|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
VIABILITY OF WETLAND TREES AFTER TWENTY YEARS ON PHOSPHATIC CLAY
SETTLING AREAS AND THEIR ROLE IN ECOSYSTEM DEVELOPMENT
WESLEY W. INGWERSEN
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
I would like to thank and acknowledge the following persons and institutions: my advisor,
Dr. Mark Brown, who made possible my involvement in the project, provided me a rich
theoretical background from which to draw questions for research, and guided me through each
successive stage of this thesis; the Florida Institute of Phosphate Research (FIPR) for its generous
support of the Wetlands on Clay project; Daniel McLaughlin for his help in site exploration and
every aspect of organizing and carrying out the field work and data input necessary for this thesis;
Sean King and Tyler Hollingsworth for their hard work in the field and lab during the summer of
2005; Betty Rushton, whose planting trials on clay settling areas and dissertation made possible
this investigation; my committee members, Dr. Clay Montague and Dr. Wendell Cropper, for
their advice and support; CF-Industries, the Mosaic Company, the Teneroc State Reserve, and the
Homeland Office of the Florida DEP for permission to repeatedly access the clay settling areas
used in this study; and my family and close friends who encouraged me to pursue my interest and
supported me as I did.
TABLE OF CONTENTS
ACKNOWLEDGMENTS ................. ................. ........ ......... ........ ......... ii
LIST OF TABLES ................. ................. ........ ......... ........ ......... ...._ v
LIST OF FIGURES ................. ...................... .................. ...................vii
ABSTRACT ................. ................. ........ ......... ........ ......... ........ .. ix
INTRODUCTION ............... .................... 1
Statement of the Problem ................. .......... .................. ....................... 1
Background ................ .................... 2
Clay Settling Areas................ ................. 2
Wetlands on Clay Settling Areas ............... .................... 2
Planting of Wetland Species on Clay ................. .............. ......... ........ ...... 4
Recruitment ........._.._......._._.....__ ...._._........__ ......_._........... 5
Ecosystem Development ........_._.......... __ ........._._.....__ ...._._........ 6
Plan of Study ........._.._.......__ ...._. .....__ ....__ ...._._ ........._..... 7
METHODOLOGY ............... .................... 9
Site and Plot Selection ............... .................... 9
Field Data Collection ................ ................... 11
Topography ............... .................... 11
Hydrology ............... .................... 11
Planted Trees ................. ............. ......... ........ ......... ........ ...... 12
Other Tree Species ................ ................... 12
Recruited Trees ............... ... ... .... .. ..... ... ........... 12
Additional Measures of Ecosystem Development: Shrub and Understory Layers:
Soils: Canopy Photos ............... .................... 12
Site H stories ................ ................... 13
Data Analysis ................... ..... .............. 13
Topography and Water Levels ................ ................... 13
Tree and Plot Basal Area................ ................. 14
Tree Growth Comparisons ................ ................... 15
Population Size Class Distributions ................ ................... 15
Canopy Photos ............... .................... 16
Understory Vegetation ...................... ... .. ......... 16
Ordination of Plots by Prevalent Understory Species .................. ......................... 18
Correlation Matrices of Ecosystem Development Variables ................. ............... .... 18
Tree Population Model................ ................. 18
RESULTS ................ .................... 29
Tree Populations in Relation to Environmental Factors ............... ................ .... 29
Tree Survival by Site and Species .......__ ............_ ......._._ ....._ .......... 29
Hydrology ............... ..... .. .. .................. 30
Initial Tree Growth and 20-year Tree Survival ................. ............. ......... ....... 34
Site Disturbance and Tree Survival ................. .............. ......... ........ ...... 34
Recruited Trees ............... .................... 35
New Seedling Survival ..................... ............................ ............. ....... 36
Tree Population Size Class Distributions ......... ........ ................. ................ 36
Tree Population M odel............... ............ ................ 37
Ecosystem Development in Rushton and Reference Plots. ................. ............... ...._... 38
Topographic Comparison of Rushton and Reference Plots ................. ........_..._. ...... 39
Plot Basal Area in Rushton and Reference Plots ............... ................... 39
Percent Canopy Cover ................. ....................__ ...._._................. 39
Soil Organic Matter................ ................. 40
Understory Vegetation ............... .................... 40
DI SCU SSION ................. ................ ........ ......... ........ ......... ........ .. 74
Sum m ary .................... ..... ... ........ ...........74
Tree Populations in Relation to Environmental Factors ............... ................ .... 75
Tree Survival By Site and Species ................ ....................75
Tree Survival and Hydrology ............................... .......... .... ......... ................ 76
Tree Growth Comparison Between Sand-Clay and Clay Sites ................. ................. 78
Recruited Trees ............... .....................79
Tree Population Model................ ..................... 80
Characteristics of Successful Species on CSAs ..................................... 81
Ecosystem Development in Rushton and Reference Plots. ................. ............... ...._... 83
Structural Differences ............... .................... 85
Soil Organic Matter................ ................. 85
Understory Vegetation ............... .... .. .. ................... 86
Relationships Among Measures of Ecosystem Development................... ................ .. 88
SUPPLEMENTAL FIGURES, TABLES, AND CODE ................ .................... 90
LIST OF REFERENCES .................. ................... .................. ................ 114
BIOGRAPHICAL SKETCH .............. .................... 118
1. Site summary table................ ................. 20
2. Species list for cypress-gum plots................ ................. 22
3. Species list for "wet" and "transitional" hydric swamp plots ................ .................... 23
4. Size class key used in tree size class distributions ................. .................. .............. 28
5. Tree survival from initial planting in 25 sampled cypress-gum plots ................. ................ .. 43
6. Tree survival from initial planting in 12 sampled hydric swamp plots................ .................. 43
7. Comparison of trees growing in different soil media by species ................ .................... 55
8. Results of a two-way ANOVA comparing the effect of two soil types (clay and sand-clay)
and two water levels (shallow and deep) on Fraximes pennsylvanica growth. ................... 55
9. Results of a two-way ANOVA comparing the effect of two soil types (clay and sand-clay)
and two water levels (shallow and deep) on Taxodium distichum growth. ......................... 55
10. Site disturbance record................ ................. 56
11i. Plots with potential offspring of planted trees ordered by reproductive ratio ................... ...... 57
12. Rushton plots/subplots ranked by planted tree basal area ............... ................ .... 66
13. Topography and water level comparison of Rushton and reference plots .................. ............ 67
14. Plot-scale basal area comparison in Rushton and corresponding reference plots ................... 67
15. Percent canopy cover comparison in Rushton and corresponding reference plots ................. 69
16. Soil percent organic matter comparison in Rushton and reference plots. ................... ............ 70
17. Soil percent organic matter summarized by site and plot type. .............. ..................... 70
18. Average percent understory cover comparison between Rushton and reference plots........... 71
19. Species richness and evenness comparison in Rushton and reference plots ................... ........ 71
20. Correlation matrices for ecosystem development variables by site ................ .....................73
LIST OF TABLES
21. Understory species in pair 1 (CFI) ranked by Importance Value (IV)................ .................. 96
22. Understory species in pair 2 (HOM) ranked by Importance Value (IV) ............... ................. 97
23. Understory species in pair 3 (OHW) ranked by Importance Value (IV) ................................ 98
24. Understory species in pair 4 (OHW) ranked by Importance Value (IV) ................................ 98
25. Understory species in pair 5 (OHW) ranked by Importance Value (IV) ................................ 99
26. Understory species in pair 6 (PRP) ranked by Importance Value (IV).................. ................ 99
27. Understory species in pair 7 (TEN) ranked by Importance Value (IV) ................ ............... 100
28. Understory species in pair 8 (TEN) ranked by Importance Value (IV) ................ ............... 100
29. Understory species in pair 9 (TEN) ranked by Importance Value (IV) ................ ............... 100
30. Understory species in pair 10 (TEN) ranked by Importance Value (IV) .............................. 101
31. Autecological characteristics of species prevalent in understory ............... .................... 102
1. Study site locations.. .............. ..................... 21
2. Cypress-gum plot layout ................. ................. ......... ......... ........ ........22
3. Hydric swamp plot layout .................. ....___......_ ....___......_ .............. 23
4. Elevation diagram for a cypress-gum plot ................. .............. ......... ........ ...... 24
5. Elevation diagram for a hydric swamp plot ................. ................. ........ ......... .. 25
6. Soil, understory, shrub, and canopy photo sampling scheme for cypress-gum plots. .............. 26
7. Soil, understory, shrub, and canopy photo sampling scheme for hydric swamp plots ............. 27
8. The format for the transition, A, matrix for a matrix population model ............... .................. 28
9. Percentage of planted trees surviving in cypress-gum plots ................ ............... ..... 44
10. Percentage of planted trees surviving in hydric-swamp plots................ ................. 45
11. Percentage of plot inundated for all sampled months on cypress-gum plots. ......................... 46
12. Percentage of plot inundated for all sampled months on hydric swamp plots. ....................... 47
13. Distribution of average water depth at planted and surviving tree locations in cypress-gum
plots ................ .................... 48
14. Number of planted trees that died between years 1 and 20 and trees alive in 2005, in 0.1m
depth classes on CFI (sand-clay) on plots R1-R6 ......____ ... .... ..__ .. ......__......... 49
15. Number of planted trees that died between years 1 and 20 and trees still alive, in 0.1Im
depth classes on OHW (clay) plots R2A and R2B. ................. .............. ......... ..... 50
16. Number of planted trees that died between years 1 and 20 and trees still alive, in 0.1Im
depth classes on TEN (clay) plots R5A, R5B, R6A, R6B, R7A, and R7B. ........................ 51
17. Distribution of average water depth at planted and surviving tree locations in hydric
swamp plots ................. ................. ......... ........ ......... ........ ...._ 52
18 Percentage of planted trees surviving in cypress-gum plots by soil type after approximately
1 (Rushton 1988), 3 (Rushton and Paulic 2001), and 20 years. ................ .............. ...... 53
LIST OF FIGURES
19. Distribution of average water depth at surviving tree locations in cypress-gum plots
grouped by soil type. ................. .......... ........ ......... ........ ............... 54
20. Size class distributions of Taxodium distichum seedlings at Ten H3 counted in June and
November, 2005 .......... ................ ............... .................. .............. 58
21. Size class distribution of Taxodium distichum in 6 basins on five CSAs ............................... 59
22. Size class distributions ofNyssa aquatica in five basins on four CSAs................ ..... ..........._ 60
23. Size class distribution of Fraxinus pennsylvanica in five basins on four CSAs. .................. .. 6 1
24. Size class distribution of Fraxinus carobiniana in two basins on two CSAs .................. ........ 62
25. Transition matrix for CFI SP-1 Taxodium distichum population model................ ................ 63
26. Model predicted population change of CFI SP-1 Taxodium distichum ................ ..................63
27. Transition matrix for OH Wright Taxodium distichum population model.............................. 64
28. Model predicted population change of OH Wright Taxodium distichum ............................... 64
29. Model elasticity values showing sensitivity of different parameters ................ ................... 65
30. Subplot basal area and percent canopy cover at HOM. ............... .................. .. 69
31i. NMDS plot of understory species assemblages ................. .................. .............._72
32. Succession in a forested system ................. ................. .................. ........... 89
33. 2005 water depth in a well at CFI measured by continuous data logger ................................ 91
34. 2005 water depth in a well at TEN measured by continuous data logger. .................. ........... 92
35. Distribution of Fraxinus pennsylvanica basal area by average 2005 sampled water level
for clay, sand-clay, and sand cap sites. ...._.._.._ .... ... .___ ....._.._......_.._.......... 93
36. Distribution of Nyssa aquatica basal area by average 2005 sampled water level for clay,
sand-clay, and sand cap sites. ........................... ........94
37. Distribution of Taxodium distichum basal area by average 2005 sampled water level for
clay, sand-clay, and sand cap sites. .............. ..................... 94
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
VIABILITY OF WETLAND TREES AFTER TWENTY YEARS ON PHOSPHATIC CLAY
SETTLING AREAS AND THEIR ROLE IN ECOSYSTEM DEVELOPMENT
Chair: Mark T. Brown
Major Department: Environmental Engineering Sciences
Clay settling areas (CSAs) are constructed on about 2,000 acres of land every year to
contain waste clays following phosphate mining. The reclamation of CSAs to foster wetland
ecosystems has been proposed for these areas but not yet demonstrated as a viable alternative, due
to the lack of natural colonization of species typical of mature wetlands. Clay settling areas
planted with wetland trees in an early test of forested wetland viability were revisited after twenty
years. Survival and growth of species typical of riverine swamps demonstrated the suitability of
planted trees in seasonally wet areas, but the general lack of recruitment does not assure long-
term sustainability of the populations. After twenty years planted trees provide additional canopy
structure but they are less influential in the development of soil and understory ecosystem
components than site-specific exogenous factors. Engineering of CSAs to promote hydrology
typical of natural wetlands and supplementing tree planting with understory species are likely to
lead to more persistent and diverse wetland communities.
Statement of the Problem
Phosphate mining has been a major industry in central Florida for the past 60 years.
Currently about 5,000 acres of land are mined every year (Richardson 2005). Clay settling areas
(CSAs) are dominant features of the post-mining landscape that comprise about 40% of the post-
mining area. The land use options for CSAs after they have been filled are partially limited due
to the unstable nature of the consolidating ground surface.
The design and planting of these areas to create wetland ecosystems is one option the
industry and state are still exploring for the use of abandoned CSAs. Species characteristic of
wetlands naturally colonize depressional areas on CSAs in the years following abandonment.
Recognition of the potential for wetland establishment led to attempts to augment the
composition of wetland species on these areas. In an attempt to determine if forested wetland
ecosystems will persist on CSAs, a limited number of CSAs have been planted with wetland
trees. But the success of these plantings has not been evaluated after the initial few years of
establishment. The long-term development and viability of forested wetland ecosystems on CSAs
are critical to the determination of the suitability of wetlands on CSAs. In this study CSAs
planted with wetland trees were evaluated after 20 years in one attempt to evaluate forested
wetland development and viability on CSAs. The questions explored in this study have been
grouped under three foci:
1. How have the planted trees fared over time; what are the primary factors influencing tree
growth, survival, and recruitment?
2. How might the tree populations change in the future?
3. Are differences discernable in the ecosystem development of areas planted with trees and
those not planted?
