where mesic to hydric pine lands and then marl prairie historically occurred, further delineated the extent to which S outh Florida slash pine c ould be reestablished. Aerial photographic interpreta tion of historical aerial photography, in conjunction with current elevation survey, was used to select an e xperimental area removed from pine seed source in order to advance the reestab lishment of South Florida slash pine. My working hypothesis was that hydroperiod was the cont rolling factor for potential South Florida slash pine reestablishment. In order to test this hypothesi s, the objectives of this research were to (1) interpret historical land cover in the Hole-in-the-Donut circa 1940 using aerial photographs, (2) select study plots using cu rrent elevation data, and (3) characterize the site with soil analyses and rainy season data as a function of elevation range in order to observe trends by elevation which could additionally affect sl ash pine seedling survival. With an understanding of the edaphic and hydrological conditions as related to elevation data and historical occurrence, it was possible to recommend favorable areas for the reestablishment of Pinus elliottii var. densa 36

PAGE 37

Methods Aerial Interpretation of 1940s Photos Aerial photographs of the Ev erglades from 1940 were downl oaded from the USGS SOFIA database (2007) and the digital images were geo-referenced into the NAD UTM coordinate system. Specifically, Sections 32 and 33, Li ne 13, Images 13-12, 13-13, and 13-14, were downloaded at 800 dpi in georefer enced Tagged Image File Format (TIFF). These images were brought into ArcMap and were fu rther georeferenced to assure that landmarks such as roads matched precisely. ArcMap was used to interpre t and delineate land cover in the HID based on a scale of 50 map units (m2), such that a given land cover was defined within a minimum area of 50 square meters. Land cover cate gories included the historical extents of pine rockland, marsh, hardwood hammock, and agriculture. Total ha were calculated for each land cover, and divided by an HID footprint of 3,065 ha in order to determ ine the percentage of each land cover in 1940. Given that agriculture was limited to marsh befo re the invention of the rockplow during the 1950s, it was possible to determine all historic la nd cover within the Hole-in-the-Donut prior to farming. Research Plot Establishment The historical transition between pine rock land and marsh, in conjunction with current elevation data, were both used as guidelines for research plot establishment. A 32-hectare research area was selected in the northeastern portion of the 2001 restored area. Five elevation ranges were chosen (60-70 cm, 70-80 cm, 8090 cm, 90-100 cm, 100-110 cm). Then thirty random points were selected using the Random Point in Poly gon Generation Program (VBA Macro) in ArcMap. A high-accuracy Topcon slope laser was then used to select the first four random points in each elevation that fit the elevation range. It was necessary to place the laser in the center of the research area in order to reach all coverage area. 37

PAGE 38

The random points, in UTMs, marked the northwest corner of the 2 m2 plot, a square that was 141.42 cm by 141.42 cm. By fitting the elevati on range, all four poten tial plot corners and the center were checked to assure that the plot fit within the elevation range. If not, the plot was thrown out and the plot selection continued in or der of random points. The laser was tested for accuracy and checked using two known benchmarks on the road, alongside the restoration area. The first twenty plots in this study were configur ed to the elevation range using relative data, hence the reason the ranges do not fall within exact 10-cm intervals (i.e. 63-71 cm, 73-81 cm, 83-91 cm, 93-101 cm, 103-111 cm). A gap of tw o cm was left between elevation ranges to minimize, if not prevent, overlap between ranges. The coordinates for the soil sampling plots are listed in Table A-1 of the Appendix. The result was four random plots per elevation treatment, for a total of 20 plots. This was done twice. The first set of 2 m2 research plots was established fo r the direct seeding effort in March 2007, as well as the pine planting in De cember 2007. The second set of 20 pine plots was established in December 2007 for the s econd direct pine seeding effort. It is worth noting that in February of 2007, the site was pre-treated with fire All plots in the four highest elevation ranges were burned; however, none of the low elevation plots burned. They were pre-treated with a Stihl gas powered string trimmer instead. Soil Analyses Soil samples were collected with a trowel from occasional pockets of soil in each of the 20 pine plots on March 13, 2008. Most research plots contained e xposed limestone and a very thin soil layer. Soils were air-dried, sieved through a two mm screen, and shaker ground for 10 minutes. Laboratory tests were performed for pH, total-phosphorus, tota l-nitrogen, and totalcarbon between March 28 and April 11, 2008. There we re four samples per elevation, except for the lowest elevation, where four periphyton sample s were also analyzed for pH. Here it was 38

PAGE 39

necessary to separate the detrita l floc (suspended organic aggregat es) from the mineral soil layer, as the 63 to 71 cm elevation range had a hydroperiod which supporte d a periphyton layer. Soil depth Depth of soil was measured in the fiel d on March 13, 2008. Ten random points were selected within each of the 20 research plots usin g a random numbers table. Data were recorded to the nearest millimeter. When exposed limesto ne bedrock was encountered, a reading of zero soil depth was recorded. In the lowest elevat ion, soil depth was measured independent of the periphyton layer. Average depth for each plot was calculated, as well as for each elevation range. Data comparisons were made across the five elevation ranges. Soil pH The pH was measured for soil collected in al l 20 pine plots using a glass electrode. The air-dried ground soil was weighed at 10 grams into a 50 ml beaker. Then 20 ml of DDI (double deionized) water was added to the beaker, the mixture was stirred with a glass rod, and the suspension was left to stand for 10 minutes. For periphyton, ten milligrams of water was added to two grams of periphyton. Once the machine was calibrated, the electrode was placed into the suspension directly above the sedimented soil and the pH was recorded. The glass electrode was rinsed with distilled wate r between each reading. Total phosphorus Since this was a calcareous so il, the modified Bowman method via sulfuric acid digestion was used to extract total phosphorus in the soil. Before initiating the analysis the digestion block was pre-heated to 340 C, with foil atop the block to keep the heat in. Approximately 0.2 grams of the ground soil was weighed into digestion tubes for each sample. Two milliliters of H2SO4 was added to the tubes and digested for 30 minutes. Then the samples were removed from the block and allowed to cool for 10-15 minutes. This was followed with the addition of 0.5 ml of a 39

PAGE 40

30% solution of H2O2 and tubes were replaced onto the bl ock for 15 minutes. Again the tubes were removed and allowed to cool. The addition of H2O2 was repeated until the solution was clear. The cleared tubes were placed back onto the block and di gested for 45 minutes to ensure that all the H2O2 evaporated. Then the tubes were remove d from the block and allowed to cool. In order to prepare the samples for measurem ent, a few ml of DDI water was added and vortexed. The solution was transferred to 25 ml volumetrics using Whatman #5 filter paper. The tubes and funnels were rinsed with DDI wate r and the volumetric was brought to volume. Murphy and Rileys Ascorbic Acid Method was us ed to analyze the concentration of phosphorus in solution (Schoenau and OHalloran 2007). Once the spectrometer was calibrated with the standards, all 20 samples were read at 880 nm Three blanks and peach leaves of known concentration were also sampled for comparison. Total carbon and total nitrogen Sample combustion took place in a Shimadzu SSM-5000A Solid Sample Module. The total carbon was measured by the Shimadzu TOC-Vcph Total Organic Carbon Analyzer and the total nitrogen was measured by the Shimadzu TNM-1 Total Nitrogen Measurement Unit. Shimadzu Scientific Instruments (SSI) in Columbia, Maryland is the American subsidiary of Shimadzu Corporation, headquartered in Kyoto, Japan. The machine was calibrated with carbon and nitrogen standards with an R2 of 0.995 or higher, at 900 C. Approximately, 0.0175 grams of sample were measured into the burning tray. Each sample was placed into the machine and areas under the curves for both carbon and nitrog en were displayed on the computer screen. The linear equation from the standard was solved for each x-coordinate using area as the y-value. This was divided by the grams of sample to obt ain either mg/g of carbon or ug/g of nitrogen. 40

PAGE 41

Statistical analyses SPSS was used to check the data for normality before analyses, using a one-way Analysis of Variance (ANOVA), with Tuke ys post-hoc comparison of mean s with a significance level of alpha = 0.05. A logarithmic transformation, excl uding the zero measurements, was necessary for soil depth. All replicates in the data set were av eraged to obtain a mean value for each elevation treatment. Mean data and standard errors fo r each dependent variable were inputted into SigmaPlot in order to illustrate the desired soil ch aracteristic as a function of elevation treatment, as well as compare the significance of the data between treatments. A best-fit linear regression line was applied to pH data. Logarithmic mean s were back-transformed from the log10 values using the equation: z= exp[y*ln10+0.5*variance2*(ln10)2], where z is the back-transformed value, y is the log10 value, and is the variance of the log10 va lue (Webster and Oliver 2001). Hydrological Measurements Water depth was taken once a week during th e rainy season for all plots that were inundated with surface water. Water levels were observed from the onset of the rainy season in April through draw-down of surface water below th e water table in November. For the 63 to 71 cm elevation data, a hydroperiod graph was cons tructed in Microsoft Office Excel using water depths for both the 2007 and 2008 summer wet seasons. Results By 1940, interpretation of aerial photography sh owed that agriculture had occurred on 663 hectares within the Hole-in-the-Donut restoration project footprint (Figure 2-1). At this time, it was also determined that there were 1940 ha of ma rl prairie, 449 ha of pine rockland, and 12 ha of hardwood hammock. According to Table 21, farmland encompassed 21.6% of the Hole-inthe-Donut, while the remaining la nd cover consisted of natural Everglades ecosystems. The aerial photographs further demonstrated that in 1940, 63.3% of the HID remained as marl prairie 41

PAGE 42

wetland, while 14.7% was pine rockland, as dom inated by a mix of dense and sparse Pinus elliottii var. densa The remaining 0.4% was hardwood hammock. Since Everglades vegetation communities were attributed in part to elevation, true elevation measurements for pine research pl ots were calculated according to benchmarks submitted to HID on November of 2007 (Table 2-2). The lowest elevation treatment was found to range from 63 to 71 cm, and all subsequently higher elevation treatments were classified at eight-centimeter increments every ten centimeters, such that the highest elevation plots ranged from 103 to 111 cm. When curren t plot locations were compar ed with the 1940s images, the highest four elevations were lo cated in former pine rockland, while all plots in the lowest elevation range occurred in historic marsh (Figure 2-2). As elevation decreased, soil depth increased (F igure 2-3). The data were not normal, as associated with the exposed limestone rock outcrops that covered approximately 5% of the lower elevated plots and up to 30% in the highest elevation (Table 2-3). The highest elevation treatment had an average soil depth of 4 mm, while the lowest had a soil depth of 14 mm, including measurements of limestone rock. When the limestone rock was not accounted for, soil depths ranged from 6 mm in the highest elevatio n to 15 mm in the lowest. When the logarithms of these data were analyzed, the 63 to 71 cm elevation had a significantly greater average soil depth than the 93 to 101 and 103 to 111 cm elevations (Figure 2-3). Though less soil depth, the 103 to 111 cm el evation range had the highest pH of approximately 8.00, statistically similar to the periphyton layer discerned at 8.12 in the 63 to 71 cm range (Figure 2-4). Given a five percent st andard error, elevation ranges 103 to 111, 93 to 101, and 83 to 91 cm were statistically similar, as were the 83 to 91, 73 to 81, and 63 to 71 cm elevation ranges. Therefore, the soil pH of the two highest elevation treatments was statistically 42

PAGE 43

different than the pH of lowest two elevation ranges, with pH of 7.84 (73-81 cm) and 7.85 (63 71 cm). There was a direct correlation between soil pH and elevation with an r2 of 0.680. Total phosphorus (T-P) varied among elevations, with levels significantly lower in the 63 to 71 cm elevation range at 159 mg TP/kg soil, and significantly higher in the 73 to 81 and 83 to 91 cm mid-elevation ranges, at 414 and 549 mg TP/ kg soil, respectively (Figure 2-5). There was overlap between high and low total phosphorus in the upper elevation ra nges, with statistical similarity of the 93 to 101 and 103 to 111 cm elev ation ranges to the lowe r elevation treatments. As for total carbon (T-C) and total nitrogen (T-N ), there were no significant differences for both elements between elevation treatments. Ther e were fluctuations in the carbon data, with a high of 210 mg T-C /g soil at 63 to 71 cm and a low of 110 mg T-C / g soil at 103 to 111 cm; however, no substantial correlati on was found among the data points (Figure 2-6). Average total nitrogen also increased with a decrease in elev ation, yet there was no sign ificance in the data. An average high of 2178 ug T-N / g soil was computed for 63 to 71 cm elevation, while a low of 1746 ug T-N / g soil was computed at 103 to 111 cm in elevation. Carbon had a greater slope than nitrogen as elevation decreased, such that the T-C:T-N ratio ranged from 64:1 in the highest elevation to 96:1 in the lowest elevation (Table 2-4). Average T-N:T-P wa s highest for the 63 to 71 cm range with a ratio of 16:1, while the higher elevations had T-N:T-P ra tios of 3:1 to 6:1. The 2007 and 2008 rainy season hydroperiod data fo r surface water observed at the lowest elevation treatment is illustrated in Figure 2-7. Hydroperiod differed in terms of duration, frequency, and even seasonality between years, ye t was more consistent in water depth. Water depths fluctuated, but peaked in early Octobe r around weeks 25 and 26 at approximately 25 cm. During 2007, the 63 to 71 cm elevation range resear ch plots were inundated for 22 weeks, with a frequency of wetting seven times between drawdown conditions. In 2008, the research plots 43

PAGE 44

were inundated for 15 weeks with a frequency of two wetting events. Flooding began earlier in the year during 2007, yet complete draw-down occurre d about the same time in early November. However, water depths were higher in 2008 during the latter part of the wet season. Discussion It was possible to distinguish land cover in the Hole-in-the-Donut circa 1940, as well as determine the historical extent of vegetation communities that existed prior to anthropogenic influence. Interpretation of the 1940s aerial phot ographs showed that farming was limited to the marl prairie wetlands, an indicati on that prior to farming about 85% of the HID was freshwater marsh. This inference was based upon the fact that these lands were developed before the invention of the rockplow in the 1950s; theref ore, farming was limited to the deeper soils found in the marl prairie (Ewel et al. 1982, Krauss 1987). Prior to la rge-scale farming, fields were small and irregularly shaped to fit the deeper area s of soil. There were larger areas of farmland within the marl prairies of the southeastern HI D near Taylor Slough, yet li mited farms within the fingerglade marshes that existed am ong the ridges of pine rockland. The ridges of pine rockland would have occurred at slightly higher el evations, with shorter hydroperiods, minimal soil, and exposed limestone rock (OHare and Dalrymple 2006). Though rockplowing permanently lowered elevations in the Hole-in-the-Donut, the objective to select study plots at varying elevation gradients was achie ved using current elevati on data. Most of the research plots for this study were located in former pine rockla nd, except the lowest 63 to 71 cm elevation range. The transition from high to low elevation was also associated with a transition from pine rockland to marl prairie, which explained the re lationship between soil depth and elevation. The lack of soil found at higher elevations could potentially limit pine growth and survival in the long-term. The marl prairie wetland in the 63 to 71 cm range had a longer hydroperiod where 44

PAGE 45

soil accretion occurred as a result of weathering and periphy ton associations. The marl accumulation associated with periphyton wetland processes suggested that conditions at the lowest elevation could be detrimental to South Florida slash pine survival, causing flood-related mortality. The periphyton layer in the lowest elevation plots had a higher average pH than soils collected from all pine plots. This occurred when periphyton cyanobact eria production became greater than respiration, such that pH increa sed as calcium carbonate precipitated, which then caused the formation of marl (Dav is et al. 2005). There was no standing water present during the time of sampling, so there was no calcium carbona te in the water column to confound soil pH results. It was also determined that the peri phyton layer was too thin to create detrital floc between the periphyton and the underlying soil surface. The pH sampled across all elevations ranged from 7.84 to 8.00, which was consistent with farmed HID lands. Li and Norla nd (2001) discovered that the pH of disturbed rockplowed and non-rockplowed land in the HID was 7.9 and HID restored lands had a pH of 7.8, which was significantly higher than the pH of 7.6 for undisturbed land in the HID surrounds. This was attributed to greater accumulati on of organic carbon in the undist urbed marl prairie, which would have lowered soil pH. Bachoon and Jones (1992 ) also sampled undisturbed marl soils in a sawgrass wetland, obtaining a pH of 7.6 at the surf ace that decreased to a pH of about 6.0 with an increase in depth of 18 to 20 cm below the surface. In this case, pH served as an indicator of hydrological conditions in restored HID, which may be used as a guide for pine reestablishment. The lowest elevation did have hi gher total carbon and lowe r soil pH, which could be attributed to the flooding regime. 45

PAGE 46

Soils in the lowest elevation were also was found to contain a highe r, but not significant, average nitrogen content, though nitrogen concentr ations in restored HID were initially low (Inglett et al. 2006). In fact, undi sturbed lands were found to cont ain significantly higher total soil nitrogen concentrations as attributed to organi c nitrogen, when comp ared to rockplowed lands (Li and Norland 2001). Restored lands we re found to have statis tically similar total nitrogen to undisturbed and rockplowed lands at 4.87 g T-N / kg soil, though much lower than the 8.12 g/kg found in the undisturbed land, and slight ly higher than the 4.45 g/kg in rockplowed soils. It was the high C:N ratios which suggested n itrogen-limitation and caused nitrogen to be limiting to South Florida slash pine growth, partic ularly in the lower elev ations, where nitrogen was relatively unavailable due to immobilizatio n (Havlin et al. 2005). Given the duration of flooding observed in the 63 to 71 cm elevation ra nge, it was possible that microbes were using NO3as the terminal electron ac ceptor under anaerobic conditi ons, indicative of low oxygen levels that could limit South Florida slash pine occurrence. The periphyt on layer had an effect on the C:N ratio, by contributing to the precipitati on of calcium carbonate to the soil as marl over time (Davis et al. 2005). Add itionally, the poorly drained wetl and soils accumulated organic matter which increased carbon concentrations (Reddy and DeLaune 2008). Since there was a greater depth of carbon accumulation and marl formation in the lower elevations, there was even more carbon, and hence less available nitrogen. This nitrogen limitation was also seen in nitrogen to phosphorus ratios, where higher elevations in restored HID we re a nitrogen-limited system (Koerselman and Meuleman 1996). The lowest elevation was nearly phosphorus-lim ited and more typical of the oligotrophic Everglades marsh (Mitsch and Gosselink 2000). Th e area located in histor ic marl prairie also 46

PAGE 47

had comparable total phosphorus to the 219 mg P/ kg soil observed in south Everglades marl soil (Chen et al. 2000). This lower elev ation was in an area of deeper marl soil that was historically marl prairie, as compared to th e higher elevations that were fo rmerly pine rockland, where more rock plowing would have been necessary to cr eate enough soil for farming. Li and Norland (2001) also found 3.5 times greate r concentration for total phosphor us when rockplowed soils were compared to non-rockplowed soils in the HID. The differences in phosphorus concentrations may have also been attributed to the homogenized limestone layer that resulted fr om rockplowing followed by restorative land clearing. The breaking and scra ping of limestone rock increas ed the fractions of calcium carbonate that existed in the soil, which then reacted with the phosphorus fertilizer, thereby forming a precipitate that decreased solubility and leachability, and hence, increased phosphorus concentrations in the soil (Orth and Conover 1975 ). Chen et al. (2000) established background phosphorus concentrations for Everglades marl soils at 515 mg P/ kg soil, which included a merged data set of disturbed and undisturbed so ils. Iron was considerably high at 6,600 mg Fe/ kg soil, but not necessarily available. While there were background data for marl we tland soils, there were minimal data that existed for the pine rockland soils. The color of the soils in the lowest elevation plots was a lighter whitish-gray when compared to the hi gher elevated soils that were brown-gray in coloration. Perhaps these lower elevated soils reduced more iron and there was greater iron mobility than the higher elevated soils. The hi ghest elevation remained free of inundation for most of the rainy season, while the lowest elev ation treatment remained flooded for about four months. This suggested that pi ne rockland soils could become fo list in nature if left unburned. 47

