The Florida Wetland Condition Index (FWCI): Preliminary Development of Biological Indicators for Forested Strand and Flo...

MISSING IMAGE

Material Information

Title:
The Florida Wetland Condition Index (FWCI): Preliminary Development of Biological Indicators for Forested Strand and Floodplain Wetlands
Physical Description:
Report - Pilot Study
Language:
English
Creator:
Reiss, Kelly Chinners
Brown, Mark T.
Publisher:
Center for Wetlands
Publication Date:

Subjects

Subjects / Keywords:
floodplain wetlands
forested wetlands
Florida Wetland Condition Index (FWCI)
bioassessment
biological indicators
swamps
Spatial Coverage:
United States -- Florida

Notes

General Note:
94 Pages

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
System ID:
AA00004283:00001


This item is only available as the following downloads:


Full Text











Pilot Study -
The Florida Wetland Condition Index (FWCI): Preliminary Development
of Biological Indicators for Forested Strand and Floodplain Wetlands







Report Submitted to the
Florida Department of Environmental Protection
Under Contract #WM-683











Kelly Chinners Reiss
and
Mark T. Brown






Howard T. Odum Center for Wetlands
University of Florida
Gainesville, Florida 32611-6350


June 2005









ACKNOWLEDGMENTS


Research on biological indicators was supported by a grant to Mark T. Brown,
principle investigator, from the Florida Department of Environmental Protection (FDEP).
FDEP staff provided support for this research, particularly Russ Frydenborg, Ellen
McCarron, Ashley O'Neal, Erica Hernandez, Julie Espy, Tom Frick, Joy Jackson, Liz
Miller, Johnny Richardson, and Lori Wolfe. FDEP staff served as reviewers for the draft
report, including Russ Frydenborg, Connie Bersock, Nia Wellendorf, Julie Espy, and Joy
Jackson.
Additionally, acknowledgement is due to the systems ecology research group at
the Howard T. Odum Center for Wetlands. In particular, Chuck Lane (who developed
biological indicators for Florida marshes) provided a framework for this analysis.
Assistance in field-data collection, laboratory analysis, data entry, and/or feedback on
statistical analyses from Eliana Bardi, Matt Cohen, Tony Davanzo, Melissa Friedman,
Kristina Jackson, Joanna Reilly-Brown, Vanessa Rumancik, Kris Sullivan, Jim Surdick,
and Casey Chinners Virata, was particularly valuable.
Acknowledgement is due to the Florida botanists who participated in the Floristic
Quality Assessment index surveys from 1999-2004, including: Guy Anglin, Anthony
Arcuri, Dan Austin, Keith Bradley, Kathy Burks, David Hall, Ashley O'Neal, Jim
Poppleton, Nina Raymond, Bruce Tatje, John Tobe, and Wendy Zomlefer.

























This project and the preparation of this report were funded in part by a Section 319
Nonpoint Source Management grant from the U.S. Environmental Protection Agency
through a contract with the Florida Department of Environmental Protection.












TABLE OF CONTENTS

Page

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

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

L IST O F F IG U R E S .... ...... ................................................ .. .. ..... .............. vii

E X E C U TIV E SU M M A R Y ........................................ ................................................. ix

CHAPTER

1 INTRODUCTION AND OVERVIEW ..................................... .................

Forested Strand and Floodplain Wetlands.........................................................1
H historical Perspective ...................................................................................... 5
Biological Indicators of Ecosystem Integrity ........................................ ...............6
Quantifying Anthropogenic Influence ........................................................... 8
Landscape Development Intensity Index................. ............................8
H um an D disturbance G radient ........................................... .....................9
Project O overview ....................................... ................ .. ............ 11

2 M E T H O D S ................................................................13

S tu dy A rea ....................................................... 13
Site Selection ........................................14
Gradients of Landscape Development Intensity ..................................................17
F ield-data C collection .......... ........................................................... .. .... .. .... .. 2 1
Transect Sam pling D esign ...................................................................... 21
Floristic Quality A ssessm ent .......................................... ............... 24
D ata A naly sis .......................................................................27
Sum m ary Statistics...................................................................................27
Regional Com positional Analysis................................... ............... 28
Com m unity Com position....................................... ......... ............... 29
M etric D ev elop m ent ................................................................... .....................3 0
Florida W etland Condition Index ........................................ ....... ............... 31

3 R E SU L T S ........................................... ........................... 33

Gradients of Anthropogenic Activity............... ..........................................33
Landscape Development Intensity ............................ .................... 33
H um an D disturbance G radient ........................................ ....................37









W ater Q u ality ................................................................ ............... 40
D ata A n a ly sis ................................................................................................... 4 0
Sum m ary Statistics ..................................................... ........ .... .......... 42
Regional Com positional Analysis................................... ............... 42
Com m unity Com position....................................... ......... ............... 44
M etric Selection ................................................. .. .. .. .. ........ .... 46
Tolerance m etrics ............. .. .................. ..... .... .. .......... 46
Floristic Quality Assessment Index metric............................... 49
Exotic species m etric .......................................... ............... 52
N ative perennial species m etric ..................................... .......... 52
Florida W etland Condition Index ................................... ............... ..53
C lu ster A naly sis ............... ..... ............................. .... ............. ... ........ 56
Landscape Development Intensity Index and the Florida Wetland
C condition Index .......................................................................... 58
Human Disturbance Gradient and the Stream Condition Index ..........................61

4 D ISCU SSION .................. ...................................................... 64

D escribing B biological Integrity................................................... ........ ....... 65
Richness, Evenness, and D diversity ............................................. ............... 65
Measuring Anthropogenic Activity ....................................... ............... 67
Regionalization of the Florida Wetland Condition Index..................... ......... 69
Florida Wetland Condition Index Independent of Wetland Type .........................69
Limitations and Further Research................................ ............................ 71
C on clu sion s ............. ................... ........... .... ......... ....... ............... 7 1

