<%BANNER%>

Microtopographical Effects of Clearcutting of Taxodium-Nyssa Swamps

Permanent Link: http://ufdc.ufl.edu/UFE0043013/00001

Material Information

Title: Microtopographical Effects of Clearcutting of Taxodium-Nyssa Swamps
Physical Description: 1 online resource (82 p.)
Language: english
Creator: WASHUTA,AMY BETH
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: CLEARCUTTING -- CYPRESS -- LOGGING -- MICROTOPOGRAPHY -- SPATIAL -- SWAMP -- TUPELO
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: MICROTOPOGRAPHICAL EFFECTS OF CLEARCUTTING OF TAXODIUM-NYSSA SWAMPS Hummock-hollow microtopography, common in Taxodium-Nyssa swamps of the Southeast Coastal Plain, may be an important indicator of forest wetland conditions following logging. Past studies have indicated the importance of these microtopographic features because many plant species germinate and establish preferentially on hummocks. Concern over clearcut logging in forested wetlands has prompted interest in evaluating the recovery of these systems following conventional and new logging practices. The objectives of this research were to: 1) characterize the microtopography in reference sites (in terms of prevalence, spatial organization, and elevated feature characteristics) and determine the degree to which conventional techniques (bottom logging) and new techniques (mat logging) alter microtopography; and, 2) to compare the relationships between vegetation and microtopographic position in both reference and logged swamps. Five second-growth reference swamps that were logged approximately 40-100 years ago, three bottom-logged sites logged 3-13 years ago, and two mat-logged sites logged 1-5 years ago were sampled. Elevations were recorded following a spatially-nested design in 30 x 30 m plots, and systematic surveys of vegetation and microtopographic position were conducted concurrently. Thirty transects measuring elevations across skid trails (15 mat-logged, 15 bottom-logged) were also recorded. Reference swamps displayed clear bimodality in elevational histograms related to hummock and hollow topography, which was reduced in prevalence over 50% in recently mat and bottom-logged sites. Bottom-logged sites showed trimodality, which was due to the abundance of low ruts as well as remaining elevated features. Line transects also revealed lower elevations across bottom-logged skid trails as compared to mat-logged trails. Variograms indicate differences in horizontal spatial patterns between bottom-logged, mat-logged, and reference sites, with reference and bottom-logged sites exhibiting the highest sill, and bottom-logged sites the greatest range. Relative abundances of substrate types of elevated features differed between reference and logged sites, with root/soil features relatively more common in reference sites, and stump/soil and stump features more common in mat and bottom-logged sites. Total understory stem density and species richness were greater on elevated microsites than on the swamp bottom in both reference and logged sites. The woody vine and shrub functional groups exhibited the strongest preferences for elevated features. The results of this study indicate that both logging techniques diminished the prevalence of elevated microtopographic features as well as altered the relative abundance of their substrate types, while only bottom-logging created low elevations via deeper skid trial rutting. Vegetation is clearly linked to these features, which suggests their importance in swamp plant communities. Sampling soil elevations proved an effective tool for understanding changes in the vertical structuring of microtopography across sites.
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 AMY BETH WASHUTA.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Bohn, Kimberly Kirsten.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043013:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043013/00001

Material Information

Title: Microtopographical Effects of Clearcutting of Taxodium-Nyssa Swamps
Physical Description: 1 online resource (82 p.)
Language: english
Creator: WASHUTA,AMY BETH
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: CLEARCUTTING -- CYPRESS -- LOGGING -- MICROTOPOGRAPHY -- SPATIAL -- SWAMP -- TUPELO
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: MICROTOPOGRAPHICAL EFFECTS OF CLEARCUTTING OF TAXODIUM-NYSSA SWAMPS Hummock-hollow microtopography, common in Taxodium-Nyssa swamps of the Southeast Coastal Plain, may be an important indicator of forest wetland conditions following logging. Past studies have indicated the importance of these microtopographic features because many plant species germinate and establish preferentially on hummocks. Concern over clearcut logging in forested wetlands has prompted interest in evaluating the recovery of these systems following conventional and new logging practices. The objectives of this research were to: 1) characterize the microtopography in reference sites (in terms of prevalence, spatial organization, and elevated feature characteristics) and determine the degree to which conventional techniques (bottom logging) and new techniques (mat logging) alter microtopography; and, 2) to compare the relationships between vegetation and microtopographic position in both reference and logged swamps. Five second-growth reference swamps that were logged approximately 40-100 years ago, three bottom-logged sites logged 3-13 years ago, and two mat-logged sites logged 1-5 years ago were sampled. Elevations were recorded following a spatially-nested design in 30 x 30 m plots, and systematic surveys of vegetation and microtopographic position were conducted concurrently. Thirty transects measuring elevations across skid trails (15 mat-logged, 15 bottom-logged) were also recorded. Reference swamps displayed clear bimodality in elevational histograms related to hummock and hollow topography, which was reduced in prevalence over 50% in recently mat and bottom-logged sites. Bottom-logged sites showed trimodality, which was due to the abundance of low ruts as well as remaining elevated features. Line transects also revealed lower elevations across bottom-logged skid trails as compared to mat-logged trails. Variograms indicate differences in horizontal spatial patterns between bottom-logged, mat-logged, and reference sites, with reference and bottom-logged sites exhibiting the highest sill, and bottom-logged sites the greatest range. Relative abundances of substrate types of elevated features differed between reference and logged sites, with root/soil features relatively more common in reference sites, and stump/soil and stump features more common in mat and bottom-logged sites. Total understory stem density and species richness were greater on elevated microsites than on the swamp bottom in both reference and logged sites. The woody vine and shrub functional groups exhibited the strongest preferences for elevated features. The results of this study indicate that both logging techniques diminished the prevalence of elevated microtopographic features as well as altered the relative abundance of their substrate types, while only bottom-logging created low elevations via deeper skid trial rutting. Vegetation is clearly linked to these features, which suggests their importance in swamp plant communities. Sampling soil elevations proved an effective tool for understanding changes in the vertical structuring of microtopography across sites.
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 AMY BETH WASHUTA.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Bohn, Kimberly Kirsten.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043013:00001


This item has the following downloads:


Full Text

PAGE 1

1 MICROTOPOGRAP H ICAL EFFECTS OF CLEARCUTTING OF TAXODIUM NYSSA SWAMPS By AMY B. WASHUTA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Amy B. Washuta

PAGE 3

3 To my husband and best friend, Luke Gommermann, for years of love and support

PAGE 4

4 ACKNOWLEDGMENTS I thank my mom and dad for their love and direction throughout my life and my brother, Steven for his constant companionship. I am thankful to my ever optimistic field assistant, Liz Martin for helping me in the swamps, and to my other field helpers, Luke Gommerman, Jason Gommermann, David Holden, Jennette Hozworth, and Emily Ro driquez. I thank Dr. Rob Fletcher for his help with spatial statistics and Chad Foster for helping me locate field sites. Last, I would like to thank my committee members, Dr. Jack Putz, Dr. Matt Cohen and my advisor Dr. Kimberly Bohn for technical help, emotional encouragement, and scientific inspiration.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Taxodium Nyssa Swamps ................................ ................................ ...................... 12 Logging History ................................ ................................ ................................ ....... 14 Logging Concerns ................................ ................................ ................................ ... 15 2 IMPACTS OF LOGGING ON MICROTOPOGRAPHY IN TAXODIUM NYSSA SWAMPS ................................ ................................ ................................ ................ 18 Background ................................ ................................ ................................ ............. 18 Methods ................................ ................................ ................................ .................. 24 Study Sites ................................ ................................ ................................ ....... 24 Field Methods ................................ ................................ ................................ ... 25 Water depths ................................ ................................ .............................. 25 Skid trail transects ................................ ................................ ...................... 26 Elevated microsite surveys ................................ ................................ ........ 26 Data Analysis ................................ ................................ ................................ ... 27 Water depths ................................ ................................ .............................. 27 Skid trail transects ................................ ................................ ...................... 28 Elevated microsite surveys ................................ ................................ ........ 28 Result s ................................ ................................ ................................ .................... 28 Water Depths ................................ ................................ ................................ ... 28 Vertical structure ................................ ................................ ........................ 28 Horizontal spatial structure ................................ ................................ ......... 30 Skid Trail Transects ................................ ................................ .......................... 31 Elevated Microsite Surveys ................................ ................................ .............. 31 Substrate characteristics of elevated microsites ................................ ........ 31 Size of elevated microsites ................................ ................................ ........ 32 Disc ussion ................................ ................................ ................................ .............. 33 Microtopography in Reference Conditions ................................ ........................ 33 Impact of Logging on Microtopography ................................ ............................ 35 Microsite Availability, Substrate Type, and Potential for Regeneration ............ 37 Concluding Remarks ................................ ................................ ............................... 39 3 RELATIONSHIPS BETWEEN VEGETATION AND MICROTOPOGRAPHIC POSITION IN REFERENCE AND CLEARCUT TAXODIUM NYSSA SWAMPS .... 57

PAGE 6

6 Background ................................ ................................ ................................ ............. 57 Methods ................................ ................................ ................................ .................. 60 Study Sites ................................ ................................ ................................ ....... 60 Re ference sites ................................ ................................ .......................... 61 Logged sites ................................ ................................ ............................... 62 Field Methods ................................ ................................ ................................ ... 62 Understory vegetation ................................ ................................ ................ 62 Tree surveys ................................ ................................ .............................. 62 Data Analysis ................................ ................................ ................................ ... 63 Understory vegetation ................................ ................................ ................ 63 Tree surve ys ................................ ................................ .............................. 63 Results ................................ ................................ ................................ .................... 64 Understory Vegetation ................................ ................................ ...................... 64 Species richness ................................ ................................ ........................ 64 Stem density ................................ ................................ .............................. 64 Tree Surveys ................................ ................................ ................................ .... 65 Discussion ................................ ................................ ................................ .............. 66 Concluding Remarks ................................ ................................ ............................... 67 APPENDIX A: E XAMPLE R CODE FOR HISTOGRAM CLUSTERING ANALYSIS ...................... 73 B: SAS CODE FOR ELEVATED MICROSITE SIZES ANALYSIS .............................. 74 C: LIST OF SPECIES FOR UNDERSTORY VEGETATION SURVEYS ..................... 75 LIST OF REFERENCES ................................ ................................ ............................... 77 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 82

PAGE 7

7 LIST OF TABLES Table page 2 1 List of study sites. ................................ ................................ ............................... 41 2 2 BIC values for unimodal, bimodal and trimodal models generated by the Mclust package. ................................ ................................ ................................ .. 42 3 1 List of study sites. ................................ ................................ ............................... 69

PAGE 8

8 LIST OF FIGURES Figure page 2 1 Skid trails created during bottom logging operations A) 10 yrs ago and B) 12 yrs ago. ................................ ................................ ................................ ............... 43 2 2 Skid trails created during mat logging operations A) 1 yr ago and B) 5 yrs ago. ................................ ................................ ................................ .................... 43 2 3 A) A water depth sampling scheme for one 30x30m frame and B) An example sampling station within the frame ................................ ......................... 44 2 4 A tree base microtopographic feature. ................................ ................................ 45 2 5 A soil microtopographic feature. ................................ ................................ ......... 45 2 6 A stump/soil microtopographic feature. ................................ .............................. 46 2 7 A root/soil microtopographic feature. ................................ ................................ .. 46 2 8 A log microtop ographic feature. ................................ ................................ .......... 47 2 9 A stump microtopographic feature. ................................ ................................ ..... 47 2 10 A tip up mound microtopographic feature. ................................ .......................... 48 2 11 A slash debris microtopographic feature. ................................ ........................... 48 2 12 Elevational histograms for frames in three reference swamps.. ......................... 49 2 13 Elevational histograms for frames in two reference swamps. ............................. 50 2 14 Elevational histograms for swamps mat logged 1 year ago (A and B) and 5 years ago (C and D). ................................ ................................ .......................... 51 2 15 Elevational histograms for swamps bottom logged 3 years ago(A and B), 7 years ago (C and D), and 13 years ago (E and F). ................................ ............. 52 2 16 Exponential theoretical variograms modeled after the composite variogram parameters for all mat logged, bottom logged and reference sites. .................... 53 2 17 Composite transects showing average elevations in 15 transects across mat logging skid trails and 15 transects across botto m logging skid trails. ................ 54 2 18 The relative proportions of various microsite substrate types in A) reference swamps, B) mat logged s wamps and C) bottom logged swamps ...................... 55