Clay Settling Areas
CSAs are depositories for residual clay separated from phosphate rock and sand in the first
stage of processing following mining. The residual phosphatic clay is then slurried for pipe
transport. The handling of the residual clays has changed during the history of phosphate mining
in Florida. In the early years of large scale mining clays were pumped into mining cuts. More
recently large impoundment areas with high walls, often 1 mile square, have been created for
disposal of the clay. Alternatively, clays are sometimes mixed with residual sand before being
pumped into settling areas. Though the name is typically reserved for impoundment areas for
unmixed clay disposal, in this study CSAs refers to all three types of depositories for residual
As clay slurry is pumped into CSAs, clay particles settle to the bottom and water is drawn
off through outfall structures. A solid crust forms on the pond surface after 3-5 years (Richardson
2005), but consolidation of clays under the surface continues for decades. The consolidated
ground surface is often at an elevation above the original ground elevation and higher than the
surrounding landscape. Rate of consolidation of clays is not even across the CSAs, which are
often built on mined land characterized by patterns of mine cuts and spoil piles, resulting in an
uneven land surface. One result of the differential consolidation is the formation of deeper
depressions that hold surface water. These depressions often sink below the elevation of the
outfall structure causing them to become hydrologically isolated such that they seasonally retain
Wetlands on Clay Settling Areas
Isolated depressions on CSAs as well as water features drained by an outfall structure
support the establishment of hydrophytic vegetation characteristic of wetlands. Vegetation begins
to colonize these areas before the slurry-water has completely drained off in a phase called
dewatering. Algae often colonizes the water surface in the initial phases, followed by wind-
dispersed herbaceous macrophytes like Typha spp. and Scirpus spp. or shrubs and small trees
like Ludwigia peneviana and Salix caroliniana. But the continuation of seral succession with the
establishment of species characteristic of mature systems is not common even on the oldest CSAs
(Rushton 1983). Rushton (1988) suggested the dominance of these early successional wetlands
are characteristic of arrested succession, whereby climax species fail to establish. However the
successional pathway of a CSA and the composition of a 'climax' system are unclear. CSAs are
examples of what some members of the scientif ic community have referred to as "emerging
ecosystems" (Odum 1971, Hobbs et al. 2006), defined as 'new" environments which result from
heavy modification of the environment by human agency. Such ecosystems lack a precedent
from which to anticipate long-term composition and dynamics.
A few obstacles hinder natural succession on these areas: (1) the landscape surrounding
CSAs has generally been cleared and modified, so the recruitment of native species is difficult
due to hydrologic isolation, above-grade elevation, and distance to seed sources (Odum et al.
1983); (2) soils on CSAs contain a high percentage of clay (60-80%) and initially lack structure,
differing significantly from soils characteristic of wetlands in central Florida, which are sandier
with developed horizons (Rushton 1988, Myers and Ewel 1990, Graetz and Reddy 1997); and (3)
the hydrologic regime of these clay depressions may be different than natural wetlands due to the
high water-holding capacity of clay, the continuing consolidation of the clays, and large
watershed:wetland ratios (Rushton 1988).
Hydrologic regime and the physical and chemical nature of the soil are important factors in
the determination of the type of wetland that may be established (Mitsch and Gosselink 1993).
Water level is perhaps the most important factor for determining if a marsh (herbaceous wetland)
or swamp (forested wetland) will establish in these areas. Though the period and depth of
inundation in CSA depressions is typically unknown, existing vegetation may provide a clue as to
what the hydroperiod is like. Areas where Salix caroliniana has established may indicate
locations appropriate for forested species. Phosphorous (P) is often the limiting nutrient in
Florida freshwater wetlands systems (Reddy et al. 1999) but residual inorganic phosphorus is
high in phosphatic clays (Rushton 1988). Phosphorous has been directly correlated with
productivity in cyress ecosystems (Brown 1981). Highly productive systems situated on a
substrate with high clay content is characteristic of some alluvial forested wetlands in the
southeastern United States (Faulkner et al. 1991). Depressional areas on CSAs may be suitable
for the establishment of forested wetlands.
Planting of Wetland Species on Clay
Planting species characteristic of mid- to late-succession is one method to direct the
successional process (Brown and Tighe 1991). Monitored field trials on CSAs using wetland tree
species began in the 1980s (Rushton 1988, Paulic and Rushton 1991a, Everett 1991), and tree
survival and growth has been documented during the initial years after planting. Water
availability, species properties, tree size, and edaphic factors including soil age and nutrient levels
have all been shown to effect tree survival on clay settling areas. The following list summarizes
findings of earlier studies of wetland trees on CSAs.
* Hydrology was more important in determining tree survival than canopy or understory
cover (Rushton 1998, Paulic and Rushton 1991b).
* Wetland trees typical of floodplain and backwater swamps of central and northern Florida
have had greater than 50% survival after 1 year on clays, including Acer rubrum, Betula
nigra, Carya aquatica, Liquidambar \ap,, at rybre, Quercus laurifolia, Quercus lyrata,
Quercus michauxii, Sabal palmetto, and Ulmus americana. (Paulic and Rushton 1991b)
* Fraxinus spp. and Taxodium spp. had high (>80%) survival after 3 years (Paulic and
Rushton 1991a, Everett 1991);
* Clay is a suitable medium for wetland species (Cates 2001);
* After three years, trees growing on a sand-clay mix and on sand had higher survival than
those on clay. Trees in clay grew faster than trees in sand (Paulic and Rushton 1991a);
* Most major nutrients are available in sufficient quantities for tree growth. Nitrogen may be
the limiting nutrient. N-fertilizer increased growth but had no effect on survival ofdcer
rubrum in a greenhouse experiment (Paulic 1991). Fertilizer enhanced growth of
Taxodium spp. in clay both in the field and in the greenhouse (Everett 1991, Paulic 1991).
* Soil age was positively correlated with Acer rubrum growth in a greenhouse experiment
* Animal grazing can reduce tree survival (Rushton 1988).
These earlier studies have censused planted and non-planted trees in a variety of
hydrologic conditions, among different vegetation communities, and on a number of CSAs.
However, these earlier studies did not census planted trees after more than a few years, and thus
could not consider longer-term survival and growth, nor the potential ecosystem function of more
mature trees on CSAs. Time until maturity for forested swamps can be as long as 250 years in a
natural environment. Long-term monitoring is necessary to understand the long-term dynamics
of a restored forested system (Clewell 1999).
An important ingredient for the sustainability of a constructed forested system and an
indicator of the appropriateness of an environment for introduced species is the ability to
propagate. Wetland trees have specific moisture requirements for successful reproduction
(Mitsch and Gosselink 1993). These requirement can be important for seed set, germination, and
establishment. Poor seed set may occur from pollen limitations (McLanahan 1986). Dispersal is
important in order for fertilized seeds to find a viable location in which to germinate. Together
water levels and microtopography are important in determining seed dispersal. Because some
seeds float in water they tend to accumulate in greatest densities near the edge of water or near
obstructions. Seeds of wetland trees do not germinate in standing water. Thus areas of
permanent standing water may preclude the emergence of new seedlings. In areas with infrequent
drawdown, seed germination may still occur but viability of seeds may be decreased by long
periods of inundation (Schneider and Sharitz 1986). If seeds are able to germinate, water
conditions during the first few months can be critical to survival. Most wetland tree seedlings
cannot survive extended periods of inundation.
The recruitment success of wetland trees characteristic of mid to late succession is
unknown on CSAs. One direct seeding experiment on phosphate mined land was largely
unsuccessful: 10 of 14 plots that were covered with litter collected from floodplains in the
vicinity failed to produce seedlings (Rushton 1988). The quantity of viable seeds in the collected
litter was unknown.
A series of gradual changes in the dominant vegetation community toward a predictable
climax state summarizes the traditional concept of succession. Numerous theories have emerged
both further elucidating the mechanisms of succession (Clements 1916, Egler 1954, Connell and
Slayter 1977), and challenging its linearity and predictability (Anand and Desrochers 2004). Yet
the changes in the composition of the vegetation community are just one aspect of alterations to
both the abiotic and biotic environment that are associated with succession. In the context of the
entire system this dynamic process has been called ecosystem development (Odum 1969).
A key aspect in the development of an ecosystem is an increasing effect of the biotic
components of the system on the modif ication of the environment and the selection of the biota.
The increasing control exerted by the biotic components is a characteristic of self-organization
(Odum 1989). The dynamics of self-organization in the "emerging" ecosystems on CSAs are
unclear. Measures of the modifications that the biota are making to the environment and the
changes in the community composition that may be resulting from those changes are potential
indicators of ecosystem development.
In forested ecosystems, trees are key agents of influence over the local environment and
thus the ecosystem. As trees mature and canopies develop, they reduce the quantity of light that
is able to penetrate to the lower vertical strata of the forest. The reduction in light penetration
alters the microclimate (notably temperature and humidity) underneath the tree canopy. These
changes to the abiotic environment imparted by the trees may in turn cause changes in the cover
and composition of the understory vegetation (Beatty 1984) and the rate of organic matter
decomposition in the soil. Trees also contribute a substantial amount of the detritus that
decomposes and becomes incorporated in soil organic matter (Rhoades et al. 1998). In a study of
carbon budgets in the Dismal Swamp, tree leaf litter and fine tree roots composed the largest
annual input to the detritus pool in both cypress-dominated swamps and mixed forested wetlands
(Megongial and Day 1988). All these effects are expected to be enhanced with increasing tree
size and dominance in the landscape.
Planted wetland trees on CSAs may serve the role of directing ecosystem development.
Restoration ecologists have traditionally looked at a spectrum of similar sites of different ages to
study the dynamics of ecosystem development. A number of studies of the progress of
restoration efforts in the phosphate mining districts have adopted this approach (Rushton 1983,
Carstenn 2000), and identified trends in ecosystem development across sites. A potential
drawback of this approach is that is overlooks the site-specific influences. The topography and its
influence over the hydrology and the proximity to seed source are unique to a site and important
external drivers of ecosystem development. These external factors may create challenges for
cross-site comparison of CSAs.
Plan of Study
Rushton planted tree species on a number of abandoned CSAs in 1985- 1986 as part of her
doctoral study (Rushton 1988). Because she published precise information on location, number,
and type of species planted as well as growth and survival rates after one year and descriptions of
sites conditions, monitoring these planted areas and adjacent non-planted areas provided an
opportunity to evaluate tree growth and ecosystem development of areas with and without planted
trees after a 20-year time period.
To evaluate how the planted trees have fared over time and to determine what factors are
influencing growth and survival, survival, size, and reproductive success of planted trees was
measured. The tree parameters were statistically evaluated in the context of site hydrology and
soils. Elevation data and water levels were collected to estimate water depths and period of
inundation in planted plots during the 2005 growing season, and data on site soils were gathered
from the Rushton study (1988). The effect of disturbance was qualitatively assessed through site
histories and field evidence.
In effort to project how the tree populations might change in the future, the tree data were
used to calibrate population models to determine future population trajectories.
To determine if the planted trees were steering ecosystem development, selected ecosystem
development measures were collected in planted and non-planted areas with similar hydrologic
conditions and external influences. Woody vegetation was measured to assess the development
of the tree and shrub strata; canopy photos were taken to estimate canopy cover; soil samples
were collected to estimate percent organic matter; and understory vegetation was sampled. The
raw data were summarized by plot and statistical techniques were then used to compare measures
of ecosystem development in planted and non-planted areas.
Site and Plot Selection
Five of the CSAs planted by Rushton were selected for study. Sites were chosen that were
currently accessible and that Rushton (1988) had determined had an average of at least 50% tree
survival after one year. Table 1 presents a summary of the selected sites. Figure 1 provides an
overview map of site locations.
CFI SP-1 (CFI) is a sand-clay mix settling area abandoned in the early 1980s with two
distinct connected lobes. Plots were planted on the fringe of the east lobe. Since the Rushton
planting, the site has been planted with additional tree and understory species and the water level
has been lowered by adjusting the weir. The upland area surrounding the wetland and adjacent to
the plots is regularly mowed and shrubs have been removed. The understory of a few of the plots
were planted with ferns on their upland half.
Homeland (HOM) is a pond formed over an old mine cut backfilled with clay and capped
with sand around 1979. The pond is surrounding by pasture that is part of the DEP Homeland
office property. Bill Hawkins planted Taxodium distichum trees in 2/3 of the pond approximately
in 1982. The Rushton plots traverse the east side of the pond.
OH Wright (OHW) is an older CSA (abandoned approximately in 1960) adjacent to the
Whidden Creek floodplain. One plot (R1A) traverses a swale just above the outfall structure,
which is still active. Four plots (R2A, R2B, H1, and H4) are on the fringe of a pond. Two other
plots (H2, H3) lie in a depression between two spoil rows.
Peace River Park (PRP) was abandoned in 1968 and leased for pasture until 1986. Two
plots are located in small depressions (H1, H6) and two are located on the edge of a pond (H4,
H5). All plots are connected by surface water when water levels are high.
Tenoroc 4 (TEN) was abandoned around 1972 and is now part of the Tenoroc Fish
Management Area. Four plots (R2A, R2B, H2, H3) were located in a depression on the NW
corner of the site. The other plots are on the north and south side of an interior spoil pile in the
north central area of the site. Prior to a ditching effort in this area to connect isolated depressions
and convey water off the site in 200 1, the seasonal water levels in the plots were likely higher.
Rushton plots. A total of 37 planted plots on 5 CSAs were selected for study. Selected
plots were located in the field from site diagrams (Rushton 1988) and matched to an original plot
number. All selected plots had at least one surviving tree at the present time. Plots were
representative of the two planting schemes used and referred to by Rushton as cypress-gum (CG)
plots and hydric-swamp (HS) plots. Figures 2 and 3 depict planting schemes for these two types
of plots. Twenty-five cypress-gum plots and 12 hydric swamp plots were included in the current
study. Species planted in the two plots types are listed in Tables 2 and 3. Cypress-gum plots
were planted with all three species except for 4 plots at Tenoroc 4 planted only with two species.
Among the 12 hydric swamp plots, 8 were planted with species with a group of 'transitional'
trees and 4 were planted with a group of 'wet' trees.