APPENDIX

A Standard O operating Procedures......................................... ......................... 72
B Coefficient of Conservatism Scores.................................. ....................... 80
C Sum m ary Statistics ......................................................................... ..... 87
D M etric Scoring Criteria .......................................................... ............... 88

REFERENCES ......................... ......... ......................... 89












LIST OF TABLES


Table Page

1-1 Non-renewable energy use and LDI coefficients per land use used in the
calculation of the LDI index ................................................ 10
1-2 Categorical scoring criteria used to calculate the Human Disturbance
G radient (H D G ) ........................... .................................... ................. 12
2-1 Site characterization for 24 freshwater forested wetlands .............................. 16
3-1 Landscape Development Intensity (LDI) index scores for 10 forested
strands using 1995 and 2000 land use coverages.......................................34
3-2 Landscape Development Intensity (LDI) index scores for 14 forested
floodplain wetlands using 1995 and 2000 land use coverages ...........................35
3-3 Human Disturbance Gradient (HDG) for 13 storet stations, which
correspond with the forested floodplains sampled.....................................39
3-4 Water quality (chemical and physical parameters) for 13 storet stations,
which correspond with the forested floodplains sampled.............................. 41
3-5 Richness, evenness, and diversity of the macrophyte assemblage among
a priori land use categories......................................................... ............... 43
3-6 Richness, evenness, and diversity of the macrophyte assemblage between
low (LDI < 2.0) and high (LDI > 2.0) LDI groups........................ ...............43
3-7 Macrophyte community composition similarity among Florida wetland
regions (Lane 2000) and bioregions (Griffith et al. 1994) with MRPP ................44
3-8 Spearman's correlation coefficients for the macrophyte metrics and
FW CI w ith LD I F/w o 200m ........................................ .......................... 46
3-9 Comparisons among five macrophyte metrics and the FWCI between
low (LDI < 2.0) and high (LDI > 2.0) LDI groups (LDI F/wo 200m) ...............47
3-10 Tolerant indicator species for forested strand and floodplain wetlands ................48
3-11 Sensitive indicator species for forested strand and floodplain wetlands ...............50
3-12 Exotic species identified at 24 forested strand and floodplain wetlands ..............54
3-13 FWCI scores and LDI values for wetland clusters based on macrophyte
community composition............................. ............................... 59
3-14 Correlations of metrics and FWCI scores with 20 variations of the LDI index ....60









3-15 Forested Wetland Condition Index (FWCI), Landscape Development
Intensity Index (LDI F/wo_200m), Human Disturbance Gradient
(HDG), and Stream Condition Index (SCI) data available for 13 forested
floodplain w wetlands .................. ............................. ........ .. ............ 62
3-16 Correlations among four measures of ecosystem condition or
anthropogenic activity, including the Human Disturbance Gradient
(HDG), Stream Condition Index (SCI), Landscape Development Intensity
Index (LDI F/wo_200m), and the Florida Wetland Condition Index (FWCI) .....63
4-1 The five metrics of the preliminary Florida Wetland Condition Index for
freshwater forested strand and floodplain wetlands based on the macrophyte
species assem blage........ ............................................................ .... .... ..... .. 65












LIST OF FIGURES


Figure Page

1-1 Photograph showing the interior of a freshwater forested strand wetland
in O sceola C county, F lorida ........................................................... .....................3
1-2 Photograph showing the interior of a freshwater forested floodplain
w etland in Polk C county, Florida ................................................................... ..... 4
2-1 Florida wetland regions defined by climatic and physical variables (solid line;
Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994) ..................14
2-2 Study site location of 24 forested wetlands in Florida ......................................15
2-3 Boundaries of the 100 and 200 m buffers drawn around the downstream
transect for the LDI_T calculation at Forested Floodplain 3 (FF3).....................18
2-4 Boundaries of 100 and 200 m buffers around a wetland feature and
designated land use ............... .................................. .. .. .. .. ...... ...... 19
2-5 Boundary of the watershed buffer around Forested Floodplain 3 (FF3)
for the LDI W S calculation....................... ..... ............................. 20
2-6 Idealized transect layout for forested strand wetlands...........................................22
2-7 Idealized transect layout for forested floodplain wetlands ...................................25
3-1 Comparison between the LDI calculations including (1995_LDI_T/w_100m)
and excluding (1995 LDI_T/wo_100m) wetland area .................................36
3-2 Comparison between LDI calculations at the transect
(1995_LDI_T/wo_200m) and feature (1995 LDIF/wo_200m) scale..................36
3-3 Comparison between LDI calculations at the watershed scale using equal
weighting for all area within a watershed (1995 LDIWS_ED/wo) and a
distance weighted approach using linear weighting (1995 LDI WS DWlin) ....38
3-4 Comparison between 1995 and 2000 LDI calculations at the feature scale
(without the wetland area) for 13 forested wetlands............... ...............38
3-5 NMDS ordination bi-plot of 24 sample wetlands in macrophyte species space....45
3-6 The proportion of tolerant indicator species at wetlands increased with
increasing develop ent intensity..................................... .......................... 49
3-7 The proportion sensitive indicator species at wetlands decreased with
increasing develop ent intensity..................................... .......................... 51
3-8 FQAI scores decreased with increasing landscape development intensity............51
3-9 The proportion of exotic species at a wetland increased with increasing
develop ent intensity ...................... .. .................... .................. ...........52









3-10 The proportion of native perennial species decreased with increasing
develop ent intensity (LD I)........................................................ ............... 55
3-11 Forested Wetland Condition Index (FWCI) scores decreased with
increasing develop ent intensity (LD I) ..................................... ............... ..55
3-12 Change in average species p-value from the randomized Monte Carlo tests
at each step in clustering .................. .................. .. ......... .......... ... 57
3-13 Change in the number of significant indicator species from the indicator
species analysis performed at each step in clustering ............................................57
3-14 FWCI scores for three wetland clusters based on macrophyte community
co m p o sitio n ............................. .................................................... ............... 5 8