PAGE 9

9 2 19 Boxplots indicating the minimum, maximum, mean, interquartile range and outliers for the area (m 2 ) of elevated microtopographic features across six study sites. ................................ ................................ ................................ .......... 56 3 1 Average vegetation richness of A) forbs and graminoids and B) woody vines, shrubs and tree seedlings ................................ ................................ ................... 70 3 2 Average plant stem density of forbs, woody vines, shrubs and tree se edlings ... 71 3 3 The proportion of A) Nyssa sylvatica (n=30 per DBH class) and B) Taxodium spp .(n=10 per DBH class) individuals grow ing on elevated microsites ............... 72

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requir ements for the Degree of Master of Science MICROTOPOGRAPHICAL EFFECTS OF CLEARCUTTING OF TAXODIUM NYSSA SWAMPS By Amy B. W ashuta May 2011 Chair: Kimberly Bohn Major: Interdisciplinary Ecology Hummock hollow microtopography, common in Taxodium Nyssa swamps of the Southeast Coastal P lain, may be an important indicator of forest wetland conditions following logging. Past studies have indicated the importance of th ese microtopographic features because many plant species germinate and establish preferen tially on hummocks. Concern over clearcut logging in forest ed wetlands has prompted interest in evaluating the recovery of these systems following conventional and new logging practices. The objectives of this research were to: 1) characterize the microt opography in reference sites (in terms of prevalence, spatial organization, and elevated feature characteristics) and determine the degree to which conventional techniques (bottom logging) and new techniques (mat logging) alt er microtopography; and, 2 ) to compare the relationships between vegetation and microtopographic position in both reference and logged swamps. Five second gro wth reference swamps that were logged approximately 40 100 years ago, three bottom logged sites logged 3 13 years ago, and two mat logged sites logged 1 5 years ago were sampled. Elevations were recorded following a spatially nested design in 30 x 30 m plots, and systematic surveys of vegetation and microtopographic position were conducted concurrently.

PAGE 11

11 Thirty transects measurin g elevations across skid trails (15 mat logged, 15 bottom logged) were also recorded. R eference swamps displayed clear bimodality in elevational histograms related to hummock and hollow topography, which was reduced in prevalence over 50% in recently mat and bottom logged sites. Bottom logged sites showed tr imodality, which was due t o the abundance of low ruts as well as remaining elevated features. Line transects also revealed lower elevations across bottom logged skid trails as compared to mat log ged t rails. Variograms indicate differences in horizontal spatial pattern s between bottom logged, mat logged, and reference sites, with reference and bottom logged sites exhibiting the highest sill, and bottom logged sites the greatest range. Relative abundan ces of substrate types of elevated features differed between reference and logged sites, with root/soil features relatively more common in reference sites, and stump/soil and stump features more common in mat and bottom logged sites. Total understory stem density and species richness were greater on elevated microsites than on the swamp bottom in both reference and logged sites. The woody vine and shrub functional groups exhibited the strongest preferences for elevated features. The results of this stud y indicate that both logging techniques diminished the prevalence of elevated microtopographic features as well as altered the relative abundance of their substrate types, while only bottom logging created low elevations via deeper skid trial rutting. Veg etation is clearly linked to these features, which suggests their importance in swamp plant communities. Sampling soil elevations proved an effective tool for understanding changes in the vertical structuring of microtopography across sites.

PAGE 12

12 CHAPTER 1 INTRODUCTION Taxodium Nyssa Swamps Swamps with an overstory dominated by c ypress ( Taxodium distichum var. nutans and var. distichum ) and swamp t upelo trees ( Nyssa sylvatica var. biflora ) are found throughout floodplain and depressional areas of the Southeast Coastal Plain (Ewel 1998). In Florida, they often occur as semi isolated depressional features within a matrix of pine flatwoods or pine pl antations (Ewel, 1998). Because hydroperiods of 10 12 months are characteristic of this forest association (Kellison and Young, 1997), decomposition of organic materia l is slow, allowing a thick organic rich soil horizon to accu mulate. The forest floor of cypress tupelo swamps often exhibits microtopographic relief (on the order of 0.5 1.5 m ) that influences the hydrology, decomposition, and plant community structure of the forest community (Brown, 1972). Several studies have demonstrated the importance of topographic variabililty in swamps as providing a range of hydrologic conditions that encourages plant diversity and shapes vegetative structure (Anderson et al., 2009, Streng, 1989, Moser et al., 2007). In other studies, the presence of raised microt opographic features, specifically, has been shown to encourage woody plant regeneration (Titus, 1990, Huenneke and Sharitz, 1986) and enhance plant productivity (Jones et al., 2000). In many cypress tupelo swamps, higher topographic positions support a de nse shrub layer and a number of herbaceous species (Ewel, 1998). Depressional cypress t upelo swamps coincide with an exposed surfic al aquifer (Crownover et al, 1995) and can affect regional hydrology by acting as groundwater

PAGE 13

13 recharge areas, aiding in flood prevention in surrounding areas (Ewel, 1998) and influencing nitrogen and phosphorus dynamics (Walbridge and Lockaby, 1984). Water table fluctuations can result in periodic exposure of portions of the forest floor during crucial regeneration times for tr ees and other plant species. Almost all woody species in these swamps inc luding cypress and tupelo require unflooded sites to ger minate (Myers and Ewel, 1990). Cypress t upelo swamps are an important component of the larger landscape in which they reside for reasons beyond hydrology. The ecotone between c ypress and hardwood swamps and upland forests are believed to be hot spots for wildlife diversity in Fl orida (Myers and Ewel, 1990). Cypress t upelo swamps provide food and cover for numerous animal speci es -especially amphibians, reptiles and birds. Because they are often nestled within pine plantations, the swamps act as refugia for displa ced wildlife after pine logging (Ewel, 1998). Perturbations, including natural flooding, fire and hurricanes as we ll as logging play an important role in shaping the ecological trajectory of swamps. It is thought that cypress tupelo swamps may be mostly dominated over time by regenerative processes that occur during recovery from minor or major perturbations (Odum and Ewel, 1984). Flooding is a frequent and mild pulse of energy that can affect plant regeneration and productivity as well as soil nutrient cycling. According to Myers and Ewel (1990), cypress tupelo swamps may experien ce several fires per century, and thes es events undoubtedly shape the long term plant succession of these swamps and possibly even plant evolution. Also, when fire occurs during dry periods, the peat layer can burn, and i t is thought that fire may be what limits the thickness of this layer (E wel, 1998 ).

PAGE 14

14 Hurricanes exert selective pressures on swamp vegetation (Putz and Sharitz, 1990), and logging drastically alters the condition of the swamp and may initiate changes in the type of ecosystem present. Interestingly, cypress tupelo swamps are tho ught to be subclimax communities that are maintained by natural perturbations that prevent the encroachment of other, less adapted species (Gunderson, 1977). Logging History Since the early 1800s, the logging industry in cypress t upelo swamps has been an import ant component of the economy of southern states of the Coastal Plain (Ewel, 1998 ). It is estimated that most cypress t upelo swamps in Florida have been selectively logged or clearcut at least one time for the valuable, rot resistant heartwood found in old growth cypress trees. The peak of c ypress logging occurred in the early 1900s, followed by collapse due to lack of accessible old growth trees and other economic constraints (Duryea and Hermansen, 1998). In the past few decades, there has been a r esurgence of cypress logging in the second growth swamps, in part because of new or more profitable markets. For example, demand for c ypress mulch for landscaping has risen such that as of 1996, 47% of harvested cy press wood was used to produce it (Duryea and Hermansen, 1998). Toda t upelo swamps is to complete a seed tree cut with the intention that natural regeneration will then occur (Ewel, 1998 where all harvesting equipment drove directly on the soil surface, causing soil rutting and destruction of tree stumps that may have otherwise sprouted. Also, during each operation, a large percentage of the swamp was affected by skid trails because deeply rutted trails were routinely abandoned in favor of routes through untouched soils.

PAGE 15

15 Bottom logging has thus received much scrutiny by a number of groups, including non governmental environmental organizations as well as the Environmental Protection Agency and the Division of Forestry. The result has been an impetus to develop alternative harvest strategies that may better protect water quality and otherwise better ma intain the ecological integrity of southern swamps. A number of alternative harvesting strategies are listed in the Florida Silvicultural Best Management Practices (BMP) Manual, which is an evolving document that includes guidelines for harvesting near rip arian areas and in wetlands, road and culvert design, site preparation and logging techniques. This manual is the product of the Silvicultural BMPs program that was started by the Florida Division of Forestry in 1979 to improve ecosystem protection related to forestry operations. It is continually updated via a technical advisory committee that includes state, federal, university, industry and environmental representatives (Karels, 2009). In the spring of 2007, an alternative swamp logging technique, refe impacts. In this technique, logs are cut and laid down length and skidding traffic is then c onfined to these routes during the logging operation. However, the impacts of this technique relative to former methods like bottom logging have not been quantified. Logging Concerns There are a number of areas of environmental concern and uncertainty in cypress tupelo swamp recovery after logging, including slow or undesirable vegetative succession, hydrologic alterations, physical and chemical changes in soils, and microtopographic changes. Cypress tupelo swamps are characteristically slow to

PAGE 16

16 regenerate to a desired vegetative composition (Kellison and Young, 1997) due primarily to the hindrance of woody plant regeneration under flooded conditions (Laderman, 1998) and secondarily to the paucity of advance regeneration at the time of harvest (Meadows and Stanturf, 1997). Also, while coppicing of both cypress and tupelo is common and seed production can begin as soon as two years after harvest (Ewel, 1998), long term survival of sprouts and their contribution to the future forest canopy remains uncertain (R andall et al., 2005, Conner and Buford, 1998). A recent study found that the percentage of cut cypress stumps to have live sprouts ten years after clearcutting ranged from 10 41% in seven North Florida domes (Ricci, 2010). In a number of case studies, swa mps quickly succeeded to dense shrubby cover following disturbance and subsequent tree regeneration was slow, unreliable, and tended to occur via cohort regeneration in years when conditions were dry and otherwise suitable (Dunn and Sharitz, 1987, Spencer et al., 2001). Cypress trees, in particular, have low seed viability and a narrow range of germination and establishment requirements, including open and unflooded conditions (Gunderson, 1977) and may only regenerate in large numbers when a number of crit eria are simultaneously met. While vegetation may be slow to recover after clearcutting of cypress tupelo swamps, the impacts of logging on soils and hydrology have been found to be relatively minor and short lived (returning to pre logging conditions wi thin 1 2 years) in a number of cases (i.e. Sun et al., 2001). The most commonly reported hydrologic change following logging in swamps is a temporary rise in the water table, presumably due to a decrease in evapotranspiration (Lockaby et al., 1997, Sun et al., 2000). This effect may be most pronounced when logging occurs during dry periods (Sun et al., 2000).

PAGE 17

17 Changes in soil properties tend to be spatially variable across logged sites, with impacts concentrated near skid trails (Sun et al., 2001). In parti cular, soil bulk density may increase (Grace et al., 2006) and hydraulic conductivity decrease (Gardiner et al., 1994). Soil organic matter also commonly decreases after logging (Aust et al.,1991), which can be attributed to both higher soil temperatures ( Trettin et al.,1996) driving higher rates of decomposition and a lack of organic inputs from vegetation. While the soil and hydrologic impacts of logging cypress tupelo swamps are relatively consistent and understood, considerable uncertainty remains reg arding logging impacts on the land surface character, and, specifically, the microtopography. Perhaps the most influencial stressor in swamp ecosystems is flooding, which is an obstacle to woody plant regeneration. Therefore, small topographic changes in these systems can have drastic consequences for vegetation. Microtopographic variability has been shown to influence plant regeneration (e.g.,Titus, 1990) and productivity (Jones et al., 2000) in swamps. The soft, organic surface soils are susceptible to compaction by heavy harvesting vehicles (Casey and Ewel, 1998) and removal of logs may also abrade the surface. Therefore, alteration of this surface could substantially alter the rate and/or mode of recovery of the swamp after logging. The objectives of this study were to compare 1) the vertical structuring of microtopography, 2) the horizontal structuring of microtopography, 3) the raised microtopographic feature substrate composition and size distribution between mat logged, bottom logged and reference sites, 4) skid trail dimensions between mat logged and bottom logged sites, and 5) the relationships between vegetation and microtopographic position in reference and logged swamps.