Reference plots. In order to compare the ecosystem development on non-planted areas that
were similar to the Rushton plots, adjacent 'reference' plots of equal dimensions to the Rushton
plots were selected. Reference plots were at least 25 meters away from Rushton plots to
minimize potential influence from Rushton plots. A single reference plot was designated for all
plots that shared connection to a water feature. Reference plot selection was random provided
that a plot met the following conditions: (1) it was adjacent to the same water feature as a
Rushton plot; (2) the topography was such that a hydrologic regime similar to the Rushton plot
could be inferred. An exception to the first condition occurred at Homeland, where the reference
plot was located in a pond fed by a ditch from the pond containing the Rushton plots, because not
enough non-planted area within the pond with the Rushton plots was available.
Field Data Collection
A laser level was used to determine elevations within the plot relative to the water level at
the time of first visit. Figure 4 shows where data were collected in cypress-gum plots. For these
plots, elevation was recorded every meter along a 42 m longitudinal axis which traversed the
planted area as well as 6m in front and back of it. The plots were originally laid out such that this
axis ran parallel to the elevation gradient. Additionally, elevation data were recorded from spots
6m to each side and at the beginning, middle, and end of the longitudinal axis. Figure 5 shows
where data were collected in hydric swamp plots. For hydric swamp plots, elevation data were
recorded every two meters along two perpendicular axes crossing from 6m away from the edge
through the center of the plot to 6m beyond the far edge. In these plots elevation data were also
collected at the soil and plant sample points within the planted plot, and at the four planted plot
For reference plots, elevation data were collected in the same manner, except in these plots
only data within the plot boundary were collected.
Water levels at a point of recorded elevation were manually measured to the nearest
centimeter each month through October 2005after the initial visit to a plot in the spring or early
summer of 2005.
On CFI SP-1 and Tenoroc 4, continuous digital data loggers were installed close to or
within Rushton plots to record hourly water levels. On these sites, one surface water well within
the water feature and one ground water well 25m into the upland were equipped with loggers.
The loggers were operational from the date of installation in the early part of the growing season
of 2005 through the end of October 2005.
Planted trees were identified by location and species. Using a two-dimensional grid, X,Y
plot location was recorded for planted trees to the nearest meter. Diameter at 1.5 meters (DBH)
was recorded to the nearest centimeter for all stems originating below that height. If no stems
reached 1.5m, height of the tallest stem was recorded to the nearest centimeter.
Other Tree Species
Each tree within the planted plot of a species not planted was identified to species and its
DBH was recorded if it had reached 1.5m in height. Woody plants were classified as trees or
shrubs according to Tobe et al. (1998). For Salix caroliniana, which is classified as a tree or
shrub, individuals with at least one stem with a DBH 15cm were classified as trees. In cypress-
gum plots, the 10m segment (0-10,10-20,20-30) that a tree was found in was noted.
Recruited trees are defined in this study as individuals of the same species as planted trees
not occurring in originally planted locations, irrespective of the size of the individuals. X,Y plot
location, species, and DBH or height were recorded for recruited trees inside or within 6m of the
In the plot on TEN where the greatest number of seedlings emerged, the seedlings were
resampled at the end of the growing season to determine the survival rate.
Additional Measures of Ecosystem Development: Shrub and Understory Layers; Soils;
Figure 6 and 7 show the standardized sampling locations for shrubs, understory vegetation,
soil, and canopy photos for cypress-gum and hydric swamp plots. Three 3x3m subplots within
each plot were sampled for shrubs. DBH and species were recorded for all stems 1.5m in
height. Nine 1xlm subplots within each plot were used to sample all understory macrophytes
with stem heights < 1.5m. Each species occurring was identified and the coverage of each
species was estimated into one of five possible coverage classes: 1: 1-10%, 2: 10-25%, 3: 25-
50%, 4: 50-75%, 5:75-100%. Coverage was defined as the percentage of the 1xlm horizontal
plot area covered by the plant. In the case where different species occupied the same horizontal
location but different vertical strata, both species were counted. Cores of the top 10cm of the soil
were collected with a 7.6 cm-diameter auger within all 1xlm understory sampling plots. To
estimate canopy cover, hemispherical photographs were taken using a Nikon digital camera, with
180 degree "fish-eye" lens. Inside all plots, photos were taken in 3 equidistant understory
subplots. For the Rushton plots, photos were also taken from the understory subplots outside of
the canopy. The camera was placed on a tripod approximately 50 cm above the ground or
slightly above the surface of the water, whichever was higher. The camera was then leveled with
the lens pointing up, oriented so the back of the camera faced north, and zoomed out to 100%.
When possible photos were taken close to dawn or dusk or on overcast days to avoid distortion
from direct sunlight.
Information about possible disturbance or site modification during the 20 year period since
the trees were planted was collected from site managers, from the Rushton dissertation, through
consultation with Betty Rushton, or through inference from evidence found in the plot in 2005
such as burnt stems or plot markers.
Topography and Water Levels
Topographic data collected were input in X-Y-Z form into Surfer surface mapping
software, from which a kriging function was used to create a surface map. From this interpolated
map, relative elevations were output for every square meter. Using these elevation data and the
monthly water level data, water levels were calculated for the entire sampling area for every date
water level was recorded. 'Average water depth' as referred to in the remainder of the study
refers to the average of these monthly water levels.
When the water level was below the ground, the level measured in one location was
assumed to be the same across the plot because of the small area of the plots and the small
differences in ground elevation across the plots.
On the two sites with continuous data-logging water level recorders (CFI and TEN), the
average of the sampled monthly water levels was compared with the average of all the hourly
water levels recorded by the data-loggers to determine if monthly water levels accurately
approximated hourly water levels on those sites.
Average change in elevation was computed for each plot as the average change in elevation
along the longest axis of the plot. Percent inundation was calculated as the area of the plot
covered by water at the time of sampling divided by the total plot area.
Elevation data for every planted tree along with monthly water level measurements
allowed for determination of the average sampled depth of water for every tree and at every
location where soils, shrubs, understory vegetation, and canopy photo sampling occurred. Box
plots were created to show the distribution of all trees along the average water depth at the tree
Tree and Plot Basal Area
Basal area, BA(cm2), was calculated for trees and shrubs as the sum of the all stem area at
1.5m for an individual according to the following equation:
BA = C x DBH2 2I
Plot basal area (m2/hec) was the sum of the tree and shrub basal area (m2) divided by the
plot area (hectares). Plot basal area was calculated for every 10m section of cypress-gum plots as
well as for the entire plot, but only for the entire plot in hydric swamp plots because trees were
not sub-sampled in these plots.
Tree Growth Comparisons
Basal area of all surviving Rushton trees was compared by species and soil type for trees
with an average water level in the range of -0.75 to 0.25m. Trees with a basal area < 7.8 cm2
were assumed to be resprouts, and they were eliminated from the growth comparison because the
stem age was unknown. Tree basal area of the remaining trees was then log-transformed for
normality. T-Tests were conducted to compare the effect of two soil types in areas with similar
average water depths, assuming that similar hydrologic regimes can be inferred from similar
average water depth at the tree base during the 2005 season. Two way ANOVA was used to
simultaneously compare the effect of water level, soil type, and soil type-water interactions on
tree growth. In the two way ANOVA test trees were split into shallow and deep water levels by
species based on the median water level of surviving trees.
Soil Percent Organic Matter
Soil cores were manually homogenized and three 40g samples of each core were dried a
minimum of 48 hours at 300 C. The ignition method without rehydration was then used to
estimate % organic matter (% OM) Dried samples were ground with a mortar and pestle and
three 1 g sub-samples were ashed in a muffle furnace for 6 hours at 4500 C. This temperature was
deemed appropriate for burning off the organic matter without removing inorganic carbon
(CaCO3). The following equation is used to calculate percent organic matter:
((dry weight -ashed weight) / dry ~ight) 100% = % organic matter 
Population Size Class Distributions
All surviving planted trees and offspring were placed into size classes that represented 5 or
10 cm DBH intervals (Table 4). Classification was done by basal area to accommodate multiple
stem trees where summation of DBH would have resulted in inflated values and inconsistent
classification.' Classified trees were then grouped by species and by basins to define a
For example, a tree with two 5 cm DBH stems has less basal area (39.4 cm2) than a tree with
one 10 cm DBH stem (78.5 cm2)
population. Basins are defined as areas where multiple plots are adjacent to the same body of
water and no plot is more than 50 meters away from its nearest neighbor. The sampled area of
each basin represented the sum of the seedling sampling areas of every plot within the basin; not
the area of the entire basin.
Canopy photos were analyzed in Adobe E Photoshop software. Photos were transformed
into 2-color black and white images using the Threshold function. The threshold level was
subjectively chosen to yield the most accurate conversion of vegetation pixels to black and sky
pixels to white. Before transformation images were cleaned up with editing tools to remove
shadows, clouds, sun spots, glare, or other aspects of the image that would been incorrectly
assigned to black or white. After transformation, the black and white pixels were counted in
Keigan Systems E MFworks software. The percent canopy cover was then calculated as the sum
of black pixels divided by the sum of black and white pixels.
Cover for all understory vegetation in a plot was estimated using the mean of the coverage
class. The classes thus corresponded to the following percentages: Class 1: 5.0%; Class 2:
17.5%; Class 3: 37.5%; Class 4: 62.5%; Class 5: 87.5%.
Species richness was calculated for all plots as the sum of the unique species occurring.
Species evenness, a measure of the evenness of the distribution of species, was calculated with
the Shannon evenness formula (Gurevitch et al. 2002):
E =H/In(S) 
H = 1(p, In(p;))
where evenness, E, is equal to the Shannon-Wiener index, H, divided by the natural log of the
total number of species, S. The Shannon-Wiener index was calculated as in Equation 4.
Importance Value is a metric that combines the relative frequency and relative cover in
order to consider together both characteristics of a species presence in an understory (Cole 1978).
Importance Value for species occurring in the understory were calculated using the following
IT/= rfs + rcs 
where Importance Value of a species, IV,, is equal to the sum of the relative frequency, rfs ,and
relative cover, rcs, of that species. Relative frequency was calculated using the following
s= > 
J = os q 
where relative frequency is equal to the frequency of a species, s, divided by the sum of the
frequency of species encountered on a plot. The frequency of a species was calculated by the
number of a 1m2 quadrats in which species s occurred, os, divided by the number of 1m2 quadrats,
q, in a plot.
Relative cover was calculated using following equations:
'= > 
where the relative cover of a species, rcs, is the cover of a species divided by the sum of the cover
all species, n, in a plot. The cover of a species, cs, is equal to the sum of the mean cover of a
species, s, in all 1m2 quadrats, q. Because a cover class was assigned to a species rather than a
mean cover, each cover class was translated to a mean cover (reference on method) as follows: 1:
5%, 2: 17.5%, 3: 37.5%, 4: 62.5%, 5: 87.5%.
Ordination of Plots by Prevalent Understory Species
In order to visualize the differences in the cover of prevalent understory species between
plots, the Nonmetric Multidimensional Scaling (NMDS) ordination technique was applied. The
prevalent understory species were those with a Importance Value of > 0.10 (out of a possible
2.0) for a plot. The NMDS method does not require assumptions that the data fit a normal
distribution nor that the data fit a linear pattern (Faith et al. 1987, McCune and Grace 2002). The
NDMS was run on a (n x p) contingency table of average species cover in a matrix where the
rows, n, were plots, and the columns, p, were species. The data were first standardized using a
Wisconsin double standardization and then square-root transformed. A Bray-Curtis dissimilarity
method was used as to create the dissimilarity matrix necessary to rank plots by dissimilarity and
to position the points along the two principal component axes, so that the ordination could be
shown in two-dimensional space.
Correlation Matrices of Ecosystem Development Variables
To find patterns in the relationship between Rushton trees and total basal area, canopy
cover, understory cover, understory species richness, understory species evenness, and soil
organic matter, correlation matrices were created using R statistical software. Pearson's formula
was the correlation method used to produce the matrices.
Tree Population Model
In order to predict the population trajectory of a planted tree population, a size class matrix
population model was constructed for populations of planted Taxodium distichum at CFI and in
one basin at OHW.
Size class matrix population models use principles of matrix algebra to estimate changes in
population distribution over a time series as well as the steady-state population distribution and
growth rate (Caswell 2001). Size class bins are determined and individuals are classified into size
classes. A transition matrix, A, is constructed by determining probabilities after a year that a tree
will remain in a size class, P,, transition, G,, and/or reproduce, F, (Figure 8). The transition
matrix is multiplied by a vector of the number of individuals in each size class, N,, to determine
the number of individuals in each size class after one time increment, N,+;. According to matrix
theory the transition matrix alone determines the long-term population state. Mathematically
decomposing the transition matrix, A, yields a vector of eigenvalues and their associated
eigenvectors. The dominant eigenvalue of A, h, gives the population growth rate when there is a
stable population distribution. The stable population distribution is given by the right eigenvector
of the transition matrix.
Customarily tracking the growth, survival, and seed production of a cohort of trees over a
period of years provides the data from which transition probabilities are calculated. In this case,
empirical time series data was not available for the entire period. Using data from the most
current year and incorporating data on survival and growth after 1 and 3 years, growth of
individual trees were interpolated by fitting a curve based on the growth rate of other Taxodium
distichum in the phosphate mining area (Miller 1983). Mortality after years 1,3, and 20 years
were used to estimate mortalities of the given size classes, with the assumption that slower-
growing trees were more likely to die. Reproductive probabilities were calculated based on the
ratio of first year seedlings to mature adults, distributing this probability among the mature size
classes such that each successively larger size class had a greater reproductive probability. The
matrix populations models were created in the Python 2.3 programming language. The model
was programmed to estimate population change over a 50 year period. An elasticity analysis
(Caswell 200 1) of the model was conducted to estimate the relative sensitivity of the model to
the changes in the probability values of the transition matrix, A. The code for the population
model is included at the end of the Appendix.
Table 1. Site summary table
Abandoned # Cypress-gum # Hydric Swamp
Site Name Symbol (Estimated) Type Plots Plots
CFI-SP 1 CFI 23 Sand-Clay 6 0
Homeland HOM 46 Sand Cap 8 0
0.H. Wright OHW 46 Clay 3 4
Peace River Park PRP 38 Clay 0 4
Teneroc 4 TEN 34 Clay 8 4
Figure 1. Study site locations. Map adapted from Rushton (1988).
OF IND. HA DE E
Table 2. Species list for cypress-gum plots
Fraximes pennsylvanica FRPA
Nyssa aquatica NYAQ
Taxodium distichitm TADI
0 0 09 Ra* a 0 0 01
60 0 e Ispl 0 0
Figure 2. Cypress-gum plot layout from Rushton (1988). Two plots are pictured. Each plots was
planted with 93 seedlings.