EXECUTIVE SUMMARY


PILOT STUDY
THE FLORIDA WETLAND CONDITION INDEX (FWCI):
PRELIMINARY DEVELOPMENT OF BIOLOGICAL INDICATORS
FOR FORESTED STRAND AND FLOODPLAIN WETLANDS

Over 30 years ago, the federal Water Pollution and Control Act obliged states to
protect and restore the chemical, physical, and biological integrity of waters, and charged
states with establishing water quality standards for all waters within state boundaries
including wetlands. Criteria for defining water quality could be narrative or numeric, and
it could be addressed through chemical, physical, or biological standards. Initially, states
used chemical and physical criteria (testing waters for chemical concentrations or
physical conditions that exceeded criteria), assuming losses in ecosystem integrity if the
criteria were exceeded (Danielson 1998). The United States Environmental Protection
Agency (USEPA) recognized the potential of biological criteria to assess water quality
standards and in the late 1980s required states to use biological indicators to accomplish
the goals of the Clean Water Act (USEPA 1990). In effect, biological assessment has
evolved into one of the standard monitoring tools of water resource-protection agencies
over the past 2 decades (Gerristen et al. 2000). Such biological assessment programs
have been created for lakes and streams throughout the United States (Barbour et al.
1996a; Karr and Chu 1999; Gerristen et al. 2000), and more recently efforts to assess
wetland condition have been initiated (Mack 2001; USEPA 2002). Within Florida,
biological indices have been created based on macroinvertebrate community composition
for streams (Barbour et al. 1996a; Fore 2004), lakes (Gerristen and White 1997), and
isolated depressional freshwater herbaceous (Lane et al. 2003) and forested (Reiss and
Brown 2005) wetlands. Biological indices have also been created based on the
community composition of the diatom and macrophyte assemblages for Florida
freshwater herbaceous and forested wetlands (Lane et al. 2003; Reiss and Brown 2005).
The primary objective of this research was to develop a preliminary Florida
Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland
study sites were sought in various a priori designated land use categories that included
natural, agricultural, and urban land uses. An independent measure of anthropogenic
activity in the landscape was calculated for each wetland using the Landscape
Development Intensity index (LDI) (Brown and Vivas 2005). The contribution of this
research to our understanding of changes in the macrophyte community composition of
forested strand and floodplain wetlands in relation to different anthropogenic activities in
the surrounding landscape can be summarized in five main points:

1. Five macrophyte based metrics including proportion tolerant indicator species,
proportion sensitive indicator species, Floristic Quality Assessment Index
(FQAI) score, proportion exotic species, and proportion native perennial
species, were useful biological indicators for defining biological integrity for
forested strand and floodplain wetland vegetation;









2. Vegetation richness, evenness, and diversity were not sensitive to a priori
land use categories or development intensities in the surrounding landscape
for forested strand and floodplain wetlands;
3. The Landscape Development Intensity (LDI) index was a useful tool
correlating with the measured biological condition of vegetation for forested
strand and floodplain wetlands;
4. Regional species lists for metrics would enhance the forested strand and
floodplain Florida Wetland Condition Index (FWCI);
5. An FWCI with a set of core metrics could be developed for Florida freshwater
wetlands, which includes separate species lists for indicator species by
wetland type and ecoregions and separate Floristic Quality Assessment Index
(FQAI) scores for species by wetland type.

The FWCI provided a quantitative measure of the biological integrity of forested
strand and floodplain wetlands in Florida. Comprised of five metrics, the FWCI was
developed based on the community composition of the macrophyte species assemblages.
Metrics were selected for inclusion in the FWCI based on the correlation (nonparametric
Spearmans correlation coefficient) of each metric with a quantitative gradient of
Landscape Development Intensity (LDI); based on a metrics visually distinguishable
correlation with LDI in a scatter plot; and based on a statistical difference of metric
values between low and high LDI groups (Mann-Whitney U-test). The FWCI was
composed of individual metrics, which were scaled and added together, creating the
preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest
score of 50 reflecting the highest biological integrity and the lowest score of zero
reflecting a lack of biological integrity or no similarity to the reference wetland condition.
The five macrophyte metrics that met the three selection criteria (Spearmans
correlation coefficient (Irl > 0.50, p < 0.05), visually distinguishable scatter plots, and
Mann-Whitney U-test between LDI groups (p<0.10)) were proportion tolerant indicator
species; proportion sensitive indicator species; Floristic Quality Assessment Index
(FQAI) score; proportion exotic species; and proportion native perennial species.

Forested Strand and Floodplain FWCI Metrics for Wetland Vegetation
1. Proportion Tolerant Indicator Species
2. Proportion Sensitive Indicator Species
3. Floristic Quality Assessment Index (FQAI) Score
4. Proportion Exotic Species
5. Proportion Native Perennial Species

The variable sensitivities of three different independently derived indices
compared to the forested strand and floodplain FWCI, including the Landscape
Development Intensity index (LDI; Lane et al. 2003; Brown and Vivas 2005), the Human
Disturbance Gradient (HDG; Fore 2004), and the Stream Condition Index (SCI; Fore
2004), suggest that multiple measures of biological integrity may be more effective at
describing ecosystem wide biological integrity than any single measure based on an
individual species assemblage or surrounding land use activity. However, the strong