PAGE 48

Pine Rockland soils were described in the Soil Survey of Dade County (USDA 1996) as being of the Cardsound Series, and were closely associated with th e Krome Series soils. The soil transition between marl prairie and pineland is not well known, t hough it is likely that the pine rockland soils contain more organi c carbon than their more mineral marl counterparts, given that soils of the Biscayne Series contain 70% to nearly 100% calcium carbonate in the A-horizon and 80% to nearly 100% calcium carbonate equivalencies in the Cg horizon (USDA 1996). Also, the soils in the pine rockland solu tion holes more closely resemble d the Biscayne Soil Series and were of a different nature than the upland so ils, merely due to hydrol ogical influences. Of course, fire burning back to ba re rock would decrease the or ganic carbon content of pine rockland soils, where the rooting media was found to contain 30% to 50 % organic matter (Snyder et al. 1990). In the future, it is wort h investigating inorgani c carbon verses organic carbon contents in pine rockland soils and further comparing the pine rockland soils to marl soils and restored HID soils. Restored HID demonstrated soil and hydrological patterns characteristic of an upland to wetland transitional area. Water depths and durations in the 63 to 71 cm elevation plots were found to be characteristic of a marl prairi e wetland over both the 2007 and 2008 rainy seasons, despite fluctuations between years. Even though there were differences in frequency of flooding, duration, and seasonality between y ears, all plots in the lowest el evations remained inundated for 15 weeks or more and there was a consistent relationship with eleva tion. For hydrology and soils data, only the first set of established research plots was discussed here Further elaborations on HIDs hydroperiod for both sets of research plot s, as well as eleva tion ranges, will be discussed in Chapters 3 and 4. Chapter 3 si mulated the flooding data collected in this 48

PAGE 49

experiment to test the effects of hydroperiod on S outh Florida slash pine sa plings in mesocosms, while Chapter 4 examined South Florida slash pi ne seedling survival and growth in the HID. Conclusion Though previous attempts to restore pine in the HID were unsuccessful, aerial interpretation in conjunction with current elev ation survey, soils, and hydrological data were used to characterize potentially suitable areas for pine reestablishment. It was determined that current research plots were all in former pine rockland except for the lowest elevation range, which was found where marl prairie historically occurred. Indirect relationships were discovered in the soils, as depth, total ca rbon, and total nitrogen increased as elevation decreased. Though no relationship was found for total phosphorus in soils, pH increased with an increase in elevation. There were observable differen ces in hydroperiod over the 2007 and 2008 rainy seasons; however, the lowest elevation supporte d a hydroperiod typical of marl prairie wetland. The following chapters will disc uss the effects of hydrology on Pinus elliottii var. densa growth, as well as the species reestablishment in restored areas as attributed to historic, edaphic, and hydrologic variables. This information was then u tilized to create a suitab le guide for restoration of South Florida slash pine in the Hole-in-the-Donut a nd South Florida. 49

PAGE 50

Table 2-1. Hole-in-the-Donut land cover in 1940, as interpreted using historic aerial photography. Land Cover Hectares Percent Cover (%) Agriculture 663 21.6 Marl Prairie 1940 63.3 Pine Rockland 449 14.7 Hardwood Hammock 12 0.4 Table 2-2. True elevation range s for pine plots within the HID restoration area, as eight-cm increments every ten centimeters, across five elevation ranges. Plots Elevation Range (cm) Elevation 1 103-111 Elevation 2 93-101 Elevation 3 83-91 Elevation 4 73-81 Elevation 5 63-71 Table 2-3. Percent of exposed limestone bedr ock and average soil depth, with and without limestone rock measurements, as associated with each elevation range. Elevation (cm) Exposed Limestone (%) Soil Depth (mm) Rock Included Rock Excluded 63 to 71 5.0 14 15 73 to 81 5.0 9 10 83 to 91 7.5 8 9 93 to 101 17.5 7 8 103 to 111 30.0 4 6 Table 2-4. C:N and N:P ratios acco rding to elevation. Means without letters or followed by the same letter are not significant for alpha at 0.05. C:N ratios are statistically similar Elevation (cm) N o. C:N N :P 103 111 4 64:1 4:1a 93 101 4 86:1 5:1a 83 91 4 72:1 3:1a 73 81 4 86:1 6:1a 63 71 4 96:1 16:1 b 50

PAGE 51

Figure 2-1. Aerial interpretation of the Hole-in-the-Donut by ve getation cover circa 1940. Note the ridges of pine rockland and fingers of freshwater marsh to the north. Approximately 15% of the Hole-in-the-D onut was dominated by pine rockland. 51

PAGE 52

Figure 2-2. Plot locations for the fi rst set of research plots. Note the majority of research plots in the upper four elevations were located in what was pine rockland during 1940. All 63 to 71 cm plots (lowest elevation) were located in historic marl prairie. GIS coordinates are listed in the Appendix, Table A-1. 52

PAGE 53

Elevation Range (cm) 63-7173-8183-9193-101103-111 Depth (mm) 0 2 4 6 8 10 12 14 a ab ab b b Figure 2-3. Soil depth as a func tion of elevation range, excludi ng exposed bedrock. Logarithmic means were back-transformed to arithme tic units, with arithme tic error bars one standard error above the mean. Means that share similar letters are statistically similar for a log10 (soil depth +1) transformation. 53

PAGE 54

Figure 2-4. Soil pH verses elevati on range for the first set of pine research plots. Note that periphyton pH was analyzed at the 63 to 71 cm elevation range. Mean separations are at one standard error above the mean. 54

PAGE 55

Elevation Range (cm) 103-111 93-101 83-91 73-81 63-71 Soil TP (mg P/kg soil) 0 100 200 300 400 500 600 700 ab ab a a b Figure 2-5. Distribution of tota l-phosphorus in soils as related to elevation ranges for pine research plots. Mean soil phosphorus followed by the same letters were not significantly different for alpha = 0.05. Error bars are at one sta ndard error above the mean. 55

PAGE 56

Elevation Range (cm) 63-7173-8183-9193-101103-111 Total Carbon (mg C / g soil) 80 100 120 140 160 180 200 220 240 260 Total Nitrogen (ug N / g soil) 1400 1600 1800 2000 2200 2400 Figure 2-6. Total-carbon and totalnitrogen concentrati ons as a function of elevation, with error bars at one standard error above the mean and an alpha = 0.05. 56

PAGE 57

57 0 5 10 15 20 25 301 3 5 7 9 11 13 1 5 1 7 1 9 2 1 2 3 25 27 2 9 3 1Week (mid-April through mid-November)Average Water Depth (cm) 2007 2008 Figure 2-7. A comparison of wate r levels in the 63 to 71 cm elevation treatment between the 2007 and 2008 rainy seasons. These data were co llected from the first set of research plots, where pine was seeded and planted.

PAGE 58

CHAPTER 3 THE EFFECTS OF FLOODING TREATMENT ON PINUS ELLIOTTII VAR. DENSA SAPLINGS USING FIELD-SIMULA TED HYDROPERIOD DEPTHS AND RECOMMENDATIONS FOR RESTORATION Introduction The vegetation changes along the transition from pine rockland to marl prairie occur where increases in depth and durati on of flooding during the rainy s eason are observed. Though the Pinus elliottii var. densa (South Florida slash pine) which are scattered amongst the ecotone that distinguishes these two vegetation communities are hydric pines, there are limits to which this subspecies has adapted to survive flooded c onditions over time (USFWS 1998b). When the Hole-in-the-Donut (HID) lands were left fallow in 1975, the majority of the HID footprint had been farmed, as the rockplow allowed farmers to break-up the underlying rock and intermix it with the shallow soil layer above (Ewel et al. 1982, Krauss 1987). Since restoration of these lands removed disturbed soils down to the limestone bedrock, the elevations of restored lands were permanently lower than historical occurrence. In comparison to farmed lands, the hydroperiod of restored areas was longer and the invasion of Brazilian pepper ( Schinus terebinthifolius) was limited, making the wetland portion of the restoration successful (Li and Norland 2001). It was unclear how hydrological records compared to historic hydrological conditions fo r this area; therefore, it was important to investigate the potential for Pinus elliottii var. densa occurrence in the HID at present. Under flooded conditions and without need le exposure to the air, oxygen becomes limited and stomata do not gain carbon for growth or maintenance processes (Pallardy 2008). This could be detrimental to South Florida slash pine reestablishment. There were pine planting failu res in HID at lower eleva tions during the early 1970s, where 60% of newly established pine seedlings had died after four months of flooding across 58

PAGE 59

half the study area (Bringman 1976). Prior to the rainy season, 95% of those pines had survived the dry season and survival leveled off once flooding subsided. Following restoration of the HID adjacent to pine rockland, survival of naturally recruited pines in the re stored area was linked to higher elevations where inundation did not cause mortality, or that seedlings grew large enough to withstand short periods of i nundation (OHare and Dalrymple 2006). In Pinus elliottii var. elliottii, bedding has been used to enhance surviv al on poorly drained so ils, by elevating the roots into a zone of aeration above th e water table (Duncan and Terry 1983). Though limited by hydroperiod, both Pinus elliottii subspecies have adapted to poorly drained soil conditions and surviv ed best along the edges of in undated areas, like swamps and streams or on flatwood sites (Langdon 1963, L ohrey and Kossuth 1990). On flatwood poorlydrained soils, needle and fine root respiration influenced the Pinus elliottii var. elliottii carbon budget, such that fine root resp iration encompassed about two-thirds of the total evolved soil carbon dioxide. Therefore, the total below-ground carbon allocati on was greater than one-third of the gross primary production (Ewel et al. 1987, Cropper and Gholz 1991). Large Pinus elliottii var. elliottii trees extended taproots and ve rtical sinker roots into saturated soils for months at a time (Fisher 1989). Within the wood of sinker roots, there was an open pathway where oxygen diffusion occurred. Adapte d to a high water table, the air content in the green rootwood of Pinus elliottii var. elliottii ranged from 48 to 69% when saplings were grown in an inundated peat-sand mixture (Fis her and Stone 1990a). There were aerenchyma found only within five cm of root apices, which was similar to observations of flooded Pinus taeda seedlings (McKevlin et al. 1987, Fisher and Stone 1990a). On poorly drained Leon fine sand, Pinus elliottii var. elliottii seedlings were exposed to two constant water table depths of 46 cm and 92 cm as well as a fluctuating water table for five 59

PAGE 60

years (White and Pritchett 1970). The fluctuat ing, natural water table in a given year ranged from a depth of 38 to 76 cm for four to five m onths, was within 100 cm for at least nine months, increased below 38 cm for less than two months, and periodically dropped to a depth greater than 76 cm during the dry season. The extension of Pinus elliottii var. elliottii sapling root systems were limited by distance to the high water table, su ch that roots in the fluctuating water table reached 36 cm, while roots in the 46 cm and 92 cm water tables reached 56 cm and 107 cm, respectively. Additionally, the he ights of slash pine saplings were influenced more by water table depth than fertilization. Height growth was highest in th e 46 cm stabilized water table of the spodic horizon, slightly lower in the 92 cm water table, and lo west in the fluctuating, more inundated root system. Pinus elliottii var. densa was even more tolerant of flooding and drought than Pinus elliottii var. elliottii (Abrahamson and Hartnett 1990). Ew e et al. (1999) found a greater decrease in predawn water potential from wet to dry seas on in hammock species, as compared to pine rockland species, such that Pinus elliottii var. densa and Myrica cerifera were less affected by the dry season. While hammock species were mo re dependent on soil water than pine rockland species, pine rockland species were dependent on a more stable groundwater supply and regional hydrologic patterns. Pinus elliottii var. densa utilized deeper groundwater sources in the Florida Keys, such that a 15 cm rise in sea level over 70 years increased salt water intrusion and soil salinities to the point where the pine rockland co mmunity was reduced by at least 16 ha (Ross et al. 1994). Prior to 1935, an additional 42 ha of those pine rocklands were lost. Water stress has also been observed for Pinus elliottii var. densa in terms of water shortage. A decline in growth was found in Pinus elliottii var. densa stands where the regional water table had been lowered over time (Oberbauer et al. 1997). Higher el evations may in fact 60

PAGE 61

have limited germination, given a greater depth to the water table (OH are and Dalrymple 2006). Also, chlorosis in South Florida slash pine has often led to mortality because iron was relatively unavailable in the calcareous soils, especially wh ere pine root systems were unable to tap into the water table (Barnard an d Greenstein 1995). Although Pinus elliottii var. densa were capable of withstanding flooding, the extent that South Florida slash pine growth and nutrient characteristics have reacted to different flooding regimes over time is less known. In order to investigate the autecology of South Florida sl ash pine saplings as related to hydroperiod, a mesocosm experiment was used to simulate three HID elevations, with defined hydroperiods linked to the field si tuation. In Chapter 2, the aver age hydroperiod was determined for research plots within fi ve elevations ranges, where Pinus elliottii var. densa were directly seeded and monitored throughout the 2007 rainy season (Chapter 4). The high, mid, and low hydroperiod data were used in this experiment to simulate HID elevati ons with potential to support South Florida slash pine. Saplings were us ed in lieu of newly germinated seedlings in order to show more pronounced effects of flooding on growth and nutrient characteristics over time. Young seedlings were more susceptible to flooding, as discussed in the field trials of Chapter 4. The hypothesis for this mesocosm e xperiment was that hydroperiod limited Pinus elliottii var. densa growth and survival at lower elevations a nd was the driving factor for South Florida slash pine reestablishment in the Hole-in-the-D onut restoration area. The objectives of this mesocosm experiment were to 1) determin e how exposure to diffe rent flooding treatments affected survival of Pinus elliottii var. densa saplings, 2) examine how flooding affected growth rates and biomass of pine over the growing season, and 3) conduct tissue analyses of pine saplings to determine how chemicals and nutri ents were affected by flooding. Since, these 61

PAGE 62

flooding treatments represented the range of hydrope riods, or elevations, currently available in the HID, it was possible to incorporate the mesocosm data with the HID Pinus elliottii var. densa field data of Chapter 4, and determine the hydroper iod criteria necessary for short-term survival and successful establishment of this species within HID restored areas. A hydroperiod map of flooding durat ions by elevation was constructed for the Hole-in-theDonut using research plot water depths in conjun ction with local hydrologi cal station data. From there, restored areas were identi fied as suitable for South Florida slash pine reestablishment, marginal, or unsuitable. The results of this mesocosm experiment were useful in targeting favorable areas for restoration, such that South Florida slash pine will be reestablished in areas where root systems are capable of withstandi ng rainy season flooding, yet still have sufficient growth. Unfavorable areas, where flooding cause d poor growth or South Florida slash pine mortality will be avoided in future restoration efforts. Materials and Methods Study Site The mesocosm experiment was conducted outside in full sunlight at the Institute of Food and Agricultural Sciences extension TREC (T ropical Research and Education Center) in Homestead, Florida, commencing on June 27, 2008 and ending November 14, 2008. Over the course of the 20-week e xperiment, a total of 54 Pinus elliottii var. densa saplings were exposed to two flooding treatments and a control of no flooding. The South Flor ida slash pines were purchased through Vebers Jungle Garden, Inc. on May 19, 2008 in Homestead. At this time, the South Florida slash pine saplings were estimated at 2.5 years of age and had a watering schedule of one hour a day during the afternoon when there was no rain. 62

PAGE 63

Mesocosms All pines were immediately re potted May 20 into 11liter pots using Fafard Growing Mix No. 2 potting soil and covered with three to five cm of 57-limestone rock fragments to serve as weights under flooded conditions. Pines were standa rdized in the repotting process, such that confounds of inundated root space were avoided. The potted pines were fertilized once with a low nutrient solution of 0.5 grams of NH3 granule and iron powder per liter of water, and left to acclimate to the transplant for th ree weeks on a drip-irrigation sy stem in the shadehouse. The watering schedule during this time was 7:00 am and 7:00 pm for 10 minutes. In late June, pines were taken off the drip sy stem and moved into three mesocosm pools, as a randomized complete block design with th ree flooding treatments. The mesocosms were approximately 91 cm wide, 152 cm long, and 61 cm in height. They were constructed with cinder blocks, and lined with 305 cm by 366 cm wa terproof PVC tarps. The pools were filled with groundwater and then 18 pines were exposed to each of three treatments, as six subsamples per treatment in all three blocks. The flooding tr eatments consisted of 1) pots and root system fully flooded in water, 2) partially flooded pots w ith roots partially submerged, and 3) a control of no flooding. Fully flooded pines represented the hydroperiod at the lowest elevation in the HID, partially submerged represented the mid-point 20 cm higher, and a control of no flooding was representative of the highest elev ation (theoretically another 20 cm higher). The water table was dynamic and changed on a weekly basis in or der to reflect the 2007 HID hydropattern where pines were planted and seeds were scattered in the field. For the flooded treatment, pine root systems were fully submerged over the course of the experiment and water depth was at or above the surface as modeled over the rainy season. Par tially flooded pines at the mid-point had their 63

PAGE 64

root systems inundated to differe nt degrees as water levels fl uctuated each week, while root systems for control pines only re ceived water as precipitation. All South Florida slash pine sa plings were randomly tagged fo r blocks 1-3 and treatments 1-3. In order to represent the three elevations in the HID, pines were separated at 20 cm intervals. Flooded pines were placed in the bottom of the pool, while partially flooded pines were placed into a stack of 3gallon pots inside the mesocosm, which raised the partially flooded pines 20 cm above the flooded pines. The control pines were placed just outside the mesocosm on ground cloth to represent non-flooded South Fl orida slash pines. Water levels were equivalent to those observed in the HID at the lowest plots during the 2007 rainy season once flooding remained consistent. The mesocosm fl ooding regime was initiated in the mesocosms within two weeks of the actua l flooding date, as modeled afte r the 2007 HID rainy season data (Figure 3-1). Since pines were exposed to the rainy season, watering to field capacity was only necessary when the pines went three days wit hout water. There was a rain gauge within 500 meters of the experiment, monitored hourly online by Everglades National Park (Figure 3-2). In-Situ Measurements and Growth Every two weeks, visual evaluations for chlo rosis and percent dead (brown) needles were taken, including initial and fina l observations. On July 16, 2008, the temperature of the water in all mesocosms was 26 C. Electrical conductivity readings for 30 ml water samples for each mesocosm were 657 s/cm, 702 s/cm, and 683 s/cm, for blocks 1, 2, and 3, respectively. Initial and final measurements were also taken for height, root collar, and health before and after flooding treatment. Biomass Upon harvest on November 14, 2008, the pine s were removed from the pools. All recoverable dead and brown needles on each pine were collected and bagged. After a portion of 64

PAGE 65

the needles were removed for tissue samples, the remaining needles were stripped from the tree stem, bagged, and weighed for fresh/wet weights in grams. These wet needle measurements had a ratio with the tissue wet weights that was useful in later estimating the dry wet total. Needles for the biomass measurements were rinsed with DDI water, padded against paper towels, placed into new bags, and into the drying oven at 70 C for one week. The stems were then severed from the roots and fresh weights were taken for a ll pine stems. Lastly, the roots were removed from the pots and all excess soil was washed away from the roots. Dunking buckets were used to remove excess potting media and a sieve was used to collect any floati ng, detached roots. Tissue samples, bulk needles, dead needles, stems, and roots were all placed into the drying oven at 70 C for one week. Dry weights were taken for n eedles, stem and branches, and roots, and dry root to shoot ratios were calculate d. Biomass results were analyzed in SPSS as a block design and Analysis of Variance by floodi ng treatment. Log root was graphed as a function of log shoot. Tissue Analyses Fresh green needle samples were collected from the active growing portion of each tree (approximately five to 10 cm along the apical bu d) before the tree was harvested of all green needles. These samples were then used to anal yze chlorophyll a and b, carbon (C), nitrogen (N), phosphorus (P), and iron (Fe). The wet weights were taken befo re five fresh needles were utilized for chlorophyll measuremen ts. Chlorophyll samples were pl aced in the refrigerator for four days until sampled. Dried, ground needle sa mples were used to analyze carbon, nitrogen, phosphorus, and iron. Chlorophyll a and b Fresh, wet samples were kept cold and dark until November 18 and 19, 2008, when chlorophyll was determined according to Abadia and Abadia (1993). Approximately 0.05 gram 65