O PVC PrE
" + + + + +1
a c 9m I +
Table 3. Species list for 'wet' and 'transitional' hydric swamp plots
"Wet" Plots Transitional' Plots
Species Symbol Species Symbol
Fraximes caroliniana FRCA Acer rubrum ACRU
Nssa sylvatica NY SY Gordonia lasianthus GOLA
Persea pahtstris PEPA Nssa sylvatica NY SY
Quercus laurifolia QULA Quercus laurifolia QULA
Taxodium distichum TADI Sabal palmetto SAPA
Ulmus americana ULAM Taxodium distichum TADI
Figure 3. Hydric swamp plot layout from Rushton (1988). Two plots are pictured. Each plots
was planted with 108 seedlings.
is 6m ,\E6m E/
E- Gm E 6m
Figure 4. Elevation diagram for a cypress-gum plot. Numbers are in meters.
6 14) ($ 6 14) (1
3 5 1
Elevation taken evey 2 m
Retainted Iree: samplilng area
Figure 5. Elevation diagram for a hydric swamp plot. Numbers are in meters.
H = 1m2 understory sampling plot
S = 10cm soil core
P = Canopy photo
Figure 6. Soil, understory, shrub, and canopy photo sampling scheme for cypress-gum plots.
Numbers are in meters.
Planted plot Shrub3
-1 3 5 7 9 11
H = 1m2 understory sampling plot
S = -10cm soil core
P = Canopy photo
Figure 7. Soil, understory, shrub, and canopy photo sampling scheme for hydric swamp plots.
Numbers are in meters.
Table 4. Size class key used in tree size class distributions.
size class DBH(cm) BA(cm2)
0 NA 0
2 5-10 19.7-78.5
3 10-15 78.6-176.7
4 15-20 176.8-314.2
5 20-30 314.3-706.9
6 30-40 707-1256.6
7 >40 >1256.6
'Po Fl F2 F3 F4 F5 F6 F7
Go P1 0 0 0 0 0 0
0 G1P2 0 0 0 0 0 O
0 0 G2 P3 0 0 0 0
0 0 0 G3 P4 0 0 0
0 0 0 0 O G4 Ps 0 0
0 0 0 0 O Gs P6 0O
0 0 0 0 0 0 G6 P7
Figure 8. The format for the transition matrix, A, for the matrix population model. The figure
above is a matrix for a population with eight size classes (0-7). The P values along
the diagonal represent probabilities of remaining in the same size class; the G values
represent the probability of advancing into the next class, and the F, values represent
the probability of successful reproduction.
Tree Populations in Relation to Environmental Factors
Survival of trees planted by Rushton is first summarized. Hydroperiod of the planted plots
is compared and tree survival is examined across different water depths and on three soil types.
The consequences of initial tree growth and site disturbances are considered. Population scale
questions are approached by looking at populations of recruited trees within and on the periphery
of plots, tree size class distributions, and population models of Taxodium distichum on two sites.
Tree Survival by Site and Species
Cypress-gum plots. Table 5 summarizes the planted tree survival percentages after 1, 3,
and 20 years. Aggregating all sites, Taxodium distichum survived best after 20 years (34%),
though Fraxinus pennslyvanica had the best survival at the end of the three years (70%).
Figure 9 presents survival trends by site and species in cypress-gum plots. Aggregating all
three species, trees at the CFI site had the highest survival after 20 years (50%), and trees at TEN
had the lowest survival (9%). More Fraxinus pennslyvanica were found after three years than
after one year at CFI and OHW, most likely due to resprouting. Survival of Fraxinus
pennslyvanica after 20 years was the poorest at HOM (8%), but highest of the three species at
CFI (70%) and TEN (27%), two of the four sites with cypress-gum plots. Survival of Taxodium
distichum was greater than 98% year' (indicated by the slope of the trend line) between years 3
and 20 at all but the TEN site. Nyssa aquatica had poorest survival in the initial year, but the
survival rate between years 3 and 20 was the best of the three species at OHW and TEN, and
better than Fraxinus pennslyvanica at CFI and HOM. Compared with the survival rate during the
first year, all species had improved annual survival rates between years 3 and 20.
Hydric swamp plots. Table 6 summarizes tree survival in hydric swamp plots after 1 and 19
years. Acer rubrum, Fraxinus caroliniana, and Taxodium distichum, and Ulmus americana were
the only species present in sampled plots after 19 years. No individuals of Gordonia lasianthus,
Nyssa sylvatica, Persea palustris, Quercus laurifolia, or Sabal palmetto were found surviving in
any of the plots after 19 years.
Figure 10 shows tree survival in hydric swamp plots by site and species. Only one individual
of Ulmus americana survived 19 years and is not depicted. Of the three other surviving species,
total survival after the first year for each was greater than 80% (see Table 2 for survival data by
species). Survival ofdcer rubrum after 19 years was 20% or less at all sites, with no surviving
individuals found at TEN. Fraxinus caroliniana had the best survival in HS plots. At both OHW
and TEN, all individuals survived after 19 years, a few having resprouted after the original stem
died during the initial year. About half of Taxodium distichum trees that were surviving after 1
year survived 19 years, except at OHW where 20 year survival was only 12%, due to high
mortality in two plots.
Average hourly water levels on the two sites where continuous data loggers were installed
were within 3 cm of the average monthly water level measurements. Appendix Figures 33 and 34
show the continuous recorded levels and the monthly sampled levels at CFI and TEN.
Cypress-gum plots. Figure 11 shows the percentage of a plot that was inundated at the
time of monthly water level sampling. Variation of inundated area occurs within and between
sites, with some obvious trends apparent. Plots at CFI demonstrate a range of inundation, varying
from R1, which was almost totally inundated on all dates, to plot R6, which was at most 15%
inundated. Thus all trees at R1 stood in standing water much of the season, whereas water level
was below ground for most trees in R6. Nearly all eight plots at HOM were inundated upon
every visit. At OHW, plots R2A and B, adjacent plots on a pond fringe, were more than 50%
inundated in 4 of 5 months sampled, whereas about one third R1A, which crosses a drainage
channel, was consistently covered in water. At TEN all plots were dry in May but for most of the
season more than 50% of R2A and R2B were inundated. R5A, R5B, R6A, R6B, R7A, and R7B
are on a pond fringe, and all plots were mostly inundated when sampled in July and August, but
on visits earlier and later in the season were wet only in the deepest ends, if at all.
Hydric swamp plots. Occupying less of an elevation gradient than cypress-gum plots,
hydric swamp plots exhibit a more uniform response to water level than cypress-gum plots (see
Figure 12). Many plots were inundated through the season, including all plots at PRP and H1 and
H4 at OHW, whereas others, such as OHW H2, were dry at every sampling. Sites at TEN all
were completely dry when sampled during May, and only H2 and H3 had a small area inundated
at the September sampling, but during other months H2 and H3 were completely inundated. TEN
and OHW both had two rather wet and two dry sites, whereas at PRP, all sites were wet.
Tree Survival and Hydrology
Cypress-gum plots. A box plot (Figure 13) showing the average water depth for the
planted trees by species are shown in comparison with a box (first from left) showing the average
water depth for the plots. As all species were initially planted along the entire water level
gradient in a plot, this box represents the distribution of water depths at all original planting
locations. A comparison of this first box of all planting locations with plots of surviving
individuals of each species shows where tree survived along the water level gradient. The range
of surviving Fraxinus pennslyvanica extends from a water depth of 0.5 to -1.0 m, excluding the
deeper portion of the original range. The population of surviving Nyssa aquatica and Taxodium
distichum withstood more inundation than the population of surviving Fraxinus pennslyvanica.
Only a few outliers of the two populations occur where the average water level was below -0.6
meters. Taxodium distichum, which had the highest survival, occurs along a broader continuum
of water depths than Nyssa aquatica. No individuals of any of the three species survived in the
deepest part of the originally planted range.
In Figures 14-16, tree survival after 1 and 20 years is compared by species for all cypress-
gum plots within the same site. For instance, in the bottom chart in Figure 14, Taxodium
distichum are split into those surviving after 20 years and those that died between years 1 and 20.
These two groups are then classified by average water depth either at the tree base, or the former
location of the tree for those that died between years 1 and 20. At CFI, the range of water depth
in which all three species survived did not change between years 1 and 20. More Fraxinus
pennsylvanica and Nyssa aquatica trees died than lived in the shallowest water depths at this site.
Once established, Taxodium distichum at CFI appears to be capable of tolerating the entire water
level range over which the trees were planted. At OHW (Figure 15), Fraxinus pennsylvanica
appears to have a much more limited water tolerance range, as only trees with an average water
depth of 0.2-0.3 meters survived. Only a few Nyssa aquatica survived and they appear to have
tolerated depths between 0.2 and 0.4 meters, as Taxodium distichum appears to have tolerated
those depths as well as 0.0-0.2 meters. At TEN (Figure 16) Fraxinus pennslyvanica tolerated the
drier locations where it established, but not in locations with average water levels above the
ground surface (0.0 meters). Taxodium distichum survived where water levels were higher than -
0.3 meters. Nyssa aquatica survival was poor across the range.
Hydric swamp plots. Figure 17 shows distributions of Taxodium distichum in hydric
swamp plots where average depths at the surviving trees ranged from -0.5 to 0.9 meters. This
species was not found in drier locations from -0.75 to -0.5 meters and not in the wettest locations
where average depth was >0.9 meters. The range of original planting locations of Fraxinus
caroliniana were similar to that for Taxodium distichum but not drier than -0.3 meters because it
was not planted in the drier plots. Surviving individuals were not found where average depths
were < -0.2 or >0.9 meters.
Tree Survival and Soil Type
Cypress-gum plots. Figure 18 summarizes trees survival on the sand-clay, sand-capped,
and 3 clay sites. Trees growing on sites with clay soils had the lowest survival after 20 years.
CFI, the sand-clay site, had the best overall survival. Though Nyssa aquatica survived poorly on
the clay sites after the first year, the survival rate between years 3 and 20 on clay was better than
on the sand-cap site (HOM) and similar to sand-clay site (CFI). The slope of the trend line can be
used to estimate annual survival rates of species. Taxodium distichum average survival rate
between years 3 and 20 was poorest on the clay sites at about 97% yr ', and high on both the
sand-cap and sand-clay site, at >99% year '. The population of Fraxinus pennslyvanica declined
about 50% on the clay and sand-cap site between years 3 and 20. Due in part to resprouting,
almost as many Fraxinus pennslyvanica trees were alive at CFI after 20 years as there were after
1 year, where a very high percentage (70%) survived.
Tree Growth Comparison Between Sand-Clay and Clay Sites
Tree populations in clay and sand-clay were compared to examine the effects of soil
medium on tree growth. In 2005, all surviving trees on clay occurred within the range of water
depths to which trees growing in sand-clay were exposed (see Figure 19). Results of t-tests to
determine if a significant difference existed between growth of trees on clay and sand-clay are
presented in Table 7. Taxodium distichum trees from both cypress-gum plots and hydric swamp
plots were considered in the analysis. Growth of Fraxinus pennslyvanica and Taxodium
distichum on clay and sand-clay was not statistically different. Growth of Nyssa aquatica was
better (at a 95% confidence level) on clay, however there were only 13 Nyssa aquatica trees
surviving on clay, a very small percentage of those originally planted.
Results of the two-way ANOVAs performed to simultaneously compare the effect of water
level and soil type on tree growth for trees growing in clay and sand-clay are presented in Tables
8 and 9. Trees on the sand-cap site (HOM) were eliminated from consideration because of higher
water levels. For Fraxinus pennslyvanica, trees with an average water depth of less than -0.25m
were grouped as 'shallow' and those with a water depth greater than -0.25 were grouped as
'deep'. Fraxinus pennslyvanica did not show a significant difference for either the soil type,
water level, or interaction of the two. Taxodium distichum trees were split into 'shallow' and
'deep' classes using the median average water level of0O.0m. This test showed a significant
effect for water level and for the interaction of water level and soil type. Trees in deep water had
an average basal area of 5.4 cm2, .4cm cm2 greater than trees in shallow water, but the variance in
basal area was also much higher for deep trees (1.53 to 1.19). Though planted on both soils,
survival of Nyssa aquatica in clay was too low to allow for a comparison of the effects of soil
type and water level on growth for this species.
Initial Tree Growth and 20-year Tree Survival
Records of tree height on cypress gum plots after 1 year were paired with tree survival
records within the same plot to determine if trees that grew faster during the 1st year were more
likely to survive 20 years. Tree height records after one year were available for 6 plots on CFI, 2
plots on OHW, and 6 plots on TEN. Of the trees with a height record, 296 were surviving in
2005 and 408 were dead. A T-Test was performed to determine if the heights of the trees after
one year were different for these two groups, after the height was square root-transformed to
satisfy the condition of similar between-group variance. The outcome, a p-value of 2.2E-16,
indicated with a very high level of confidence that the surviving trees had a greater height after 1
year than the trees that died between 1 and 20 years.
Among the six plots on TEN, the average height of planted trees after one year was 35 cm,
in comparison with 95 cm at CFI. Twenty-year survival of the TEN trees was 17%, versus 54%
at CFI. Among these plots there is a strong correspondence between tree height after 1 year and
Site Disturbance and Tree Survival
On a number of sites, disturbance factors directly caused mortality or damage to the planted
trees within the initial year of establishment or in years since. Where records of these
disturbances exist, they are presented in Table 10. Fire, heavy grazing, and mechanical
disturbance (tractors, etc.) are known to have influenced a number of plots. A fire occurred in
two hydric swamp plots (as well as in a number of cypress-gum plots not monitored in this study)
that lie within a gully between two spoil piles on OHW. Multiple fires burned into all four of the
hydric swamp plots in PRP, where dead tree trunks blackened from burning still stand as
evidence. On HOM, four transects were subjected to grazing by cattle during their early years.
In one basin of TEN, heavy herbivory negatively effected tree growth and survival during the first
year (Rushton 1988). Segments of a few transects were damaged by earth-moving equipment,
including the first 8 meters of CFI R2 and the first few meters of both TEN 5A and 5B.
Numerous other disturbances may have occurred without leaving any direct or anecdotal
evidence, including prolonged flood events, drought or heavy winds.