correlations among FWCI, LDI, and HDG (Spearmans correlation coefficient Irl > 0.58, p
< 0.05), and lack of correlations of SCI with both FWCI and LDI, suggest that in-stream
macroinvertebrate based measures of biological condition and surrounding forested
wetland macrophyte based measures of biological condition did not respond in a
consistent manner to changes in anthropogenic activity. Using both the in-stream
macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte
FWCI biological assessment methods may provide a more complete picture of the overall
condition of a wetland and associated stream at a particular spatial location. While
agreement in the ranking of the biological condition of study wetlands using the FWCI
and SCI was anticipated, discrepancies among the ranking from the different assemblages
may provide great insight into biological condition as different species assemblages
respond to changes in anthropogenic activities and the associated changes in inflows (e.g.
nutrient enrichment) over different time scales. Additionally, use of the forested strand
and floodplain FWCI may lead to specific conclusions as to the biological condition of
local or nearby anthropogenic activity, while use of the SCI may enhance understanding
of larger watershed scale influences from anthropogenic activity (i.e. due to the
convergence of water within the watershed associated with stream flow).
The quantitative score of biological integrity established through the FWCI can be
used as an objective, quantitative means of comparing changes in macrophyte community
composition for wetlands, including those impacted by varying degrees of anthropogenic
influence. While the forested strand and floodplain FWCI for flowing water systems can
not be used to predict changes in the physical and chemical parameters of a wetland, its
strength lies in providing an overview of biological integrity through the integration of
changes in macrophyte community composition from cumulative effects.











CHAPTER 1
INTRODUCTION AND OVERVIEW

Assessment techniques for categorizing ecosystem condition have been
established for many Florida ecosystems, including freshwater lakes (Lake Condition
Index, LCI: Gerristen and White 1997), streams (BioRecon and Stream Condition Index,
SCI: Fore 2004), and depressional freshwater wetlands (Florida Wetland Condition
Index, Lane et al. 2003; Reiss and Brown 2005). This research furthers the development
of the Florida Wetland Condition Index (FWCI), establishing preliminary metrics with
species lists specific to forested strand and floodplain wetlands, referred to as flowing
water systems.
The overall goal of the FWCI is to use changes in community composition to
characterize the biological condition of wetland ecosystems. Wetland condition is scored
on a numeric scale developed from the addition of scores from individual metrics. A
metric is defined as a biological attribute having a consistent and predictable response to
anthropogenic activities (Karr and Chu 1997). Each metric represents an indication of
biological integrity, or a signal of ecosystem condition, based on a change in community
composition from the reference standard condition. The reference standard condition is
defined as the condition of wetlands surrounded by undeveloped landscapes and without
apparent human induced alterations. By designating a measure of ecosystem condition
we refer to what others have described as ecosystem integrity, defined by Karr and
Dudley (1981) as "the ability of an aquatic ecosystem to support and maintain a balanced,
integrated, adaptive community of organisms having a species composition, diversity,
and functional organization comparable to that of the natural habitats of the region."

Forested Strand and Floodplain Wetlands

Wetlands have been categorized in many different ways based on any number of
community attributes including, but not limited to, dominant vegetation, hydrology, soil
type, and location in the landscape (Mitsch and Gosselink 1993; Keddy 2000; Kent
2000). One of the most widely recognized classification systems in North America is
that by Cowardin et al. (1979). Our study focused on what Cowardin et al. (1979)
categorize palustrine wetlands, commonly described as freshwater marshes and swamps.
More specifically, palustrine wetlands are defined as nontidal wetland ecosystems with
trees, shrubs, persistent emergents, or emergent mosses or lichens as the dominant
vegetation type, and tidal wetlands with these vegetation types with ocean-derived
salinity levels below 0.5%o (Cowardin et al. 1979). Palustrine wetlands occur throughout
the landscape as small, shallow, permanent or intermittent water bodies; shoreward of
lakes, river channels, or estuaries; on river floodplains; in isolated catchments; on slopes;
or as islands in lakes or rivers (Cowardin et al. 1979). The category of palustrine
wetlands includes eight classes: aquatic bed, emergent, forested, moss-lichen, rock
bottom, scrub-shrub, unconsolidated bottom, and unconsolidated shore.
Wetlands for this study were primarily in the forested wetland class, which
includes wetlands in all water regimes, except subtidal wetlands, that are characterized by









woody vegetation 6m tall or greater. The structure of palustrine forested wetlands
typically includes an overstory of trees with an understory of young trees and shrubs and
an understory of herbaceous species (Cowardin et al. 1979). The class of palustrine
forested wetlands includes six subclasses and dominance types, including broad-leaved
deciduous, needle-leaved deciduous, broad-leaved evergreen, needle-leaved evergreen,
dead, and indeterminate deciduous. Four categories of water modifiers are used to
describe the palustrine forested wetlands in this study (as categorized on the National
Wetlands Inventory GIS coverage available from the Florida Geographic Data Library at
http://www.fgdl.org) including temporarily flooded, seasonally flooded, semipermanently
flooded, and permanently flooded. During the growing season, temporarily flooded
wetlands have surface water present for brief periods with a water table typically well
below the soil surface. Vegetation in temporarily flooded wetlands consists of facultative
species, including those that grow in both uplands and wetlands. Early in the growing
season seasonally flooded wetlands have surface water standing for extended periods
most often without surface water late in the season but with a water table near the soil
surface. Similarly, semipermanently flooded forested wetlands generally have standing
surface water throughout the growing season and a water table at or near the soil surface
when not flooded. At the flood extreme, permanently flooded wetlands have standing
surface water throughout the year with a vegetation community of obligate wetland
species (Cowardin et al. 1979).
Within the text Ecosystems of Florida (Myers and Ewel, eds. 1990), Ewel (1990)
describes approximately 10 distinctive types of swamps. The freshwater forested strands
in this study most closely resemble wetlands in the cypress pond and strand category.
Wharton et al. (1976) described cypress strands as "a diffuse freshwater stream flowing
through a shallow depression on a greatly sloping plain." While Mitsch and Gosselink
(1993) suggest that cypress strands are found primarily in south Florida, Ewel (1990)
notes that cypress strands are common throughout Florida and are found where water
flow is sufficient to create a depression channel in areas with little slope but where actual
flow is seldom observed. The definition of forested strands for this study broadly
encompasses all of these definitions with the primary distinction of forested strands
including evidence of channelized flow (Figure 1-1), though actual flow was rarely
observed at any of the strands during the 2003 growing season during the period of
sampling.
The forested floodplain wetlands in this study most closely resemble the river
swamp category in Ecosystems of Florida (Myers and Ewel, eds. 1990); however the
floodplain forests in this study were associated with smaller river systems than those
typically characterized by Ewel (1990). Forested floodplain wetlands in this study were
associated with low order streams and rivers and were not associated with the main
channels of the largest river systems in Florida (ex. Apalachicola, Suwannee, etc.). The
forested floodplain wetlands in this study can also be categorized as riparian wetlands
such named for the influence on the wetland environment by the adjacent stream or river
system. Mitsch and Gosselink (1993) note vast differences among riparian wetlands,
with the common link being the interconnection between the riparian zone, the river or
stream, and the adjacent upland environment. Riparian wetlands in the southeastern
United States are characterized by low-lying, low slope, and broad floodplain areas with
seasonally pulsing hydrologic influences on well developed soils (Mitsch and Gosselink
