PAGE 66

of cut needle sample was placed in a mortar with just enough sand to help extract the chlorophyll from the plant and a small amount of calcium carbonate to neutralize the acetone used in the extraction. Working under a fume h ood, the needles were crushed with a pestle and chlorophyll extracted with an acetone solution. Samples were rinsed with acetone until clear, then brought up to 5 ml in a test tube with a pi pette. Chlorophyll a a nd b were read in the Beckman DU-640 spectrophotometer at 662 nm and 645 nm wavelengths, respectively. Beckman Coulter, Inc. is headquartered in Fullerton, CA. Total-nitrogen, total-carbon, total-phosphorus, and iron For each tree, approximately 200 milligrams of tissue was weighed. Elementar Americas, Inc., located in Mt. Laurel, New Jersey, produc ed the vario MAX CNS, which was used to directly analyze totalcarbon and total-nitrogen. In order to prepare P and Fe samples, the digestion procedure by Ha nlon et al. (1994) was followed. Approximately 0.2 g of each dried, gro und tissue sample was weighed and placed into a 20 ml glass vial. Additionally, two replicat es, 0.1 g peach leaves, and two blanks were sampled. All samples were placed into a muffle furnace at 250 C for 30 minutes, and then the temperature was raised to 550 C for four hours. Once cooled, the tissue samples were digested with 2 ml blank solution (formulated as 100 ml 6 N HCl / 2500 ml DDI water) and 18 ml DDI water. Tubes were rinsed with DDI water and brought to 50 ml. Then the tubes were covered, shaken well, and filtered into new 50 ml tubes. For phosphorus, all samples were diluted ten times with DDI water, or 0.5 ml of sample to 4.5 ml DDI water, in order to obtain the correct pH and fit the samples with the standard. Given extraction with AB-DTPA, SAA color reagent was used to measure tissue total-phosphorus brought into solution using the Beckman DU-640 spectrophotomet er at 880 nm (Soltanpour and Schwab 1977). For iron, digested samples were poured into 15 ml tubes and analyzed with the 66

PAGE 67

Shimadzu AA-6300 atomic absorption spectrophot ometer, a company headquartered in Kyoto, Japan, and located in Columbia, Maryland. All values were calculated in mg/kg. Hydroperiod and Pine Suitability Maps There were two hydrostations located in the Hole-in-th e-Donut; DO-1 and DO-3, which tracked daily stage measurements. DataForEVE R was used to acquire the data for 2007 and 2008, using the browser: 10.72.34.58 (Kahn and Serra 2008) In order to calculate water depth, the stage reading was subtracted from the gr ound surface elevation. Ground surface elevation was given for DO-1; however, it was necessary to obtain a ground surface elevation for DO-3 according to Everglades hydrological protocol, us ing a reference elevation and laser level. Once all depths were calculated, the hydrostatio n data was examined weekly, in order to calculate the total number of weeks flooded (as duration, not frequency) for the given elevation range where the hydrostation was located. Number of weeks flooded by elevation was averaged for 2007 and 2008. Duration of flooding was also calculated for pine plots by elevation and averaged for 2007 and 2008. This information was then used to construct a hydroperiod map in ArcGIS for flooding durations for Hole-in-theDonut restored sites, as based upon weekly flooding and high-accuracy 10-cm elevation data ranges. Given pine survival data, a pine suitability map was also created and recommendations were made for areas of suitable, marginal and unsuitable pine reestablishment. Statistics All growth, biomass, and tissue results we re analyzed with Analysis of Variance (ANOVA) in SPSS as a blocked design. This requir ed univariate for gene ral linear model, with a dependent variable (growth, biomass, chemical), and treatment and block as fixed factors in a custom model with main effects. Descriptive st atistics were also computed with Post hoc Tukey by treatment at an alpha of 0.05. Visual evaluati ons for chlorosis and dead needles were run as 67

PAGE 68

repeated measures by treatment as eleven va lues over twenty week s. Where necessary, transformations were made to normalize the data before analyses, and logarithmic means were back-transformed from the log10 values using the equation: z= exp[y*ln10+0.5*variance2*(ln10)2] where z was the back-transformed value, y was the log10 value, and variance was of the l og10 value (Webster and Oliver 2001). Total-carbon data was fit with an arcsin transformation, while total-nitrogen and C:N (total-carbon:total-nitrogen) data were normal. In order to describe chlo rophyll a to b ratios as normal, an arcsin transformation was necessary. Analysis of Covarian ce (ANCOVA) tests were performed using initial height a nd root collar diameter measurem ents as covariates for growth and biomass data, with flooding as a fixed factor. Overall, mortality was low over the course of the experiment, as only one pine out of 54 died and was therefore excluded from the analyses. Results Tree Growth and Survival All growth rates passed the normality test and Tukey showed no significant differences by block for all growth variables, as illustrated by averages for all treatments by block (Table 3-1). For the three flooding treatments, there were no signi ficant differences in initial height and root collar diameter measurements, which ranged from 28.3 to 29.1 mm in root collar diameter and 61.8 to 64.1 cm in height. After 20 weeks of gr owth, final height remained similar among all treatments, ranging from 74.1 to 82.2 cm (Figure 3-3) Though the final root collar was similar between the flooded and control treatments at 29.5 and 31.8 mm, respectively, there was a significant difference in the final root collar diameter of the partially flooded treatment, with the highest average measurement of 38.7 mm in diameter (Figure 3-4). It was apparent that initial root collar diam eter was a significant covariate for final root collar diameter measurement, in relation to flood ing treatment (Table 3-2). Initial height was 68

PAGE 69

also a significant covariate with flooding treatment for final height measurement. As for growth over the course of the experiment initial height and root colla r diameter were not significant cofactors with flooding for height and root collar growth rates. However, the partially flooded treatment had a significantly higher growth rate for both root collar diameter and height when compared to the flooded and c ontrol treatments (Figure 3-5). Only one tree suffered mortality in the fully flooded treatment. There was no significant difference for dead needles collected and weighe d at the end of the experiment; however, the flooded treatment had the highest percent dead needle weight. It was important to note that there were dead needles that had fallen off the flooded trees over the course of the experiment that were not collected and analyzed. Visual evaluations over time indicated a higher observable percentage of dead needles in the fully flooded treatment as compared to the partially flooded and control treatments (Figure 3-6). By week six, the logarithmic per cent dead needles was statistically similar for the partially flooded and control treatments, and yet significantly lower than the flooded treatment for the remainder of the experiment. There was no significant difference in chlorosis by treatment over time, given a 1 to 5 scale of yellow to green, respectively (Figure 3-7). Severe yellowing, or chlorosis, was judged as a low value, while a high value was representati ve of dark green. When each treatment was compared over time, average chlorosis remained relatively similar for the control, yet pines became greener under fully flooded and partially flooded conditions over the course of the experiment. Biomass Measurements and Calculations All mesocosm pools were analyzed as blocks which were statisti cally similar for all biomass measurements. Dry weights for root, sh oot (stem + needle), stem, needle, and sum of plant parts were normally distribu ted. As for the dry biomass, th e roots were significantly lower 69

PAGE 70

in the flooded treatment at 33.1 g (Table 3-3). Co nversely, dry shoot weight, as attributed to the stem, was significantly higher in the partially flooded pines wh en compared to the other treatments. Needle weights were similar and ra nged from 75.6 g in the flooded treatment to 103.6 g in the partially flooded pines. The total dry biomass was also statistically higher for the pines in the partially fl ooded treatment, at 285.1 g. When total biomass, root, shoot, stem, and n eedle measurements were dependent on initial height as a covariate with flooding treatment, th ere were significant in teractions for these biomass measures (Table 3-4). In itial root collar diameter was al so a significant covariate, along with flooding treatment for total biomass, roots, shoot, and stem measurements (Table 3-5). Therefore, initial measurements for height and r oot collar diameter taken at the beginning of the experiment had effects on biomass measurements. As for the dry root to shoot ratio, control pines had an average ratio of 0.32, which was significantly higher than both floodi ng treatments (Figure 3-8). Since control pines had a higher root:shoot ratio, the slope of logarithmic root values as a f unction of logarithmic shoot was the greatest, with an equation of y = 2.250ln(x) 0.075, and a logarithmic regression of R2 = 0.772 (Figure 3-9). Flooded pines had a much lowe r slope of 1.313, with a logarithmic regression value of 0.789. The partially flooded treatment had a slope of 1.101 and a weak correlation where R2 = 0.363. Chemical Analyses There were no significant differe nces by block for all chemical analyses. There were also no significant differences in percent carbon am ong treatments, with values ranging from 49.1% to 49.2% (Table 3-6). Nitrogen was 0.61% in the partially flooded treatment and 0.75% for the non-flooded control, with a signifi cantly higher measurement for th e control, when compared to the other flooded treatments. As for C:N, the pa rtially flooded treatment had the highest ratio of 70

PAGE 71

82:1, which was significantly higher than the control at 67:1. Th e flooded treatment was similar to both the control and pa rtially flooded pines. The flooded treatment was also similar to the other treatments for logarithmic totalphosphorus levels in needle tissu e (Figure 3-10). As with n itrogen, total-phosphorus was also significantly higher in th e control when compared with the partially flooded treatment. The average actual phosphorus content in the control was 854 mg P/ kg tissue (or ppm), 741 ppm for the flooded treatment, and 629 ppm fo r the partially flooded treatment. Iron was also converted to logarithmic values, where the normalized data were significantly similar among treatmen ts (Figure 3-11). Actual aver age concentrations for iron in plant tissue were 39 ppm for fl ooded conditions, 46 ppm for part ially flooded conditions, and 40 ppm for the control. Though not significant, iron was highest in the partially flooded treatment. A covariate of visual chlorosi s evaluations with iron was not significant, though South Florida slash pines coloration from yello w to green increased over time. There were no significant differences among flooding treatments for chlorophyll a, chlorophyll b, or the chlorophyll a: b ratio. The control had the highest average of 477 mg/g chlorophyll a, while the partially flooded pines had an aver age of 388 mg/g (Table 3-7). Chlorophyll b only ranged from a high of 110 mg/g in the flooded treatment to a low of 107 mg/g in the non-flooded control. The actual chlo rophyll a:b in pine tissue ranges from 4.1 in the partially flooded treatment to 5.6 in the control. Discussion Though almost all pines survived the simu lated HID 2007 rainy season data for depth, frequency, and duration, there were indications that flooding would have affected South Florida slash pine survival over time. The higher per centage of dead needles found in flooded pines indicated that photosynthesis was less optimal, wh ich may induce more stress in flooded pines 71

PAGE 72

over time (Pallardy 2008). The pine that died was only 40 cm in length and suffered because most of the needles were fully inundated with wate r. This was the case with smaller seedlings in the HID field studies (Chapter 4), and carbon-13 isotope data illustrated that there was less flooding stress for older tissue (Obe rbauer et al. 1997). A portion of the dead needles were balanced by growth over the course of the experime nt, that as pine grew in all treatments, the lower needles were shaded from light. It was possible to achieve the objective a nd demonstrate the effects of flooding on Pinus elliottii var. densa growth rates and biomass differences ov er the course of one growing season. The partially flooded treatment had the best grow th rate, given a partially flooded root system, where oxygen and moisture levels were adequate, if not optimal, for transport and growth processes. Though initial height and root collar diameters of pine s had an effect on final growth measurements as related to ontogeny, these measur ements did not have an effect on the growth rates, an indication that height and root collar diameter grow th rates were related to flooding. Initial height and root collar diameter also ha d effects as covariates for dry shoot and root biomass measurements; however, the slope of log root / log shoot among treatments demonstrated how root to shoot ratios changed with development and removed the ontogeny effect. As pines grew bigger in the contro l of no flooding, carbon was fixed nearly two times more into the roots as to the shoots when compared to the partially flooded and flooded treatments (Figure 3-9). Clearl y, the control in this experiment invested a higher biomass into the roots, as further indicated by th e dry mass root to shoot ratio. When comparisons were made for dry root mass alone, the flooded treatment had a lower root biomass, similar to that obser ved by White and Pritchett (1970) for Pinus elliottii var. elliottii five years following water table treatments. Fluctuating inundation and greater water72

PAGE 73

logging of the root system decreased slash pine roots as well as shoots when compared to two constantly regulated water tables. Additionally, they found that Pinus elliottii var. elliottii grown in the constant 46 cm water table of the spodic horizon had greater root and shoot masses and growth than the 92 cm better-drained water tabl e. In both slash pine variety experiments, allocations to the root system further expl ained South Florida slas h pine dependence on groundwater, as well as the lower growth rates obse rved in control pines (Ewe et al. 1999). Fisher and Stone (1990a) found that Pinus elliottii var. elliottii root systems did not extend vertical roots as far downward, had increased branching, and higher cross section to length proportions when grown under wet site conditions; an adaptation for an inundated root system to effectively diffuse internal oxygen. Therefore, it was possible that Sout h Florida slash pine developed hypertrophied lenticels in the bark of the stem or aerenchyma in the root system for storage and diffusion of oxygen, a nd hence, survival under flooded conditions. It would have been necessary to take mass verses length to in fer the presence of aerenchyma; therefore, the lower mass of the flooded roots may simply indi cate that this treatm ent had less roots. Surface water was found to slow the growth of South Florida slash pine seedlings in the Hole-in-the-Donut, if not cause mortality over time (OHare and Dalrymple 2006). The data used to simulate the 2007 hydroperiod was based on a drier year (Kahn a nd Serra 2008). Wetter years and 46-cm rain events like Hurricane Katrina were fatal to pine s at slightly higher elevations that had es tablished during drier years, give n lower flooding durations between wetting events and lower water depths during those dry years (OHare and Dalrymple 2006). Also, when the roots of Pinus elliottii var. elliottii were exposed to low levels of oxygen, there was a decrease in nutrient uptake, though not ev idenced in the fully flooded tissue analyses conducted in this study (Shoul ders and Ralston 1975). 73

PAGE 74

Water stress, as excess or more limited water availability, apparently limited growth even more than nitrogen concentrations. There were lower nitrogen and phosphorus concentrations in the needle tissues of South Flor ida slash pines in th e partially flooded treatment. The lower nitrogen levels in both fl ooding treatments indicated that Pinus elliottii var. densa was competing with microbes for nitrogen, as nitrogen serv ed as a terminal electron acceptor under flooded conditions (Mitsch and Gosselink 2000). There were no fertilizer additions to Pinus elliottii var. densa during flooding, making this somewhat of a cl osed system, especially in comparison to field conditions. As pines grew, nitrogen was not replaced as it may have been in the field. White and Pritchett (1970) also obser ved lower nitrogen in unfertilized Pinus elliottii var. elliottii exposed to fluctuating and higher water tables, with the fluctuating water table at 0.88% N, the 46 cm water table at 1.01% N, and the 96 cm water table at 1.11% N. Despite any nutrient effects from the potting soil, these South Florida slash pines appeared nitrogen-deficient at 0.75% when compared to Pinus elliottii var. elliottii. However, Pinus elliottii var. densa leaf nitrogen and carbon concentrations for all mesocosm treatments, as well as C:N in the flooded and partially flooded pines were similar to t hose observed for South Florida slash pine by Oberbauer et al. (1997) in Everglades National Park and were indi cative of more water stress. This was in reference to lower C:N ratios found outside the park, which were comparable to the control. According to Pritchett and Comerford (1983), Pinus elliottii var. elliottii had a critical foliar phosphorus range of 800 to 900 mg/kg. Th e control was in this range; however, the partially flooded treatment had lower phosphorus, either caused by additional stem growth or water stress. It appeared that the Pinus elliottii var. densa subspecies was adapted to lower phosphorus concentrations than Pinus elliottii var. elliottii, as tissue phosphorus was 74

PAGE 75

approximately 400 mg/kg in the Ev erglades and upwards of 700 mg/ kg in areas outside the park, as influenced by either soil type or agriculture (O berbauer et al. 1997). Af ter all, the Everglades are a phosphorus-limited system and Pinus elliottii var. densa does have a lower site index than Pinus elliottii var. elliottii (Lohrey and Kossuth 1990, Mitsch and Gosselink 2000). It was possible that phosphorus depletion at th e root tips was responsible for lower tissue phosphorus concentrations in the partially floode d root systems. Phosphorus depletion of Pinus elliottii var elliottii in air, following nitrogen gas treatmen ts was not significantly different, such that phosphorus depletion leveled out after 42 h ours and that the uptake of phosphorus was not inhibited after three days under hypoxia (Escam illa and Comerford 1998). However, despite a full recovery of phosphorus uptake upon the rem oval of oxygen-deprived conditions, phosphorus uptake was reduced by 28% during a second exposure to nitrogen gas. The flooded pines in this mesocosm experiment were only exposed to hypo xic conditions once, while the partially flooded root system had periodic draw-down and rewet ting of the root system that may have made phosphorus less available for uptake into the plant. Field conditions would differ, as inundation in the fully flooded root system, as represented in the 63 to 71 cm elevation, would repeatedly occur each rainy season, possibly deplet ing the root system of phosphorus. This phosphorus uptake at the roots described biomass phosphorus content, which may not necessarily have been indicative of foliage phosphorus content. Interestingly, White and Pritchett (1970) described higher tissue phosphorus for Pinus elliottii var. elliottii in the betterdrained soil, similar to that of the mesocosm control. Also, the fluctuating water table that provided more flooded root conditions was in the middle, like that of the mesocosm, while roots that tapped the spodic 46 cm water table were lowest, like that of the partially flooded treatment. Phosphorus contained in the crown may have been lost when needles dropped into the flooded 75

PAGE 76

conditions of the mesocosm, as opposed to needles in the control that fell back into the pots. In fact, the total labile P of Oi horizon needles for Pinus elliottii var. elliottii was 146 mg/kg in an unfertilized reference treatment (Polglase et al 1992). Regardless, th ere was adequate tissue phosphorus to provide energy for Pinus elliottii var. densa growing in the Everglades. Chlorophyll measurements and decreases in chlo rosis also indicated that the majority of the South Florida slash pines had adequate iron th roughout the study, as they were fertilized with iron while on the drip system. Waterlogging of these soils rende red iron more available, and though not significant, pines in both flooding treatm ents did become greener over four months, while the control remained the same. It was like ly that the primary tissue of South Florida slash pine roots had an oxidized rhizosphere in anaerob ic conditions, similar to that of 1oblolly pine (Fisher and Stone 1990a). The oxyge n-transporting lacunate roots oxi dized iron in loblolly pine and were not found in well-drained soils with adequate oxygen (McKevlin et al. 1987). Despite similar chlorophyll, there were less needles on the flooded Sout h Florida slash pines, which indicated lower growth rates and potential survival. Given survival and growth characteristics of South Florida slash pine in the mesocosm and Hole-in-the-Donut restoration, it was possible to develop a hydrope riod map and a plan for pine reestablishment in the HID and South Florida. In the Hole-in-the-Donut, average flooding durations were examined for pine research plots and two hydrostations over the 2007 and 2008 rainy seasons. Pine research plots indicated that elevations gr eater than 90 centimeters had a hydroperiod of two weeks or less flooding durati on (Figure 3-12). There was a flooding duration of two to five weeks at elevations of 81 to 90, five to ten weeks of flooding in the 71 to 80 cm range, and ten to 20 weeks of flooding at 51 to 70 centimeters, which may have included several wetting events. Hydrostation DO-1 was found at 56.7 cm NAVD-88 and flooded an average 76

PAGE 77

duration of 11 weeks, while hydrostation DO-3 was located at approx imately 41.3 cm, and had water above the surface for an average of 22 weeks. It was then estimated that water was present above the ground surface for 20 to 40 weeks from el evations slightly below sea level to 50 cm, while elevations below .10 cm had 40 weeks or more flooding, with inundation possible for the entire year in the lowest elevat ions of the southeastern HID. The flooding regime for 2007 that was simulated in this mesocosm experiment was based on a relatively dry year for the Hole-in-the-Donut, which raised concern fo r pine reestablishment in the lowest 63-71 cm elevation range. In fact, hydr ostation DO-1 had lower periods of constant inundation during 2007 when wetting events were compared from 2000 through 2008 (Table 38). Although the number of discontinuous number of days DO-1 had water levels above the surface was near average at 90 days, the average number of days for each wetting event was only 10 days, while the 2000 to 2008 average was 23 days. The maximum inundation period for a continuous wetting event during 2007 was 25 days, versus an average 54 day maximum, continuous inundation over the past nine years. Therefore, it was highly unlikely that South Fl orida slash pine seedlings would survive in the 60 to 70 cm range without se veral consecutive years of abnormally dry rainy seasons. It was probable that hydric pine would exist in the 70 to 80 cm range where a much thinner layer of periphyton was present. In fact, survival of Pinus elliottii var. elliottii saplings after five years was highest for a constant 46 cm water table, when compared to a higher fluctuating water table and a better-drained 96 cm water table (White and Pritchett 1970). Of the 1,540 ha of restored areas in the Hole-in-the-Donut, approximately 65 aces were found to be currently suitable for South Florida slash pine (Figur e 3-13). There were about 73 ha considered marginal for Pinus elliottii var. densa growth, where pine may or may not survive over time. The remainder of 77