PAGE 18

18 CHAPTER 2 IMPACTS OF LOGGING O N MICROTOPOGRAPHY IN TAXODIUM NYSSA S WAMPS Background Swamps with an overstory dominated by c ypress ( Taxodium distichum var. nutans and var. distichum ) and swamp t upelo trees ( Nyssa sylvatica var. biflora ) are found throughout floodplain and depressional areas of the Southeast Coastal Plain ( Ewel 1998). In Florida, they often occur as semi isolated depressional features within a matrix of pine flatwoods or pine plantations (Ewel, 1998). Because hydroperiods of 10 12 months are characteristic of this forest association (Kellison and Young, 1 997), decomposition of organic materia l is slow, allowing a thick organic rich soil horizon to accu mulate. Microtopography and flucuations in water depth (Odum and Ewel, 1986) act to create a spatial and temporal variety of hydrologic niches that allow fo r woody plant regeneration. In addition to natural stressors, anthropogenic disturbances such as logging and hydrologic modifications have historically had large impacts on cypress tupelo swamps (Duryea and Hermansen, 1998). Over the past 200 years, logg ing has been a pervasive disturbance, and most cypress t upelo swamps in Florida and many across the southeast coastal plain have been selectively logged or clear cut at least on c e for the valuable, rot resistant heartwood found in old growth cypress trees (Odum and Ewel, 1986). The peak of c ypress logging occurred in the early 1900s, followed by collapse due to lack of accessible old growth trees and other economic constraints (Duryea and Hermansen, 1998). In the past few decades, there has been a resurgen ce of cypress logging in second growth swamps, in part because of new or more profitable markets. For example, demand for c ypress mulch for landscaping has risen such that as of 1996,

PAGE 19

19 47% of harvested cy press wood was used to produce it (Duryea and Herman sen, 1998). Currently, an estimated 1.3 million h a of cypress tupelo swamps are managed for timber across 13 states of the S outheastern U.S., and over half of the acreage occurs in Florida and Louisiana (USFS Forest Inventory and Analysis in Faulkner et al., 2009). A resurgance in cypress logging has renewed concerns about the impacts of this activity on ecosystem structure and function of these swamps. In recent decades, swamp logging has been conducted using a method know as where all harvesting equipment drove directly on the soil surface, causing soil rutting in skid trails (Figure 2 1) and destruction of tree stumps that ma y have otherwise sprouted. Also, during each operation, a large percentage of the swamp was affected by skid trails because deeply rutted trails were routinely abandoned in favor of routes through untouched soils. Bottom logging has thus received much sc rutiny by a number of groups, including the E nvironmental P rotection A gency due to these impacts and concerns about their long term effects. In the spring of 2007, an alternative swamp logging technique developed by loggers and referred to loggin ilvicultural B est M anagement P ractice manual for it s purported reductions in soil rutting (Karels, 2009) In this technique, logs are cut and laid down length skiddi ng traffic is then confined to these routes during the logging operation. After the operation is complete, all merchantable logs are removed from the skid trails (Figure 2 2). Today, this technique has all but replaced bottom logging for most loggers. How ever, the impacts of this technique relative to former methods like b ottom logging

PAGE 20

20 remain uncertain; perhaps most conspicuously, it is still not clear whether mat logging techniques results in a reduction in soil surface rutting. Microtopographic variatio n characterizes the surface of many Southeastern swamps, and clearcut logging may alter this character. Importantly, microtopography has been linked to enhanced botanical diversity (Moser et al., 2007, Bukata and Sloan, 2002, Simmons et al., 2009) and fin e root productivity (Jones et al., 2000), as well as having an influence over biogeochemical functions and hydrology. Additionally, elevated microsites have been shown to aid in seed entrapment (Bukata and Sloan, 2002), enhance wetland water storage and f iltration capacity (Bukata and Sloan, 2002), affect nitrogen mineralization and cycling (Bruland and Richardson, 2005), and provide aerobic conditions and subsequently increase success of vesicular arbuscular mycorrhizae (Cantelmo and Ehrenfeld, 1999). Co nsequently, microtopographic features may be useful as a metric for assessing the magnitude of physical damage or recovery after disturbance, and also for making predictions about the vegetative succession during recovery. Cypress tupelo swamps are often slow to recover vegetatively from disturbances, both natural and anthropogenic, as most woody plants require unflooded conditions for germination (Titus, 1990, Anderson, 2009), and many species, even those adapted to wetland conditions, grow optimally with some relief from flooding (Dickson and Broyer, 1971). Exacerbating the delay of arboreal regeneration are elevated water tables that can persist for months after logging due to decreased evapotranspiration (e.g. Sun et al. 2000) and soil compaction that can reduce local soil elevations and hydraulic conductivity, and increase bulk density (Gardiner et al., 1994, Grace lll et al., 2006).

PAGE 21

21 While the seeds of many wetland species can remain viable during prolonged submergence (Burns and Honkala, 1990) and cypress and tupelo readily sprout from coppice and can produce seed in under three years (Ewel, 1998), considerable uncertainty remains regarding long term regenerative capacities of swamp trees (especially cypress) due to po or initiation of cypress sprouts over time (Ricci, 2010). Microtopographic relief in swamp environments creates a variety of hydrologic niches that influence the vegetation dynamics following disturbance by providing microsites with different flooding re gimes. Woody vegetation is often preferentially associated with the elevated microtopographic features due to differential seed dispersal and germination as well as growth restrictions caused by flooding (Titus, 1990, Bukata and Sloan, 2002). A recent stu dy by Anderson (2009) highlighted the impacts of elevated microsites on the distribution of a woody shrub, Itea virginica Though the adults of this species tolerate flooded conditions, they are almost always found on elevated microsites due to restriction s they faced as seeds and seedlings. This and other evidence suggests that m icrotopography can be important in determining vegetation patterns in swamps (Moser et al., 2007). The arrangement and heterogeneity of microtopographic features across space m ay offer additional clues to the ability of a swamp to recover ecologically after disturbances. In other systems that contain biogenic microtopographic features, regular patterns have been observed, suggesting that scale dependent feedbacks, which include local positive feedbacks, but negative feedbacks at greater distances are controlling the spatial distribution of ecosystem features. For example, the Everglades in South Florida have elongated high ridges and low sloughs that are regularly spaced

PAGE 22

22 (Watts e t al., 2010). In Siberia, peatlands have been shown to have a regular, maze like sereis of ridges that support woody vegetation otherwise unsuccessful at colonizing inundated hollows (Eppinga et al., 2009). Evidence suggests that in both of these systems, higher evaportranspiration on ridges may be leading to higher nutrient concentrations on ridges, thus further enhancing productivity and evapotranspiration. This then leads to a paucity of nutrients at some distance, which perpetuates the hollow state. T hus, the arrangement of microtropographic features and measurements of disturbance effects on this may provide valuable information about the carbon dynamics of these systems. Also, the amount of topographic heterogeneity across horizontal space may allow for predictions of biotic recovery after logging based on the amount of physical complexity in the ecosystem. In addition to size and configuration of microtopography, substrate composition and stability can also be important to predicting plant regenerat ion responses. Microsites that are composed of logs or other woody debris are not as stable as those of soil, roots or stumps and may be less likely to support long term woody plant growth (Huenneke and Sharitz, 1990). Interestingly, plant species can be found preferentially with particular substrate types, suggesting that they benefit from particular characteristics of a substrate. For instance, in a South Carolina swamp, woody seedlings were found more often than expected on mucky substrates that were c lose to fixed objects, such as trees or roots, and were found less than would be expected on loose muck or muck that overlaid logs and/or branches (Huenneke and Sharitz, 1986). While emergent substrates can be ideal spots for seedlings, they may render adu lt trees more likely to be uprooted in storms, as was seen in perched Acer rubrum trees in a

PAGE 23

23 South Carolina swamp after Hurricane Hugo (Putz and Sharitz, 1991). It has also been found that disturbance can alter the abundance of particular substrate types of microsites (Huenneke and Sharitz, 1986), which may influence the composition of the regeneration plant community. A number of approaches have been used to quantify and characterize swamp microtopography. Commonly, mathematical measures of topographic he terogeneity such as vertical relief, rugosity (wrinkledness in a plane), tortuosity (ratio of over surface distance to the straight line path), and surface roughness (frequency of microtopographic features along a transect) have been employed to quantify t he degree of microtopography present (Bukata, 1999, Sloan, 1998, Moser et al.,2007). Others have tallied individual microtopographic features, including soil features as well as those composed of decomposing stumps, logs, roots and other elevated features, focusing on the abundance of these features, their suitability as germination sites, and plant preferences for particular features (Titus, 1990, Huenneke and Shartiz, 1986). This research draws from each of these approaches to allow for multiple comparis ons of microtopographic characteristics across recently logged and reference sites. The objectives of this study were to compare 1) the vertical structuring of microtopography 2) the horizontal spatial structuring of microtopography 3) the raised microt opographic feature substrate composition and size distribution between mat logged, bottom logged and reference sites; and, 4) skid trail dimensions between mat logged and bottom logged sites. These results may provide insight into the environmental soundne ss of two harvesting systems, as indicated by changes in microtopography and the potential for ecosystem recovery after logging.

PAGE 24

24 Methods Study Sites Sites were located within backwater floodplain, strand, and depressional Taxodium Nyssa swamps on industri al (agricultural/silvicultural) and federal lands in Alachua, Gilchrist, Bradford and Baker counties of North Central Florida between 29 2 1). ver age annual precipitation is 1250 mm, with a verage annual high and low temperatures of 26.6C and 14.2 C (NOAA, 2010). Soils are very poorly drained fine sands and sandy loams (Monteocha and Mascotte series), very poorly drained fine sands with a mucky fine sand surface layer (Lynn Haven series) an d very poorly drained mucks with loamy substratums (Pamlico series) (NRCS 2010 ) Depth to water table for each of these series is zero inches, and each are found in depressions in the landscape. Pine plantations and isolated wetland features represent the m ain type of land use in the area (Table 2 1). Tree canopies were most often dominated by T distichum var. nutans and N. sylvatica var. biflora with other species such as Magnolia virginiana, Gordonia lasianthus and Persea borbonia sometimes common or even co dominant in the canopy. Microtopographic relief had an elevational range of 0.75 1.4 m within each site. Sampling occurred between January and May of 2010, during which monthly rainfall was near average and air temperatures were colder than av erage by 3.3C in January through March (NOAA, 2010). In most sites, all soil except the raised microtopographic features was submerged in tannic water for the duration of the study.

PAGE 25

25 In total, 5 reference swamps were sampled. The sites were chosen based on their similarity in species composition and hydrology to sampled logged sites and often were in close proximity (< 1 km) to the logged sites. Each of these sites exhibited evidence of a history of at least some selective logging within the past century swamps that had escaped all logging and/or hydrologic modification are rare in North Florida. Basal areas exceeded 34.4 m 2 /ha in these sites, though it was difficult to measure accurately due to buttressing. The understory vegetation wa s characteristically dense and was often dominated by woody shrubs such as Lyonia lucid a and Itea virginica, among others. Additionally, ferns, forbs and vines were also common in some sites. Five recently logged sites were sampled, 2 of which had been mat logged (1 and 5 years ago), and 3 of which had been bottom logged (3,7 and 13 years ago). These sites ranged from having sparse or very young tree regeneration (basal area of 0.5 1.2 m 2 /ha) to dense regeneration of coppice and/or seedling recruits (bas al area of 4.6 9.2 m 2 /ha). Field Methods Water depths In each study swamp, I randomly located two 30 x 30 m sampling frames (except swamps by satellite imagery. I di vided each frame into four quadrants and randomly located two non overlapping sampling stations in each quadrant, for a total of eight sampling stations per frame (Figure 2 3A). A sampling station consisted of 16 water depth measurements, four taken at ra ndom distances between 0.5 4 m along each cardinal direction from the center of the station (Figure 2 3B). When raised