Figure 1-1. Photograph showing the interior of a freshwater forested strand wetland in
Osceola County, Florida. The dark organic layer in the center of the photo shows
evidence of flowing water during times of high water.


1993). The floodplain wetlands in this study are similar to the strands, however standing
water was always observed in the channelized stream (Figure 1-2).
The average hydroperiod (the seasonal pattern and length of saturated soils or
standing water level during a year (Ewel 1990; Mitsch and Gosselink 1993)) varies
among strand and floodplain wetlands, with strands having a moderate length
hydroperiod with saturated soils or standing water for six to nine months a year and
floodplain forests having a short length hydroperiod with generally less than six months
of saturated soils or standing water during a year (Ewel 1990). Fire frequency in forested
strand and floodplain wetlands ranges from moderate frequency (approximately one per
20 years) for strands to low frequency (approximately one per 100 years) for floodplain
wetlands (Ewel 1990). Strands and floodplains also differ in their organic matter
accumulation depths, with strands having high organic matter accumulation with an
organic soil layer greater than 1 m deep and floodplain wetlands having low organic
matter accumulation with an organic soil layer less than 1 m deep (Ewel 1990).
Additionally, strand and floodplain wetlands have different primary water sources with
strands receiving most water from shallow groundwater sources and the main source of
water for floodplain wetlands from surface water originating from the associated stream
or river (Ewel 1990).
Despite these differences in fire frequency, organic matter accumulation, and
water source, the species composition is similar among forested strand and floodplain
wetlands. Common shared tree species in strand and floodplain wetlands include Acer




























i -.


Figure 1-2. Photograph showing the interior of a freshwater forested floodplain wetland
in Polk County, Florida. The channelized stream is visible in the center of the photo
showing the presence of a permanently flooded and flowing stream adjacent to the
floodplain wetland.


rubrum (red maple), Fraxinus caroliniana (water ash), Gordonia lasianthus (loblolly
bay), Liquidambar styraciflua (sweetgum), Magnolia virginiana (sweet bay), Nyssa
sylvatica (black gum), Persea palustris (swamp bay), Quercus laurifolia (swamp laurel
oak), Sabal palmetto (cabbage palm), Salix caroliniana (coastal plain willow), and
Taxodium distichum (baldcypress). Additional tree species common to strands include
Annona glabra (pond apple), Pinus elliottii (slash pine), and Pinus palustris (longleaf
pine). The tree stratum of floodplain wetlands has greater species richness with
additional common tree species including Alnus serrulata (hazel alder), Betula nigra
(river birch), Carpinus caroliniana (American hornbeam), Carya aquatica (water
hickory), Carya glabra pignutt hickory), Celtis laevigata (hackberry), Chamaecyparis
tyoid le (Atlantic white cedar), Diospyros virginiana (persimmon), Fraxinus
pennsylvanica (green ash, red ash), Fraxinus profunda (pumpkin ash), Gleditsia aquatica
(water locust), Magnolia grandiflora (southern magnolia), Nyssa aquatica (water
tupelo), Pinus glabra (spruce pine), Pinus taeda (loblolly pine), Planera aquatica (planer
tree), Platanus occidentalis (American sycamore), Quercus lyrata (overcup oak),
Quercus michauxii (basket oak, swamp chestnut oak), Quercus nigra (water oak),
Quercus virginiana (live oak), Rhapidophyllum hystrix (needle palm), Sabal minor
(bluestem, dwarf palmetto), Salix nigra (black willow), and Ulmus americana (American
elm).









Forested strand and floodplain wetlands also share a number of species in the
shrub stratum including Cephalantus occidentalis (buttonbush), Clethra alnifolia (sweet
pepperbush), Cliftonia monophylla (black titi), Cyrilla racemiflora (titi), Ilex cassine
(dahoon holly), Itea virginica (Virginia willow), Lyonia lucida fetterbushh), Myrica
cerifera (wax myrtle), and Rubus argutus (blackberry). Additional common shrub
species in forested strands include Chrysobalanus icaco (coco plum), Ilex glabra
gallberryy), Leucothoe racemosa fetterbushh), Myrica heterophylla (northern bayberry),
Myrsine guianensis (myrsine), Psychotria sulzneri (wild coffee), Psychotria undata (wild
coffee), and Vaccinium arboretum sparkleberryy). Other common shrub species in river
swamps include Aronia arbutifolia (red chokeberry), Crataegus marshallii (parsley haw),
Ilex decidua (possum haw), Ilex vomitoria yauponn), Leucothoe axillaries (dog-hobble),
Rhododendron viscosum (swamp honeysuckle), Rubus betulifolius (blackberry),
Sambucus canadensis (elderberry), Sebastiana fruticosa (Sebastian bush), Viburnum
nudum (swamp haw), and Viburnum obovatum (small viburnum, black haw).
The species composition of woody vines in strand and floodplain wetlands vary a
great deal according to Ewel (1990), as there is only one shared species, Smilax laurifolia
(bamboo-vine, catbrier). Common woody vine species in strands include Ampelopsis
arborea (pepper vine), Ficus aurea (strangler fig), Ficus citrifolia (wild banyan tree),
Vitis aestivalis (summer grape), and Vitis \inudeii i wthii (calusa grape); and common
woody vine species in floodplain wetlands include Ampelopsis arborea (pepper vine),
Aster carolinianus (climbing aster), Smilax walteri (coral greenbrier), Toxicodendron
radicans (poison ivy), and Vitis rotundifolia (muscadine grape). Species lists for
common tree, shrub, and woody vine species were adopted from Ewel (1990).
Forested strand and floodplain wetlands provide important habitat for wildlife
such as invertebrates, amphibians, reptiles, birds, and mammals. Benthic invertebrates
form the base of the forested wetland food chain, and water quality is strongly related to
the diversity of the benthic macroinvertebrate community (Ewel 1990). Strands and
floodplain wetlands provide valuable habitat for bird and mammal species characterized
by low vegetation density and high cavity density. Though these wetlands differ
somewhat in their relative contributions to bird and mammal habitat as strands have low
canopy insect production, low production of edible fruits and seeds, and high presence of
water, whereas forested floodplain wetlands have high canopy insect production, high
production of edible fruits and seeds, and low presence of water (Ewel 1990).