PAGE 78

restored areas located in the southern and wester n-most portions of HID we re unsuitable for pine reestablishment. Pinus elliottii var. densa was suitable for elevations greater than 80 cm in areas that were historically pine rocklan d. Pines that are able to tap the gr oundwater supply above this elevation will likely have the highest growth rates over time. The future transitiona l ecotone for marginal growth is likely between 70 and 80 cm, as pine seedlings in the HID 73 to 81 cm elevation range showed higher mortalities over th e second, more typical rainy seas on (Chapter 4). These hydric, fringe pines may or may not survive and will probab ly be stunted in growth, given a total of five to ten weeks flooding duration among wetting events. Conditions were unsuitable for slash pine survival below 70 cm. Despite permanent elevation changes and lowe r present-day water levels in Taylor Slough, the extent to where slash pine historically occu rred was similar to present-day reestablishment. Therefore, it will be possible that marginal acreag e may be converted to unsuitable over time if the hydropattern at the 70 to 80 cm range become s wetter than pre-farming conditions, given the permanently lowered elevation which resulted from rockplowing and rest oration processes. Despite pine survival in the fully flooded mesoco sm experiment, it was unlikely that these pines would survive at the 60 to 70 cm elevation range in HID after a few rainy seasons. Even if pines survived at the lowest elevation over time, th ere would be no point in establishing pine where pine rockland understory cannot follow via hydrology. Though some of the understory plants that comprise the pine rockland ecosystem are better able to tolerate flooding, there are many that cannot tolerate the hydric soils found at the 60-70 cm, incl uding rare and endangered pine rockland species such as Ipomoea microdactyla and Chamaesyce pinetorum 78

PAGE 79

The next challenge for the Hole-in-the-Donut wi ll be to pursue the reestablishment of such species in these suitable pine rockland areas. In the meantime, Chapter 4 demonstrated that early pine growth ceased as newly germinated or young planted seedlings succumbed to flooding when the water depths of the 63 to 71 cm elevation range completely submerged the needles. Chapter 4 also showed that the planted pines survived longer periods of inundation than younger scattered seedlings, while the mesocosm flooded sa plings survived even longer. Therefore, larger pine saplings might better survive the flooded conditions of the Hole-in-the-Donut. In any case, the effects of flooding on nutrient uptake an d growth that were demonstrated in this experiment furthered our unders tanding of the autecology of Pinus elliottii var. densa. Conclusion A mesocosm that simulated the high, low, and midpoint of Hole-i n-the-Donut elevation ranges and hydroperiods for South Florida slash pine reestablishment revealed that partially flooded pines had higher growth and biomass afte r one growing season. Pines were able to survive the range of flooded, or non-flooded treatm ents, an indication that South Florida slash pine may have developed aerenchyma or hypertr ophied lenticels, similar to those found in Pinus elliottii var. elliottii. Nitrogen and C:N tissue levels were similar to Pinus elliottii var. densa tissue samples in the Everglades and this subspe cies has apparently adapted to survive with lower phosphorus levels than Pinus elliottii var. elliottii Iron was probably more available under flooded conditions, as flooded pines were observed as less chlorotic over time. While it was apparent that South Florida slash pine could be reestablished in similar areas to the historical occurrence, pine adaptations to the cu rrent hydrology will sort out the current Pinus elliottii var. densa range in time. South Florida slash pine is only one species that comprises the pine rockland ecosystem and it is important to gear restoration efforts towards protected upland species as well. 79

PAGE 80

Table 3-1. Root collar diameter height, and growth measurem ents by treatment by block for pine seedlings before and after flooding treatments. Root collar diameter (mm) Height (cm) TX BLOCK June November Growth June November Growth Flooded 1 26.8 28.9 2.1 59.9 73.1 13.2 2 29.8 30.4 0.6 60.0 72.2 12.2 3 28.6 29.2 0.6 65.2 76.6 11.4 Total 28.3 29.5 1.1 61.8 74.1 12.3 Partial 1 30.2 37.3 7.1 63.8 84.8 20.9 2 28.4 38.0 9.6 66.3 83.3 17.0 3 30.6 40.6 10.1 61.5 78.6 17.1 Total 29.7 38.6 8.9 63.9 82.2 18.3 Control 1 28.0 31.3 3.3 64.2 81.6 17.5 2 31.4 33.4 2.0 62.3 73.6 11.2 3 27.7 30.6 2.8 65.7 78.0 12.4 Total 29.0 31.8 2.7 64.0 77.7 13.7 Total 1 28.4 32.5 4.2 62.6 79.8 17.2 2 29.9 34.2 4.3 63.0 76.6 13.5 3 29.0 33.5 4.5 64.1 77.7 13.6 Total 29.0 33.4 4.3 63.3 78.1 14.8 Table 3-2. Covariate effects of in itial root collar diameter or height measurements on dependent variables of final root colla r diameter or height measurements, respectively. Flooding treatment was considered as a fixed fact or. Values are significant at alpha = 0.05. Final Root Collar (mm) Final Height (cm) Source df F Sig. df F Sig. Corrected Model 3 43.836 0.000 3 96.697 0.000 Intercept 1 7.866 0.007 1 7.149 0.010 Initial Root Collar (mm) 1 55.496 0.000 Initial Height (cm) 1 270.414 0.000 Flooding Treatment 2 29.259 0.000 2 5.797 0.005 Error 49 49 Total 53 53 Corrected Total 52 52 80

PAGE 81

Table 3-3. Dry weights for Pinus elliottii var. dens a root, stem, and needles by flooding treatment. Similar letters indicate st atistical significance for a given biomass measurement. Flooded Partial Control Root a33.1 b 49.4 b 51.6 Shoot a161.5 b 235.6a161.2 1. Stem a85.9 b 132.0a84.1 2. Needles 75.6 103.6 77.0 Total Biomass a194.6 b 285.1a212.8 Table 3-4. Covariate effects of in itial height measurements of S outh Florida slash pine saplings on dependent variables for total biomass, r oot, shoot, stem, and needles. All weights for dry mass, with a fixed factor of flooding treatment. Values are significant at alpha = 0.05. Total Biomass (g) Root (g) Shoot (g) Stem (g) Needles (g) Source df F Sig. F Sig. F Sig. F Sig. F Sig. Corrected Model 3 29.692 0.000 16.764 0.000 27.839 0.000 44.257 0.000 12.316 0.000 Intercept 1 6.483 0.014 1.151 0.289 6.746 0.012 16.055 0.000 1.186 0.282 Height (cm) 1 64.091 0.000 30.683 0.000 57.568 0.000 90.088 0.000 26.089 0.000 Flooding Treatment 2 10.628 0.000 7.713 0.001 11.731 0.000 19.477 0.000 4.828 0.012 Error 49 Total 53 Corrected Total 52 81

PAGE 82

Table 3-5. Covariate effects of in itial root collar diameter measurements of South Florida slash pine saplings on dependent variables for total biomass, root, shoot, and stem. All weights for dry mass, with a fixed factor of flooding treatment. Values are significant at alpha = 0.05. Total Biomass (g) Roots (g) Shoot (g) Stem (g) Source df F Sig. F Sig. F Sig. F Sig. Corrected Model 3 8.180 0.000 8.139 0.000 8.059 0.000 12.005 0.000 Intercept 1 0.140 0.710 .072 0.789 0.136 0.714 1.458 0.233 Root Collar 1 11.233 0.002 9.917 0.003 9.847 0.003 16.052 0.000 Flooding Treatment 2 5.055 0.010 5.908 0.005 5.748 0.006 7.912 0.001 Error 49 Total 53 Corrected Total 52 Table 3-6. Carbon, nitrogen, and C:N in Pinus elliottii var. densa tissue following 20 weeks of flooding treatment. Similar letters indicate statis tical significance for a given nutrient or ratio. %C %N C:N Flooded 49.2 a0.66 a b 77:1 Partial 49.2 a0.61 b 82:1 Control 49.1 b 0.75 a67:1 Table 3-7. Chlorophyll A and B concentratio ns (mg/g), and chlorophyll A:B after flooding treatment for Pinus elliottii var. densa tissue. Chlorophyll A (mg/g) Chlorophyll B (mg/g) Chlorophyll A:B Flooded 427 110 4.2:1 Partial 388 108 4.1:1 Control 477 107 5.6:1 82

PAGE 83

Table 3-8. Hydroperiod data fo r Station DO-1 in the Hole-inthe-Donut, Everglades National Park. Discontinuous, minimum, average, a nd maximum readings are given in days. Year Wetting Discontinuous Minimum Average Maximum 2000 5 91 5 18.2 45 2001 4 99 3 24.8 67 2002 9 88 1 9.8 45 2003 6 137 2 22.8 82 2004 3 68 2 22.7 49 2005 2 108 30 54.0 78 2006 4 81 2 20.2 39 2007 9 90 2 10.0 25 2008 3 75 7 25.0 58 Average 5 93 6 23.1 54 83

PAGE 84

Initiate Flooding Figure 3-1. Rainy season average weekly wate r levels for 2007 by elevation range, for plots seeded with Pinus elliottii var. densa Arrow delineates flooding initiation of the mesocosm field-simulated hydrological regime. 84

PAGE 85

Figure 3-2. Total daily rainfall at the IFS hydr ological station rain gauge for 20 weeks of mesocosm flooding treatment during the 2008 rainy season. 85

PAGE 86

Flooding Treatment FloodedPartialControl Height (cm) 0 20 40 60 80 100 Initial Height Final Height Figure 3-3. Initial and fina l height measurements of Pinus elliottii var. densa saplings before and after flooding treatment. Error bars are one standard error above the mean for separate comparisons between initial he ight or final height measurements. 86

PAGE 87

Flooding Treatment FloodedPartialControl Root Collar Diameter (mm) 0 10 20 30 40 50 Intial Root Collar Final Root Collar a a b Figure 3-4. Initial root collar and final root collar diameter measurements by flooding treatment for Pinus elliottii var. densa saplings. Error bars are one standard error above the mean. Similar letters shown in the final root collar treatment are statistically similar. 87

PAGE 88

a b a a a b Flooding Treatment FloodedPartialControl Height Growth (cm), Root Collar Growth (mm) 0 5 10 15 20 25 Height Growth Root Collar Growth Figure 3-5. Height growth and root collar diameter growth after 20 weeks of flooding treatment for Pinus elliottii var. densa saplings Error bars are one st andard error above the mean, while similar letters among height or root collar are sta tistically similar. 88

PAGE 89

Week of Treatment 048121620 Percent Dead Needles 0 10 20 30 40 50 Flooded Partial Control a a a a a a a a a a a ab ab b b b b b b b b b b Figure 3-6. Repeated measures over 20 weeks fo r percent dead needles by flooding treatment for Pinus elliottii var. densa saplings. Logarithmic data were back-transformed to original arithmetic means. Error bars are one standard error above the mean and means denoted with the same letter are statistically similar within five percent error. 89

PAGE 90

Week of Treatment 048121620 Chlorosis Visual Evaluation (Color Scale 1-5, Yellow-Green) 4.0 4.2 4.4 4.6 4.8 5.0 Flooded Partial Control Figure 3-7. Bi-weekly repeated measures of ch lorosis estimates for various flooding treatments in a mesocosm study of Pinus elliottii var. densa saplings. Lower values were more chlorotic (yellow) than highe r ones (green). Error bars are one standard error above the mean. 90

PAGE 91

Flooding Treatment FloodedPartialControl Root:Shoot Dry 0.0 0.1 0.2 0.3 0.4 b a a Figure 3-8. Root:shoot ratios of dried Pinus elliottii var. densa saplings according to flooding treatment. Error bars are one standard e rror above the mean, with similar letters of statistical similarity. 91

PAGE 92

Figure 3-9. Logarithmic root mass as a func tion of logarithmic shoot mass (in grams) by flooding treatment. The equations are plotted for each flooding treatment, along with logarithmic regression values. 92

PAGE 93

Flooding Treatment FloodedPartialControl Total Phosphorus (mg P / kg tissue) 0 100 200 300 400 500 600 700 800 900 1000 ab b a Figure 3-10. Total tissue phosphorus of South Florida slash pine needles by mesocosm flooding treatment. Means were back-transformed fr om logarithmic to arithmetic units. Error bars are arithmetic and mean s with the same letter are statistically similar for normally distributed logarithmic data. 93

PAGE 94

Flooding Treatment FloodedPartialControl Iron (mg Fe / kg tissue) 0 10 20 30 40 50 Figure 3-11. Iron tissue measuremen ts verses flooding treatment for Pinus elliottii var. densa Means are shown as back-transformed logarithmic to arithmetic values, while significance was based on logarithmic result s. Error bars are one standard error above the mean and shown as arithmetic units. 94

PAGE 95

Figure 3-12. Hydroperiod map for the Hole-i n-the-Donut, as based upon number of weeks flooded in one year. The 2007 and 2008 flooding duration data were averaged by elevation. 95

PAGE 96

96 Figure 3-13. Pine suitability map for the Hole-in-the-Donut, as based upon number of weeks flooded and elevation data. Areas are identifi ed as suitable for optimal pine growth, marginal where pines may or may not survive, or unsuitable.

PAGE 97

CHAPTER 4 THE REESTABLISHMENT AND MONITORI NG OF SOUTH FLORIDA SLASH PINE SEEDLINGS IN A RESTORED AGRICULT URAL AREA WITHIN EVERGLADES NATIONAL PARK Introduction Before agriculture and restorative land-clear ing, the Hole-in-the-Donut area of Everglades National Park was dominated by pine rockland and marl prairie. In 1940, aerial photographs revealed that 449 ha, or 14.7% of the land cover in the Hole-in-theDonut (HID) was pine rockland (Chapter 2). Prior to Hurricane Andr ew, Doren et al. (1993) conducted a study on the density and size class distribution of South Florida slash pine ( Pinus elliottii var. densa) logged and unlogged stands that were eith er unmanaged or managed by fire The results were variable; however, the density of the unmanaged, unlogged site was 672 live pines per hectare and the stand had an apparent uneven-age d structure. The other study sites had even-aged size class distributions, with the fire ma naged, unlogged stand at a density of 479 trees per hectare and the fire managed, logged stand was composed of 1,146 lives pines per hectare. Management consisted of fire suppression, th at when followed by intense dry season fires most likely converted the unlogged st and to even-aged structure (Doren et al. 1993). As a result, recommendations were made for intense, numerous early wet season fires resembling presettlement conditions, in order to encourage r ecruitment and uneven-aged management of these South Florida slash pine forests. Snyder et al. (2006) further i nvestigated the response of pine rocklands to seasonality of burning and determined that summer burns cau sed higher mortality in South Florida slash pine trees th an winter burns. Other species growth stages showed different responses to seasonality of the bur n; however, this may have also been related to fire intensity, as summer burns had a higher fire intensity when compared with winter burns. 97

PAGE 98

The species that compose a forest dominated by South Florida slash pine are adapted to and even dependent on frequent fires, as lack of fire can lead to the en croachment of hardwoods, thereby displacing pine rockland under-story pl ants and shifting community composition (Loope and Dunevitz 1981). Histori cally, the fire regime of Pinus elliottii var. densa was a frequency of every three to seven years, sim ilar to that of longleaf pine ( Pinus palustrus) (Long et al. 2004). The grass stage of South Florida slash pine has served as an adaptation that makes Pinus elliottii var. densa more resistant to fire than Pinus elliottii var. elliottii (Lohrey and Kossuth 1990). Pinus elliottii var. elliottii required a longer fire interval b ecause the seedlings were likely to succumb to fire during the first four years of gr owth (Long et al. 2004). Since winter dry season backfires exhibited homogeneity and were more subject to restrictiv e conditions, summer wet season burns became the norm for Everglades National Park in 1981 (Doren and Rochefort 1984). Prescribed wet season head fires were more favorable in that fi re behavior was more variable and coincided with th e lightning-ignition fire season. The canopy of Pinus elliottii var. densa was open, such that pine needles and other accumulated litter dried out so quickly that fire s have occurred as soon as one day following a rain event (Robertson 1954). Given enough rain flooding does occur in the pine rocklands. South Florida has a subtropical climate, with a dry season from November until May that suddenly commences into the rainy season. Ther efore, South Florida slash pine was more tolerant of floodi ng and drought than Pinus elliottii var. elliottii (Abrahamson and Hartnett 1990). With observable differences in the structur al characteristics of these two trees, Langdon (1963) defined the range of bot h slash pine varieties. Pinus elliottii var. elliottii was found along the coastal plain from South Carolina to Miss issippi and southward through central Florida, 98

PAGE 99

while Pinus elliottii var. densa was found only in central and Sout h Florida, including the Florida Keys. Despite these different environmental cond itions, both slash pine varieties were adapted to tolerate poorly drained conditions, whethe r in the flatwoods of Georgia or amongst the solution holes of the Miami Rock Ridge. In fact, slash pine growth was optimal along pond margins where soil moisture content was plentifu l, but not excessive (L ohrey and Kossuth 1990). Squillace (1966) found clinal variation existed throughout the slash pine range, with genetic traits such as germinability, total height, and thickness of hypode rm showing a north-south trend in variation. In areas where slash pine range s overlapped, there were no clear distinction of types, and according to Lohrey and Kossuth (1 990), hybridization does occur naturally between Pinus elliottii var. densa and Pinus elliottii var. elliottii In 1966 when Squillace described his genetic da ta, there was question and still is debate over dividing this species at the taxonomic leve l. For instance, the Biota of North America Program under Kartesz ( 1999) lumped slash pine, while Wunderlin and Hansen (2003) separated Pinus elliottii into two varieties. Given that the rema ining pine rocklands currently exist as fragments, genetic mapping of Pinus elliottii var. densa was used to evaluate the quality of seed sources for current pine rockland restoration proj ects (Williams et al. 2007). There was a lack of genetic erosion and inbreeding w ithin pine rockland fragments, indicating that seed could be collected from a single fragment; however, it woul d be best to collect seeds from as many fragments as possible. It was suggested that within fragments, seed should be collected from trees separated by at least 15 m, in order to increase genetic dive rsity for restoration efforts. While this was applicable to the 126 small fragments, or 1.8% of pine rockland that remains in the 920 ha within Miami-Dade County, th e research staff at Everglades National Park are concerned with maintaining th e genetic diversity of South Flor ida slash pine occurrence in 99

PAGE 100

the more than 4,250 ha of pine rockland ecosystem that exists within the park boundary (OHare and Dalrymple 2006). Since the slash pine popula tion in the Everglades is adapted to more mesic conditions than the surr ounding populations in the Miami Rock Ridge, the National Park Service focused restoration efforts in-house, with the collection and rais ing of their own seed stock for planting and reforestation purposes. Ec otype differences in South Florida slash pine have been observed in the soils that support Pinus elliottii var. densa growth. When a non-local sandy seed source was compared to a local rocky seed source, planted seedlings from the sandy seed source grew slower and e xperienced higher mortality on th e pine rockland habitat following fire, though there was only 0.4% genetic differe nce between the sandy and rocky seed types (Hamrick 1995, Mayo 2000). The Reforestation Project dur ing the 1970s was the first doc umented attempt to restore pine in the Hole-in-the-Donut (Bancroft 1973). In 1972, seed was collected using a tree shaker, harvested, and grown by the Florida Department of Agriculture and Consumer Services, Division of Forestry. Then, a D-8 was used to clear th e shrubby landscape for plan ting in one of the HID management units that was abandoned in the 194 0s. This involved the removal of Brazilian pepper ( Schinus terebinthifolius ), but left the disturbed soil in place. Reforestation study plots encompassed about a 405 ha area where native transplants, bare root seedlings, containerized seedlings, and potted pines were introduced. After three years of reestablishment, survival of bareroot hand plantings was 3% and potted pines was 25%. Cont ainerized seedlings and native transplant survival after six months was 11% and 0%, respectively. The outcome was unsuccessful and it was determined that pine was not a practical way to reclaim the Hole-in-theDonut. 100