PAGE 26

26 microtopographic features were encountered at sampling points, their height above the water surface was recorded as a negative water de pth. For each water depth measurement, I recorded its location as being in one of three categories: skid trail (in a rutted area where a skidder had driven), swamp bottom (flat, low, and often flooded soil surface) and raised microtopographic features (di stinctly higher than the surrounding soil surface, often unflooded). The stratified locations of the sampling stations across a sampling frame ensured that an adequate number of point pairs were sampled for each lag distance to conduct a spatial autocorrel ation analysis. Skid trail transects To compare the impacts of the two logging methods, fifteen transects in two mat logged sites (seven in a site logged one year ago and eight in a site logged five years ago) and fifteen transects in two bottom logged s ites (seven in a site logged seven years ago and eight in a site logged thirteen years ago) were traversed perpendicular to skid trails, with water depth measurements recorded every 0.5 m. Each transect extended for an additional 2 5 m beyond each trail e dge. Trail edges were defined based on visual and elevational evidence present at the site. Elevated microsite surveys A survey of elevated microsites was conducted in two reference sites, two mat logged sites and two bottom logged sites. A total of 291 r andomly located features in reference, 134 in mat logged and 128 in bottom logged were surveyed. Each elevated feature within each plot was measured for maximum length and width and classified by type as follows: tree base expanded bases of live trees wi th some litter accumulation (Figure 2 4), soil predominantly organic and/or inorganic soil material (Figure 2 5), stump/soil decomposing stump with substantial accumulations of litter and

PAGE 27

27 decomposed organic material (Figure 2 6), roots/soil exposed t ree roots with substantial accumulation of litter and decomposed organic material (Figure 2 7), log dead wood (Figure 2 8), stump tree stump with little or no decomposed organics or litter associated with it (Figure 2 9), tip up mound root ball and a ssociated soil from up turned trees (Figure 2 10), and slash debris dense accumulation of course woody debris (Figure 2 11). Data Analysis Water depths Water depths were converted to relative elevations by setting the deepest observation in each frame equal to zero, and then histograms of elevation were created for each. To model the data according to elevational clusters, the univariate model based clustering feature of Mclust, a package for the statistical software R, was used to generate best fit mod els for the set of observations (128) in each sampling frame across all sites. For univariate data, two types of models were possible: equal variance and unequal variance. Upon model selection, Mclust then found the number of components (or normal distrib utions) that resulted in the best fit Bayesian Information Criterion (BIC) value, incurring a penalty that increased based on the number of components (Fraley and Raftery, 2006) (Appendix A). To verify the presence of sp atial autocorrelation within the si tes, a maximum likeli hood analysis (AIC) was used to determine if an exponential, spherical, or Gaussian spatial variogram model ex plained more variation among sites than a non spatial model Using a mixed model (proc mixed) in SAS statistical software, a verage sill, range, and nugget values for reference, mat, and bottom logged sites were obtained. Variograms were then drawn based on these values and the most

PAGE 28

28 appropriate spatial variogram model (previously obtained via AIC values). Correlograms were also constructed using GS+ software for each frame in order to compare spatial correlations for each site. Skid trail t ransects Average skid trail depths, measured for each transect as the difference in elevation between on and off trail measurements, were calculated for mat and bottom test. Average skid trail widths were also compared test. A composite skid trail profile was created by aligning all 15 transects from each logging treatment at the trail centers and obtaining an average elevation at each measurement point. Elevated microsite surveys The size of each el evated microsite was estimated as the area (m2) of an ellipse. The median and interquartile range of microsite areas were calculated for each surveyed site. Features are characteristically steep sided, so change in area with water level was assumed to be negligible for comparing general characteristics of elevated microsites across si tes. For each substrate type, I test ed for differences in sample proportions between each site treatment (reference, mat logged, bottom logged) (alpha = 0.05). The mean sizes of elevated features were compared between sites using the Bonferroni multiple comparison method in an ANOVA (alpha = 0.05) (Appendix B). Results Water Depths Vertical s tructure In reference sites, two elevational modes (bimodality) was the best fit mod el for seven of nine frames (Figures 2 12 and 2 13, Table 2 2). The other two frames were

PAGE 29

29 trimodal and pentamodel. (Figure 2 13). In each of the nine reference frames, the highest elevational mode included a surprisingly consistent average of 22.6% ( 1.67 1%) of observations for that frame in the model output. In each frame, this mode was composed almost entirely of observations whose topographic positions had been ref erence frame also displayed a lower elevational mode that contained the majority of observations. For each frame, this mode was composed almost entirely of observations results suggest that reference swamps are characterized by two elevational modes: the swamp bottom, and distinctly higher microtopographic features. In mat logged sites, bimodality was the best fit model in three of the four sampling frames, though one of the f our frames was unimodal (Figure 2 14, Table 2 2). As seen in reference sites, the majority of the observations were found in the lower elevational mode, and a smaller proportion were found in the higher elevational mode. However, the higher elevational mo de for mat logged frames contained an average of only 10.6% ( 4.345%) of the observations, which was lower than that of reference sites. The vast majority of observations in the higher elevational mode corresponded to field topographic positions. The decrease in the number of observations in the higher elevational mode (or, in one case, the absen ce of this mode) as compared to reference frames suggests a loss in the area of elevated microtopographic features relative to swamp bottom.

PAGE 30

30 In bottom logged sites, one frame was trimodal, two were quadramodal, and three were bimodal (Figure 2 15, Table 2 2). Five of the six bottom logged frames had a these frames. Unlike mat logged fra mes, four of the six bottom logged frames also observations according to field classification. As in other sites, the majority of elevational mode. These results suggest that bottom logging results, not only in the loss of elevated microtopographic features, but the creation of deeply rutted areas that are distinctly lower than the swamp bottom. In co mparing mat and bottom logged frames, was variable across logged sites, and did not appear to be a function of logging treatment and years since harvest Horizontal s patial structure The exponential spatial variogram model was consistently the best fit model (lowest AIC value) for explaining the distribution of soil elevations for each site. Average e xperimental variogram parameter s differed between reference, bottom logged and mat logged sites. Reference sites had the highest sill or variance (782 cm 2 ) relative to that bottom logged sites (763 cm 2 ) and mat logged sites (420 cm 2 ) (Figure 2 16). Bottom logged sites exhibited the great est range of spatial autocorrelation (501cm), while reference sites had a range of 200 cm and mat logged sites had a range of 157 cm. All sites had low nugget variance, indicating that most of the variation in soil

PAGE 31

31 elevations was explained by location in space. Correlograms revealed no consistent or strong patterns in microtopographic variance across sites. Skid Trail Transects Skid trails resulting from bottom logging operations were, on average 34.7 cm ( 2.350) deep relative to the surrounding land, wh ich was deeper than trails following mat logging operations, which were 17.5 cm ( 1.450) deep (p <.0001) (Figure 2 17). Mat logging and bottom logging trails were not different in width, averaging 5.2 m ( 0.332) m and 4.5 m ( 0.276), respectively. Ti me since logging (up to 13 years) did not appear to effect the depth or width of skid trails. Elevated Microsite Surveys Substrate characteristics of elevated microsites The relative proportions of microtopographic feature substrate type s varied between reference, mat logged, and bottom logged sites (Figure 2 18). In mat and bottom logged sites, a lower proportion of features were root/soil (24.5% and 11.7%, respectively) as compared to 52% in reference (p <.0001, p <.0001). Though relatively uncommon i n all sites, tip up mounds were most prevalent in reference sites, comprising 2.9% of elevated features, were rarer in mat logged sites (p = 0.0440), comprising 0.5% of features (Figure 2 18B), and were not encountered in bottom logged sites (Figure 2 18C) Stump microsites were rarer in reference sites (0.4%) than in mat and bottom logged sites (p <.0001, p <.0001), comprising 15.2% and 30.5 % of microsites, respectively. Decaying stump/soil microsites were also proportionally more common in both mat logg ed sites (43.2%) and in bottom logged sites (50.0%) than in reference (31.5%) (p <.0001, p <.0001). Decaying log microsites and tree base microsites were only found in reference sites, making up 1.6% and 5.0% of elevated

PAGE 32

32 features, respectively. Slash debri s was only found in mat logged sites and made up 6.1% elevated microsites. Differences between logging types included a greater proportion of stump (30%) and soil features (8%) in bottom logged sites and a lower proportion of root/soil features (12%) than in mat logged (15%, 3% and 25% respectively). Size of elevated microsites The size distribution of microtopographic features varied from site to site, and no clear differences in the average size or the size range of these features were observed between r eference and logged sites (Figure 2 19). Rather, size distribution and range appeared to be a function of proximity of sites to each other, regardless of logging history. For example, sites 4R (reference) and 5M (mat logged 5 years ago) were located adjac ent to one another in a backwater riverine swamp and had mean elevated microsite areas of 1.39 m 2 and 1.83 m 2 respectively, which did not differ from one another, but differed from other sites that were compared. Similarly, sites 3R (reference) and 1M (ma t logged 1 year ago) were located adjacent to one another in the same depressional complex, and had mean elevated microsite areas that did not differ from one another (0.48 m 2 and 0.41 m 2 respectively). Sites 7B (bottom logged 7 years ago) and 13B (bottom logged 13 years ago), though not adjacent, were within 5 km of each other as part of a contiguous, managed pine flatwood and swamp mosaic, and had mean elevated microsite areas that did not differ (0.63 m 2 and 0.85 m 2 respectively, alpha = 0.05), although they also did not differ from sites 3R and 1M.

PAGE 33

33 Discussion Microtopography in Reference C onditions There was a striking similarity in the frequency and distribution of elevated microsites across reference stands measured in our study. Microtopographic f eatures represented a distinct elevational state from the swamp bottom. Reference sites showed a strong bimodality in elevations, and except for one site (3R; Figure 2 12), the elevational distributions appearred quite similar. There was also a striking co nsistency in the proportion of observations in each elevational mode across all reference sites, with about 1/5 th of all observations falling into the higher elevational mode. Interestingly, Bukata and Sloan (2002) predicted this to be the ideal proportio n of elevated features in terms of productivity and water storage in a model developed based on Florida swamps. Swamp trees will organize over time to fill all canopy space, and this proportion of elevated features may correspond to the amount of adventiti ous and aerial root mass that the limited biomass of trees can produce considering continual losses due to organic matter oxidation. These features are biogenic, having formed directly from living plant tissues and from organic materials accumulated from t he death of these tissues, and it is likely that environmental factors such as hydrology, nutrient inputs and disturbances govern the formation and maintenance of microtopographic features. Elevational bimodality, which implies that there are two stable elevations, has been observed in a number of wetland ecosystems, from ombrotrophic peatlands in Siberia (Eppinga et al., 2008) to the Everglades marshes of South Florida (Watts et al., 2010). In these systems and likely in Southeastern swamps, short range positive feedback between productivity and organic matter accretion cause the formation of elevated features (Eppinga et al., 2008). Higher elevations have higher productivity than

PAGE 34

34 lower (Jones et al., 2000), which leads to accretion of organic materials, creating a positive feedback. By contrast, low elevations that experience prolonged flooding have lower productivity, which prevents the accretion of organic materials, thus perpetuating the low elevational state. In cypress tupelo swamps, this mechanism may contribute to the formation of steep sided and distinct microtopographic features, though further studies are needed to elucidate detils of this process. In contrast, results from our correlograms indicated no regular horizontal spatial patterning in microtopography across the landscape, which suggests a lack of strong negative productivity feedbacks at some distance away from elevated features, a prerequisite for the formation of spatial elevational patterns (Reitkert and Koppel, 2007). Regular spaci ng of elevated features, as seen in some peatlands (Eppinga et al., 2008) would be predicted only if the formation of elevated features effectively inhibited the formation of additional features for some distance. It is possible that weak, scale dependent feedbacks are present, the self organization of microtopographic features in swamps is a very gradual process, and only in remote swamps that have not experienced large disturbances for long periods of time would regular microtopographic patterns become d etectable. determing the arrangement of elevated features as they nucleate the features. There is evidence from other Southeastern swamps that as trees of become larger, they tend towards regularity in spacing, suggesting that old forests may be most likely to exhibit overall regularly in tree patterning (Good and Whipple, 1982). Spatial structuring of microtopography would also be predicted over time based on this mechanism. If this is

PAGE 35

35 th e case, the reference sites in this study may not be developed to this point since most were previously selectively logged or have experienced some degree of hydrologic modification. Additionally, selective logging opens the tree canopy and originates stu mp and log features, which may alter the natural arrangement of microtopgraphy. Impact of Logging on M icrotopography Both mat logging and bottom logging alter the frequency and distribution of elevated microsites, though to varying degrees as compared to reference sites. Mat logged sites appeared to have weaker bimodality as compared to reference sites, a much smaller number of observations in the higher elevational mode, and a smaller height difference over the swamp bottom than the corresponding mode in reference sites. This homogeneity in elevational ranges following mat logging was also detected by the much smaller sill of the variograms as compared to reference sites. However, the relatively short range of spatial autocorrelation in both reference and mat logged sites indicates that microtopographic variability occurs on the scale of individual trees in both types of sties. Both mat and bottom logging resulted in the loss of microtopographic features and/or in the height of these features. During harvest operations, uprooting of trees by heavy machinery and compression by skidder traffic has the potential to destroy microtopographic features. Additionally, dragging of logs may abrade microtopographic features as was seen in a study of Canadian pea tlands where logging resulted in reduced heights of hummocks (Locky and Bayley, 2006). In the years following logging, additional losses in microtopographic features by decomposition could potentially result from a lack of carbon inputs and death of suppo rting tissues such as roots. Conversely,

PAGE 36

36 some new elevated features are created during logging in the form of stumps and logs that are left after harvest and soil mounded by skidders, but the proportions of such sites is nevertheless much reduced due to lo gging. Temporary rises in the water table following logging may act to slow decomposition and thus protect remaining organic features. Bottom logging had similar impacts as mat logging to the frequency and distribution of elevated microsites, though mat logging appeared to minimize the impact of skid trails on the swamp bottom as compared to bottom logging. In mat logged sites, skid trail observations were always included in the lower elevational mode, indicating that, across frames, they were not suffic iently rutted to be clustered as a separate elevational mode from the swamp bottom. On the other hand, skid trails produced by bottom logging were often substantially deeper than the swamp bottom, as evidenced by the trimodality of elevation modes due to a very low elevational mode classified separately from the swamp bottom. Interestingly, the sill of the variograms in bottom logged sites were more similar to reference conditions than mat logged sites. It may have been possible that the older bottom logg ed sites have had more time to recover towards reference conditions. However, given the evidence of deeper skid trails as measured in the line transects and detected in the trimodality of elevational distributions, it is more likely that bottom logged site s appear more similar to reference conditions due to the heterogeneity introduced by deeply rutted skid trails. This suggests thatvariograms provided only limited ecological information as overall heterogeneity measures did not necessarily correlate to im proved site conditions.