In a few cases, seedlings and mature trees of the same species as planted trees ('recruited
trees') were found in abundance inside seedling sample plots, whereas in some plots no recruited
trees were found. Tree populations in plots are presented in Table 11, where they are ranked by
the ratio of the number of surviving planted trees to the number of recruited trees (reproductive
ratio). Populations are defined in this table as all trees of a given species within the seedling
sampling area of a plot. Only populations with at least one surviving tree and one planted tree are
listed; 30 populations met this criterion. Where another plausible source for the recruited trees
exists, this source is mentioned in the table. In nine populations, the number of recruited trees
was greater than or equal to the number of planted trees. In two of these populations, the number
of recruited trees was approximately 100 times greater than the number of planted trees. But in
both of these two populations, there are clear seed sources other than the planted trees.
Additional plantings of Taxodium distichum adjacent to or within sampling areas since 1985
occurred at CFI and HOM, but locations of those plantings were not available and thus trees not
planted by Rushton could have either been planted later or are offspring of trees from another
New Seedling Survival
The Taxodium distichum seedling (0-100 cm in height) population at TEN H3 was the
largest of any plot sampled in June with 128 individuals. In November, the population had been
reduced to 52 individuals. As location of the seedlings was noted only to the nearest meter and
seedlings were not tagged, it was not possible to track individual seedling growth with certainty.
But size class distributions of the seedling populations during both periods reveal in which
segments of the population mortality occurred (Figure 20). A comparative look at the two
distributions reveals a close match between trees in classes > 20 cm, but there are many more
trees in the first two classes in June than in November. In June there were a total of 87 trees in
the first two classes, whereas there were only 10 in November. The size of class '3' in November
indicated that only a few of these trees likely grew into a larger size class during this period. The
water level record reveals that the water was between -0.5 and the ground surface in May at the
locations were the 87 individuals less than 20 cm stood in June. Of those seedlings, 72 were
completely inundated in water during the June and July sampling.
Tree Population Size Class Distributions
Figure 21-24 show size class distributions of Taxodium distichum, Nyssa aquatica,
Fraxinus pennsylvanica, and Fraxinus caroliniana. The composition of each size class is split
into planted and recruited trees. Populations of Taxodium distichum are shown in six basins in
Figure 21. Trees at CFI are the most evenly distributed across size classes. Recruited trees at
CFI appear in the first four size classes. At HOM there is a more normal-shaped distribution,
with obvious omissions in the seedling class (class 0). At OHW, PRP, and TEN there are fewer
trees, in part because some of the plots were hydric swamp plots, where fewer trees of a species
were planted, and in part because of lower survival. The first basin at TEN had an exceptionally
high number of seedlings (see Table 1 1, row 1). Four trees in classes 4, 5, and 6 in this basin
appear as 'recruits' but are actually trees planted by Rushton in a plot not included in this study
that overlapped with the recruited tree sampling area.
Nyssa aquatica populations were too small in basins at OHW and TEN such that trees were
only distributed between 2-3 middle range size classes (Figure 22). CFI has a small number of
seedlings but the approximately the same number relative to other size classes in comparison with
its Taxodium distichum population.
The CFI basin had six times as many surviving Fraxinus pennslyvanica as the other basins
and a normal shaped population distribution (Figure 23), but the distributions of the populations
are similar in the other basins, albeit they were lacking in smaller trees.
Only a small number of Fraxinus caroliniana were planted in two basins and in both cases
there are more individuals than originally planted (Figure 24).
Tree Population Model
The model for Taxodium distichum at CFI used the records of 266 trees to construct the
transition matrix (Figure 25). The h of this transition matrix was 1.005; the model predicts that if
the population were to obtain a stable population distribution, it will increase but at a slow pace.
The population projection for the next 50 years shows at first a slowing decline from 150 to a low
of about 120 trees after 20 years, but then growing again to 130 at the end of 50 years (Figure 26).
The model for the Taxodium distichum population on the OHW basin used records of 106 trees
for construction of the transition matrix (Figure 27), with no trees presently in the largest size
class (7). The h of this transition matrix was .991, indicating a slow long-term population
decline. After 50 years the model predicted that the tree population would fall from 36 to 16 trees
in the basin (Figure 28). Though the h values represent potential opposite long-term projections
for the two populations, the model does not predict drastic population change for either basin
within the next 50 years.
Relative to the mature tree population size, the larger number of new seedlings at CFI
compared to OHW resulted in slightly higher fecundity values, or the probability of creating a
successful offspring. These values are depicted in the first row of the transition matrices.
The stasis values, or the probability of remaining in the same size class over the year, are
presented along the diagonal. These values are similar for the two sites. Predicted growth values
(the value below the diagonal) were also similar at both sites. Because no trees were present in
the largest size class at OHW, there was no probability of advancement into the largest size class
at OHW, which does not represent a realistic scenario.
Figure 29 shows the results of the elasticity analysis of the CFI model. The elasticity
analysis was nearly identical for the OHW model. This analysis shows the chief importance of
the stasis values for the largest three size classes. Though there are different growth rates for the
two populations, the stasis values for the last size class were 0.99 for both models, suggesting that
99 of 100 trees in the largest size class are likely to survive a given year. This value was,
according to the sensitivity analysis, nearly five times as important as any other value in the
Ecosystem Development in Rushton and Reference Plots
Comparisons between pairs of one or more Rushton and a reference plot were made based
on the canopy cover, plot vegetation including trees, shrubs, and understory vegetation, and soil
percent organic matter. Samples from Rushton plots were only considered when basal area
density of Rushton trees was > 10 m2/hec in the sample area.
Selection of Plots for Comparison
Table 12 presents all the Rushton plots and subplots ordered by basal area (m2/110C) Of
Rushton trees. The plots/subplots considered in the comparative analysis with reference plots are
those listed above the dotted line. A distinction was drawn at a basal area of 10 m2/hec below
which survival in plots was so poor as to potentially nullify the effect of planted species on the
surrounding environment. This distinction was drawn based on an arbitrary but clear break in the
basal area in plots/subplots between the plot with a basal area of approximately 13 m2/hec and
the next lowest with a basal area of approximately 8 m2/hec. Five hydric swamp plots and 1
complete cypress-gum plot along with portions of five others were thus removed from
consideration in the following comparative analysis.
In addition to the Rushton plots removed from consideration, one subplot of the reference
plot at CFI was removed from consideration upon realizing that this segment had been subjected
to repeated disturbance from mowing and would not be representative of reference conditions.
Topographic Comparison of Rushton and Reference Plots
Table 13 shows a comparison of topography and water levels in Rushton plots and their
corresponding reference plots, which are the highlighted items appearing at the bottom of the
groups of Rushton plots. In most cases all reference plot variables including average change in
elevation, average water depth, minimum and maximum water depth fell within 3 standard
errors of the mean of the variable for the corresponding Rushton plots.
Plot Basal Area in Rushton and Reference Plots
Table 14 provides data on plot basal area from Rushton and reference plots. Plot basal area
includes the total basal area of all trees and shrubs. For all but TEN R2A and R2B, the plot basal
area (m2/hec) in reference plots was less than in Rushton plots. The mean plot basal area in
Rushton plots was up to 12 times greater than in corresponding reference plots. Typically the
difference in plot basal area between Rushton and reference plots grew as planted species made
up a larger portion of the plot basal area in a Rushton plot.
Percent Canopy Cover
Table 15 compares percent canopy cover determined from canopy photos in Rushton and
reference plots. In 7 of 10 pairs Rushton plots had greater canopy cover than corresponding
reference plots. In the remaining 3 pairs, reference plots' canopy cover were within 1% of
Rushton plots. Except at HOM, there was not a difference between the canopy cover in Rushton
and reference plots of more than 10%. Figure 30 demonstrates the trend in canopy cover as
subplot basal area increases at HOM, which is typical of other sites. As subplot basal area
increases, the canopy cover increases steeply and then levels out between 80 and 90%.
Soil Organic Matter
Table 16 provides a comparison of the percent soil organic matter found in samples of the
top 10cm of the soil in Rushton and reference plots. At CFI, HOM, and PRP, soil organic matter
was greater in Rushton plots, but in most pairings at the older sites of OHW and TEN, percent
soil organic matter was higher in reference plots. In all cases the differences between the
Rushton and reference plots as indicated by T-tests were significant at the 90% confidence level.
At HOM there was a very wide range of organic matter within the Rushton plots, not present at
the other sites.
Table 17 compares Rushton and reference plot percent organic matter by site. The
variation between reference plots on different sites is greater than the variation between Rushton
plots on different sites. Excluding HOM, the average %OM in Rushton sites varies between 9
Table 18 presents a comparison of the understory coverage in Rushton and reference plots.
Inconsistent differences occur between the Rushton and reference plots. Among the Rushton
plots, the highest cover occurs at CFI, where ferns were planted underneath the drier portions of
the plots. Understory coverage at OHW is consistent around 30% for Rushton plots, lower than
at other sites.
Table 19 summarizes species richness and evenness among pairs of Rushton and
references plots. No consistent signal of a difference in richness and evenness is apparent
between Rushton and reference plots. The average number of species occurring in Rushton plots
is never more than 13, whereas reference plots at CFI and TEN have as many as 21 and 20
species. Species evenness follows a similar trend to species richness when comparing within
Rushton and reference pairs.
The range of both richness and evenness is greater in the reference than in the Rushton
In order to determine the dominant species in the understory assemblage within each plot,
Importance Values were calculated for each species. Lists of the most prevalent species for each
plot determined by Importance Values can be found in Appendix Tables 21-30. Each table
includes a list of prevalent species for every plot in a comparison pair.
The ordination of species assemblages based on the average cover of species can be a
useful means of visualizing the similarity of assemblages in different plots. Figure 3 1 presents
the result of an Nonmetric Multidimensional Scaling (NMDS) of the most prevalent species in the
plots. The diagram shows a clear separation of sites and pairs. CFI reference plots are clustered
on the left side, with the drier plots R-6 and R-4 close together and R1, the wettest site, on the
other end. The CFI reference plots are closer to the HOM Rushton plot. All the HOM Rushton
plots (names starting with '2') are clustered among themselves and with the 3 PRP sites (names
starting with '6'). The HOM reference plot is isolated from the other groups. All the OHW
(names starting with '3','4', and '5') and TEN plots are clustered within their respective sites.
Overall there is a much greater difference in species assemblies between sites than within sites or
Relationship among measures of ecosystem development. Table 20 contains
correlations among selected ecosystem development variables by site. Rushton and reference
plots are combined in this analysis by site. Differences in the relationship strength and the
direction of the relationships between these variables occur between different sites.
Two hydrologic variables average depth and range of average depth are included in the
correlations, along with the total Rushton tree basal area (RushBA). The response variables
included are total basal area, canopy cover, understory cover, understory richness, understory
evenness, and soil percent organic matter. The relationship of the response variables to
RushBA is of primary interest, though the correlations between response variables are also
At all sites RushBA is strongly positively correlated with total basal area, as was apparent
in Table 14, which showed that Rushton trees made up the majority of total basal area in most
Rushton plots. However the correlation with canopy cover is less clear. At CFI correlation is
nearly absent, because all plots including Rushton and reference have very similar canopy
coverage (see Table 15). The trend is more positive at the sites where reference plots have less
canopy cover. The correlations between RushBA and understory cover are mostly negative,
except at CFI where understory planting occurred, though the relationship is weak at the older
sites of OHW and TEN. RushBA ranges from being strongly negatively correlated with
understory richness at PRP to strongly positively correlated at OHW. The correlations between
understory evenness and also range from strong negative to strong positive.
OWH and TEN show the same direction of correlation for all response variables. HOM
and PRP, the wettest sites, also show the same direction of correlation in all variables but species
Table 5. Tree survival from initial planting. in 25 sampled cypress-gum plots.
No. % Survival
Planted lyr 3yrs 20yrs
Fraxinus pennsylvanica 651 72% 70% 29%
Nyssa aquatica 837 44% 34% 18%
Taxodium distichum 837 66% 55% 34%
Acer rubrum 126 94% 6%
Fraxinus caroliniana 72 99% 82%
Taxodium distichum 216 89% 31%
Table 6. Tree survival from initial planting in 12 sampled hydric swamp plots.
12 3 4 5 6 7 8 9 10 11 12 131415 16 1718 19 20
0 E---- ....-
0 ** .......
12 3 4 5 6 7 8 9 10 11 12 131415 16 1718 19 20
4000' %-- .
1 2 3 4 5 6 7 8 9 10 11 12 131415 16 1718 19 20
700 o z
1000- .* -
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
-4- Fraxinus pennsylvanica -- Nyssa aquatica 5-- Taxodium distichum
Figure 9. Percentage of planted trees surviving by site and species in cypress-gum plots after 1
year (Rushton 1988), 3 years (Paulic and Rushton 1991a), and 20 years. The dashed
line represents a hypothetical trend in between the sampled years.
Figure 10. Percentage of planted trees surviving by site and species in hydric-swamp plots after 1
year (Rushton 1988) and 19 years. The dashed line represents a hypothetical trend in
between the sampled years.
12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
9 0 o
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
- -X- Acer rubrum -A- Fraxinus carohiniana -- Taxodium distichum
CFI SP-1. Homeland
a aR4 m 5R4
O R5 I Illllil l lililll R a
O R6 O 1111111111111111 R6
3 4 5 6 7 8 9 10 3; 4 6 7 8 9 10
OH WPright Teneoc 4
SR1A 5 R2A
R2A III R2B
0 O R2B 0 R5A
m m R5B
6 7 8 9 10 4 5 6 7 8 9 10
Figure 11. Percentage of plot inundated at time of monthly sampling during the period of record
on cypress-gum plots. The numbers on the x-axis represent month of the year (e.g. 3
w rim ii
6 7 X 9 10
Figure 12. Percentage of plot inundated at time of monthly sampling during the period of record
on hydric swamp plots. The numbers on the x-axis represent month of the year (e.g.
3 = March).
Figure 13. Distribution of average water depth inside original plot boundaries (Original) of
cypress-gum plots, and at the locations of surviving trees for each of the species
planted (Fraxinus pennsylvanica, Nyssa aquatica, and Taxodium distichum). The
distributions are presented as box plots that break the data into four quartiles. The
middle box represents the 25-75th percentiles, with includes the median value
represented by the middle line. The upper and lower hashes represent the 0 and 100
percentiles. The circles beyond the lower hash are outliers.