Historical Perspective

Over 30 years ago, the Water Pollution and Control Act (later referred to as the
Clean Water Act, 1972) required states to "restore and maintain the chemical, physical,
and biological integrity of the Nation's waters" (USEPA 1990). This legislation included
establishing water quality standards for all waters within state boundaries, including
wetlands. Such water quality criteria could be qualitative or quantitative, and it could be
addressed through chemical, physical, or biological standards. Initially, states used
chemical and physical criteria (testing waters for chemical concentrations or physical
conditions that exceeded known standards), assuming losses in ecosystem integrity if
these standards were exceeded (Danielson 1998).









Several shortcomings have been noted when deriving ecosystem integrity based
on exceeding established limits for chemical and physical parameters. Such criteria have
been considered incomplete in their ability to reflect more than the temporal
concentration of substances within a water body (Karr 1993). For instance, the use of
toxicity parameters for determining ecosystem integrity may falsely indicate high
ecosystem integrity when a single toxicity parameter went overlooked. This same water
body could have elevated levels of other toxins or metals that went untested, or be
physically altered so that it has lost functions typically associated with a fully functioning
water body (Karr and Chu 1997). Furthermore, chemical and physical sampling may not
occur during specific loading events and may therefore incompletely describe the
ecological condition of the system. Adams (2002) points out that other environmental
factors such as sedimentation, alterations to habitat, varying temperature and oxygen
levels, and changes in ecological aspects like food availability and predator-prey
relationships are not reflected with chemical criteria alone. James and Kleinow (1994)
note that different organisms respond in different ways to the amount, persistence, and
exposure of chemical compounds otherwise foreign to an organism; and single-valued
chemical and physical criteria of water quality may overlook important biological
implications.
Alternatively, biological indicators integrate the spatial and temporal effects of
the environment on resident organisms, and are suitable for assessing the possible effects
of multifaceted changes in ecosystems (Adams 2002). Karr and Chu (1997) and Adams
(2002) note that biological indicators signal changes in the environment that might
otherwise be overlooked or underestimated by methods that depend on chemical criteria
alone. Organisms have an intricate relationship with their environment, reflecting current
and cumulative ecosystem condition (Karr 1981). The presence of biological organisms
reveals chemical exposure, expressing changes in the physical, chemical, and biological
components of the ecosystem through changes in community composition (Adams 2002).
The United States Environmental Protection Agency (USEPA) recognized the
potential of biological criteria to assess water quality standards and in the late 1980s
required states to use biological indicators to accomplish the goals of the Clean Water
Act (USEPA 1990). In effect, biological assessment has evolved into one of the standard
monitoring tools of water resource protection agencies over the last two decades
(Gerristen et al. 2000). Biological criteria and monitoring programs through the USEPA
have been created for lakes and streams throughout the United States (Barbour et al.
1996a; Karr and Chu 1999; Gerristen et al. 2000), and more recently efforts to assess
wetland condition have been initiated (USEPA 2002).

Indicators of Biological Integrity

Biological monitoring to assess ecosystem condition has been applied widely in
ecological research. The primary aim of biological monitoring is to detect changes in
abundance, structure, and diversity of target species assemblages. One trend in biological
monitoring has led to the development of indices of biotic integrity (referred to as IBIs),
for different species assemblages including diatoms (Fore and Grafe 2002; Fore 2004);
macrophytes (Galatowitsch et al. 1999a; Gernes and Helgen 1999; Mack 2001; Lane