PAGE 37

37 Microsite Availability, S ubstrate Type, and Potential for Regeneration The loss of elevated microtopography following logging may impede recolonization by plant species that are the least tolerant of flooded conditions. For exam ple, species such as Ilex spp are often confined to the higher elevations in swamps and may be at a disadvantage, whereas other species such as Cephalanthus occidentalis and Nyssa spp. are found at lower elevations (Titus, 1990) and may be better suited t o post harvest conditions. Most shrubs are confined to elevated areas (i.e. Anderson et al., 2009); thus, the loss of these microsites may alter the structure of the vegetation in cypress tupelo swamps by decreasing the sites for shrub regeneration. Coho rt regeneration during periodic dry downs may be a means of tree canopy recovery in some swamps (Dunn and Sharitz, 1987). However, this was not observed in the study sties, which generally were dominated by shrub, small tree and coppice regeneration. Mat logging may offer an advantage over bottom logging in the reestablishment of woody vegetation following logging by preventing the formation of deeply rutted, permanently flooded areas, though it remains to be seen if this will foster tree canopy recovery. The substrate composition of microtopographic features is an important component of their quality for woody plant regeneration and survival. Studies have shown that woody seedlings in Southeastern swamps are distributed nonrandomly across microsite types with stable substrates such as soil imme diately adjacent to trees, aerial roots, and knees harboring more seedlings than expected given the abundance of these sites (Titus, 1990, Huenneke and Sharitz, 1986). This pattern may be due to uneven numbers of seeds arriv ing at these sites and/or higher germination and/or survival of seedlings at these sites Long lived plants such as trees may stand a

PAGE 38

38 better chance at reaching maturity on sites that are not likely to float away, deteriorate or erode. In cont rast, vines and shrubs with shorter life spans are spread more evenly across substrates types and are found more often on less permanent substrates s uch as logs and stumps than trees (Titus, 1990). Tip up mounds support greater amounts of Rubus spp. and Vitis spp. than would be expected given the abundance of these features in studied sites (Titus, 1990). A study of Nyssa sylvatica revealed each substrate type to prese nt advantages and disadvantages for regeneration. While emergent sites such as stumps, knees, and tree bases had much higher rates of germination than lower microsites (28% vs. <1%), predation was also much greater in emergent sties (Huenneke and Sharitz, 1990). The ideal microsite is likely to be species dependent and also contingent of a variety of environmental conditions including flooding regime, predators, competition and light conditions. Shifts in the composition of microtopographic features therefore has the potential to influence the succession of the plant community in swamps. L ive tree roots, including cypress knees, shallow root masses of tupelo and an important source of stable microtopographic features in reference sites that could potentially support tree regenera tion. Stump/soil features were also common due to both natural tree death and perhaps stumps that remained from past selective logging operations, though these features are prone to decomposition and are perhaps more short lived than live root features. T he relative loss of root/soil features in both mat and bottom logged sites, which is perhaps only due to the large increase in stump features, or possibly due to the loss of root features via uprooting during logging or slow deterioration due to death of

PAGE 39

39 r oots, indicates that the features that remain after logging may not be as stable or long lived as those of reference sites. The creation of slash debris as a result of mat logging could be due to the nature of the logging technique, which may cause more br anches to be left in the vicinity of skid trails from the mats used to cover the trails. Alternatively, the slash debris, which likely decomposes relatively rapidly, may have been present only in mat logged sites because they were younger in age than the bottom logged sites. Conclu ding Remarks Water depth surveys and subsequent creation of elevational histograms was an effective way to get a sense of the topographic variability at each site. This method is effective if vegetation is not too thick and if th e area is mostly flooded. However, for dry sites, similar data can be collected via a laser level and statia rod design (e.g. Bukata and Sloan, 2002). Skid trail transects proved an effective and easily interpreted method to assess soil rutting relative to the surrounding land surface, and a surprising level of detail was visible in composite transects, including individual tire tracks. Bimodality in elevations, relatively high spatial variance and a short range of autocorrelation across reference sites hi ghlights the presence of distinct, elevated microtopographic features at the scale of individual plants. These features may be maintained via positive productivity feedbacks and limited by continual carbon oxidation and discrete tree productivity, as evid enced by the similar prevalence of elevated observations across all reference sites. A lack of evidence for spatial patterning suggests a lack of strong negative distal feedback that would regulate elevated feature spacing. These features undoubtedly play an important role in shaping plant communities and promoting diversity and productivity in many cypress tupelo swamps.

PAGE 40

40 Reduced frequency and height of elevated microtopographic features and altered substrate compositions of these features that result from mat and bottom logging have the potential to alter the ability of cypress tupelo swamps to support tree regeneration and could alter biogeochemical cycling via changes in productivity and/or plant communities. Reductions in distinct, raised microtopograp hy were often near 50% in recently clearcut sites, which could favor species that tolerate flooding and decrease regeneration of those who need dry ground for germination and/or growth. Because bottom logging maintains overall microtopographic heterogeneit y across space, it may actually support a relatively diverse plant assemblage in comparison to mat logging. H owever, if site recovery toward historical conditions is desired, mat logging, while causing a loss in microtopographic heterogeneity, may have eq ual or better potential to regenerate over time due the absence of deep skid trails. Specifically, mat logging offers an advantage over bottom logging by substantially reducing the depth of soil rutting, thereby avoiding long term flooding in trails. This will encourage woody plant regeneration (including valuable timber trees), which is required of ongoing forestry operations under the Federal Clean Water Act. It is not clear yet if mat logging provides a long term advantage over bottom logging for full e cosystem recovery. There has scarcely been enough time since mat logging and even bottom logging began across large areas to understand the long term consequences; thus, it will be important to continue to track recovery of these ecosystems over greater t ime scales. The styles used in this document are called paragraph styles. Paragraph styles are used to format the entire text within a paragraph. To apply a style, follow these instructions:

PAGE 41

41 Table 2 1. List of study sites. ( *Basal areas were estimated > 34.4 m 2 /ha for reference sites, however, accurate measurements could not be made due to buttressing ) Site # treatment yrs since logging Landscape position Surrounding land use Basal area (m 2 /ha) County Location 1R Reference NA depression residential/agriculture -Alachua 2937'43.60"N, 8217'54.14"W 2R Reference NA Strand national forest/wetlands -Baker 3015'27.30"N, 8224'30.40"W 3R Reference NA depression pine plantation/wetlands -Alachua 2946'2.13"N, 8213'7.35"W 4R Reference NA backwater floodplain pine plantation/river floodplain -Bradford 2950'8.89"N, 82 9'30.92"W 5R Reference NA depression pine plantation/wetlands -Alachua 2947'7.95"N, 8216'8.54"W 1M Mat logged 1 depression pine plantations/wetlands 0 Alachua 2945'57.88"N, 8212'53.16"W 5M Mat logged 5 backwater floodplain pine plantations/river floodplain 1.0 Bradford 2950'26.23"N, 82 9'13.21"W 3B Bottom logged 3 depression pine plantations/wetlands 0.8 Alachua 2946'50.77"N, 8216'0.92"W 7B Bottom logged 7 depression pine plantations/wetlands 3.4 Gilchrist 2947'26.29"N, 8248'33.91"W 13B Bottom logged 13 depression pine plantations/wetlands 5.9 Gilchrist 2948'12.55"N, 8247'14.76"W

PAGE 42

42 Table 2 2. BIC values for unimodal, bimodal and trimodal models generated by the the best fit model. site sampling frame m odel variance components unimodal bimodal trimodal Reference 1 1 equal 2 1271.0 1206.0 1215.3 Reference 1 2 equal 2 1255.0 1232.5 1242.5 Reference 2 1 unequal 2 1088.0 1052.9 1058.5 Reference 3 1 unequal 5 893.0 849.3 849.3 Reference 3 2 equal 3 900.5 882.9 880.6 Reference 4 1 unequal 2 1180.0 1138.1 1145.0 Ref erence 4 2 equal 2 1185.0 1165.1 1174.9 Reference 5 1 unequal 2 1158.5 1105.3 1119.1 Reference 5 2 equal 2 1236.0 1204.2 1213.0 Bottom logged 3 1 unequal 2 1011.0 980.7 991.8 Bottom logged 7 1 unequal 2 1143.5 1136.3 1136.6 Botto m logged 7 2 equal 4 1161.0 1158.0 1120.0 Bottom logged 13 1 unequal 4 1257.0 1213.4 1216.8 Bottom logged 13 2 equal 2 1163.2 1162.9 1165.0 Mat logged 1 1 unequal 2 1157.0 1089.1 1097.0 Mat logged 1 2 equal 1 1140.0 1145.2 1149.0 Mat logged 5 1 equal 2 1101.9 1099.8 1107.0 Mat logged 5 2 equal 2 1072.0 1053.9 1060.9

PAGE 43

43 Figure 2 1. Skid trails created during bottom logging operations A) 10 yrs ago and B) 12 yrs ago. Figure 2 2. Skid trails created during mat logging operations A) 1 yr ago and B) 5 yrs ago. A A B B Photo by Amy Washuta Photo by Amy Washuta Photo by Amy Washuta Photo by Amy Washuta

PAGE 44

44 Figure 2 3. A ) A water depth sampling scheme for one 30x30m frame and B) An example sampling station within the frame with 16 sampling points distributed in four cardinal directions. N 4m 4m B N A

PAGE 45

45 Figure 2 4. A tree base microtopographic feature. so Figure 2 5. A soil microtopographic feature. Photo by Amy Washuta Photo by Amy Washuta

PAGE 46

46 Figure 2 6. A stump/soil microtopographic feature. Figure 2 7. A root/soil microtopographic feature. Photo by Amy Washuta Photo by Amy Washuta

PAGE 47

47 Figure 2 8. A log microtopographic feature. Figure 2 9. A stump microtopographic feature. Photo by Amy Washuta Photo by Amy Washuta

PAGE 48

48 Figure 2 1 0. A tip up mound microtopographic feature. Figure 2 1 1. A slash debris microtopographic feature. a) Photo by Amy Washuta Photo by Amy Washuta

PAGE 49

49 Figure 2 1 2 Elevational histograms for frames in three reference swamps. Shades indicate elevational modes modeled by the Mclust package (R software). 0 5 10 15 0 20 40 60 80 100 120 Frequency Elevation (cm) 1R (2) 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 1R (1) C B A D 0 5 10 15 20 25 30 0 20 40 60 80 100 120 Frequency Elevation (cm) 2R (1) 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 3R (1) Swamp bottom 4 Swamp bottom 3 Swamp bottom 2 Elevated microsites Swamp bottom 1 0 5 10 15 0 20 40 60 80 100 120 Frequency Elevation (cm) 3R (2) E D C