PAGE 101

While pine is not the way to restore all of the HID, there certainly are reclaimed areas where South Florida slash pine is naturally recruiting from seed following restoration. Looking back on the Reforestation Projec t, pine was planted in dist urbed soil and seasonally flooded areas. In restored areas of the HID that were cleared of disturbed so ils and were higher in elevation, a total of 106 out of 200 marked seedlings greater th an ten centimeters in height remained alive after five years of growth, despite Hurricane Katrina which dropped 46 cm of rainfall over a 24-hour period (OH are and Dalrymple 2006). Most of the seedling survival was limited to sites with elevation of 1.1 m or higher and less than 25 m distance from the current pine rockland habitat, hence South Florida slash pi ne seed source. Pines were able to germinate below 1.1 m in elevation; however, these pines s howed higher mortality and slower growth rates over the first three years. OHare and Dalrym ple (2006) described elevations below 98.8 cm as less suitable, with 21% of pines o ccurring in 67% of the study area. At least 90% of slash pine seed falls within 46 m of the parent tree (Lohrey and Kossuth 1990). Slash pine seed collec tion typically begins in Sept ember once seeds have ripened, although the majority of seeds sown in early fa ll were found to overwinter, with germination having occurred in February (Derr and Mann 1971) Following germination the probability for quick growth and survival decreased in seeds co llected along the continuu m of genetic variation that existed between Pinus elliottii var. elliottii and Pinus elliottii var. densa moving south along the Florida peninsula (Squillace 1966). Recommendations for South Florida slash pine germination in cluded cold stratification of fresh or stored seed for 30 days, with 32-82% ge rmination after 30 days, given 8 to 16 hours of light at 22 C (USDA 1974). According to the Fl orida Division of Forestry, South Florida slash pine germination should occur within 10 to 20 da ys of planting, with a few as late as 30 to 40 101

PAGE 102

days afterwards (Steve Gilly 2006, personal communication). Although Pinus elliottii var. densa will germinate on its own, germination of South Florida slash pine was enhanced when cold storage techniques were used as a pretreatment before planting or direct seeding. Direct seeding and planting of Pinus elliottii var. densa was used to test the working hypothesis that pine survival was at tributable to elevation, as related to hydroperiod. In order to test this, the objectives of this research were to (1) obtain locally genetic seed that was viable in the greenhouse and could be used in seeding a nd planting efforts, (2) test the effects of hydroperiod on slash pine reestablishment by sca ttering and planting South Florida slash pine across five elevation ranges in areas removed fr om seed source, and (3) evaluate whether the HID had the potential to support pine in restored areas and ac hieve a reclamation strategy for pine reestablishment. The early growth of South Florida slash pine was monitored in the restored areas for germination and surv ival over the 2007 and 2008 rainy seasons. Materials & Methods Seed Collection, Harvest, and Utilization Seeds of Pinus elliottii var. densa were collected from within Everglades National Park, mainly from the Long Pine Key and Pine Island ar eas. There were two seed collections. The first occurred from October 11 through Oct 24, 2006 for the first direct seeding and planting studies, while the second seed collection occurred during October of 2007 for the repeated seed scatter. Following cone drop, the cones were collected from the ground and stored in paper bags until harvested. Once harvested from the cones, the seeds were kept in paper envelopes until cold stratification. In order to co ld stratify the seeds, all seeds we re soaked overnight in water at room temperature, separated as floaters and sinkers, drained of excess water, de-winged, and placed into a sealed plastic bag. Seeds remained moist in the refrigerator at approximately 5 C for one week and were flipped every few days. 102

PAGE 103

Of the seeds collected and harvested in 2006, one hundred were used in a greenhouse viability study in January 2007, two thousand pine seeds were used in the first direct seeding study in March 2007, and the remaining 304 seeds were then germinated and planted in December 2007. Cones were collected again during October of 2007 and the seeds were extracted, cold stratified, and direct seedi ng was repeated in December 2007 using 1600 seeds. Weekly water levels were taken for all inund ated plots during the 2007 and 2008 rainy seasons. South Florida slash pine germination and survival were monitored in the field and later analyzed in relation to elev ation or hydrology. Viability Study On December 20, 2006, Pinus elliottii var. densa seeds were soaked overnight, cold stratified, and planted in cone-shaped trays on December 28, 2006. The 100 seeds were sown no more than six mm from the soil surface using Fafard Growers Mix #2 potting soil. The soil remained moist, but not flooded, with daytime temperatures around 21C in the shadehouse. Germination rates were kept by date, until it was cer tain that germination was no longer possible. According to Table 4-1, the germinative capacity of South Florida slash pine seeds was 36% at 21 days following planting. The seed was consid ered viable for this research study. Another viability study was conducted in March 2007 usin g 304 seeds that were grown in the greenhouse for the planting study that occurred in December 2007. Study Site A 32 ha experimental area of suitable el evation was chosen where pine rockland historically occurred in the HID and was rem oved from seed source. On February 1, 2007, the study area was pre-treated with fire. Approxi mately 40% of the site burned, exposing the mineral soil. There were equal conditions of soil exposure am ong plots, except for the lowest elevation treatment which did not burn due to a more moist periphyton layer atop the soil. In 103

PAGE 104

lieu of fire, unburned plots were pr e-treated with a Stihl gas power ed string trimmer. In March 2007, a total of 20 research plots were randomly esta blished across five elev ation treatments with four replicated research plots per elevation range (Figure 4-1). These plots were used for the first direct seeding event and the pine plantings. This was repeated in December 2007 when twenty new research plots were established for the second direct seeding. GIS coordinates for the 40 South Florida slash pine direct seed scatte r and planting plots are listed in Table A-2 of the Appendix. All research plots were 2 m2 quadrats (141.42 cm x 141.42 cm). A map with highaccuracy photogrammetric data +/-0.1 m was used to identify random plots across consecutive elevation ranges, from 60-70 cm, through 100-110 cm. Actual research plots were selected at relative elevations usi ng a Topcon laser level to the neares t centimeter of elevation. Plots encompassed a range of eight centimeters, so that each plot fit within the five elevation ranges with minimal error. There was a two-centimeter gap between each ten cm elevation range as follows: 63-71 cm, 73-81 cm, 83-91 cm, 93-101 cm, and 103-111 cm. True elevations were referenced to a road elevation marker at a later date, hence the discrepancy in elevation ranges as compared to the map. The northwest corner wa s marked with rebar and pvc, while the other three corners were flagged. Germinated a nd planted seedlings were mapped on an x,ycoordinate grid, with the scale beginning at 0,0 from the northwest ern corner pole (Figure 4-2). Then, monthly germination and presence/absence survival observations were accounted for. Direct Seeding Seed scatter one Following the pretreatment burn, 2,000 cold stratified Pinus elliottii var. densa seeds were divided into 80 vials of 25 s eeds each for scatter on March 14, 2007. There were 20 floating seeds to five sinking seeds per vial, such that 25 seeds were scattered into each quadrant of the 2 104

PAGE 105

m2 research quadrat, for a total of 100 seeds per plot. Since there were 20 research plots distributed across five treatments, with four rep licate plots per treatment, there were 400 seeds scattered per elevation range. South Florida slas h pine seeds were scattered within the plot where there was soil, avoiding areas of exposed roc k. Height and root collar measurements were taken every three months, beginning at nine m onths post-scatter and ending in December of 2008. Height measurements for all pines in di rect seed scatter and planting studies were measured to the tallest green needle. Seed scatter two On Dec 19, 2007, a laser level was used to establish 20 new replicate plots and Pinus elliottii var. densa seeds from the Fall 2007 crop were cold st ratified. Due to lower cone drop by November, 80 seeds were scattere d per plot on December 27. This consisted of 80 vials of 20 seeds for each plot quadrant, with 16 floaters and four sinkers per vial. Therefore, 320 seeds were scattered per elevation treatment, for a to tal of 1,600 seeds over 20 plots. Germination and survival were measured every month, as with the first scatter, and height and root collar measurements were taken once at 12 months in December 2008. Comparing seeding events It is important to note that despite different numbers of s cattered seeds between scattering events, the results of both seeding events were normalized for floate rs or sinkers, given that both scatters contained 20% sinkers a nd 80% floaters. It was then possible to obtain a weighted, expected germinative capacity of 48% for the direct seeding events, as taken from 41% x 0.80 floaters + 72% x 0.20 sinkers. This was for optimal greenhous e conditions; not the field. Pine Planting Following the first scatter, South Florida slash pine seeds were planted and grown in cone trays in the shadehouse for nine months. These plantings were monitored for germination in the 105

PAGE 106

second greenhouse viability study during March and April of 2007. Pinus elliottii var. densa seedlings were fertilized in the greenhouse once per month, at a rate of 5.0 ml Miracle Grow fertilizer per 3.8 L of water, from May through N ovember of 2007. Since the first direct seeding germination results were extremely low, eight pines were planted into each of the 20 original research plots, as two pines per quadrant on December 17, 2008. In order to plant 160 South Florida slash pi nes in the limestone bedrock, a jackhammer was used to drill the holes. The pines were pl anted in a thin layer of restored soil and the limestone rock below, such that the roots did not J-hook and seedlin gs were flush with the rock. Prior to planting, height, root co llar diameter, and health measurements were recorded. Survival, height, and root collar diameter were record ed every three months following planting for one year. Growth rates were calcula ted and health was recorded as live, poor, or dead. Chlorosis was noted if present. Water Levels Water level depths were recorded in centimet ers for all pertinent research plots once a week over the 2007 and 2008 rainy seasons. Seedlings in the first direct seeding event had two seasons of inundation, while the planted and sec ond direct seed scattered pines underwent one season of flooding during 2008. Research plots in the lowest elevation were inundated periodically throughout the enti re rainy season of May through November, while the highest research plots remained relativ ely free of standing water. Statistics Germination rates for both seed scattering events were run as one-way Analysis of Variance (ANOVA) statistics in SPS S with elevation as the indepe ndent variable, and also as a two-way ANOVA with elevation and time of seed scat ter as fixed factors. For the first direct seed scatter, one-way ANOVAs were run in or der to determine if there were significant 106

PAGE 107

differences in height or root collar diameter data by elevation treatment at 12 months and 18 months, and if there was signifi cant growth from nine to 18 m onths using elevation as the independent variable. One-way ANOVAs were al so run for the second direct seed scatter measurements of height and root collar at 12 months. The 12 month results were compared for both seed scatters in a two-way ANOVA, which used elevation and seed scatter as independent variables. As for planted pines, one-way ANOVAs were run for height a nd root collar diameter measurements at 18 months, as well as for growth rates over nine to 18 months, with elevation as the independent variable. In order to compar e planting with seed sc atter, two-way ANOVAs were used to compare height, root collar, and grow th rates of height and r oot collar diameter as the dependent variables, while elevation and met hod of establishment (i.e. plant, scatter) served as fixed factors. All one-way ANOVAs used Tukeys post-hoc te st and all significant differences were tested at alpha = 0.05. The data were log transf ormed to normalize data sets for seedling root collar measurements among direct seed scatters at twelve months, height a nd root collar diameter measurements at 18 months, and height growth be tween methods of establishment at 18 months. Logarithmic means were back-transformed fr om the log10 values using the equation: z= exp[y*ln10+0.5*variance2*(ln10)2] where z was the back-transformed value, y was the log10 value, and variance was of the log10 value (Web ster and Oliver 2001). Nonparametric tests were necessary to analyze survival rates at 12 m onths for direct seed scatters and 21 months for plantings, following one season of inundation. An alysis of Covariance (ANCOVA) tests were analyzed in SPSS as univariate general linear models with initial height a nd initial root collar diameter measurements for planted pines as covari ates for dependent variables of survival at 21 months, final height and final r oot collar diameter measurements at 18 months, and height and 107

PAGE 108

root collar diameter growth rates over nine months Elevation served as the independent variable in all cases. Results Greenhouse Viability Study By Day 8 of the greenhouse viability study init iated at the end of 2006, four seeds out of the 100 that were planted had ge rminated (Table 4-1). By Day 21 on January 17, 2007, the seeds had reached germinative capacity because 36% of the seed had germinated and no more seeds germinated past that date. Another viabi lity study was conducted in March of 2007, when 304 seeds were planted in the greenhouse and grown as stock for the pine plantings in December of 2007. Half of the seeds that were planted were floaters, which were skimmed off the surface of the water following the overnight soak prior to co ld stratification. The other half of the seeds were sinkers that had submerged to the bottom of the can overnight. After ten days, on March 25, germinative energy (as 50% germination) was reached for the sinkers at 57 germinated seeds (Figure 4-3). By day 13 floaters reached germinative energy with 30 seeds germinated. Day 35 was the last observa tion that any new seedlings had germinated as both sinkers and floaters, such that germinative capacity was reached at 109 and 63, respectively. This date was April 19, when 41% of floaters and 72% sinkers had germinated. Survival of these germinated pine seedlings through August 21, 2007 was 97% for floaters and 98% for sinkers. Therefore, survival was strong enough that 160 of these pine seedlings were planted in the first 20 research plots in December 2007. Seeds collected in the October 2007 crop were viable and used for the second direct seed ing that also occurre d during December 2007. Germination in the Field The number of germinants in each eleva tion range and germinative capacity for both seeding events is shown in Table 4-2. For the first seed scatter, germination was a minimum 108

PAGE 109

average of 2% in the highest 103 to 111 cm eleva tion range and a maximum of 8% in the lowest 63 to 71 cm elevation range. The second seed sc atter also had low germ ination rates, with a minimum of 2% germination at the 93 to 101 cm elevation and a maximum average of 4% in the 63 to 71 cm elevations. It was rare that any ne w pines were discovered in the research plots after three months post-scatter. There were no sign ificant differences found between treatments or among direct seeding events. That is, one-way ANOVAs for perc ent germination as a function of elevation for scatter one and scatter two s howed no significance differences. There was also no significant difference for germination when a two-way ANOVA was performed using elevation and seasonality as fixed factors (Table 4-3). Survival After one rainy season in the HID, both direct scatters at 12 months and planted pines at 21 months suffered complete mortality at the lowe st 63 to 71 cm elevation, with greater than 62% survival in all subsequently high er elevation ranges (Table 4-4) Survival ranged from 62% to 100% for the direct seed scatters at 12 months of age and 62 % to 66% for plantings at 21 months of age. No pines at the lowest elevation survived inundation over th e 2007 rainy season, as illustrated by the mortality curve shown in Figure 4-4 for 32 planted pines per elevation. Several attempts to transform the data using log and arc sin were unsuccessful, so a non-parametric Kruskal-Wallis test was conducted to generalize survival trends. There were significant differences for survival between the lowest el evation range for the tw o scatter and planting efforts, as evidenced by the assigned ranks (Table 4-5). The ranks of the upper four elevation ranges were considerably greater than the ra nk for the 63 to 71 cm elevation range. Direct Seeding Comparisons When height and root collar measurements we re compared at 12 months of age using a two-way ANOVA, there was no significant differen ce between pine seed s catters that occurred 109

PAGE 110

in March of 2007 or December of 2008, or between the four highest elevations when height and root collar diameter were examined as the dependent variables (Table 4-6). There were also no significant differences for one-way ANOVAs of height and root collar when dependent upon elevation for each direct seeding event for pines at 12 months. Heights ranged from an average of 7.2 to 8.9 cm for the first seed scatter, and 9.1 to 10.8 cm for the second direct seeding (Figure 4-5), while root collar measurements ranged fr om 1.2 to 1.5 mm for the first seed scatter and 1.3 to 1.7 mm for the second seed scatter (Figure 4-6). Planting verses Direct Seeding The two-way ANOVA that compared elevati on and method of esta blishment (planting verses seeding) indicated that there were significant differences between scatter and planting measurements for logarithmic values of height and root collar diameter at 18 months of age, as well as root collar growth of pines (Table 4-7). Similarities occurred among height growth for scattered and planted pine treatments and among elevation treatments for height, root collar, growth of height, and growth of root collar. One-way ANOVAs also supported these inferences of similarity between elevations for both plan ting and scatter measurements, when height, root collar diameter, and growth rates se rved as the dependent variables. After 18 months, the average heights of seeded pines ranged from 7.4 cm to 9.7 cm, while the root collar diameter ranged from 1.7 mm to 2.5 mm. Planted pine measurements were significantly higher than those th at were seeded, with height measurements ranging from 14.1 cm to 15.6 cm, and root collars of 4.7 mm to 5.4 mm (Figure 4-7). Over ni ne months of growth, height did not change much, if at all, for seeded nor planted pines. Av erage height growth was negligible, ranging from .7 to 1.2 cm for directly seeded pines, and .0 to 1.9 cm for planted pines (Figure 4-8). However, root collar diameter growth rates did significantly differ between establishment methods, ranging from 0.02 to 1.1 mm fo r directly seeded pine s, as compared with 110

PAGE 111

an average of 2.2 to 3.0 mm root collar growth rates among planted pi nes. A logarithmic transformation of root collar gr owth rates (as log(growth rate + 0.8)) indicated no significant differences for one-way ANOVAs of growth rate s by elevation for scat ter one or planting; however, the two-way ANOVA was si gnificant for method of establishment, such that the growth rate for root collar diamet er of planted pines was significan tly higher than the growth rate of scattered pines (Table 4-7). Planting Covariate Effects on Survival and Growth When initial measurements of height and root collar diameter for pines at planting were used as covariates for survival, there were si gnificant effects on survival, given an alpha of 0.05 (Table 4-8). Initial height measurements also ha d a significant effect as a covariate for both the final height measurement at 18 months and heig ht growth over nine months following planting (Table 4-9). Similar results were found for root collar diameter growth, as the initial root collar diameter measurement had a significant effect as a covariate when compared with the final root collar diameter and root collar diameter growth as dependent variables. For all ANCOVAs, elevation was not significant as the independent variable, as previously discussed. Hydrology and Survival Despite differences in hydroperiod seasonal ity and frequency, water depth and duration appeared to affect seedling survival, particularly in the lowest elevated research plots. During the 2007 rainy season, scattered pine seedlings withstood about four weeks of constant inundation above needle height befo re all pines suffered mortality (F igure 4-9). The same is true for the second round of scattered seeds, wh ich underwent inundati on during the 2008 rainy season (Figure 4-10). Planted pines about 1.5 years of age were able to w ithstand at most eight weeks of constant inundation at or above needle height before all pines were considered dead (Figure 4-11). Pines in the 73 to 81 cm range survived the first year of inundation, where there 111

PAGE 112

was a high frequency of inundation of flooding durations that were two weeks or less, with water depths below seedling height. Scattered seeds th at were exposed to two seasons of inundation in the 73 to 81 cm elevation range had only a 5% surv ival, compared with planted pines of the same age that had a 62% survival, yet underwent only one season of i nundation (Table 4-4). Discussion Everglades National Park achieved the goal of obtaining and rearing viable, local genetic seed stock for reestablishment of Pinus elliottii var. densa in the Hole-in-the-Donut restoration area. Overall, germinative capacity results were similar to the 47% seed viability observed by Mayo (2000) for Pinus elliottii var. densa under greenhouse conditions. The higher germinative capacity of seeds that sank was comparable to Pinus elliottii var. elliottii germination rates when stored from three to five weeks (McLemore 1975) On-the-contrary, floaters were generally smaller, of yellow coloration, and drier; an i ndication that there were empty seeds (Barnett 1996). The germinative energy observed in floaters also took three days longer to reach when compared to sinkers. Though germ inative capacity and energy of floaters was lower, utilizing the small-seed fraction may actually increase gene tic variation, which could be important for the reestablishment of Pinus elliottii var. densa (Silen and Osterhans 1979, Barnett 1996). A future recommendation would be to plant floaters as two per container in th e greenhouse and to not scatter floating seeds in the restored area. While empty and less viable floating seed s may have played a role in poor field germination, the lack of seed treatment and harsh field conditions also affected seedling germination and survival. There was no fungicide applied to the seeds, an indication that herbivory by birds, small mammals, and ants, or drought may have hindered germination (Derr and Mann 1971). Mayo (2000) found that rocky and sandy seed sources of Pinus elliottii var. densa respectively had 36% and 33% establishment rates after three years following the 112