CFI SP-1 Fraznns pennsylvnica
-09 -07 -0. -03~ -01.1 0 01i 01 03 0.4 03d 06 0.7 0.8 09 1
CFI SP-1 Nyssa aquatica
I died btwn earIad 20
-09 -0.7 -0.5 -03 -01 0 01i 01 03 0.4 0.5 0.6 0.7 0.5 0)9 1
cH SP-1 Tazedium distichum
a died btn age Iand 20
a9 4.7 4.5 -03 -0.1 0 0.1 01 03 0.4 0.5 0.6 0.7; 0.8 09 i
Figure 14. Number of planted trees that died between years 1 and 20 and trees alive in 2005, in
0.1m depth classes on CFI (sand-clay) on plots R1-R6.
OH WrCright Fraznus penusylvanica
-09 -0.8 -0)T .7 -06 ) -0 4 -03 -0 2 -0.1 0) 0.1 01.2 0r3 0.4 0 5 0~ 60.7 0l.B 09 I
Figure 15. Number of planted trees that died between years 1 and 20 and trees still alive, in 0.1m
depth classes on OHW (clay) plots R2A and R2B.
OH Wright Taxdium disticham
W died bt~wnage Iand 2
-09 -0.8 -0.7 -0.6 -05 -0.4 -03 -0.2 -0.1 0 0.1 0.2 03 0.4 0.5 0 6 0.7 0.B 0.9 I
Teneme 4 Frazinus pennsylanica
-a I dial Itwnage I ad 20
-09 -0.7 -0. -03 -01.1 0 01i 01 03 0.4 03d 0.6 03 0 B 09 1
Tenerre 4 Nyssa sqluatica
Died btwn eagIad 20
0 0.1 0t 2 3 0.4 0).5 0.6 0).7 0). B 09 1
Tenerere 4 Tnaxeium d~s~tichumt
a died htn age Iand 20
a9 a.7 4.5 -03 -0.1 0 0.1 0 2 03 0.4 0.5 0.6 0.; 7 .8 09 i
Figure 16. Number of planted trees that died between years 1 and 20 and trees still alive, in 0.1m
depth classes on TEN (clay) plots R5A, R5B, R6A, R6B, R7A, and R7B.
-09 -0.7 -0.5 -03 -01
Original Aiker 20Years
After 20 Years
Figure 17. Distribution of average water depth inside plots boundaries (Original) of hydric
swamp plots and at locations of surviving trees. See Figure 3 for explanation of box
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1718 19 20
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Figure 18 Percentage of planted trees surviving in cypress-gum plots by soil type after
approximately 1 (Rushton 1988), 3 (Paulic and Rushton 1991a), and 20 years.
Figure 19. Distribution of average water depth in cypress-gum plots grouped by soil type.
7. Comparison of trees growing in different soil media by species, among those with
similar average water deph
No. of Trees Mean of
log(Basal Area) p-value
Spces Clay ISand-Clay Clay ISand-Cla
Fraxinus pennslyvanica 45 89 4.24 4.44 0.20
Nssa aquatica 13 37 4.61 3.96 0.02*
Taxodium distichum 82 67 5.23 5.44 0.63
*Significantly different at the 95% confidence level
Table 8. Results of a two-way ANOVA comparing the effect of two soil types (clay and sand-
clay) and two water levels (shallow and deep) on Fraxinus pennsylvanica growth.
Variable p -value
Soil Type 0.39
Water Level 0.52
Table 9. Results of a two-way ANOVA comparing the effect of two soil types (clay and sand-
clay) and two water levels (shallow and deep) on Taxodium distichum growth.
Variable p -value
Soil Type 0.40
Water Level 0.02*
*Significantly different at the 95% confidence level
Table 10. Site Disturbance Record
Site Plot(s) Fire Heavy Grazing Mechanical
PRP H1,H2, +
TEN 5A,5B- + +
+ Record of incidence
- No record of incidence
source for non-
other plots 1
other planting 3
other planting 6
other planting 8
other planting 13
other planting 15
other planting 16
other planting 25
other planting 29
Table 11. Plots with potential offspring of planted trees ordered by reproductive ratio
# Planted # Recruited
TEN H3 June
TEN H3 Nove~mber
0 20 40 0 80
Figure 20. Size class distributions of Taxodium distichum seedlings at Ten H3 counted in June
and November, 2005.
CHF SP-1 167 trees
Homelnd 173 rees
OH Wright 34C trees
Peace ark 35 trees
enerer 4 142 trees
Tenemw 4 36 trees
0 1 23 4 6
01 1 11 2 3 67
S1 2 34 5 6
Figure 21. Size class distribution of Taxodium distichum in 6 basins on five CSAs. Light
sections represent recruited trees: dark sections planted trees. The size classes
represent the following DBH ranges: 0:no DBH: 1: 0-5cm: 2: 5-10cm: 3: 10-15cm: 4:
15-20cm: 5: 20-30cm: 6: 30-40cm: 7: >40cm.
CH SP-1 70 trees
Homeland 70 tre
OH Wrright 2 tres
Tenerwe 4 4 trees
Tenesee 4~ 5 trees
Figure 22. Size class distributions ofNyssa aquatica in five basins on four CSAs. Light sections
represent recruited trees: dark sections planted trees. The size classes represent the
following DBH ranges: 0:no DBH: 1: 0-5cm: 2: 5-10cm: 3: 10-15cm: 4: 15-20cm: 5:
20-30cm: 6: 30-40cm: 7: >40cm.
CH SP-1 124 trees
Hesmeland -17 tees
OII WRright IStrees
0 41 2 34 56
TenrcF 4 15 trBRE
Teneswe 4 18 trees
Figure 23. Size class distribution of Fraxinus pennsylvanica in five basins on four CSAs. Light
sections represent non-planted trees: dark sections planted trees. The size classes
represent the following DBH ranges: 0:no DBH: 1: 0-5cm: 2: 5-10cm: 3: 10-15cm: 4:
15-20cm: 5: 20-30cm: 6: 30-40cm: 7: >40cm.
OH Wright 49 ~trees
peace Prk 21 ~trees
Figure 24. Size class distribution of Fraxinus caroliniana in two basins on two CSAs. The size
classes represent the following DBH ranges: 0:no DBH; 1: 0-5cm; 2: 5-10cm; 3: 10-
15cm; 4: 15-20cm; 5: 20-30cm; 6: 30-40cm; 7: >40cm.
1 2 3
4 5 6 7
0.023 0.034 0.051 0.076]
0. 0. 0. 0.]
0. 0. 0. 0.]
0. 0. 0. 0.]
0.725 0. 0. 0.]
0.244 0.902 0. 0.]
0. 0.078 0.946 0.]
0. 0. 0.04 0.991]
0. 0.139 0.731
0. 0. 0.223
0. 0. 0.
0. 0. 0.
0. 0. 0.
Figure 25. Transition matrix for CFI SP-1 Taxodium distichum population model
CFI SP-1 Taxodium dist. Population
Years frorn the Present
Figure 26. Model predicted population change of CFI SP-1 Taxodium distichum
CLASS 0 1 2 3 4 5 6
0[ 0.67 0.005 0.006 0.004 0.006 0.01 0.015
1[ 0.247 0.775 0. 0. 0. 0. 0.
2[ 0. 0.194 0.799 0. 0. 0. 0.
3[ 0. 0. 0.109 0.732 0. 0. 0.
4[ 0. 0. 0. 0.188 0.78 0. 0.
5[ 0. 0. 0. 0. 0.189 0.938 0.
6[ 0. 0. 0. 0. 0. 0.042 0.986
7[ 0. 0. 0. 0. 0. 0. 0.
Figure 27. Transition matrix for OH Wright Taxodium distichun2 population model
OH Wright Taxodium dist. Population
Years from the Present
Figure 28. Model predicted population change of OH Wright Taxodium distichun2
Model Elasticity Values
o.s~ m Fecundity
Figure 29. Model elasticity values showing sensitivity of different parameters. For each
parameter type (stasis, growth, fecundity) the first bar from the left represents size
class 0 with the bars to the right corresponding to size class 1,2,3... to 7.
Table 12. Rushton plots/subplots ranked by planted tree basal area (m /hec)
Rushton Tree Basal
Site Plot Subplot Type Area (mz hec)
CHI k1 1 CG 226
HOhi R2 2 CG 158
CFI R3 3 CG 154
HOhi R1 2 CG 142
CFI R3 2 CG 140
TEN H6 NA HS 128
CFI R1 3 CG 108
CFI R2 3 CG 107
CFI R2 1 CG 105
CFI R6 2 CG 96
CFI R6 1 CG 96
HOhi R1 3 CG 92
OHW R2A 1 CG 90
HOhi R4 2 CG 88
CFI R2 2 CG 87
CFI R6 3 CG 87
CFI R3 1 CG 86
CFI R5 1 CG 86
HOhi R4 1 CG 82
CFI R1 2 CG 82
CFI R5 2 CG 80
CFI R4 3 CG 79
HOhi R6 1 CG 77
HOhi R3 2 CG 76
OHW R2B 1 CG 75
CFI R5 3 CG 65
HOhi R1 1 CG 63
HOhi R5 1 CG 63
HOhi R7 3 CG 61
HOhi R7 1 CG 58
OHW R2A 2 CG 57
HOhi R6 2 CG 55
HOhi R2 3 CG 50
HOhi R7 2 CG 48
TEN H5 NA HS 42
CFI R4 2 CG 41
HOhi R6 3 CG 36
OHW R2B 2 CG 35
HOhi R3 3 CG 35
TEN R2B 1 CG 32
HOhi R5 3 CG 31
HOhi R5 2 CG 26
OHW H4 NA HS 24
OHW H1 NA HS 24
HOhi R4 3 CG 22
TEN H2 NA HS 21
PRP H5 NA HS 19
HOhi R2 1 CG 18
TEN R2B 2 CG 17
TEN R2B 3 CG 16
HOhi R3 1 CG 15
OHW R1A 1 CG 14
PRP H1 NA HS 14
TEN R2A 1 CG 13
TEN H3 NA HS 8
OHW H2 NA HS 7
CFI R4 1 CG 7
HOhi R8 2 CG 6
OHW H3 NA HS 4
TEN R2A 3 CG 4
PRP H6 NA HS 4
PRP H4 NA HS 3
OHW R1A 3 CG 2
OHW R1A 2 CG1
HOhi R8 3 CG 1
OHW R2A 3 CG 0
OHW R2B 3 CG 0
TEN R2A 2 CG 0
Table 13. Topography and water level" comparison of Rushton and reference plots
Pair Site Plot Plot Type
1CFI R1 CG
1CFI R2 CG
1 CFI R3 CG
1CFI R4 CG
1CFI R5 CG
1 CFI R6 CG
1 CFI 5 CG-Ref
2 HOM R1 CG
2 HOM R2 CG
2 HOM R3 CG
2 HOM R4 CG
2 HOM R5 CG
2 HOM R6 CG
2 HOM R7 CG
2 HOM Tl CG-Ref
3 OHW H1 HS
3 OHW H4 HS
3 OHW H1R HS-Ref
4 OHW R1A CG
4 OHW Tl CG-Ref
5 OHW R2A CG
5 OHW R2B CG
5 OHW T2 CG-Ref
6 PRP H1 HS
6 PRP H5 HS
6 PRP H1R HS-Ref
7 TEN H2 HS
7 TEN H2R HS-Ref
8 TEN H5 HS
8 TEN H5R HS-Ref
9 TEN H6 HS
9 TEN H6R HS-Ref
10 TEN R2A CG
10 TEN R2B CG
10 TEN Tl CG-Ref
aWater depth data from July for all plots
Avg A Eley.(m)
bore than 3 standard errors from the mean of Rushton plots
CLess than 3 standard errors from the mean of Rushton Plots
Table 14. Plot-scale basal area comparison in Rushton and corresponding reference plots.
Percent of BA from Deviation
Plots Rushton trees Mean BA(m2/hec)a BA(m2/hec)
Pair Site Rushton Ref Rushton Ref Rushton Ref Rushton
1 CFI R1 R2 R3 R4 R5 R6 T5 93 NA 107 25 26
2 HOM R1 R2 R3 R4 R5 R6 R7 Tl 97 NA 64 6 23
3 OHW H1 H4 HIR 82 NA 29 7 5
4 OHW R1A Tl 30 NA 48 16 NA
5 OHW R2A R2B T2 71 NA 90 7 13
6 PRP H1 H5 HIR 62 NA 27 11 4
7 TEN H2 H2R 47 NA 45 17 NA
8 TEN H5 HR 79 NA 53 20 NA
9 TEN H6 H6R 98 NA 131 11 NA
10 TEN R2A R2B Tl 51 NA 33 47 13
a Bolded numbers indicate a difference of more than 1 standard deviation
a Bolded numbers indicate a difference of more than 1 standard deviation
80. 00%* *
0 20 40 60 sil 100 120 140
*Rushton Plot o Reference Plot
Figure 30. Subplot basal area and percent canopy cover at HOM. The first 3 (from the left)
Rushton subplots had < 10 m2/hec Rushton tree basal area, but are included to help
illustrate a continuous trend.
Table 15. Percent canopy cover comparison in Rushton and corresponding reference plots
Plots Mean SD
R1 R2 R3 R4 R5 R6
R1 R2 R3 R4 R5 R6 R7
Table 16. Soil percent organic matter comparison in Rushton and corresponding reference plots.
Plots Samples Mean %OMa Deviation %OM P-value
Pair Site Rushton Ref Rushton Ref Rushton Ref Rushton Ref from T-test
1 CFI R1 R2 R3 R4 R5 R6 T5 153 18 9.06 7.47 3.39 1.96 0.01
2 HOM R1 R2 R3 R4 R5R6R7 Tl 57 27 8.78 4.7 11.24 3.06 0.01
3 OHW H1 H4 HIR 54 27 10.19 9.2 1.65 1.2 3.00E-03
4 OHW R1A Tl 9 18 10.31 13.4 3.59 5.11 0.08
5 OHW R2A R2B T2 36 18 10.8 13.14 3.82 3.92 0.04
6 PRP H1 H5 HIR 36 27 10.21 7.7 3 1.54 6.00E-05
7 TEN H2 H2R 27 27 9.26 11.58 1.25 2.49 1.00E-04
8 TEN H5 H5R 27 27 8.06 14.06 1.69 2.94 1.60E-11
9 TEN H6 H6R 27 27 11.48 14.84 3.2 2.68 1.00E-04
10 TEN R2A R2B Tl 36 18 8.1 10.28 3.1 4.29 0.07
a Bolded numbers indicate a difference of more than 1 standard deviation
Table 17. Soil percent organic matter summarized by site and plot type.