PAGE 50

50 Figure 2 13 Elevational histograms for frames in reference swamp 4R (A and B) and 5R (C and D) Shades indicate elevational modes modeled by the Mclust package (R software). 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 4R (2) 0 5 10 15 20 25 0 20 40 60 80 100 120 Frequency Elevation (cm) 5R (1) 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 5R (2) 0 5 10 15 20 25 0 20 40 60 80 100 120 Frequency Elevation (cm) 4R (1) Swamp bottom 1 Elevated microsites A B C D

PAGE 51

51 Figure 2 14 Elevational histogr ams for swamps mat logged 1 year ago (A and B) and 5 years ago (C and D) Shades indicate elevational modes modeled by the Mclust (R software.) 0 5 10 15 20 25 0 20 40 60 80 100 120 Frequency Elevation (cm) 1M (2) B 0 5 10 15 20 25 30 0 20 40 60 80 100 120 Frequency Elevation (cm) 1M (1) Swamp bottom 1 Elevated microsites A C 0 5 10 15 20 25 0 20 40 60 80 100 120 Frequency Elevation (cm) 5M (1) 0 5 10 15 20 25 0 20 40 60 80 100 120 Frequency Elevation (cm) 5M (2) D C

PAGE 52

52 Figure 2 15 E levational histograms for swamps bottom logge d 3 years ago(A and B), 7 years ago (C and D), and 13 years ago (E and F) Colors differentiate elevational modes modeled by the Mclust package (R software). 0 10 20 30 40 50 0 20 40 60 80 100 120 Frequency Elevation (cm) 3B (1) 0 10 20 30 40 0 20 40 60 80 100 120 Frequency Elevation (cm) 3B (2) 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 7B (1) 0 5 10 15 20 25 30 5 25 45 65 85 105 125 Frequency Elevation (cm) 7B (2) Skid trails Swamp bottom 2 Elevated microsites Swamp bottom 1 0 5 10 15 20 25 0 25 50 75 100 125 150 175 Frequency Elevation (cm) 13B (1) 0 5 10 15 20 0 20 40 60 80 100 120 Frequency Elevation (cm) 13B (2) A B C D E F

PAGE 53

53 Figure 2 16 Exponential theoretic al variograms modeled after the composite variogram parameters for all mat logged bottom logged and reference sites 0 100 200 300 400 500 600 700 800 900 50 200 350 500 650 800 950 1100 1250 1400 1550 Variance average lag distance for class (cm) Mat logged Bottom logged Reference

PAGE 54

54 Figure 2 17 Composite transects showing average elevations in 15 transects acros s mat logging skid trails and 15 transects across bottom logging skid trails 95% confidence envelopes are shown with each composite transect. Transects are superimposed so that the average off trail elevations for mat and bottom logging trails are equal. 45 40 35 30 25 20 15 10 5 0 5 10 15 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Relative elevation (cm) meters across stand Mat logging ave. Bottom logging ave.

PAGE 55

55 Figure 2 18 The relative proportions of various microsite substrate types in A) refer ence swamps, B ) mat logged swamps and C) bottom logged swamps wit h (*) indicating a difference in p roportion from reference swamps, and (~) indicating a difference in proportion between mat and bottom logged swamps 5% 6% 32% 52% 2% 0.4% 3% Tree base Soil Stump/soil Roots/soil Log Stump Tip up mound 0.5% 3% 51% 25% 6% 15% Soil ~ Stump/soil* Roots/soil* ~ Slash debris* ~ Stump* ~ Tip up mound* 8% 50% 12% 30% Soil ~ Stump/soil* Roots/soil* ~ Stump* ~ A B C

PAGE 56

56 Figure 2 19 Boxplots indicating the minimum, maximum, mean, interquartile range and outliers for the area (m 2 ) of elevated microtopographic features across six A A A A B B

PAGE 57

57 CHAP TE R 3 RELATIONSHIPS BETWEE N VEGETATION AND MICROTOPOGRAPHIC POS ITION IN REFERENCE AND CLE ARCUT T AXODIUM NYSSA SWAMPS Background Flooded conditions pose problems for terrestrial woody plants as they greatly reduce oxygen available to the roots and allow a build up of toxins tha t would normally be degraded under aerobic conditions. Also, this reducing environment can alter the form in which nutrients exist and thus affect their availability to plants (Conner and Bufo rd, 1998). Even the most flood tolerant species, such as cypress ( Taxodium distichum ) and tupelo ( Nyssa sp p .) have been shown to perform b etter in wet, but aerated conditions than under anaerobic saturation such as that found on swamp bottoms (Dickso n and Broyer, 1971). T hese species as well as most other woody plants that thrive in swamps cannot generally germinate in submerged soils (Mye rs and Ewel, 1990, Titus, 1990). Disturbances such as clearcutting in these swamps are often followed by slow regeneration due to the hindrance of woody plant germination and establishment in flooded conditions (Laderman, 1998) and can exacerbate flooding due to decreased evapotranspiration (Sun et al., 2000), making microtopographical highs potentially important regeneration microsites in these swamps. In addition to flooding, to the paucity of advance reg eneration at the time of clearcutting may contribute to slow regeneration of desired woody species (Meadows and Stanturf, 1997). Also, while coppicing of both cypress and tupelo is common and seed production can begin as quickly as two years after harvest (Ewel, 1998), long term survival of sprout s and their contribution to the future forest canopy remains uncertain (Randall et al., 2005, Conner and Buford, 1998). A recent study found 10 41% of cut cypress stumps to have live sprouts ten years after clearcutting in eight North Florida

PAGE 58

58 domes (Ricci 2010), and early studies revealed survival rates between 17 23% for cypress stump sprouts (Conner et al., 1986). In a number of case studies, swamps quickly succeeded to dense shr ubby cover following clearcutting or thermal effluent wastewater loading but subsequent tree regeneration was often slow, unreliable and tended to occur via cohort regeneration in years when conditions were dry and appropriate (Dunn and Sharitz, 1987, Spencer et al., 2001). Cypress trees, in particular, have low seed viability and a narrow range of ideal germination and growth requirements, including open and unflooded conditions (Gunderson, 1977) and may only regenerate in large numbers in when a number of c riteria are simultaneously met. Microtopographic variability provides a variety of hydrologic niches for plants and is likely a key structural component regulating the recovery of cypress tupelo swamps. The microtopography in many southeastern swamps has an elevational range on the order of 0.5 1.5 m (Titus, 1990), with elev ated features rising above a lower swamp bottom. Substrate conditions on elevated microtopographic features differ from that of the swamp bottom in that they are more aerobic, potentially more productive (Jones et al., 2000), harbor more arbuscular mycorrh izae structures (Cantelmo and Ehrenfeld, 1999), and potentially contribute disproportionately to ammonification and nitrification (Bruland and Richardson, 2005). Also, m icrotopograp hic features have been shown to influence the distribution of swamp plant species over the life cycle of the plants. Woody seedlings are often preferentially associated with elevated features (Titus, 1990, Anderson, 2009) due to factors including differential seedfall, seed trapping, germination and survival across microsites ( Schneider and Sharitz, 1988, Titus, 1991 ). Individual plant species can show strong preferences for a particular and often unique elevational

PAGE 59

59 range, with many woody and rare species favoring elevated mi c rosites (Vivian Smith, 1997, Simmons et al.,2009 Tit us, 1990). For example, in a study of woody seedlings in a Florida floodplain swamp, Ilex cassine V accinium elliottii and Lyonia lucida were found at mean elevations about 20 cm higher than that of Cephalanthus occidentalis Taxodium distichum and Diosp yros virginiana (Titus, 1990). In general, early successional species have been shown to exhibit wide r tolerance ranges for microtopographic positions in comparison to later successional species, which often thrive in narrower elevational ranges (Simmons et al., 2009). Beyond individual plant species, microtopographic heterogeneity has been correlated to enhanced botanical diversity in both observational and experimental wetland studies ( Vivian Smith, 1997, Moser et al.,2007, Sloan, 1998 ), and in particula r, elevated features harbor greater plant diversity than lower elevations ( Bukata and Sloan, 2002, Bukata, 1999, Simmons et al., 2009 ). Not all plant species have the same relationship with microtopographic features in swamps. These features are biogenic, and some species play a role in their creation, while others simply take advantage of their existence as suitable germination sites. The tree and shrub species that contribute to the formation of microtopographic features can be thought of as autogenic ec osystem engineers that modify the elevation of soil by virtue of their own morphology (Jones et al.,1994). These species undoubtedly alter the distribution and movement of resources in swamps, which can affect entire communities of organisms and overall ec osystem functioning (Gilad et al.,2004). Still, little is known about the nature and extent of these effects in southern swamps.

PAGE 60

60 Given the important relationships between microtopography and swamp plant communities, it follows that microtopography plays a n important role in the ability of these systems to regenerate after disturbance (Bukata and Sloan, 2002), and conversely, the regeneration of some key species may be important in the recovery of microtopography after disturbance. Characteristically slow regeneration of swamps after clearcutting (Kellison and Young, 1997) makes understanding limitations to regeneration of particular interest. Diminished prevalence and height of microtopographic features has been observed following clearcutting (Locky and Bayley, 2006), which may delay subsequent development of microtopography (Bukata and Sloan, 2002). Additionally, the creation of soil ruts from heavy skidding equipment during harvest may further reduce the suitable sites for woody plant regeneration. Th e first objective of this study was to compare the relationships between the density and diversity of plant species and microtopographic position in both reference and clearcut swamps. The second objective was to determine the relationships, if any, betwee n the size of individual trees of various species and microtopographic position in reference swamps. The results provide insight into the differences in plant composition and density between microtopographic highs and lows in sites with open and closed tre e canopies and allow for regeneration predictions based on site microtopography. Methods Study Sites Sites were located within backwater floodplain, strand and depressional Taxodium Nyssa swamps on industrial and federal lands in Alachua, Gilchrist, Bradfo rd and Baker counties of 3 1) Aver age annual precipitation is 1250 mm, with average

PAGE 61

61 annual high and low temperatures of 26.6C and 14.2C (NOAA). Sampling occurred in Ap ril and May of 2010, during which rainfall was near average and air temperatures were warmer than average by 1.1C (NOAA, 2010). Soils were very poorly drained fine sands and sandy loams (Monteocha and Mascotte series), very poorly drained fine sands wi th a mucky fine sand surface layer (Lynn Haven series) and very poorly drained mucks with loamy substratums (Pamlico series) (NRCS 2010 ) Depth to water table for each of these series is zero inches, and each are found in depressions in the landscape. Pine plantations and isolated wetland features represent the main type of surrounding land use (Table 3 1) Reference s ites Six reference swamps were sampled in all, 3 were sampled for the understory vegetation analysis and 5 were sampled for the tree an d elevated microsite surveys (Table 3 1). The sites were chosen based on their similarity in species composition and hydrology to sampled logged sites and often were in close proximity to the logged sites. Each of these sites exhibited evidence of a histor y of minor to moderate incomplete or logging and/or hydrologic modification could not easily be found in North Florida. Each of the 5 reference sites were similar in that they were all dominated by a small suite of canopy trees that included primarily Nyssa sylvatica, Taxodium ascendens, and Taxodium distichum, and secondarily Gordonia lasianthus, Magnolia virginiana, and Persea borbonia The understory vegetation was characteristically dense and shrubby and was often dominated by Lyonia lucid a and Itea virginica, among others. Additionally, ferns, herbs and vines were also common in some sites.