PAGE 113

broadcast of seeds coated with Gustafon-42 bird and rodent repellent. Additionally, this study found that South Florida slash pine seeds that we re not treated had only a <1% establishment of viable seed after two years, as attributed to predation. The us e of Gustafon-42 required a special permit to be used in the Everglad es and was not an option at the time seeds were scattered. It was also possible that damping off occurred in the scattered seedlings ; however, no signs of seedling die-off were observed in the monthly inventories. Al though pines were mapped such that the mortalities of recorded germinants were logged for individua l seedlings, damping off disappearance of newly germinated, unrecorded seedlings could have occurred during the windows between inventories. Though the direct scattering was co mpletely replicated, seasonality of seed scatter and seed respiration were not significant factors among scatter one in March 2007 and scatter two in December 2007. Phenologically, seedfall of South Flor ida slash pine in the Everglades begins in October, similar to Pinus elliottii var. elliottii (Lohrey and Kossuth 1990). Slash pine seeds may fall as late as March, with germination occu rring within two weeks, given adequate soil moisture. In the Everglades, germination would therefore occur during the dry season. Despite observations that the soil was dr ier at the time of scatter in March verses December, seed germination rates were similar, o ccurring within three months. As for respiration, five weeks of Pinus elliottii var. elliottii storage following cone collection did increase germination from 49% at one week to 98% at five weeks (McLemore 1975). Therefore, those seeds scattered in Decem ber were at an optimal point for germination, while those in March were exposed to three additional months of seed respiration, which apparently did not have an effect on germination. To scatter at the onset of the rainy season in late spring or early summer may be beneficial at the highest elevations, yet risky in lower 113

PAGE 114

elevation where flooding could affect germination and initial survival. While the effects of respiration may have been obscured by the eff ects of phenology, there was also high variability of tree heights and root collar di ameters for both direct seedings. Before growth rates were taken in the 63 to 71 cm elevation, hydroperiod did have an effect on slash pine survival and reestablishm ent for both direct seeding and planting, which indicated that pine survival ove r time was possible in higher elevations for areas removed from a pine seed source. Clearly, the lowest elevation was unfavorable for pine reestablishment, as none of the South Florida slash pine seedlings survived the rainy season of the year they were reestablished in the field. Though not significantly different, germination rates were highest at the 63 to 71 cm elevation range for both direct scatters, where periphyton may have acted as a nurse bed. Of course, these pines did not su rvive inundation through the rainy season and reestablishment was limited by hydrology as related to elevation. It is exp ected that the 73 to 81 cm range will be a transitional area, given low pi ne survival of scattere d pines after the second growing season. Recently germinated pine seed lings in a pine remnant of HID also had mortalities upwards of 60% to 70% as a result of flooding (OHare and Dalrymple 2005). For scattered seeds and plantings that did survive 18 months in the field, there was minimal or negative height growth for Pinus elliottii var. densa seedlings despite consistency in measurements taken from the tallest green (or live) needle of the seedlin g. Negative growth was therefore attributed to the loss of the initial embryoni c needles for replacement by apical seedling needles, as well as browning of needle tips that occurred over the course of this study. Those pines that had a negative r oot collar diameter growth ove r the growing season typically succumbed to mortality, as it is important that pi ne seedlings in the grass stage develop a thick bark rather than height growt h, given this is an adaptation to fire (Little and Dorman 1952). 114

PAGE 115

Initial seedling size as covariates for height and root collar diam eter at the time of planting had effects on the growth and survival of Pinus elliottii var. densa in the HID that were unrelated to elevation. That is, larger seed lings had higher growth and surviv al rates than smaller seedlings. In stands of Pinus elliottii var. elliottii herbicide site preparat ion and release treatments were used to control weeds and re-allocate nutrie nts, while fertilizer additions further improved slash pine yield. Though fire management and mechanical weed removal was used for site preparation of research plots, which controlle d competition of grasses and understory species like Baccharis fertilizer was not applied (Wade 1983). It has been shown that herbicide release in conjunction with fertilizer increased slash pi ne production on Pomona fi ne sands (Colbert et al. 1990). Annual fertilization and weed c ontrol can yield a stand volume of 55.6 m3/ha on fouryear old slash pine. These values were comparable to volumes of 30.1 m3/ha and 35.9 m3/ha for fertilization or herbicide alone, respectively, and 6.9 m3/ha for a control treatment of no fertilizer and no herbicide. Fertilizers also enhanced site quality and have been shown to temporarily shift the site index of slash pine at 18 years from 22 to 26 me ters with a treatment of fertilizer and weed control (Jokela 2004). Given a control value of 22 meters, Pinus elliottii var. densa does not grow as tall as Pinus elliottii var. elliottii reaching approximately 17 meters in height (Landers 1991). This was perhaps an adaptation of Pinus elliottii var. dens a to minimize high wind damage from storms and hurricanes, a frequent disturbance to South Florida. Due to exotic plant invasions and the pine rockland ha bitat listed as an Environmenta lly Endangered Land, stands of South Florida slash pine are typi cally not fertilized, despite the fact that growth and problems with chlorosis could potentially be improved. As demonstrated in Chapter 2, the HID restored soils may be limited in total nitrogen; however, they remain higher than normal in total 115

PAGE 116

phosphorus, such that phosphorus additions could lead to the re-invasion of Schinus terebinthifolius, thereby defeating the purpose of restoration. The Hole-in-the-Donut restoration was capable of supporting South Fl orida slash pine and this study provided insight for a restoration strate gy that the National Park Service can utilize. The survival of South Florida slash pine over the rainy season indicated th at it was possible to reestablish pine within the historic range, despite restorativ e lowering of the elevation and increased hydroperiod following restoration. Ther e may be a balancing act that has occurred here, given that adjacent Taylor Slough currently r eceives less water than indicated by historical records (Armentano et al. 2006). Given that rece nt hydrological changes were affected by water management, future changes in water manageme nt could play a role in the hydroperiod over time. Additionally, there was a pooling effect along HID edges, where water flow met the natural area that was 15 to 30 cm higher in elevat ion. Lastly, as these tr ees grow there could be an evapotranspiration effect that further lowers the water table; the opposite effect observed when pine flatwoods are harvested and the wa ter table rises (Sun et al. 2001). Time and additional restoration effo rts will provide more insight on exactly where Pinus elliottii var. densa will occur and form a dominant canopy in the restored HID. As for a pine reestablishment strategy in th e HID, planting was a more efficient technique than seed scatter. After one y ear of growth in HID, the plante d pines had substantially greater height and root collar diameters, as well as ro ot collar growth rates. This was because these pines were germinated and grown under optimal conditions in the greenhouse for nine months before planting, whereas the direct ly seeded pines germinated onsite last March without adequate water or fertilizer. Pinus elliottii var. elliottii was shown to have a better start when planted, as there were more planted pines in the larger diameter classes when compared to seeded pines at 116

PAGE 117

15 years of age (Campbell 1980). For Pinus elliottii var. densa in the HID, 500 trees per hectare would be a desired density after 25 years and sh ould be accounted for when plantings are made on a large-scale reestablishment effort. A prior study in Miami-Dade County found th at after three years of growth on pine rockland with fire treatment, there was 76% su rvival of planted South Florida slash pine seedlings and 36% survival of direct seeded se edlings from a local rocky South Florida slash pine seed source (Mayo 2000). Ecotype also differed, as seeds of a rocky source averaged 23.6 cm in height verses 12.5 cm for a sandy seed source after three years of growth on a rocky substrate. Therefore, autecol ogical and genetic differences were important considerations when restoring South Florida slash pine, such that the Hole-in-the-Donut Project has been committed to reestablishing and preservi ng the local, rocky, more mesic to hydric seed source found along the southern portions of the Miami Rock Ridge. Planting native, better-adapted ecotypes like South Florida slash pine will offer protection from hurricanes, adaptations to fire, and tole rance to flooding. Whether in the Hole-in-theDonut, Miami-Dade County, Big Cypre ss, or the Florida Keys, there is an appropriate restoration strategy for South Florida slash pi ne in rocklands or sandy areas, as well as for hydric, mesic and xeric conditions. Where fire is not an option, mechanical and chemical site preparation may decrease the vulnerability of invasion by exotic species, thereby promoti ng regeneration of slash pine rockland remnants. Short hydroperiods of ten weeks or less were optimal for Pinus elliottii var. densa seedling survival and early gr owth toward reestablishment, particularly for one-year old containerized seedlings. The National Park Service will adapt these recommendations for South Florida slash pine in future reestablishment efforts, which are further summarized in Chapter 5. Outside of the 117

PAGE 118

Everglades, only 920 ha, or <2% of pine rock land remain in Miami Dade County, as 126 fragments (OHare and Dalrymple 2006). South Florida slash pine is only one component of these ecosystems, as there are rare and endange red shrubs and herbaceous species which must also be preserved and restored in order to sustain the ecological integrity of this critical ecosystem. With time and fire management, pine rockland species diversity should increase in areas within the Hole-in-the-Donut considered suitable for pine, as illustrated in Chapter 3. Conclusion Preliminary data support the hypothesis that Sout h Florida slash pine is more likely to survive at high elevations, given that elevation was reflective of hydroperiod and that flooding caused mortality of all South Florida slash pine seedlings in the lowest 60 70 cm elevation. Germination rates, though less than eight percent, were similar fo r seeds scattered in March and December; however, planted pines had significantly greater height and root collar growth when compared to seeded pines of the same age. Th e survival of South Florida slash pine across the higher elevation ranges was similar to the historical occurrence of Pinus elliottii var. densa that existed in the Hole-in-the-Donut prior to farmi ng, despite lower elevations associated with the restoration process. Although survival by elevation was not significa nt in the upper four elevation ranges, growth rates and survival over time may reflect a transitional area in the 70 to 80 cm range more characteristic of a hydric pineland community. Everglades National Park recommends that the reestablishment of Pinus elliottii var. densa in the Hole-in-the-Donut involve the planting of the local, rocky, mesic seed source as containerized seedlings and that the appropriate South Florida slash pine ecotype be considered fo r establishment throughout the species range. 118

PAGE 119

Table 4-1. Germination of South Florida slash pine during the initial greenhouse viability study conducted in January 2007. Date Day Number of seedlings January 4 8 4 January 6 10 11 January 7 11 16 January 8 12 23 January 9 13 24 January 12 16 31 January 13 17 32 January 14 18 35 January 17 21 36 Table 4-2. Number of germinants and germ inative capacity between elevations and among direct seeding events. Direct Seed Scatter 1 Direct Seed Scatter 2 Elevation (cm) Number of Germinants Germinative Capacity(%) Number of Germinants Germinative Capacity (%) 103-111 9 2.2 11 3.4 93-101 13 3.2 6 1.9 83-91 10 2.5 10 3.1 73-81 21 5.2 11 3.4 63-71 33 8.2 12 3.8 Table 4-3. Two-way ANOVA of germinative cap acity for direct seed scatters 1 and 2 by elevation treatment, with percent germina tion as the dependent variable. Elevation and seed scatter were fixed factors. Seed scatter 1 occurred in March 2007 and seed scatter 2 in December 2007. Source df F Sig. Corrected Model 9 0.879 0.554 Intercept 1 35.561 0.000 Elevation 4 1.099 0.375 Seed Scatter 1 0.891 0.353 Elevation Seed Scatter 4 0.655 0.628 Error 30 Total 40 Corrected Total 39 119

PAGE 120

Table 4-4. Survival data of Sout h Florida slash pine for direct se eding scatters one and two at 12 months, as well as scatter one and pl anted pines at 21 months of age. Scatter 2 at 12 months Scatter 1 at 12 months Scatter 1 at 21 months Plant at 21 months Elevation (cm) No. alive Survival (%) No. alive Survival (%) No. alive Survival (%) No. alive Survival (%) 103-111 10 96 9 100 4 48 21 66 93-101 5 92 10 80 2 30 21 66 83-91 10 100 10 100 8 86 21 66 73-81 9 90 18 62 2 5 20 62 63-71 0 0* 0 0* 0 0* 0 0* All pines in the lowest elevation did not surviv e. There were 320 seeds per elevation in scatter 2, 400 seeds per elevation in scatter 1, and 32 seedlings planted per elevation. Table 4-5. Kruskal-Wallis test results for surviv al rates of seeded and planted South Florida slash pines following exposure to one season of inundation. Scatter 1 Scatter 2 Planting Elevation (cm) N (16 Plots) Rank N (19 plot s) Rank N (20 Plots) Rank 103-111 3 13.0 4 11.9 4 12.6 93-101 4 8.5 3 11.0 4 12.6 83-91 2 13.0 4 13.5 4 13.6 73-81 3 8.3 4 11.4 4 11.1 63-71 4 3.0 4 2.5 4 2.5 Asymp. Sig. 0.03 0.01 0.04 Pines were 12 months of age for direct scatter survival and 21 m onths of age for plantings. All pines had been reestablished in the HID for 12 months when survival data were taken. 120

PAGE 121

Table 4-6. Height and root co llar comparisons in a two-way ANOVA, with elevation and timing of seed scatter as fixed factors. Dependent variables were height at 12 months or log 10 root collar diameter at 12 months. Re sults were significant for an alpha = 0.05. Height at 12 months Root collar diameter at 12 months Source df F Sig. F Sig. Corrected Model 7 0.763 0.620 1.505 0.179 Intercept 1 471.849 0.000 90.133 0.000 Elevation 3 0.598 0.618 1.450 0.235 Seed Scatter 1 2.666 0.107 2.824 0.097 Elevation Seed Scatter 3 0.366 0.778 0.514 0.674 Error 73 Total 81 Corrected Total 80 Table 4-7. Two-way Analysis of Variance for height and root collar measurements of scattered and planted South Florida slas h pines at 18 months, as well as growth over 9 months. Dependent variables were log 10 height, log 10 root collar diameter, or log 10 root collar diameter growth, while elevation and method of establishment (i.e. direct seed scatter or planting) were fixed factors. Results were significant for an alpha = 0.05. Height at 18 months Root collar diameter at 18 months Root collar diameter growth after 9 months Source df F Sig. F Sig. F Sig. Corrected Model 7 6.120 0.000 7.236 0.000 5.114 0.000 Intercept 1 1979.674 0.000 239.262 0.000 39.633 0.000 Elevation 3 0.628 0.599 0.121 0.947 0.436 0.727 Method 1 28.307 0.000 40.557 0.000 24.916 0.000 Elevation Method 3 0.255 0.857 0.465 0.707 1.182 0.320 Error 107 Total 115 Corrected Total 114 121

PAGE 122

Table 4-8. Covariate effects of initial height or log 10 of initial root collar diameter measurements with elevation treatment. The dependent variable was survival, while elevation and initial measurements were fixed factors. Results were significant for an alpha = 0.05. Survival Survival Source df F Sig. df F Sig. Corrected Model 4 4.743 0.001 4 3.064 0.019 Intercept 1 0.013 0.911 1 10.290 0.002 Initial Height (cm) 1 18.857 0.000 Log 10 of Initial Root Collar Diameter (mm) 1 12.147 0.001 Elevation Treatment 3 0.051 0.985 3 0.114 0.952 Error 123 123 Total 128 128 Corrected Total 127 127 Table 4-9. Covariate effects of initial height or log 10 of initial root collar diameter measurements with elevation treatment. The dependent variables for initial height were log ten of final height or height grow th, while the dependent variables for log 10 initial root collar diameter we re log 10 final root collar diam eter or log 10 root collar diameter growth. Elevation and initial m easurements were fixed factors and results were significant for an alpha = 0.05. Final Height (cm) Height Growth (cm) Final Root Collar (mm) Root Collar Growth (mm) Source df F Sig. F Sig. df F Sig. F Sig. Corrected Model 4 3.731 0.008 2.513 0.047 4 8.675 0.000 2.786 0.032 Intercept 1 118.759 0.000 5.784 0.018 1 51.079 0.000 0.504 0.480 Initial Height (cm) 1 12.118 0.001 4.993 0.028 Log 10 Initial Root Collar Diameter (mm) 1 32.258 0.000 9.082 0.003 Elevation 3 1.675 0.179 1.227 0.305 3 1.430 0.240 1.202 0.314 Error 84 84 Total 89 89 Corrected Total 88 88 122

PAGE 123

Figure 4-1. Locations of the two sets of 20 res earch plots by elevation ra nge in the Hole-in-theDonut study area. There were four plots established across five elevation treatments for each set of research plots. GIS coordi nates are listed in the Appendix, Table A-2. 123

PAGE 124

0 144 01x-coordinate (cm)y-coordinate (cm4 4 ) Figure 4-2. An example of the x-y coordinate grid used to map pine seedlings in all study plots. This figure shows the direct seeding results after three months in 73-81 cm, Elevation 4 Plot 2. Greenhouse Seedling Germination and Survival Mar. Aug. 20070 20 40 60 80 100 12016-M a r 5 -A p r 25-Apr 1 5Ma y 4-Jun 2 4J un 14-Ju l 3 -Aug 23-AugDateNumber of Seedlings floaters sinkers Figure 4-3. Cumulative numb er of germinated South Florida slash pine seedlings and survival in the greenhouse over nine months. These seedlings were used in the second viability study and for planting in the first set of research plots in December 2007. 124

PAGE 125

0 4 8 12 16 20 24 28 32D e c 0 7 J a n 0 8 Fe b 0 8 M a r 0 8 A p r 0 8 M a y 0 8 J u n 0 8 J u l0 8 A u g 0 8 Se p 0 8 O c t0 8 N o v 0 8 D e c 0 8Date (month-year)Number of Live Seedlings 63-71 73-81 83-91 93-101 103-111 Figure 4-4. Survival of pine plantings by elevation following one year in the ground. In December 2008, pines were approximately 21 months old. 125

PAGE 126

Elevation (cm) 73-8183-9193-101101-113 Seedling Height (cm) 0 2 4 6 8 10 12 14 Scatter 1 Height Scatter 2 Height Figure 4-5. A comparison of height between elevati ons and direct seed scat ters 1 and 2 at twelve months after seeding of Pinus elliottii var. densa Error bars are one standard error above the mean. Seed scatter 1 occurred in March 2007, while scatter 2 occurred in December 2007. 126

PAGE 127

Elevation Range (cm) 73-8183-9193-101103-111 Root Collar Diameter (mm) 0.0 0.5 1.0 1.5 2.0 2.5 Scatter 1 Root Collar Scatter 2 Root Collar Figure 4-6. A comparison of root collar diameter between elevations and direct seed scatters 1 and 2 twelve months after seeding of Pinus elliottii var. densa Logarithmic means were back-transformed to arithmetic units and standard error bars are in arithmetic units. Data were statistically si milar at a 0.05 significance level. 127

PAGE 128

Root Collar Diameter (mm) 0 2 4 6 Scatter Root Collar Plant Root Collar Elevation Range (cm) 73-8183-9193-101103-111 Seedling Height (cm) 0 2 4 6 8 10 12 14 16 Scatter Height Plant Height a b b b a a a b a a a a b b b b Figure 4-7. Pinus elliottii var. densa height and root collar me asurements for scattered and planted pines by elevation treatment at 18 months of age. Logarithmic values were converted to arithmetic units and error bars are one standard er ror above the mean. Height and root collar measurements are compared separately, with similar letters statistically similar at alpha = 0.05. 128

PAGE 129

Elevation Range (cm) 73-8183-9193-101101-113 Growth of Height (cm), Root Collar Diameter (mm) -3 -2 -1 0 1 2 3 4 Plant Height Growth Scatter Height Growth Plant Root Collar Growth Scatter Root Collar Growth Figure 4-8. Growth rates of planted and seeded South Florida slash pines by elevation over nine months. Height growth and root collar diameter growth are depicted here, without statistics. All pines were m easured at 18 months of age. 129

PAGE 130

0 5 10 15 20 25 30 351 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 23 2 5 27 2 9 3 1Mid-April through mid-November (time in weeks)Water depth (cm) Number of pine seedlings Water Depth (cm) Pine Survival Figure 4-9. A comparison of water depth and total live seedlings in the 63 to 71 cm elevation range for the first direct seeding of Pinus elliottii var. densa during the 2007 rainy season. 0 5 10 15 20 251 3 5 7 9 11 1 3 15 1 7 19 2 1 23 2 5 27 2 9 3 1Mid-April through mid-November (time in weeks)Water depth (cm) Number of pine seedlings Water Depth (cm) Pine Survival Figure 4-10. A comparison of water depth and total live seedlings in the 63 to 71 cm elevation range for the second direct seeding of Pinus elliottii var. densa during the 2008 rainy season. 130

PAGE 131

131 0 4 8 12 16 20 24 28 321 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31Mid-April through mid-November (time in weeks)Water depth (cm) Number of pine seedlings Water Depth (cm) Pine Survival Figure 4-11. A comparison of water depth and total live seedlings in the 63 to 71 cm elevation range for planted Pinus elliottii var. densa seedlings during the 2008 rainy season.