Site Rushton Ref
CFI 9.06 7.47
HOM 8.78 4.70
OHW 10.42 11.52
PRP 10.21 7.70
TEN 9.16 12.91
Table 18. Average percent understory cover comparison between Rushton and reference plots
Average Cover %a
Plots Samples Mean SD
Pair Site Rushton Ref Rushton Ref Rushton Ref Rush
1 CFI R1 R2 R3 R4 R5 R6 5 34 6 0.96 0.84 0. 18
2 HOM R1 R2 R3 R4 R5 R6 R7 T1 63 9 0.94 0.86 0. 17
3 OHW H11 H14 H1R 18 9 0.32 0.22 0.20
4 OHW R1A T1 3 6 0.35 0.25 NA
5 OHW R2A R2B T2 12 5 0.32 0.86 0. 12
6 PRP H11 H15 H1R 12 9 0.83 1.20 0.20
7 TEN H12 H2R 9 9 0.59 0.62 NA
8 TEN H15 H5R 9 9 0.68 0.58 NA
9 TEN H16 H6R 9 8 0.38 0. 15 NA
10 TEN R2A R2B T1 12 6 0.47 0.37 0.21
a Bolded numbers indicate a difference of more than 1 standard deviation
Table 19. Species richness and evenness comparison in Rushton and reference plots
Species Richnessa Species Evenness
Plots Mean SD Mean SD
Pair Site Rushton Ref Rushton Ref Rush Rushton Ref Rush
1CFI R1 R2 R3 R4 R5 R6 5 12 21 3.9 0.59 0.70 0.15
2 HOM R1 R2 R3 R4 R5 R6 R7 T1 12 12 3.4 0.51 0.86 0.12
3 OHW H11 H14 H1R 7 9 0.7 0.71 0.86 0.06
4 OHW R1A T1 9 8 NA 0.81 0.85 NA
5 OHW R2A R2B T2 12 7 2.1 0.84 0.45 0.00
6 PRP H11 H15 H1R 5 7 0.7 0.70 0.53 0.02
7 TEN H12 H2R 4 3 NA 0.51 0.27 NA
8 TEN H15 H5R 13 20 NA 0.80 0.77 NA
9 TEN H16 H6R 10 5 NA 0.78 0.77 NA
10 TEN R2A R2B T1 9 12 2.1 0.72 0.77 0.05
a Bolded numbers indicate a difference of more than 1 standard deviation
Figure 31. NMDS plot of understory species assemblages. Plot names are condensed to 'pair-plot name' with a '*' added to the reference
plots. The greater the distance between the plots, the less similar their species assemblages.
Table 20. Correlation matrices for ecosystem development variables by site. Correlations
between Rushton BA and response variables (last five) are highlighted in gray.
Evaluating the progress of a created wetland after a long period of time (20 years in this
study) is valuable for determining which species are appropriately adapted for site conditions and
what the role of these species may be in the development of an ecosystem.
Though tree survival was higher on the sand-clay mix soil than on pure clay or sand,
hydrology and site disturbances were more important factors than soil type in determination of
tree survival. Each wetland tree species survived in positions along a hydrologic gradient that fit
a species-specific tolerance range for inundation. This positional range was more apparent for a
species after twenty years than it was after 1 or 3 years. This information provides a good
indicator of long-term hydrology within these plots and would be valuable for future planting
efforts on these sites.
Tree growth among surviving individuals was just as high on pure clay soils as on the
sand-clay surface. By this measure, established trees were successful on clay. Nevertheless, the
sustainability of planted tree populations on CSAs is uncertain. In most cases offspring of the
planted trees were scarce after twenty years. Models showed that the size of tree populations on
two sites, each with few offspring, will not grow significantly or possibly decline after 50 years,
assuming high survival of current mature trees. The causes) of the low numbers of new
seedlings still needs to be clarified. The presence of a high number of seedlings on one clay site
proved, however, that the clay soils alone do not prohibit seedling establishment.
Plants plots are more structurally mature than non-planted area and this is promoting the
accumulation of soil organic matter, but the rate of accumulation does not always exceed
accumulation under other CSA communities. No strong relationship between planted plots and
understory vegetation has yet emerged on the selected CSAs. The assemblage of understory
vegetation appears to be more strongly determined by the site surroundings and the plot
hydrology. The influence of the trees may become stronger as the trees continue to mature. For
sites planted with trees, the intentional introduction of additional species in the understory could
provide the source for a more diverse community.
Tree Populations in Relation to Environmental Factors
Tree Survival By Site and Species
The presence of wetland trees planted 20 years ago on multiple CSAs is an indication that
conditions alone in the planted plots are within the range of conditions in which these wetland
trees have evolved to persist. Taxodium distichum, both Fraxinus caroliniana and pennsylvanica,
and Nyssa aquatica survived on all sites chosen for the study, though not in equal percentages.
Though the typical lifespan of trees of these species is much greater than 20 years, their growth
and healthy condition on some sites herald continued persistence. For the species that did not
survive at any of the sites, questions remain as to the site factors that they were unable to tolerate.
Mature trees of some of the other species planted by Rushton were present on one or more of the
study sites. Acer rubrum, which survived in small numbers at some sites, was dominant in the
understory under canopies with many mature individuals at OHW, and to a lesser extent at TEN.
This species also occurs in high densities in some areas of CFI. Ulmus americana has also been
recruited on some of these same sites, though to a lesser extent than Acer rubrum. Quercus
laurifolia is not uncommon at CFI and OHW. Isolated individuals of Persea palustris were
found outside the sampled area at OHW. The failure of these species to persist in these planted
plots does not preclude their capacity to survive on CSAs, but does indicate a relatively poorer
survival capacity in the conditions to which the plots were subjected.
Overall Fraxinus caroliniana had a very high survival rate, though it was planted in a
limited range of water depth and a smaller number of individuals were planted. Taxodium
distichum was planted more than any of the aforementioned four species and over a range of
water depths, most of which it tolerated. In terms of survival it was the most successful species
of the three in the cypress-gum plots.
One and three-year survival was a good predictor of 20-year survival for these four
species. Though annual survival rate generally improved as mortality was more common in the
first three years, the change in the percentage of each species surviving relative to the other
species was relatively consistent across sites. In other words, a similar survival trend was present
for these species, and the species with the highest survival after three years was most likely the
species with the highest survival after 20 years. This perhaps indicates a similar response to
environmental stresses among the species.
Though individuals become more resistant to environmental stress with age, assuming that
the same regime of environmental conditions persisted from years 3 to 20 as did from years 0 to
3, notwithstanding sporadic disturbances, trees likely succumbed to the same pressures during
Tree Survival and Hydrology
Time allows for a clear determination of a suitable landscape position of a wetland species
relative to its period of exposure to saturated conditions and the depth of inundation. At CFI,
water depth did not preclude 20-year survival among the trees living after year 1, however,
hydrological factors may have had an effect on likelihood of survival of Nyssa aquatica and
OHW was likely affected by a disturbance event that affected the drier end of plots R2A
and R2B (see Table 10), thus water was not likely the key factor in mortality of the trees in the
drier area. Fraxinus pennslyvanica did not tolerate the wetter locations of this transect, though it
appears to have tolerated the same average depth at CFI.
That Fraxinus pennsylvanica did not tolerate locations where the average depth was 0.2
and 0.3m at OHW ,though it did tolerate those depths at CFI, could be interpreted as a greater
tolerance for standing water in sand-clay than in clay. But there are likely differences in the
hydrologic regime between the two sites that could have effected tree survival. In both plots,
trees are growing on the fringe of a pond where surface water outfall occurs at a given depth.
Though data are not available to determine at what water level relative to the trees that surface
outflow occurs in the two basins, it is possible that water could be retained longer at the same
average water level depths at OHW, increasing the period of inundation. This case exemplifies
the difficulty of inferring hydrologic similarity from monthly measurements over a single
growing season. Inside the plots perhaps the location of surviving trees relative to one another is
a better indication of hydrology than monthly water level measurements, but was an assumption
that could not be made within this study.
The shallower water depth distribution of surviving trees in the TEN basin (Figure 16)
likely does not represent the average depth of water trees were exposed to before the basin was
ditched in 2001, when average seasonal depths for all trees were likely greater. Hydrologic
factors may have impacted mortality, as the less tolerant Fraxinus pennsylvanica did not survive
in the deeper part of the range where Taxodium distichum did, but the animal grazing (see Table
10) noted during the initial years of establishment was likely also major factor in the high tree
mortality in this basin.
The long-term change in the hydrology on CSAs due to the continuing settling of the
clays is a challenge to long-term wetland creation unique to CSAs. But this study only revealed
anecdotal indications of an effect of clay consolidation and resultant hydrologic alteration on
planted trees. At TEN, laterally branching roots of Taxodium distichum and Fraxinus caroliniana
with rigid epidermal cells not typically found above the ground surface were found in two basins.
Faint clay stains were present on these roots which were as much as 3.5 feet above the ground
surface. These root features are potentially signs of clay consolidation, but since the basin
hydrology was altered by ditching in 2001, they could also be remnants of a dramatic decrease in
Tree Growth Comparison Between Sand-Clay and Clay Sites
Comparison of the effects of soil medium on tree growth could not include sand-capped
sites because there was no control for the effect of water level on tree growth.
The data clearly indicate that trees survived in greater numbers after 20 yrs on the sand-
clay site than in the clay, despite similar survival after 1 year for Taxodium distichum and
Fraxinus pennsylvanica. For all three species planted in cypress-gum plots survival was better on
the sand-clay site. Nevertheless, the soil medium is not the most probable explanation for this
difference. Initial growth after one year was similar for cypress-gum plots on CFI and OHW, on
which no notable growth occurred. Twenty year survival on OHW R2A and R2B was effected
by the death of all trees in 1/3 of the plot. Because death occurred for all species and because the
area experienced similar water level conditions to part of CFI on which some all species survived,
it is probable that one or multiple disturbance events, likely fire, caused the mortality rather than
the water level or the clay soil The domination of that area now by a fire-adapted species,
Imperata cylindrica, and reported fires that consumed trees in nearby plots provide further
evidence of this mortality hypothesis on OHW.
On TEN, the poor initial survival of some of the trees in the basin used in the survival
figure (Figure 16) was reported to be partially due to heavy grazing. Grazing significantly
reduced initial tree growth, an important indicator of future survival, and thus likely was the
principle cause the high mortality in the proceeding years. However, water levels possibly
resulted in Fraxinus pennsylvanica death in the deeper areas and all species in the extreme dry
areas. Though Fraxinus pennsylvanica survived into a much deeper average water depth on other
sites, it is likely that this basin stored more water before the hydrology was altered in 2001 and
that these trees then earlier were subjected to more frequent inundation.
High survival percentages in other plots in clay where the same trees were planted, like
TEN H6, is further evidence that, given appropriate hydroperiods and freedom from devastating
disturbance, viability of Taxodium distichum and Fraxinus spp. species on clay is good.
Across the board, Nyssa aquatica had poor initial survival in the clay sites, as the clay
possibly impeded the establishment process. Yet once the species established, it grew as well or
better and survived at a similar rate on the clay.
The scarcity of recruited trees in the periphery of most plots made it impossible to make
broad inferences about the conditions appropriate for seedling establishment on CSAs. Lack of
data on seed production, germination success, and seedling survival did not enable a
determination of the causes of absence of recruits.
First-year Taxodium spp. seedlings cannot tolerate long periods under water (Wilhite and
Toliver 1990). On Ten H3, a particular abundance of new seedlings emerged in the spring and
early summer 2005, where in May water levels dropped below ground but remained close enough
to the surface to maintain saturated conditions appropriate for germination. However water levels
rose and likely remained high enough to completely inundate 72 of 85 of these seedlings. This
rise in water level is the most likely explanation for the high mortality among these first-year
seedlings. If those seedlings that were inundated are assumed to have died during the period, the
survival rate for the remaining seedlings up to 100 cm would be close to 90%. Though unique in
the density of seedlings in this study, this plot provides evidence that given the presence of viable
seed and appropriate water levels, Taxodium distichum can germinate and establish on a CSA,
and that water levels are of critical importance in the establishment process.
The source of seedlings present on some of the sites was impossible to establish when
other mature trees had been planted by other parties. At CFI, >500 Taxodium distichum seedlings
had been planted on the site since the Rushton planting. Mature trees not planted by Rushton are
present just off the deeper margin of the plots and in between plots in cases. It could not be
determined with certainty that the recruited trees found inside the plots were offspring of the
planted trees. At HOM Taxodium distichum trees had been planted in the same basin a few years
before the Rushton plantings. The recruited trees found at HOM were perhaps planted or
offspring of trees from the previous planting.
Seedling establishment of wetland tree species depends heavily on a gentle rise in the
relative topography of the landscape. In natural floodplain systems, the extent of the spread of
the population is determined by the extent of the flood zone. In some CSAs the flood zone is
restricted due to a steep elevation gradient, often a residual of the mine cut spoil pile pre-fill
topography. This topography may restrict the area favorable to wetland tree seed establishment,
which require fluctuating water level conditions for adequate but tolerable moisture.
The size class distributions show normal to left-skewed shape distributions for most sites.
A right skewed or inverse-J shape distribution is a sign of a growing population dominated by
smaller individuals (Manabe et al. 2000). Overall scarcity of new seedlings at the sites poses
challenges to future population success. In all species o planted trees monitored in the study, at
least some individuals had reached a maturity to produce seed based on what is reported for
individuals of those species (USDA 2004). Though there was no formal collection of seed
production data, there were records of seeds present on trees or floating in water for each of the
species present. If the trees continue to survive it would be natural that they would become more
fecund as they grow.
Though this study shows that failure of seedling establishment is not endemic of CSAs,
studies need to be conducted to show if establishment presents any particular challenges. Further
study into seedling establishment and growth could reveal any obstacles exist on CSAs related to
soil clay content or vegetation cover. But to be conclusive, any such study needs to take into
account all stages of seedling establishment: including seed production, dispersion, viability,
germination and initial survival and along a variety of environmental gradients typical of CSAs.
Tree Population Model
Because this was a young population there was not good data on survival of older trees.