PAGE 62

62 Microtopographic relief had an amplitude of 0.75 1.4 m within each si te. Distinct elevated microsites formed by tree roots, stumps and soil and litter accumulations covered an average of 22.6% ( 1.671%) of the swamp floor. In most sites, all soil except the raised microtopographic features was submerged in tannic water fo r the duration of the study. Logged s ites Four sites that were recently clearcut were sampled, 3 of which were sampled for the understory vegetation analysis (Table 3 1). Tree and elevated microsite surveys were conducted in all 4 sites. These sites ran ged from having sparse or very young tree regeneration (Basal areas of 0.5 1.2 m 2 /ha) to dense regeneration of coppice and/or seedling recruits (Basal area of 4.6 9.2 m 2 /ha). Elevated microsite features covered an average of 9.9% ( 2.854%) of the ground surface. Field Methods Understory vegetation In each of 6 study sites, either 10 or 20 1m 2 plots on elevated microsites and 10 or 20 1m 2 plots on the swamp bottom were randomly located, for a total of 80 plots across 3 reference sites, and 80 plots across 3 recently logged sites. Due to lack of accessibility to some sites, sampling across sites was unequally replicated. In each plot, the number of stems of each shrub, vine and forb species and the number of tree seedlings were recorded. For multi stemmed shrubs such as Lyonia lucida every stem was counted individually. Percent cover was estimated for graminoids. Tree surveys In randomly located 5m radius plots, 1042 trees were sampled across 5 reference sites, and 742 trees were sampled across 4 logged sites. For each tree, species, DBH

PAGE 63

63 (diameter at breast height) and microsite association (on the swamp bottom or associated wi th an elevated microsite) were recorded. Data Analysis Understory vegetation All plant species that were encountered (Appendix C) were grouped into 5 functional groups: graminoids, forbs, woody vines, shrubs, and tree seedlings. Using general linear mix ed models (proc glimmix in SAS 9.2), a test for the fixed effects of microtopographic position (swamp bottom or elevated microsite), site treatment (reference or logged) and the interaction between these two explanatory variables on the response variables of stem density or percent cover and plant species richness was performed for each plant functional group and for all species combined. Site and plot were entered into each mixed model as random effects. Differences in plant functional group richness or density between microtopographic position and site treatment were Tree surveys I tested the distribution of trees against the percent that would be expected on hummocks. Specifically, a chi squared test was performed to compare the proportion of individuals of each tree species found with elevated microsites against the null hypotheses of 22.6% for reference and 9.9% for logged, which were the average swamp area composed of elevated microsites in each treatment and th us the expected proportions. This test was done for the seven most commonly encountered tree species, pooled across reference and then across logged sites. Differences between

PAGE 64

64 To determine whether relationships between tree species and microsite were related to tree size, observations were divided into DBH classes so that each class had the same number of trees. The percentages of trees in each DBH class associated with elevated microsites were calculated. Linear regressions were used to examine the relationship between tree DBH and likelihood of elevated microsite association, with DBH class as the independent variable and % elevated microsite association the dependent variable. This was do ne for the five most common trees (excluded Persea spp. and Pinus elliottii because their sample sizes were not adequate for the analysis). Results Understory Vegetation Species r ichness Across sites, overall species richness was higher on elevated micros ites (2.95/m 2 ) than on the swamp bottom (0.73/m 2 ) (p <.0001). There was no effect of site treatment (reference vs. logged). On elevated microsites, the species richness of forbs (0.50/m 2 ), woody vines (0.48/m 2 ), shrubs (1.48/m 2 ), and tree seedlings (1.48/ m 2 ) were greater than on the swamp bottom across sites (p = .0069, p <.0001, p <.0001, p <.0001, respectively) (Figure 3 1). The relationships between topographic position and species richness for these groups were similar between reference and logged site s. Graminoid species richness did not differ with topographic position (p = 0.1429), though an interaction effect (p = 0.0041) indicated that there was greater divergence in species richness richness with topographic position in logged sites than in refer ence sites. Stem density On elevated microsites, the stem density of forbs (ave. 2.09/m 2 ), woody vines (ave. 1.68/m 2 ), shrubs (ave. 14.03/m 2 ), and tree seedlings (ave. 1.81/m 2 ) were greater

PAGE 65

65 than on the swamp bottom across sites (p = .0025, p = .0310, p <. 0001, p <.0001, p = .0088, respectively) (Figure 3 2). There were no differences between site treatments. Shrub stem densities were more divergent between elevated microsites and the swamp bottom in logged sites than in reference sites (interaction effect p <.0001). Forbs, woody vines and tree seedlings did not have interaction effects between topographic position and site treatments, indicating that the relationship between topographic position and stem density was similar across all sites. Tree Surveys Fo r all tree taxa analyzed, a greater proportion were found in association with elevated microsites than would be expected under a random distribution, given the proportion of the swamp comprised of these microsites in both reference and logged sites (all p values <.0001). In reference sites, 44% of Nyssa sylvatica var. biflora (n = 561), 67% of Taxodium distichum var. nutans. (n = 55), 71% of Persea spp. (n = 21), 77% of Magnolia virginiana (n = 132), 91% of Ilex spp. (n = 124), 95% of Gordonia lasianthus (n = 138), and 100% of Pinus elliottii (n = 7) were found on elevated microsites. Similarly, in logged sites, 59% of Nyssa sylvatica var. biflora (n = 76), 74% of Taxodium distichum var. nutans. (n = 182), 93% of Persea spp. (n = 75), 46% of Magnolia virgini ana (n = 13), 82% of Ilex spp. (n = 184), 69% of Gordonia lasianthus (n = 127), and 61% of Pinus elliottii (n = 75) were found on elevated microsites. Interestingly, Taxodium distichum var. nutans showed a strong, positive relationship between DBH class an d likelihood of elevated microsite association (Figure 3 3, R2 = 0.7296), and that for Nyssa sylvatica var. biflora produced an R2 value of 0.7934. A strong relationship was not seen for Magnolia virginiana (R2 =0.3911), Ilex spp. (R2 = 0.0143), or Gordoni a lasianthus (R2 = 0.1457).

PAGE 66

66 Discussion Species richness and stem densities of vegetation were generally much greater on elevated microsites than on the swamp bottom regardless of recent logging history. For most plant growth forms, the relationship betwe en topographic position and the richness and density of vegetation was similar in logged and reference sites, indicating the importance of topographic position to plant communities in swamps in varying successional stages. Trees were preferentially associa ted with elevated microsites rather than swamp bottoms, likely because these elevated sites act as safe sites for capture of seed rain, germination, and establishment For some species, particularly those that have limited flood tolerance (i.e. Magnolia vi rginiana (Burns and Honkala 1990)), it is likely that elevated microsites are most suitable for their germination and/or establishment and thus influence adult distributions, as was seen in a study of an Itea virginica population which was almost entirely confined to elevated microsites due to limitations in early life stages (Anderson et al., 2009). In a North Florida swamp study, seed rain was found to be lower on the swamp bottom than on elevated microsites (Titus, 1991). Elevated microsites may serve as important flooding refugia for some species that might not otherwise thrive in swamp ecosystems, and are likely important regeneration sites after logging disturbance. The positive relationship between tree size and proportion of trees found with elevated microsites in both Nyssa sylvatica and Taxodium spp. may indicate that 1) those individuals that start on elevated microsites have greater survival and/or growth over time than those that start on the swamp bottom a nd/or 2) as individuals grow, they create new elevated microsites. There is some evidence that being located on an

PAGE 67

67 elevated topographic position can enhance productivity in swamps (i.e. Jones et al., 2000), which supports the first hypothesis. However, be cause these two tree species are among the most flood tolerant in the Southeast United States, it seems unlikely that they would have difficulty thriving under swamp bottom conditions over time to the point that all aged individuals would be those that beg an on elevated microsites. Both of these species produce aerial roots (knees in cypress trees) as they age if growing in flooded conditions, which could contribute to the accumulation of organic material around their bases. Additionally, Nyssa trees produc e a taproot and a swollen base up to the mean high water level and an accompanying proliferation of exposed roots immediately around their bases (Burns and Honkala, 1990). The facilitation of elevated microsite formation by Nyssa and Taxodium spp may be i mportant to the structure and composition of swamp plant communities. Thus, after clearcutting, the recovery of these species may be important to the recovery of other plant species. Concl uding Remarks The relationships between understory vegetation, tree communities and topographic positions in Southeastern swamps vary by species and undoubtedly are complex. What is clear is that there is an intimate relationship between elevated microtopographic features and many plant groups, especially woody species in both clearcut and reference swamps. The biogenic nature of elevated topographic features means that their formation is governed by the production and decomposition of organic materials and by the growth characteristics of swamp species. Some plant species contribute to the formation of elevated microsites via specialized roots while others indirectly contribute via leaf litter inputs, and simultaneously, some species are simply utilizing existing features for their own benefit. More research is needed to better

PAGE 68

68 understand the role of various plant species in the formation (or deterioration) and utilization of elevated microsites. This information will be particularly useful for predicting the recovery of swamp ecosystems following disturbances.

PAGE 69

69 Table 3 1 List of study sites. Sit e # treatment Yrs since logging Landscape position Surrounding land use County Location 1R Reference NA depression residential/agriculture Alachua 2937'43.60"N, 8217'54.14"W 2R Reference NA strand national forest/wetlands Baker 3015'27.30"N, 8224'30.40"W 3R Reference NA depression pine plantation/wetlands Alachua 2946'2.13"N, 8213'7.35"W 4R Reference NA backwater floodplain pine plantation/river floodplain Bradford 2950'8.89"N, 82 9'30.92"W 5R Reference NA depression pine plantation/wetlands Alachua 2947'7.95"N, 8216'8.54"W 6R Reference NA lacustrine conservation/pine plantation Alachua 2941'0.51"N, 8214'5.04"W 1L Logged 5 backwater floodplain pine plantations/river floodplain Bradford 2950'26.23"N, 82 9'13.21"W 2L Logged 0.5 depression pine plantations/wetlands Alachua 2946'50.77"N, 8216'0.92"W 3L Logged 7 depression pine plantations/wetlands Gilchrist 2947'26.29"N, 8248'33.91"W 4L Logged 13 depression pine plantations/wetlands Gilchrist 2948'12.55"N, 8247'14.76"W

PAGE 70

70 Figure 3 1 Average vegetation richness of A) forbs and graminoids and B) woody vines, shrubs and tree seedlings per m 2 on elevated microsites versus the swamp bottom in logged sites and reference sites 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 forbs graminoids Species/m2 Plant functional group 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 vines shrubs trees Species/m2 Plant functional group Logged: elevated microsite Logged: swamp bottom Ref: elevated microsite Ref: swamp bottom B A

PAGE 71

71 Figure 3 2 Average plant stem density of forbs, woody vines, shrubs and tree seedlings per m 2 on elevated microsites versus the swamp bottom in logged sites and reference sites. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 forbs vines shrubs trees Stems/m2 Plant functional group Logged: elevated microsite Logged: swamp bottom Ref: elevated microsite Ref: swamp bottom 9.7 18.4

PAGE 72

72 Figure 3 3 The pr oportion of A ) Nyssa sylvatica (n=30 per DBH class) and B ) Taxodium spp .(n=10 per DBH class) individuals growing on elevated microsites as a function of DBH in reference sites. R = 0.7296 0% 20% 40% 60% 80% 100% 1.25 2.6 3.5 4.3 5.3 6.5 7.4 8.4 11 22 on elevated microsites (%) DBH class (in) R = 0.7934 0% 20% 40% 60% 80% 100% 2.3 6.8 10.4 14 20.8 37 on elevated microsites (%) DBH class (in) Nyssa sylvatica Taxodium spp. A B

PAGE 73

73 APPENDIX A EXAMPLE R CODE FOR HI STOGRAM CLUSTERING A NALYSIS >Library (mclust) >fit=Mclust(Bradfordref2) >plot(fit,Bradfordref2) >print(fit) > Bradfordref2BIC=mclustBIC(Bradfordref2) > Bradfordref2Summary=summary(Bradfordref2BIC, data=Bradfordref2) > Bradfordref2Summary > Bradfordref2Mclust=Mclust(Bradfordref2) > plot(Bradfordref2Mclust, Bradfordref2, legendArgs = list(x = "bottomleft")) > Bradfordref2BIC=mclustBIC(Bradfordref2) > plot(Bradfordref2BIC)

PAGE 74

74 APPENDIX B SAS CODE FOR ELEV ATED MICROSITE SIZES ANALYSIS data size; input site size; datalines; 3 0.678535303 3 0.166549574 (ETC). run ; proc glm data=size; class site; model size=site; means site /bon; lsmeans site/pdiff; run ;

PAGE 75

75 APPENDIX C LIST OF SPECIES FOR UNDERSTORY VEGETATIO N SURVEYS Reference Sites Harvested Sites Species R3 R4 R6 L2 L3 L4 Acer rubrum x x x Andropogon glomeratus x Carex s pp x x Cephalanthus occidentalis x x Crinum americanum x Cyrilla racemiflora x x Dulichium arundinoceum x Eupatorium capillifolium x Gordonia lasianthus x x Hydrocotyle bonariensis x Ilex cassine x x x x x Ilex myrtifolia x Itea virginiana x x x x x Lachna n thes caroli ni ana x Leucothoe racemosa x Liquidambar styraciflua x Lyonia ligustrina x Lyonia lucida x x x x x x Magnolia virginiana x x x Morella cerifera x x x Osmunda regalis x Persea borbonia x x Poaceae x Quercus nigra x Rubus sp p x x Saururus cernuus x x Smilax s p p. x x x x x Taxodium ascendens x Toxicodendron radicans x x x Vitis rotundifolia x Woodwardia areolata x Woodwardia virginica x x x x x