2003); macroinvertebrates (Kerans and Karr 1994; Barbour et al. 1996b); amphibians
(Micacchion 2004); fish (Schulz et al. 1999); and birds (O'Connell et al. 1998).
Perhaps the most common species assemblage chosen for use in the development
of IBIs is the macroinvertebrate assemblage, because many of the macroinvertebrate
species rely entirely on the conditions of their aquatic environment for habitat, food, and
reproductive activities. In Florida there are currently three biological indices that use the
community composition of the macroinvertebrate assemblage to detect changes in
biological integrity including the Lake Condition Index (LCI, Gerristen and White 1997),
Stream Condition Index (SCI, Fore 2004), and the Florida Wetland Condition Index for
depressional freshwater wetlands (FWCI, Lane et al. 2003; Reiss and Brown 2005).
However, the forested strand and floodplain systems targeted in this study have varying
hydrologic regimes from temporarily to permanently flooded, complicating
macroinvertebrate collection due to variable hydrologic conditions. As such, we have
chosen the macrophyte species assemblage for use in the preliminary forested strand and
floodplain FWCI.
Wetland macrophytes are defined as aquatic emergent, submergent, or floating
plants growing in or near water (USEPA 1998); and are described as distinguishing
landscape features. The spatial distribution of macrophytes in the landscape occurs
according to a multitude of factors, including hydroperiod, water chemistry, and substrate
type, as well as other factors such as available seed source and climate. Fennessy et al.
(2001) state that the community composition of wetland macrophytes typifies the
physical, chemical, and biological wetland dynamic in time and space.
Macrophytes play a vital role in supporting the structure and function of wetlands
by providing food and habitat for other assemblages including algae, macroinvertebrates,
fish, amphibians, reptiles, birds, and mammals; and macrophyte populations can be used
as a diagnostic tool to assess other aspects of the wetland environment. Crowder and
Painter (1991) state that a lack of macrophytes where they are otherwise expected to
grow suggests reduced wildlife populations from lack of food or cover and/or water
quality concerns such as toxic chemical constituents, increased turbidity, or increased
salinity. In contrast, an overgrowth of particular macrophytes may signify increased
nutrient loading (USEPA 1998).
Many advantages of studying macrophytes as indicators of wetland condition
have been noted, including their large, obvious size; ease of identification, to at least
some useful taxonomic level; known response to toxicity tests; and general lack of ability
to move to avoid unfavorable conditions (Danielson 1998; Cronk and Fennessy 2001).
Additionally, macrophytes readily respond to changes in nutrient, light, toxic
contaminant, metal, herbicide, turbidity, water, and salt levels. They can also be sampled
in the field with transects, or remotely from aerial photography; and well-established
field methods of sampling macrophytes exist (USEPA 2003). Furthermore, the USEPA
(2003) states additional advantages of using the macrophyte assemblage, including that
they do not require laboratory analysis, can easily be used for calculating simple
abundance metrics, and are superb integrators of environmental condition. In general,
macrophytes represent a useful assemblage for describing wetland condition (Mack
2001). Schindler (1987) alleges that macrophytes can provide a more integrated picture
of wetland function than measures such as nutrient cycling, productivity, decomposition,
or chemical and physical composition.









There are however some noted shortcomings of using macrophytes as biological
indicators. These include the potential delay in response time for perennial shrub and
tree species, difficulty identifying taxa to the species level in certain seasons and for
some genera, uneven herbivory patterns, and varied pest-management practices (Cronk
and Fennessy 2001). Despite these limitations, macrophytes have provided strong signals
of anthropogenic influence (USEPA 2003). In fact, many states have begun using
macrophytes in their wetland biological assessment programs, including, Minnesota
(Galatowitsch et al. 1999a; Gernes and Helgen 1999), Montana (Apfelbeck 2000), North
Dakota (Mushet et al. 2002), Ohio (Mack 2001), and Florida (Lane et al. 2003; Reiss and
Brown 2005).

Quantifying Anthropogenic Influence

Wetlands occupy a large portion of the Florida landscape. An estimate from the
1780s reported 8,225,000 ha of wetlands in Florida (Dahl 2000). By the mid-1980s, the
National Wetlands Inventory estimated Florida had 4,467,000 ha of wetlands remaining,
translating into a loss in Florida of roughly 45% of the pre-1780s wetland area (Mitsch
and Gosselink 1993; Dahl 2000). Throughout the continental United States, similar
trends were apparent, with a drastic decline in the surface area of wetlands. More
specifically, Dahl (2000) reported that 98% of all wetland losses throughout the
continental United States from 1986 to 1997 were losses to freshwater wetlands. Of the
remaining freshwater wetlands, 40% were adjacent to agricultural lands and therefore
potentially affected by land use practices such as herbicide and pesticide application,
irrigation, livestock watering and wastes, soil erosion, and deposition. An additional 17%
were adjacent to urban or rural development. Freshwater non-tidal wetlands experienced
the greatest development pressure just inland from coastlines as the demand for housing,
transportation infrastructure, and commercial and recreational facilities increased (Dahl
2000). These changes in land use were proportionally more widespread in Florida than
much of the continental United States due to the remarkable length of coastline along
both the Atlantic Ocean and Gulf of Mexico coasts of Florida.
Agricultural and urban development activities influence an array of changes to the
physical, chemical, and biological characteristics in nearby ecosystems. There have been
numerous attempts at quantifying anthropogenic influence based on quantitative indices,
for example, the Wetland Rapid Assessment Procedure (WRAP; Miller and Boyd 1999),
the Minnesota disturbance index (Gernes and Helgen 1999), the Human Disturbance
Gradient (HDG; Fore 2004), and the Landscape Development Intensity (LDI) index
(Brown and Vivas 2005). Our study incorporates the LDI index as a measure of
anthropogenic influence. Additionally, for thirteen floodplain wetlands, HDG scores
have been obtained from previous studies (Fore 2004; Florida Department of
Environmental Protection Geographic Information Systems map layers).

Landscape Development Intensity Index
The LDI index has been used as a gauge of human activity based on a
development intensity measure derived from nonrenewable energy use in the surrounding
landscape. The underlying concept behind calculating the LDI (quantifying the
nonrenewable energy use per unit area in the surrounding landscape) stems from earlier