Reclamation of phosphatic clay settling areas did not begin until the early 1980s and therefore
there is no reference for longevity of Taxodium distichum in these areas. To fill in the data gap,
survival probability of larger trees was assumed to continue to increase in larger size classes. The
estimated survival probability of the largest size class of Taxodium distichum in the models was
consistent with the survival probability of the largest size class in models of other woody species
(Zuidema and Zagt 2000). Since the mortality of the largest size classes was the most sensitive
parameter in the model, the confidence of the model could be improved by real data of large tree
The probabilities of growth, survival, and reproduction are affected by the hydrologic
conditions. Incorporating the effect of different hydrologic regimes in the transition probabilites
of multiple transition matrices is one technique for implicitly accounting for the effect of
hydrology on a wetland tree population (Lytle and Merritt 2004). For these models, a time series
of data and a hydrologic record would be necessary to build this model.
The small changes in population size predicted by the models for trees on CFI and OHW
are a consequence of both high survival probabilities of larger trees and low reproductive
probabilities of mature trees. These same trends would have likely been present in models of a
number of the other tree populations in this study, but such trends cannot yet be generalized for
Taxodium distichum or other tree populations on CSAs.
Characteristics of Successful Species on CSAs
A common trait among the tree species that survived on multiple sites after 20 years
(Fraxinus caroliniana, Fraxinus pennsylvanica, Nyssa aquatica, Taxodium distichum) is the
ability to tolerate anaerobic conditions for an extended period of time during the growing season.
The least tolerant, Fraxinus pennsylvanica can tolerate inundation for up to 40% of the growing
season (Fowells 1965). Each of these species has special adaptations that permit extended
survival in when the root zone is saturated, including adventitious rooting and buttressing.
Also common to these species is the ability to resprout from the root stock and to coppice
resproutt from a stump) following disturbance. For environments that may be frequently exposed
to disturbances, especially fire, resprout ability could be important for long term survival (Pausas
et al. 2004). Evidence of resprouting was present in each of the four species.
These four species naturally occur in riverine swamps (Myers and Ewel 1990). Fraxinus
caroliniana and Taxodium distichum are also naturally present in a number of other forested
wetland types, such as cypress stands and lake fringe swamps. Two of the species, Fraxinus
pennsylvanica and Nyssa aquatica do not natively occur in Polk County. The southern extent of
the range of these species is in the big bend region. Among natural forested wetlands in Florida,
these two species are typically restricted to riverine swamps
The similarity of the natural habitat of these species and the CSA environment may help to
further explain their success on CSAs. Characteristics of riverine swamps, a common habitat of
these species, include a short hydroperiod and mineral soils typically containing clays. Plots in
the study had a mix of hydroperiods during the 2005 growing season, but Fraxinus pennsylvanica
and Nyssa aquatica were more successful in plots that had a short to moderate hydroperiod.
Fraxinus caroliniana and Taxodium distichum naturally occur in areas with a range of
hydroperiod and on CSAs were successful in areas with longer hydroperiods. Clay, sand-clay,
and sand capped sites in this study all had a low organic matter content at the time of planting that
would fit a mineral soil characterization. Other species found surviving or volunteering in
transitional areas including Acer rubrum, Quercus laurifolia, and Ulmus americana are also
naturally found in riverine swamps. Two species that did not survive, Gordonia lasianthus and
Sabal palmetto, are more often found in ecosystems with sandier soils and less dramatic
fluctuation in inundation.
Species characteristics are important in determining capacity to survive in the new
anthropogenic environment of CSAs, and copying species assemblages that exist in natural
wetlands with similar characteristics is a potential method for finding appropriate species. Yet
because the CSA conditions are unique, there is no perfect correlate ecosystem from which to
select appropriate species. The species that were most successful after 20 years in these plots
were those that not only occurred naturally in riverine wetlands, but also those with the most
tolerance for anaerobic conditions and the ability to resprout. Species biological characteristics
and the similarity of its native habitat are more important to tree success in CSAs than native
range, confirming an earlier finding by Paulic and Rushton (1991b).
In a 2005 survey not included in this study of Homeland FM-07, another CSA where trees
were planted in 1988 (see Paulic and Rushton 1991b for details), a similar assemblage of
surviving species was found. A species not planted by Rushton Quercus lyrata and two of the
species found surviving in this study, Fraxinus pennsylvanica and Taxodium distichum, were the
only species found Quercus lyrata is another native of north Florida riverine swamps adapted to
The hydric swamp plots in less wet to more transitional conditions at PRP, OHW, and
TEN were mostly devoid of trees or any plot boundary markings. Fire was a likely cause of death
at PRP and OHW, whereas circumstances are unclear at TEN. On PRP the transitional areas are
dominated by Imperata cylindrica. On another site mentioned in the previous paragraph (FM-07)
no trace of plots set-up in transitional areas was available and these areas were also dominated by
Imperata cylindrica. At OHW, a mixed forested canopy is now present over transitional plots H2
and H3. Schinus terebinthifolius was dominant in a the remnants of two TEN transitional plots.
Drier areas are more susceptible to fire and post-fire colonization, and overall had poorer
survival after 20 years, leaving the long-term viability of transitional tree species on CSAs
Ecosystem Development in Rushton and Reference Plots
Plot Selection and Comparison
Though the study intended to examine whether the surviving Rushton trees have played a
role in ecosystem development, there was no clear presumption of the quantity of trees, tree
biomass, or tree cover necessary to reveal an effect. It was not the purpose of this study to find a
minimum level of some quantitative measure of the trees at which an effect could be detected, but
at whether or not an effect on ecosystem development could be detected under an condition in
which trees were present. Because the measurements of ecosystem development had different
degrees of spatial precision, it was safer to assume common influence on a plot or subplot when
survival of trees was higher and thus spatially more homogenous.
Adequate descriptions of the vegetation composition in reference plots and Rushton plots
at the time of planting ( 1985- 1986) were not available to determine if the composition was
identical. By selecting areas adjacent to the same water feature with similar hydrology it was
assumed that: (1) the vegetation in the areas at the time of planting was similar; (2) the depth and
duration of flooding for the Rushton and reference plots was similar; and (3) no significant
disturbances that would radically alter the vegetation and/or soil affected the plots unevenly since
the time of planting. The comparison of Rushton and reference plots rests on these assumptions,
and plots or subplots were eliminated from the comparison if they violated one of these
Hydrology is perhaps the primary driver of wetland ecosystem development (Mitsch and
Gosselink 1993). Thus the most important criterion for selection of a reference plot within the
site was its hydrology. Though it was impossible to establish a reference plot in the same water
feature at Homeland, the reference plot was within 100m of the closest Rushton plot and had a
similar minimum, maximum, and average depth, and average change in elevation.
In some cases, there was considerable variation of water depth and percent inundation
within a group of Rushton plots. Mean water depths at CFI ranged from 0.36 m at R1 to -0.36
meters at R6. The mean water depth of the reference plot was appropriately exactly in the middle
at 0.0m, but the difference in depth and percent inundation within Rushton plots was large
enough to lead to detectable differences in ecosystem development parameters among the
Rushton plots. The R1,R2, R3, and R5 understories were dominated by floating aquatic
vegetation, whereas R4 and R6 were dominated by ferns. Yet the understories in these plots were
still more similar to one another than the reference plot (see Figure 31). There was also a -0.70
correlation (Table 20) between the between water depth and organic matter on CFI, indicating a
difference in soil OM within Rushton plots. Hydrologic variation within Rushton plots made
delineation of differences from reference plots more difficult.
The clearest distinction between Rushton and reference plots was present in the tree and
shrub strata. In planted areas with moderate to high survival of planted trees there was
significantly more structure at these levels in the plots. Rushton plots had in 9 of 10 cases a more
developed shrub and canopy layer. In plots on CFI, in TEN R2A and R2B, and in TEN H6, plot
basal area was more than twice as high as what has been found in natural forested wetland
systems, including mixed hardwood forest and cypress domes, but this difference is confined to
the narrow boundaries of the Rushton plots.
However, the estimates of canopy cover interpreted from the canopy photos showed little
difference between Rushton and reference plots. A possible explanation is the trend that occurs
with the estimation of canopy cover as plot basal area increases (Figure 30). Estimated canopy
cover increases very rapidly and then levels off as basal area continues to increase. Generally the
Rushton plots had enough structure so that all were near that asymptotic 'level' of canopy cover.
The canopy photo technique was used to estimate the proportion of light blocked by the
tree and shrub layers from reaching the understory. Because of the proximity of the shrub level to
the camera lens, also true of the understory, the shrub layer potentially had a more significant
effect on this estimation. The technique does not estimate layering in the canopy, nor the opacity
differences in different vegetative structures. Because there is more opaque, woody structure in
Rushton plots and likely more frequent overlap of structure in different strata, the differences in
the light reaching the understory could be greater than estimated in Rushton and reference plots.
Soil Organic Matter
Woody vegetation is an important contributor of litter that becomes incorporated into soil
organic matter. At TEN higher percent soil OM was found in reference plots dominated by Salix
caroliniana than in corresponding Rushton plots (pairs 7,8,9), though this was not the case at
PRP or CFI, where Salix caroliniana dominated reference plots. At OHW two reference plots
dominated Ludwigia peruviana had higher organic matter than corresponding Rushton plots.
Both of these species are characteristic of wetlands on CSAs, and may result in faster organic
matter buildup than planted species, but this trend is not consistent across all sites. Other factors,
such as fire frequency, also were important. At Peace Park, frequent fire and high tree mortality
likely caused high deposition of woody particulate matter in Rushton plots that led to high soil
organic matter. The presence of floating woody debris and burn scars on dead stumps was
qualitative evidence of this effect. Surprisingly, correlation of water depth with soil OM was
negative at most sites (see Table 20), which contradicts what is commonly found in wetland
systems, where sediment deposition is higher in lower areas (Hupp and Bazemore 1993). This
could be due to lack of vegetative colonization of deeper areas.
In wetland systems wood biomass and soil organic matter often represent the largest
storage of organic matter (Megongial and Day 1988). In Rushton plots a larger amount of total
basal area and smaller amount of a soil organic matter relative to Rushton sites indicates that
relatively more organic matter is bound up in living biomass in Rushton sites. A high percentage
of the organic matter pool tied up by living organisms has been proposed as an indicator of a
more mature ecosystem (Odum 1969). In a transition period the net production of organic matter
theoretically peaks and declines as biomass continues to increase (Figure 32). Though gross
production is likely still increasing in these systems as indicated by continual tree growth and a
greater total basal area in older sites, a greater proportion of the organic matter is being tied up in
woody biomass and less deposition to the soil is occurring.
For most plots, the coverage of plants in the understory, the species richness, and the
species evenness was similar among Rushton and reference plots. The similarity among Rushton
and corresponding reference plots was made apparent by the NMDS (Figure 31). A distinct site-
based grouping of understory assemblages emerged in this plot. CFI (without the reference plot),
OWH, and TEN are clustered by themselves, and PRP and HOM overlap. This finding
demonstrates the importance of site surroundings on understory composition. The dispersal of
propagules from outside is the only plant source in CSAs, as there is no seed bank in the clay
from which plants can emerge. Seeds must be carried in by wind or animals, and this process is
limited by the distance to the nearest seed source. Interestingly, the HOM and PRP sites, which
overlap on the NMDS, are within a mile from one another and likely share the same source (the
Peace River floodplain) ofpropagules.
Alternatively, propagules of wetland species other than trees could be brought in during the
reclamation process. This was done at CFI, where Nep~hrolep~is spp. were planted under the
canopy of Rushton trees.
There was some similarity in the understory across sites based on plot hydrology. Floating
aquatics, primarily duckweed (Lemna minor and Spirodella polyrhiza) and Salvinia minima, were
often the most prevalent vegetation on wetter transects. Where they occurred they often
accounted for the majority of cover. Though these species have limited to medium shade
tolerance, they were present in Rushton and reference plots, without a clear trend in a relationship
between basal area or canopy cover in their occurrence, except in pair 5 at OHW and pair 6 at
By and large the species found in the understory of plots have autecological characteristics
associated with plants present in early to middle succession. These characteristics include a rapid
growth rate, short lifespan, poor shade tolerance, high seed abundance, ability to spread
vegetatively, and seed dispersal via wind and or water (Odum 1969, Ricklefs 1990, Mitsch and
Gosselink 1993). Table 31 in the Appendix presents the prevalent understory species with scores
for each of species for all six autecological traits. Plant autecological characteristics can be
related to the stage of succession (Van der Valk 1981). If Rushton trees were helping to
accelerate succession on these areas, understory vegetation in Rushton plot would possess
characteristics typical of later succession. This could be occurring, but the differences between
the species assemblages in Rushton and reference plots were too small to test for differences in
the autecological characteristics of species.
An exception to the trend of similarity among species assemblages may exist at HOM,
where the species present in the reference plot were more typical of a freshwater marsh than a
shrub or tree-dominated system. The Rushton trees planted at HOM may be directing succession
toward a forested wetland whereas it otherwise might be developing into a marsh.
Relationships Among Measures of Ecosystem Development
The correlation matrices presented by site show some across-site similarity in relationship
between causal and response variables for OHW and TEN, and also for PRP and HOM.
Generally weak correlations are present between Rushton BA and the response variables This
could be because of they are older sites abutted on one side by a source of propagules, and
because the ecosystems reference plots are more developed on these sites, dampening the effect
of planted trees. Still there are large differences in total basal area and thus more organic material
stored in the living biomass in the Rushton plots on these sites, so differences do exist.
On both PRP and HOM, the Rushton plots stand out more in their structural differences
with reference plots than at other sites. These structural differences appear to have a strong effect
upon the understory vegetation, and clearly contribute to increased organic matter buildup.
Planted trees may have more detectable influence on ecosystems development on less vegetated
20 40 60 80
Figure 32. Succession in a forested system. From Odum (1969). PG=gross production: PN=net
production: R=respiration: B-total biomass.
SUPPLEMENTAL FIGURES, TABLES, AND CODE
~~ ~Lb i\4~0 i\~~
6~ b~LI ~ ~9 y~ ~h ~1~30 4~ 9'n~ ~\0~
Figure 33. 2005 water depth in a well at CFI measured by continuous data logger. Sampling times are marked with diamonds. The average of the
monthly sampled water levels was 0.66 meters, and the average of the hourly sampled water levels was 0.65 meters. The close
proximity of the monthly and hourly sampled water levels (within 1cm) indicates that the monthly sampled water level provided an
accurate average water level for the time period.