PAGE 132

CHAPTER 5 SUMMARY Reestablishment of South Florida Slash Pine Research Summary The Hole-in-the-Donut is a 2,670 hectare rest oration project, where the majority of restored lands have reverted to wetland, in both species composition and hydrology (OHare and Dalrymple 2005). Areas higher in elevation that were historica lly pine rockland have been reclaimed by species characteristic of this community type (i.e. Ardisia escallonioides, Dodonaea viscosa, Morinda royoc ) and over 3,000 South Florida slash pine seedlings recruited in lands adjacent to pine rock land (OHare and Dalrymple 2006). Areas far removed from pine seed source, yet favorable for South Florid a slash pine and potentially pine rockland reestablishment, were selected to test the effects of direct seed ing and planting of pine. This research supported the National Park Service mi ssion statement, which includes the preservation and protection of this wilderness area. It was esse ntial to reestablish this dominant overstory tree in order to further support a globally imperile d rare and endemic subt ropical plant community (Snyder et al. 1990). An understanding of the study site was neces sary before recommendations for South Florida slash pine could be made. The lands wh ich were currently driven towards pine rockland, were in fact historically pine rockland, given the aerial interpretation of historical photos. Despite the lowering of ground surface elevati ons, there were edaphi c, biological, and hydrological characteristics that in dicated and delineated favorable areas for South Florida slash pine reestablishment. Significant increases in soil depth and decreases in soil pH, with decreases in elevation, were indicative of the upland to wetland marl formation processes associated with the transition to marl prairie. Though not sign ificant, increased nitrogen limitation relative to 132

PAGE 133

carbon demonstrated that nitrogen consumption increased within the microbial community in response to anaerobic conditions. The phosphor us limitation at the lowest elevation was indicative of the oligotrophic na ture of the Everglades wetland ecosystem (Mitsch and Gosselink 2000). Soils in the higher eleva tion treatments of 83 cm and great er were upland soils, inundated less than five weeks of the year and suita ble for slash pine survival and growth. Though upland in nature, South Florida slash pine saplings grew best under poorly drained soil conditions when the root system was partially flooded, as was demonstrated in the mesocosm experiment. Growth of fully flooded South Florida slash pine saplings was slower, with greater needle loss and lower biomass measur ements. The potential nutrient limitations at the lowest elevation treatment and length of hydrol ogical wetting events fo r an average year in the HID created an unsuitable environment for Sout h Florida slash pine. Especially since newly germinated and one-year old seedlings succumbed within four and eigh t weeks of constant inundation above needle height, re spectively. This inherently de fined the hydroperiod for South Florida slash pine during the establishment phase. Given the failure of South Florida slash pine seedling survival in the HID at the 63 to 71 cm elevation treatment, elev ations below 70 cm with ten or more weeks of flooding were considered unsuitable for slash pine reestablishment in the Hole-in-the-Donut. The similar range of present-day and historic Pinus elliottii var. densa occurrence may or may not hold up with time, depending on the area slated as marginal. It is possible that the current range may remain similar to the historic pine rockland range, partially due to lower water flows through Taylor Slough as combined with the effects of lowered restored elevations across the landscape (Armentano et al. 2006). Perhaps a vigorous program of reforestation should be undertaken while the hydrology gover ning Taylor Slough is at its curr ent stage. Then pine could 133

PAGE 134

withstand increases in hydroperiod once established, particular ly if the evapotranspiration of adult South Florida slash pine has an effect on th e water table. There was also drainage in the higher elevated, northern portions of restored HID that has caus ed pooling along the eastern and southern edges of the restoration as water flowed toward the natural area. Not only did rockplowing and restoration cr eate a lowered elevation, these processes homogenized the topography by fl attening out the landscape, which may have caused the differences in historic verses current marginal and suitable areas for South Florida slash pine (Figure 5). Only 26 ha of suitable area for Pinus elliottii var. densa has been restored to date; however, there are areas of former pine rockland slated to be restored within the next five years. Once restored, it is likely that the hydrology and other site conditions will promote South Florida slash pine growth and long-term survival. Trans itional areas of marginal suitability for pine are more difficult to predict in terms of pine surviv al and time will sort out the South Florida slash pine range in the HID. Then comparisons of current presence can be made with the 15% historical occurrence. The adaptation of South Florida slash pine to a high water table has be en advantageous to growth in the partially flooded treatment, which showed lower nitrogen and phosphorus concentrations in the needles when compared to the control. The c ontrol did not have any exposure to a simulated groundwater table, which explained the lower growth at the theoretically high elevation. It is likely over time that the taproot of pines at the 103 to 111 cm elevation in the HID will penetrate the limerock down to the water table. Slash pines that were able to tap the water table invested more energy into shoot growth and less biomass to the roots, as demonstrated in the partially flooded treatm ent of the mesocosm. It was clear that Pinus elliottii var. densa was able to tolerate lowe r nutrient levels of phosphorus and nitrogen than subspecies 134

PAGE 135

Pinus elliottii var. elliottii and that there were ecotypic differe nces in nutrient uptake for more mesic Everglades slash pines, given ti ssue analyses (Oberb auer et al. 1997). Recommendations As based upon the data analyses and results described in this res earch, recommendations for the restoration of Pinus elliottii var. densa on the rocklands of South Florida are summarized in Table 5-1. Consideration of local genetic stock is the first recommendation for restoration efforts of Pinus elliottii var. densa In rocky areas like the restor ed HID, these seeds are from pine rocklands that are mesic to hydric in nature. Collection and growth as 1-0 containerized seedling stock by the Division of Forestry would be optimal given proper permitting and coordination efforts. If this is not possible, seeds should be cold -stratified, separa ted as sinkers and floaters, and grown in cone trays to one year. While in th e greenhouse, low, half doses of fertilizer once or twice a month will keep the pines alive without causing an acclimation to high nutrients not found in Florida rocklands. Before the South Florida slash pine seedlings are planted, site selec tion and preparation are important. Upland areas where pines are able to tap the groundwater table or wetland areas with shorter hydroperiods, as defined by te n weeks or less, are suitable for Pinus elliottii var. densa growth and survival. In South Fl orida, these are typical ly areas that are higher in elevation. If rockplowed, high-phosphorus soils are present on-sit e, removal of these soils to limestone rock may be appropriate for rockland pine seed source s, in order to avoid competition from exotic plants. Pre-treatment with fire is an a dvantageous site preparation technique, though mechanically trimming weeds or mowing is also effective at exposing so il, increasing light, and reducing competition. Planting is the preferred method of reestablishment, as based upon this study; however, if South Florida slash pine is di rectly seeded, the use of fungi cide may significantly improve 135

PAGE 136

germinative capacity. Planting should be done during the rainy season, between June and September, when growth is optimal. In order to plant flush with the rock, a jackhammer is less labor intensive than a planting bar. Avoid Jhooking the roots and consider amending with iron (and maybe manganese) if there are no harmful effects to the ecosystem. As with reference communities, a desirable endpoint w ould be 500 trees per hectare. To achieve this in restored areas like the HID, initia l planting density may be as high as 2,500 trees per hectare. While the size of three-year old saplings ma y foster the greatest survival at planting, one-year old seedlings are more practical (in terms of investment) and considerably more economical. Once plantings have acclimated to restored area s, the future management of South Florida slash pine stands should involve prescribed fire at interval s sufficient to promote pine reestablishment and recruitment (i.e. approxi mately every five years). Given these recommendations, the National Park Service should be able to establish Pinus elliottii var. densa to near pre-farming levels. While these r ecommendations were based upon the Hole-in-theDonut restoration, they do serve as a guide for landowners and organizations looking to restore South Florida slash pine throughout South Florida rocklands of th e Miami Ridge and the Florida Keys. In general, the evaluati on of site characteri stics and reestablishment techniques for any species is an important concern for re storation ecologists throughout the world. Reflections and Refinements South Florida rocklands were di fferent from the Hole-in-the-D onut, in that restored areas of the HID lacked the microtopographic variation characteristic of the pine rockland ecosystem and was the reason no control or reference commun ity was used in Chapter 4. The HID lacked the physical, edaphic, and biologi cal characteristics found in natural rocklands, as described in Chapter 2. There were no solution holes or rocky variation to HI D sites and it would have been difficult to obtain comparable elevations and hydroperiods for pine rock land control seedlings. 136

PAGE 137

At present, it is uncertain the degree to which pi ne rockland can return to the site and it is not out of the question to dynamite areas of the HID in order to create the mi crotopographic variation necessary to support pine rockland diversity. Th is research conducted in the HID was a unique project specifically designed to test the limits of slash pine survival and identify suitable areas for South Florida slash pine restoration. In regards to the mesocosm study, another i ssue was the decision to use potting media over rockland soil. Rockland soil was difficult to obt ain, as this was an endangered community and the soil layer was shallow. Unlike the deep sandy and acidic soils found in most of Florida, soils from nearby reference communities were found am ongst the surface of exposed limestone rock. In the pine rocklands of Long Pine Key, there we re circumneutral soils with 30 to 50% organic matter among fissures in the rock, while depressi ons in the rock contained residual soils that were a reddish-brown sandy loam with a pH of 6.0 to 6.5 and an organic matter content less than 10% (Snyder 1986, Snyder at al. 1990). In hydric pinelands, the soils were calcareous marl (USFWS 1998b). While there was plenty of av ailable soil in the HID disposal mounds, those soils were high in phosphorus and mostly soils from marl prairie. Typically, transplant mortality of South Florida slash pine into the Miami Rock Ridge is upwards of 50% (personal experience). Therefore, transplanting these pi nes into rockland soil would ha ve been a tremendous shock to the plants and survival may not have been hi gh enough to conduct the mesocosm experiment. This was a risk that could have jeopardi zed the experiment and wasted money. Lastly, the correlation of hydroperiod to elev ation and utilization of the most accurate approach for predicting HID hydr operiods was an important piece of this research. While elevation was used as an indicator for slash pi ne survival, hydroperiod was the driving force. There were variations across the landscape that allowed for slight fluctuations in hydroperiod as 137

PAGE 138

tied to elevation (i.e. ponding along restoration edges and roadsides). The Everglades Depth Estimation Network (EDEN) was an alternat ive model for defining hydroperiod, as EDEN incorporated hydrostations in relation to ground surface data that was collected at points along a 400-meter grid (USGS 2009). Unfortunately, th e ground surface elevations in this model included pre-restoration data for the HID, with higher elevations that underestimated the number of days the HID was flooded. Additionally, HID hydrostation DO-3 was not included in the model because the ground surface elevation was no t in the Data4Ever database (Kahn and Serra 2008). Therefore, EDEN was not an accurate opti on for modeling hydroperiod at the time of this study. Another option was inverse distance wei ghting, which used surrounding hydrostations within five miles of the HID boundary to predic t the average number of days flooded, as based upon data from 2000-2008. This was a less-accurate me thod as compared to the interpretation in Chapter 3, despite only two years of data in this study. The hydrostations used in the inverse distance weighting approach were too far apart to provide the accuracy of hydroperiod duration that this research analyzed and modeled from the twenty research plots and two hydrostations within the HID. For more detailed hydrological da ta, the manual recording of depths at areas of interest over time may offer im provement for the hydroperiod map. In any case, there was variability among hydrolog ical models, such that the accuracy of the hydrological models used by Everglades National Park (EDEN, South Florida Natural Resources Center models) was about 15 cm. These models were no more accurate of an approach than the hydroperiods predicted in Chapter 3, as tied to HID elevation data (slight variations with roads and restoration edges). The mode l utilized in this study was highl y accurate at higher elevations; however, there were limitations with extrapolatin g the data at the lower elevations, hence the 138

PAGE 139

increased range in number of weeks flooded. The goa l of this research was to make a suitability map for South Florida slash pine in the Hole-inthe-Donut. This was accurately accomplished. The more defined hydroperiods at lower elevations were not necessary fo r this research, but would serve useful in the future once HID current elevations and hydrostation DO-3 are incorporated into the EDEN mode l. It will be interesting to compare the models and their accuracy in the future. Future Considerations Research Investigations Continued research of South Florida slash pine and the pine rockland community will further our understanding of th is Environmentally Endangered Land. It has been shown that Pinus elliottii var. elliottii develops roots to become leaky when exposed to hypoxic conditions (Fisher and Stone 1990b, Escamilla and Comerford 1998). This was an indication that Pinus elliottii var. densa would be prone to high mortality unde r the fully flooded mesocosm treatment and reinforced the idea that the 63 to 71 cm elevation range was unsuitable for pine reestablishment in the Hole-in the-Donut, or similar wetland conditions in South Florida. With the threats of sea level rise to Florida Keys and eventually Everglades National Park, a mesocosm study that tests various levels of sa linity and flooding treatments for South Florida slash pine and other rockland species may provide an outlook for this species and ecosystem in the future. Another consideration of the pine rockland ecosystem was to compare the total-organic carbon verses inorganic carbon on target sites, as related to natura l transitional areas, in order to assist with pine rockland community delineation. Granted pine rocklands that are not treated regularly with fire are encroached upon by hardw oods and there is a build-up of organic matter, which could potentially lead to the formation of folist soils like those of hardwood hammocks 139

PAGE 140

(Loope and Dunevitz 1981, Snyder et al. 1990). The pine rockland ridge contained very shallow, non-rockplowed soils that differed from Hole-in-theDonut upland soils. It would be of interest to test these more upland soils found in restored HID and compare them to Cardsound soils of the pine rocklands. Over time it is hypothesized that the upland, pine-suita ble soils in restored HID will become more like soils of the Cardsound Series, and less like those of the Krome Series. A start to this experiment includes obtai ning baseline data of HID restored soils suitable for pine reestablishment in co mparison to pine rockland soils. Soil analyses in this study indicated a tran sitional area from wetland to upland soils. It would be beneficial to compare pine rockland soils to marl prai rie transitional soils data, and then to the results observed in this study. The lowland, hydric soils could be compared to natural solution holes among pine rockland and marl prairi e soils characteristic of the Biscayne Series. As reestablished trees age, the effects of South Florida slash pine evapotranspiration on regulation of the water table will be important in defining the slash pine range in restored HID. Future research with reference communities, pa rticularly young stands and limiting factors (i.e. manganese) would allow accurate predictions of South Florida slash pine reestablishment and pine rockland species survival in the Hole-in-the-Donut. Pine Rockland Restoration The restoration of a pine rockland community will certainly be more challenging than the reestablishment of Pinus elliottii var. densa In the Hole-in the-Donut not only was the majority of the seed source removed with the soil se ed bank, but the landscape was essentially homogenized, which thereby removed the micr otopographic solution holes typically found in pine rocklands. Microtopography has been important to the diversity of the pine rockland ecosystem, given that marl soils and organic litter were found with in the solution holes and that plants rooted into cracks in the limestone rock (Snyder et al. 1990, Ewe et al. 1999). Therefore, 140

PAGE 141

microtopographic restoration via dynamite may be the key to increasing diversity of pine rockland restoration in the Hole -in-the-Donut and areas where the rockplow has homogenized the landscape. The use of dynamite would crea te solution holes amongst the landscape, thereby restoring the microtopgraphy characteri stic of South Florida rocklands. There was a four hectare pine rockland remnan t in the HID that was never rockplowed, yet restored in October of 2005. Basi cally all vegetation (primarily Schinus terebinthifolius ) and some soil were removed from underneath a canopy of South Florida slash pine. Left behind were mature South Florida slash pine trees, expo sed limestone rock, and pockets of soil. These pockets of soil contained a seedbank, that onc e exposed, promoted the return of both common and rare pineland plants, including Guettarda scabra and Jacquemontia curtisii The homogeneity and lack of a seedbank seen elsewhere, in the majority of restored HID upland, caused a delay in the return of more common species like Serenoa repens Other species, such as Muhlenbergia capillaris, increased in cover following prescribed burning (personal observation). The introduction of South Florida slash pine to upland areas of the Hole-in-the-Donut increased the ecosystem structure and function of the restored community, and with time a more characteristic pine rockland co mmunity will develop. There cer tainly may be limits to this restoration, given the absence of solution holes, along with th e time and effort involved in bringing this community back. The reestablishment of South Florid a slash pine was a start to the overstory and adaptive management of seeding and planting other dominant and rare species will further improve the overall community. Restoration is always a challeng e; however, an effort to incr ease species structure in this area will lead to increased function and diversity, such that with time a habitat niche will develop 141

PAGE 142

for protected Everglades species, like the reintroduced bluebird and turkey. Not only will the presence of pine rockland species be aesthetic ally pleasing, there will be a restored area protected from hurricanes and adapted to fire. Therefore, the next step for Everglades National Park will be to conduct similar re storation studies on other plants in the pine rockland ecosystem and develop a map where major species can survive. The Hole-in-the-Donut currently serves as a model for restoration projects and the results of this research will benefit future restoration projects throughout South Florida. 142

PAGE 143

Table 5-1. Recommendations fo r the reestablishment of Sout h Florida slash pine on rocky substrate. Reestablishment Criteria Description Seed source Mesic to hydric rockland seed Pretreatment Fire (mechani cal if fire not possible) Preferred Method Planting 1-0 stock flush with limestone rock using jackhammer, avoid Jhooking roots, 2,500 tph Secondary Method Directly scatte r seeds treated with fungicide Density of adult trees 500 tph Hydrology 10 weeks or less di scontinuous flooding duration, sapling roots able to tap the groundwater table Soils Upland to marl transitional, avoid deeper marl soils with a thick periphyton layer Management Fire intervals sufficient to promote reestablishment and recruitment tph (trees per hectare) 143

PAGE 144

144 Figure 5-1. A comparison of current suitable and marginal habita t for South Florida slash pine with historical pine rockland o ccurrence in the Hole-in-the-Donut.