PAGE 76

76 APPENDIX D SAS CODE FOR UNDERSTORY VEGETATION ANALYSIS Data veg1; input site replication treatment $ loggedcond $ totalspecies; datalines; 1 1 H R 1 1 2 H R 2 1 3 H R 2 (ETC). run ; ods rtf; ods graphics on; proc glimmix data=veg1 plots=all; class site treatment replication loggedcond; model totalspecies= treatment loggedcond treatment*loggedcond; random site(loggedcond) replication(treatment*loggedcond); Lsmeans treatment loggedcond treatment*loggedcond / pdiff; Lsmeans treatment*loggedcond / plots=mean(CL join sliceby=treatment); run ; ods rtf close; (this procedure was repeated for each plant grouping of interest)

PAGE 77

77 LIST OF REFERENCES Anderson, J. T., Landi, A. A. & Marks, P. L. 2009. Limited flooding tolerance Of juveniles r estricts the distribution of adults in an understory shrub ( Itea v irginica ; Iteaceae). American Journal of Botany 96 : 1603 1611. Aust, W. M. & Lea, R. 1991. S oil temperature and organic matter in a disturbed forested wetland. Soil Science Society Of America Journal 55 : 1741 1746. Brandt, K. & Ewel, K.C. 1989. Ecology and management of cypress swamps: A review. University of Florida Cooperative Extension Service. Bruland, G. L. & Richardson, C. J. 2005. Hydrologic, edaphic, and vegetative responses to microtopographic re establishment in a restored wetland. Restoration Ecology 13 : 515 523. Bukata, B.J. III 1999. T he development and role of microtopography in constructed forested wetlands on phosphate University of Florida. Bukata, B.J. lll & Sloan, M. 2002. The development and role of microtopography in natural and co nstructed forested wetlands. In M.T. Brown & S.M. Carstenn (ed.) Successional development of forested wetlands on reclaimed phosphate mined lands in Florida, Uni versity of Florid a Center for Wetlands No. 03 131 193. Burns, R.M. & Honkala, B.H. 1990. Silvics of North America: Vol. 1. Conifers. United States Department of Agriculture, Forest Service, Agriculture Handbook 654. Cantelmo, A. J. & Ehrenfeld, J. G. 1999 Effects of microtopography on mycorrhizal infection in Atlantic white cedar ( Chamaecyparis thyoides (L.) Mills.). Mycorrhiza 8 : 175 180. Casey, W.P. & Ewel, K.C. 1998. Soil redox potential in small pondcypress swamps after harvesting. Forest Ecology and Management 112: 281 287. Conner, W.H. & Buford, M.A.1998. Southern deepwater swamps. p 261 287 I n M.G. Messina & W.H. Conner (eds.) Southern Forested Wetlands, Ecology and Management. CRC Press, Boca Raton, FL, USA. Conner, W.H., Toliver, J.R. & Sklar, F. R. 1986. Natural regeneration of bald cypress ( Taxodium distichum (L.) Rich) in a Louisiana swamp. Ecological Management 14:305 317. Crownover, S. H., Comerford, N. B., Neary, D. G. & Montgomery, J. 1995. H orizontal groundwater flow patterns through a cypr ess swamp pine flatwoods landscape. Soil Science Society of America Journal 59 : 1199 1205.

PAGE 78

78 De Steven, D. & Sharitz, R. R. 1997. Differential recovery of a deepwater swamp forest across a gradient of disturbance intensity. Wetlands 17 : 476 484. Dickson, R.E. & Broyer, T.C. 1972. Effects of aeration, water supply, and nitrogen source on growth and development of tupelo gum and bald cypress. Ecology 53:626 634. Dunn, C. P. & Sharitz, R. R. 1987. Revegetation of a Taxodium Nyssa forested wetland following complet e vegetation destruction. Vegetatio 72 : 151 157. Duryea, M.L & Hermansen, L.A. 1998. wetlands t rees. University of Florida IFAS Extension Circular 1186. Eppinga, M.B., Rietkerk, M., Borren, W., Lapshina, E.D., Bleu ten, W. & Wassen, M.J. 2008. Regular surface patterning of peatlands: confronting theory with field d ata. Ecosystems 11: 520 536. Ewel K.C. 1998. Pondcypress Swamps in Southern Forested Wetlands. Lewis Publishers. Boca Raton, F L, USA. Faulkner, S. P., Bhattarai, P., Allen, Y., Barras, J. & Constant, G. 2009. Identifying baldcypress water tupelo regeneration classes in forested wetlands of t he Atchafalaya Basin, Louisiana. Wetlands 29 : 809 817. Fraley, C. & Raftery, A.E. 2006. MCLUST Vers ion 3 for R: normal mixture modeling and model based clustering University of Washington, Dept. of Statistics, Techincal Report No. 504. Gilad, E., Von Hardenberg, J., Provenzale, A., Shachak, M. & Meron, E. 2004. Ecosystem engineers: From pattern forma tion to habitat creation. Physical Review Letters 93. Grace, J. M., Skaggs, R. W. & Cassel, D. K. 2006. Soil physical changes associated with forest harvesting operations on an organic soil. Soil Science Society of America Journal 70 : 503 509. Good, B.J. & Whipple, S.A. 1982. Tree spatial patterns: South Carolina bottomland and swamp forests. Bulletin of the Torrey Botanical Club 109:529 536. Gunderson, L.H. 1977. Regeneration of cypress in logged and burned strands at Corkscrew Swamp Sanctuary, Florida p. 349 357 I n K.C. EWEL & H.T. ODUM (eds.) Cypress Swamps. University of Florida Press, Gainesville, FL, USA. Huenneke, L. F. & Sharitz, R. R. 1986. Microsite abundance and distribution of woody seedlings in a South Carolina cypress tupelo s wamp. American Midland Naturalist 115 : 328 335.

PAGE 79

79 Huenneke, L. F. & Sharitz, R. R. 1990. Substrate heterogeneity and r egeneration of a swamp tree, Nyssa aquatica American Journal of Botany 77 : 413 419. Jones, C.G., Lawton, J.H. & Shachak, M. 1994. Organisms a s ecosystem engineers. Oikos 69: 373 386. Jones, R. H. & Gresham, C. A. 1985. A nalysis of composition, environmental g radients, and structure in the C o astal Plain lowland forests of South C arolina. Castanea 50 : 207 227. Jones, R. H., Henson, K. O. & Somers, G L. 2000. Spatial, seasonal, and annual variation of fine root mass in a forested wetland. Journal of the Torrey Botanical Society 127 : 107 114. Jones, R. H., Sharitz, R. R., Dixon, P. M., Segal, D. S. & Schneider, R. L. 1994. Woody plant regeneration In 4 floodplain f orests. Ecological Monographs 64 : 345 367. Karels, J.R. 2009. Silviculture best management practices. Florida Department of Agriculture and Consumer Services, Division of Forestry Kellison, R. C. & Young, M. J. 1997. The bot tomland hardwood forests of the S outhern United States. Forest Ecology and Management 90 : 101 115. Laderman, A.D. 1998. Coastally restricted f orests Oxford Univerity Press, New York, USA. Lockaby, B.G., Stanturf, J.A. & Messina, M.G. 1997. Effects of silvicultural activity on ecological processes in floodplain f orests of the Southern United States: A review of existing r eports. Forest Ecology and Management 90 : 93 100. Locky, D. A. & Bayley, S. E. 2006. Plant diversity, composition, and rarity in the southern boreal peatland s of Manitoba, Canada. Canadian Journal of Botany 84 : 940 955. Lugo, A.E., Nessel, J.K. & Hanlon, T.M Root d istribut ion in a North Central Florida cypress s trand. p. 279 285. I n K.C. EWEL & H.T. ODUM (eds.) Cypress Swamps. University of Florida Press, Gain esville, FL, USA. Meadows, J. S. & Stanturf, J. A. 1997. Silvicultural systems for southern bottomland hardwood forests. Forest Ecology and Management 90 : 127 140. Moser, K., Ahn, C. & Noe, G. 2007 Characterization of microtopography and its influence on vegetation patterns in created wetlands. Wetlands 27 : 1081 1097. Myers, R.L., & Ewel, J.J. 1990. Ecosystems of Florida. University Pres ses of Florida. Gainesville, FL,USA.

PAGE 80

80 NOAA 2010. National Oceanic and Atmospheric Administration, National Weather Service Weather Fo recast Office, Gainesville FL (accessed October 21, 2010 as http://www.srh.noaa.gov/jax/?n=climate ). NRC S 2010. Natural Resources Conservation Service, United States Department of Agriculture. Web Soil Surveys (accessed December 16, 2010 as http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx ). Odum, H.T. & Ewel, K.C. 1984. Cypress Swamps. Center for Wetlands, Universi ty Presses of Florida, Gainesville, FL,USA. Randall,C.K., M.L. Duryea, S.W. Vince, & R.J. English. 2005. Factors influencing stump sprouting by pondcypress ( Taxodium distichium var. nutans (Ait.) Sweet). New Forests 29: 245 260. Ricci, N 2010 Recovery of cypress domes in North Central F lorida 10 years after harvest Schneider, R. L. & S haritz, R. R. 1988. Hydrochory and regeneration in a bald cypress water tupelo swamp f orest. Ecology 69 : 1055 1063. Simmons, M.E., Ben Wu, X. & Whisenant, S.G. 2009. Plant and soil responses to created microtopography and soil treatments in bottomland hardwood forest r estoration Society for Ecological Restoration International. Sloan, M. 1998. The role of mi crotopographic relief in maintaining understory vegetative community diversity in forested wetlands. Gainesville, FL, USA. Spencer, D. R., Perry, J. E. & Silberhorn, G. M. 2001. Early secondary succession in bottomla nd hardwood forests of S outheastern Virginia. Environmental Managemen t 27 : 559 570. Stokes, B.J. & Schilling, A. 1997. Improved harvesting systems for wet s ites. Forest Ecology and Management 90: 155 160. Sun, G., Mcnulty, S. G., Shepard, J. P., Amatya, D. M., Riekerk, H., Comerford, N. B., Skaggs, W. & Swift, L. 2001. Effects of timber management on the hydro logy of wetland forests in the S outhern United States. Forest Ecology and Management 143 : 227 236. Sun, G., Riekerk, H. & Kornhak, L. V. 2000. Ground w ater table rise after forest harvesting on cypress pine flatwoods in Florida. Wetlands 20 : 101 112. Titus, H. J. 1990. Microtopography and woody plant regeneration in a hardwood floodplain swamp i n Florida. Bulletin of the Torrey Botanical Club 11 7 : 429 437.

PAGE 81

81 Titus, J. H. 1991 Seed bank of a hardwood floodplain swamp in Florida Castane a 56 : 117 127. Trettin, C. C., Davidian, M., Jurgensen, M. F. & Lea, R. 1996. Organic matter decomposition following harvesting and site preparation of a forested wetland. Soil Science Society of America Journal 60 : 1994 2003. Vivian Smith, G. 1997. Microtopographic heterogeneity and floristic diversity in experimental wetland communities. Journal of Ecology 85 : 71 82. Walbridge, M. R. & Lockaby, B. G. 1994. Effects of forest man agement on biogeochemical functions in southern forested w etlands. Wetland s 14 : 10 17. Watts, D.L., Cohen, M.J. & Heffernan, J.B. 2010. Hydrologic modification and the loss of self organized patterning in the ridge slough mosaic of the Everglades. Ecosystem s 13:813 827. Zedler, J. B. & Kercher, S. 2005. Wetland resources: Status, trends, ecosystem services, and restorability. Annual Revi ew of Environment and Resources 30 : 39 74.

PAGE 82

82 BIOGRAPHICAL SKETCH Amy Washuta was raised in Stuart, Florida amongst well wate red lawns and planted palms. She became interested in natural science after she enrolled in an introductory geology class while studying fine arts at the University of Florida. She went g eology from the Univers ity of Florida in 2007. After working with an environmental education program in Big Cypress National Preserve for a semester in 2008, she began graduate studies in Landscape Architecture at the University of Florida. A year later she shifted her gradu ate focus to Interdisciplinary Ecology in order to study her beloved cypress swam ps more closely, working toward a Master of Science degree.