PAGE 145

APPENDIX LOCATION OF RESEARCH PLOTS IN HID Soil Sampling Coordinates Table A-1. GPS coordinates, in UTMs, for the first set of re search plots where soil samples were collected for analyses. Elevation (cm) Soil Plot X (Easting) Y (Northing) 103-111 E1P1 532647 2808008 103-111 E1P2 532644 2808010 103-111 E1P3 532705 2808020 103-111 E1P4 532672 2808090 93-101 E2P1 532662 2808056 93-101 E2P2 532651 2807987 93-101 E2P3 532478 2807943 93-101 E2P4 532532 2808026 83-91 E3P1 532726 2807956 83-91 E3P2 532640 2807983 83-91 E3P3 532695 2807895 83-91 E3P4 532591 2807924 73-81 E4P1 532757 2807820 73-81 E4P2 532500 2807740 73-81 E4P3 532579 2807782 73-81 E4P4 532664 2807900 63-71 E5P1 532787 2807832 63-71 E5P2 532662 2807658 63-71 E5P3 532833 2807756 63-71 E5P4 532814 2807952 145

PAGE 146

146 South Florida Slash Pine Research Plot Coordinates Table A-2. GPS coordinates, in UTMs, for all South Florida slas h pine research plots in the Hole-in-the-Donut. Pines were directly seed ed and planted in the first set of plots, while pines in the second set of plots were directly seeded as scatter 2. Pine Reestablishment Method Elevati on (cm) Plot X (Easting) Y (Northing) Scatter 1 and Planting 103-111 E1P1 532647 2808008 Scatter 1 and Planting 103-111 E1P2 532644 2808010 Scatter 1 and Planting 103-111 E1P3 532705 2808020 Scatter 1 and Planting 103-111 E1P4 532672 2808090 Scatter 1 and Planting 93-101 E2P1 532662 2808056 Scatter 1 and Planting 93-101 E2P2 532651 2807987 Scatter 1 and Planting 93-101 E2P3 532478 2807943 Scatter 1 and Planting 93-101 E2P4 532532 2808026 Scatter 1 and Planting 83-91 E3P1 532726 2807956 Scatter 1 and Planting 83-91 E3P2 532640 2807983 Scatter 1 and Planting 83-91 E3P3 532695 2807895 Scatter 1 and Planting 83-91 E3P4 532591 2807924 Scatter 1 and Planting 73-81 E4P1 532757 2807820 Scatter 1 and Planting 73-81 E4P2 532500 2807740 Scatter 1 and Planting 73-81 E4P3 532579 2807782 Scatter 1 and Planting 73-81 E4P4 532664 2807900 Scatter 1 and Planting 63-71 E5P1 532787 2807832 Scatter 1 and Planting 63-71 E5P2 532662 2807658 Scatter 1 and Planting 63-71 E5P3 532833 2807756 Scatter 1 and Planting 63-71 E5P4 532814 2807952 Scatter 2 103-111 E1P1 532672 2808083 Scatter 2 103-111 E1P2 532733 2808064 Scatter 2 103-111 E1P3 532656 2808078 Scatter 2 103-111 E1P4 532660 2808073 Scatter 2 93-101 E2P1 532663 2807983 Scatter 2 93-101 E2P2 532692 2807995 Scatter 2 93-101 E2P3 532537 2807880 Scatter 2 93-101 E2P4 532490 2807888 Scatter 2 83-91 E3P1 532516 2807917 Scatter 2 83-91 E3P2 532514 2807946 Scatter 2 83-91 E3P3 532507 2807914 Scatter 2 83-91 E3P4 532764 2807965 Scatter 2 73-81 E4P1 532671 2807916 Scatter 2 73-81 E4P2 532497 2807717 Scatter 2 73-81 E4P3 532754 2807804 Scatter 2 73-81 E4P4 532541 2807767 Scatter 2 63-71 E5P1 532725 2807635 Scatter 2 63-71 E5P2 532788 2807921 Scatter 2 63-71 E5P3 532716 2807705 Scatter 2 63-71 E5P4 532846 2807776

PAGE 147

LIST OF REFERENCES Abada, J. and A. Abada. 1993. Iron and plant pigments. p. 327-343. In L. L. Barton and B. C. Hemming (eds.) Iron Chelation in Plan ts and Soil Microorganisms. Academic Press, Inc., New York, NY. Abrahamson, W. G. and D. C. Hartnett. 1990. Pine flatwoods and dry prairies. p. 103-149. In R. L. Myers and J. J. Ewel (eds.) Ecosystems of Florida. University of Central Florida Press, Orlando, FL. Armentano, T. V., Sah, J. P., Ross, S. M., Jones, D. T., Cooley, H. C., and C. S. Smith. 2006. Rapid responses of vegetation to hydrologi cal changes in Taylor Slough, Everglades National Park, Florida, USA. Hydrobiologia 569:293. Bachoon, D. and R. D. Jones. 1992. Potential rate s of methanogenesis in sawgrass marshes with peat and marl soils in the Everglades. Soil Biology and Biochemistry 24(1):21-27. Bancroft. 1973. Memos and Annual Report from Ho le-in-the-Donut. Everglades National Park Technical Report. South Florida Research Center, Homestead, FL. Barnard, E. L. and M. Greenstein. 1995. Pine chloro sis and decline in Flor ida landscapes. Forest and Shade Tree Pests: Division of Forestry Leaflet Number 12:1-2. Barnett, J. P. 1996. How seed orchard culture affects seed quality: experience with the southern pines. The Forestry Chronicle 72(5):469-473. Bringman. 1976. Hole-in the-Donut Status Report South Florida Resear ch Center, Homestead, FL. Campbell, T. E. 1980. Direct-seeded and plante d loblolly and slash pine on a hardwooddominated site yield similar 15year results. Southern Journa l of Applied Forestry 4(3): 147-151. Chen, M., Ma, L. Q., and Y. C. Li. 2000. Concentra tions of P, K, Al, Fe, Mn, Cu, Zn, and As in soils from south Everglades. Soil Crop Science Society Florida Proceedings 59:124-129. Colbert, S. R., E. J. Jokela, and D. G. Near y. 1990. Effects of annual fertilization and sustained weed control on dry matter partitioning, leaf ar ea, and growth efficiency of juvenile loblolly and slash pine. Forest Science 36:995-1014. Cornwell, G. W. and K. A. Atkins. 1975. The imp act of evicting farmers from the Hole-in-theDonut. EcoImpact, Inc., Gainesville, FL. 184 p. Cropper, W. P., Jr., and H. L. Gholz. 1991. In situ needle and fine root respiration in mature slash pine ( Pinus elliottii) trees. Canadian Journal of Forest Resources 21:1589-1595. 147

PAGE 148

Dalrymple, G. H., Doren, R. F., OHare, N. K., Norland, M. R., and T. V. Armentano. 2003. Plant colonization after comp lete and partial removal of disturbed soils for wetland restoration of former agricultural fields in Everglades National Park. Wetlands 23:10151029. Dalrymple, N. K., Dalrymple, G. H., and K. A. Fanning. 1993. Vegetation of restored rockplowed wetlands of the East Everglades. Restoration Ecology 1(4):220-225. Davis, S. M., Gaiser, E. E., Loftus, W. F., and A. E. Huffman. 2005. Southern marl prairies conceptual ecological model. Wetlands 25(4):821-831. Derr, H. J. and W. F. Mann .1971. Direct seed ing pines in the South. USDA Agriculture Handbook No. 391. U.S. Government Prin ting Office, Washington D.C. Doren, R. F., Platt, W. J., and L. D. Whiteak er. 1993. Density and size structure of slash pine stands in the Everglades re gion of south Florida. Forest Ecology and Management 59:295311. Doren, R. F. and R. M. Rochefort. 1984. Summary of fires in Everglades National Park and Big Cypress National Preserve, 1981. South Flor ida Research Center Report SFRC-84/01. National Park Service, Atlanta, GA. 60 p. Doren, R. F. and L. D. Whiteaker. 1990. Effects of fire on different size individuals of Schinus terebinthifolius. Natural Areas Journal 10:106-113. Duncan, D. V. and T. A. Terry. 1983. Water management. p. 91-111. In E. L. Stone (ed.) The Managed Slash Pine Ecosystem: Proceedings of a Symposium held at the University of Florida, June 9-11, 1981. School of Forest Re sources and Conservation, University of Florida, Gainesville, FL. Escamilla, J. A. and N. B. Comerford. 1998. P hosphorus and potassium depletion by roots of field-grown slash pine: aer obic and hypoxic conditions. Forest Ecology and Management 110:25-33. Ewe, S. M. L., Sternberg, L. d. S., and D. E. Busch. 1999. Water-use patterns of woody species in pineland and hammock communities of Sout h Florida. Forest Ecology and Management 118:139-148. Ewel, J. J., Ojima, D. S., Karl D. A., and W. F. DeBusk. 1982. Schinus in successional ecosystems of Everglades National Park. T echnical Report T-676. South Florida Research Center, Homestead, FL. 141p. Ewel, K. C., Cropper, W. P., Jr., and H. L. Gholz. 1987. Soil CO2 evolution in Florida slash pine plantations. II. Importance of r oot respiration. Canadian Jour nal of Forest Research 17: 330-333. Fisher, H. M. 1989. The aeration of slash pine ro ot systems in saturated soils. Ph.D. Thesis. University of Florida, Gainesville, FL. 130 p. 148

PAGE 149

Fisher, H. M. and E. L. Stone. 1990a. Air-conducting porosity in slash pine roots from saturated soils. Forest Science 36:18. Fisher, H. M. and E. L. Stone. 1990b. Active potassium uptake by slash pine root from O2 depleted solutions. Forest Science 36:582-598. Gilly, S. 2006. Personal Communication. Andrews Nursery. Flor ida Division of Forestry, Chiefland, FL. Hamrick, J. L. 1995. Genetic diversity within and among populations of South Florida slash pine, Pinus elliottii var. densa Final Report. Department of Environmental Resources Management, Miami, FL. Hanlon, E. A., Gonzalez, J. S., and J. M. Bartos. 1994. Standard determination of Ca, Mg, P, K, Na, Mn, Cu, Fe, Zn, and B in plant tissue. p. 47-48. In E. A. Hanlon, J. S. Gonzalez, and J. M. Bartos (eds.) IFAS Extension Soil Testing Laboratory Chemical Procedures and Training Manual. Cir 812. Florida Cooperative Extension Service, IFAS, University of Florida, Gainesville, FL. Havlin, J. L., Beaton, J. D., Tisdale, S. L., a nd W. L. Nelson. 2005. Soil Fertility and Fertilizers: An Introduction to Nutrient Management. 6th ed. Prentice Hall, Upper Saddle River, NJ. 515 p. Inglett, P. E., Reddy, K. R., Inglett, K. S., and M. W. Clark. 2006. Patterns of soil biogeochemistry in the Hole-in-the-Donut re gion with implications for restoration management. Technical Report. Soil and Wate r Science Department, Gainesville, FL. 206 p. Jokela, E. J. 2004. Nutrient manageme nt of southern pines. p. 27-35. In E. D. Dickens, J. P. Barnett, P. James, W. G. Hubbard, and E. J. Jokela (eds.) Slash Pine Still Growing and Growing!: Proceedings of the Slash Pine Symposium, Jekyll Island, Georgia, April 23-25, 2002. General Technical Report SRS-76. U.S. Depart ment of Agriculture, Forest Service, Southern Research Station, Asheville, NC. Kahn, F. A. and L.A. Serra. 2008. South Florid a Natural Resources Center. DataForEVER Dataset, Everglades National Park, Homestead, FL, using Appaserver software (http://www.appaserver.com), S acramento, CA, Public URL not currently available, please send data requests to EVER_data_request@nps.gov Last accessed December 31, 2008. Kartesz, J, T. 1999. Synthesis of the North Amer ican Flora. Chapel Hill, NC: North Carolina Botanical Garden, University of No rth Carolina at Chapel Hill, NC. Koerselman, W and A. F. M. Meuleman. 1996. The vegeta tion N:P ratio: a new tool to detect the nature of nutrient limitation. Journal of A pplied Ecology 33(6):1441-1450. Krauss, P. 1987. Old Field Succession in Evergl ades National Park. Technical Report SFRC87/03. South Florida Research Center, Homestead, FL. 62 p. 149

PAGE 150

Landers, J. L. 1991. Disturbance influences on pine traits in the southeastern United States. p. 61-95. In Proceedings, 17th Tall Timbers Fire Ecology Conference; May 18-21, 1989, Tallahassee, FL. Tall Timbers Rese arch Station, Tallahassee, FL. Langdon, O. G. 1958. Cone and seed size of Sout h Florida slash pine and their effects on seedling size and survival. Journal of Forestry 56(2):122-127. Langdon, O. G. 1963. Range of South Florida slas h pine. Journal of Forestry 61(5):384-385. Li, Y. and M. Norland. 2001. The role of soil fertility in invasion of Brazilian pepper ( Schinus terebinthifolius) in Everglades National Park, Fl orida. Soil Science 166(6):400-405. Little, E. L. and K. W. Dorman. 1952. Slash pine ( Pinus elliottii ) its nomenclature and varieties. Journal of Forestry 50:918-923. Lohrey, R. E. and S. V. Kossuth. 1990. Pinus elliottii (Engelm.). p. 338. In R. M. Burns and B. H. Honkala (eds.) Silvics of North Am erica: Volume 1 Conifers, Agriculture Handbook 654. United States Forest Se rvice, Washington D.C. Long, A. J., Behm, A., Cassidy, L., DiMartino, J., Doran, D., Freeman, D., Helmers, J., Keefe, K., Miller, A., Ranasinghe, S., Randall, C., Rasser, M., Ruth, A.,Shipley, D., and A. Van Loan. 2004. Prescribed fire and slash pine. p. 66-78. In E. D. Dickens, J. P. Barnett, P. James, W. G. Hubbard, and E. J. Jokela (e ds.) Slash Pine Still Growing and Growing!: Proceedings of the Slash Pine Symposium, Jekyll Island, Georgia, April 23-25, 2002. General Technical Report SRS-76. U.S. Depart ment of Agriculture, Forest Service, Southern Research Station, Asheville, NC. Loope, L. L. and V. L. Dunevitz. 1981. Investigations of early plant succession on abandoned farmland in Everglades National Park..Techni cal Report T-644. South Florida Research Center, Homestead, FL. 65 p. Mayo, E. 2000. Reforestation of Pinus elliottii var. densa on the Miami Rock Ridge: field experiment and economic analysis of two alternative methods. M.S. Thesis. Florida International University, Miami, FL. 55 p. McKevlin, M. R., Hook, D. D., Mckee, W. H ., Wallace, S. U., and J. R. Woodruff, Jr. 1987. Loblolly pine seedling root anatomy a nd iron accumulation as affected by soil waterlogging. Canadian Journal of Forest Resources 17:1257-1264. McLemore, B. F. 1975. Collection da te, cone storage period affect southern pine seed yields, viability. Tree Planters Notes 26(1):24-26. Meador, R. E. 1977. The role of mycorrhizae in influencing succession on abandoned Everglades farmland. M.S. Thesis. University of Florida, Gainesville, FL. 99 p. Mikola, P. 1967. The effect of mycorrhizal inoc ulation on the growth a nd root respiration of Scotch pine seedlings. Proc.14th Congr., Interrat. Union For. Res. Organiz., Munich, Part V. Section 24:100-111. 150

PAGE 151

Mitsch, W. J. and J. G. Gosselink. 2000. Wetla nds. 3rd ed. John Wiley and Sons, Inc., New York, NY. 920 p. Oberbauer, S. F., Kariyawasam, P. S., and J. N. Burch. 1997. Comparativ e analysis of growth, nutrition, carbon isotope ratios, a nd hurricane-related mortalit y of slash pines in South Florida. Florida Sc ientist 60(4):210-222. O'Hare, N. K. and G. H. Dalrymple. 2005. Final annual report: biological monitoring of restored wetlands in Everglades National Park July 1, 2003-June 30, 2004. Submitted to Everglades National Park, Homestead, FL. OHare, N. K. and G. H. Dalrymple. 2006. Growth and survival of South Florida slash pine ( Pinus elliottii var. densa ) on restored farmland in Evergl ades National Park. Ecological Restoration 24(4):242-249. Orth, P. G. and R. A. Conover. 1975. Changes in nutrients resulting from farming in the Hole-inthe-Donut, Everglades National Park. Pr oc. Fla. State. Hort. Soc. 28: 221-225. Pallardy, S. G. 2008. Physiology of Woody Plants. 3rd Ed. Academic Press, New York, NY. 454 p. Polglase, P. J., Jokela, E. J., and N. B. Comerford. 1992. Phosphorus, nitrogen, and carbon fractions in litter and soil of southern pine plantations. Soil Science Society of America Journal 56(2): 566-572. Pritchett, W. L. and N. B. Comerford. 1983. Nu trition and fertilization of slash pine. p. 69-90. In E. L. Stone (ed.) The Managed Slash Pine Ecos ystem: Proceedings of a Symposium held at the University of Florida, June 9-11, 1981. Sc hool of Forest Resources and Conservation, University of Florida, Gainesville, FL. Reddy, K. R. and R. D. DeLaune. 2008. Bioge ochemistry of Wetlands: Science and Applications. CRC Press, Boca Raton, FL. 774 p. Robertson, W. B., Jr. 1954. Evergl ades fires past, present a nd future. Everglades Natural History 2(1):9-16. Ross, M. S., Carrington, M., Flynn, L. J., and P. L. Ruiz. 2001. Forest succession in tropical hardwood hammocks of the Florida Keys: eff ects of direct mortality from Hurricane Andrew. Biotropica 33(1):23-33. Ross, M. S., O'Brien, J. J., and L. d. S. Sternberg. 1994. Sea level rise and the reduction in pine forests in the Florida Keys. Ec ological Applications 4:144-156. Schoenau, J. J. and I. P. OHalloran. 2007. S odium bicarbonate-extractable phosphorus. p. 89-95. In M. R. Carter and E. G. Gregorich (eds.) Soil Sampling and Methods of Analysis. 2nd ed. Canadian Society of Soil Science. CRC Press, Boca Raton, FL. 151

PAGE 152

Shoulders, E. and C. W. Ralston. 1975. Temperature, root aeration, and light influence slash pine nutrient uptake rates. Forest Science 21(4):401-410. Silen, R. and C. Osterhans. 1979. Reduction of ge netic base by sizing of bulk Douglas-fir seed lots. Tree Planters Notes 30(1):24-30. Snyder, J. R. 1986. The impact of wet season and dry seas on prescribed fires on Miami rock ridge pineland, Everglades National Park Technical Report SFRC-86/06. South Florida Research Center, Homestead, FL. 106 p.. Snyder, J. R., Herndon, A., and W. B. Robe rtson, Jr. 1990. South Florida rockland. p. 230-274. In R. L. Myers and J. J. Ewel (eds.) Ecosyste ms of Florida. University of Central Florida Press, Orlando, FL. Snyder, J. R., Ross, M. S., Kopter, S., and J. P. Sah. 2006. Developing ecological criteria for prescribed fires in South Florida pine ro ckland ecosystems. Technical Report. USGS, Ochopee, FL. 109 p. Soltanpour, P. N. and A. P. Schwab 1977. A new so il test for simultaneous extraction of macroand micro-nutrients in alkaline soils. Commun. Soil Sci. Plant Anal. 8:195-207. Squillace, A. E. 1966. Geographic variation in sl ash pine. Forest Science Monograph 10. Society of American Foresters, Washington, D.C. 56 p. Sun, G., McNulty, S. G., Shepard, J. P., Amatya, D. M., Riekerk, H., Comerford, N. B., Skaggs, W., and L. Swift, Jr. 2001. Effects of tim ber management on the hydrology of wetland forests in the southern United States. Forest Ecology and Management 143:227-236. USDA. 1958. Soil Survey of Dade County, Florida. Natural Resources Conservation Service. U.S. Government Printing Office, Washington, D.C. 56 p. USDA. 1974. Seeds of Woody Plants in the United States, USDA Ag Handbook #450. Natural Resources Conservation Service. U.S. Gove rnment Printing Office, Washington, D.C. 883 p. USDA. 1996. Soil Survey of Dade County, Florida. Natural Resources Conservation Service. U.S. Government Printing Office, Washington, D.C. 116 p. USFWS. 1998a. Mesic pine flatwoods. p. 195-230. In U.S. Fish and Wildlife Service Multispecies Recovery Plan for the Threatened and Endangered Species of South Florida, Volume 2 of 2, The Ecosystem. Technical/Agency Draft. USFWS Southeast Region, Atlanta, GA. USFWS. 1998b. Hydric pine flatwoods. P. 231-278. In U.S. Fish and Wildlife Service Multispecies Recovery Plan for the Threatened and Endangered Species of South Florida, Volume 2 of 2, The Ecosystem. Technical/Agency Draft. USFWS Southeast Region, Atlanta, GA. 152

PAGE 153

153 USGS. 2007. 1940 Greater Everglades and South Florida Aerial P hotoset [Internet]. Available from: http://sofia.usgs.gov/exchange/aerial-photos/ Last accessed February 1, 2007. USGS. 2009. Everglades Depth Estimation Netw ork (EDEN) for Support of Biological and Ecological Assessments [Int ernet]. Available from: http://sofia.usgs.gov/eden/ Last accessed April 1, 2009. Wade, D. D. 1983. Fire management in the slash pine ecosystem. p. 203-227. In E. L. Stone (ed.) The Managed Slash Pine Ecosystem: Proceedings of a Symposium held at the University of Florida, June 9-11, 1981. School of Forest Resources and Conservation, University of Florida, Gainesville, FL. Webster, R. and M. Oliver. 2001. Geostatistics fo r Environmental Scientists. John Wiley & Sons, LTD, New York, NY. White, E. H. and W. L. Pritchet t. 1970. Water-table c ontrol and fertilization for pine production in the flatwoods. Florida Agricultural E xperiment Station Technical Bulletin 743. University of Florida, Gainesville, FL. 41 p. Williams, D. A. Wang, Y., Borchetta, M., and M. S. Gaines. 2007. Genetic diversity and spatial structure of a keystone species in fragmented pine rock land. Biological Conservation 138: 256-268. Wunderlin, R. P. and B. F. Hansen. 2003. Guide to the Vascular Plants of Florida. 2nd ed. University Press of Florida, Gainesville, FL. 787 p.

PAGE 154

BIOGRAPHICAL SKETCH Lauren Ann Serra was born in Saratoga Springs, New York and grew up in the Adirondack Mountains. She graduated vale dictorian of her Lake George High School class in 1997 and Summa Cum Laude from the University of Al bany in 2000, with a Bachelor of Science in Biology. In 2003, Lauren completed a Master of Professional Studies in Environmental Forest Biology from the State University of New York College of Environmental Science and Forestry in Syracuse. From there, she pursued a career with the National Park Service, where she has served as a plant biologist with the Hole-in-the-Donut project in Everglades National Park for more than five years. With a passion for re storation ecology, Lauren went on to achieve her Doctor of Philosophy in the Soil and Water Science Department, th anks to the University of Florida. 154