works by Odum (1995), who pioneered emergy analysis for environmental accounting.
[Emergy is an environmental accounting term referring to expressing energy use in solar
equivalents (Odum 1995).] Brown and Ulgiati (2005) suggest that landscape condition,
or ecosystem health, is strongly related to the surrounding intensity of human activity,
and that ecological communities are affected by the direct, secondary, and cumulative
impacts of activities in the surrounding landscape. Healthy ecosystems are defined as
those with integrity and sustainability, which correlate to limited development in the
surrounding landscape and the maintenance of ecosystem structure and function, even
when stressors (e.g. flooding, drought, etc.) are present (Brown and Ulgiati 2005).
The LDI scale encompasses a gradient from completely natural to highly
developed land use intensity, and is calculated based on the percent of the area in a
particular land use within the designated area surrounding the wetland (ex. 100 meters,
200 meters, etc.) multiplied times the LDI coefficient (Table 1-1), which is defined by the
amount of nonrenewable energy use for a given land use (Brown and Vivas 2005). The
LDI coefficient does not account for any individual causal agent directly, but instead,
may represent the combined effects of air and water pollutants, physical damage, changes
in the suite of environmental conditions (ex. groundwater levels, increased flooding), or a
combination of such factors, all of which enter the natural ecological system from the
surrounding developed landscape. Wetlands surrounded by more intense activities such
as highways and multi-family residential land uses receive higher LDI index values, as
the highest LDI coefficient of 10.0 is assigned to the urban land use category of Central
Business District. Undeveloped land uses such as wetlands, lakes, and upland forests are
assigned an LDI coefficient of 1.0, the lowest possible value, based on no use of
nonrenewable energy in these ecosystems.

Human Disturbance Gradient
The Human Disturbance Gradient (HDG) is a quantitative measure used to assess
the level of human disturbance to an ecosystem based on the cumulative score of four
independent measures of the environment including ammonia concentration in the water
(as mg N/L), hydrologic index, habitat assessment, and LDI for the buffer which included
an area of 100 m on each side of the stream and 10 km upstream of the sampling point
(Fore 2004). The first measure included in the HDG, ammonia concentration, is included
as a summary variable defining water quality because of its consistent correlation with
other water quality parameters (i.e. total phosphate) and its availability in the dataset
(Fore 2004). Additionally, increases in ammonia concentration are thought to be
evidenced from both agricultural (e.g., fertilizer and farming practices) and urban
activities, whereas changes in total phosphorus concentrations are thought to be solely
associated with agricultural operations (Fore 2004), so using ammonia concentration
accounts for changes in both agricultural and urban land use activities.
The second measure included in the HDG is an estimate of hydrologic condition
of the stream site, which is scored on-site at the discretion of the biologist conducting the
sampling effort. Scores for the hydrologic condition range from 1-10, with a score of 1-2
Excellent representing a natural, undisturbed system with few impervious surfaces, high
connectivity with ground water and surface features, and a natural flow regime (Fore
2004). Intermediate categories include 3-4 Good, 5-6 Moderate, and 7-8 Poor. The final
category, 9-10 Very Poor, is reserved for those systems with a flow regime entirely









Table 1-1. Non-renewable energy use and LDI coefficients per land use used in the
calculation of the LDI index.


controlled by human modification, with a flashy hydrograph and extreme alteration of the
natural ecosystem (Fore 2004).
Habitat assessment (or habitat condition index as termed in Fore 2004) is the third
measure included in the HDG. The Florida Department of Environmental Protection
(FDEP) has developed Standard Operating Procedures (SOPs) for river and stream
habitat assessments (FDEP-SOP-001/01: Form FD 9000-5 June 1, 2001 available from
the Bureau of Laboratories at http://www.dep.state.fl.us/labs/library/forms.htm). The
habitat assessment is separated into four primary habitat components: substrate diversity
(number of diverse, productive habitats), substrate availability (percent of stream reach
composed of productive habitat), water velocity (based on the maximum observed


Land Use
Natural Land / Open Water
Pine Plantation
Low Intensity Open Space / Recreational
Unimproved Pastureland (with livestock)
Improved pasture (no livestock)
Low Intensity Pasture (with livestock)
High Intensity Pasture (with livestock)
Citrus
Medium Intensity Open Space / Recreational
Row crops
High Intensity Agriculture (dairy farm)
Single Family Residential (Low-density)
Recreational / Open Space (High-intensity)
Single Family Residential (Med-density)
Low Intensity Transportation
Single Family Residential (High-density)
Low Intensity commercial (Comm Strip)
Institutional
Highway (4 lane)
Industrial
Multi-family residential (Low rise)
High intensity commercial (Mall)
Multi-family residential (High rise)
Central Business District (Avg 2 stories)
Electric Power Facility


Nonrenewable Energy Use
(E14 solar equivalent
joules/ha/yr)
0.0
5.1
6.7
8.3
19.5
36.9
51.5
65.4
67.3
117.1
201.0
1077.0
1230.0
2461.5
3080.0
3729.5
3758.0
4042.2
5020.0
5210.6
7391.5
12661.0
1285.0
16150.3
29401.3


LDI
Coefficient
1.00
1.58
1.85
2.06
2.89
3.51
3.83
4.06
4.09
4.63
5.15
6.79
6.92
7.59
7.81
7.99
8.00
8.07
8.28
8.32
8.66
9.18
9.19
9.42
10.00









velocity, where higher velocities receive higher scores), and habitat smothering (as the
percent of the stream reach covered by sand or silt accumulation). The four secondary
habitat parameters include artificial channelization (as an assessment of anthropogenic
modification of the stream reach), bank stability (evidence of bank stability/instability
from erosion or bank failure), riparian buffer zone width (estimate of the width of the
vegetation on the least buffered side), and riparian zone vegetation quality (estimate of
community composition and structure). These eight habitat parameters are scored by the
field biologist on a scale of 1-20, with 20 representing the highest habitat quality. Three
of the secondary habitat components (bank stability, riparian buffer zone width, and
riparian zone vegetation quality) are scored separately on a scale of 1-10 (with 10
representing the highest habitat quality) for both the right and left banks (when added
together these three categories are still given equal weighting, with half of the score
reflecting the condition of each bank). Scores for each of the eight parameters are
summed, and a stream is then assigned a categorical score of optimal (134-160),
suboptimal (91-123), marginal (54-80), or poor (11-43) based on the total score (Fore
2004).
The fourth measure of the HDG is an LDI score calculated within a 100 m buffer
on each side of the stream for 10 km upstream from the sampling point (LDI BF) (Fore
2004). Each measure included in the HDG is categorically assigned a score based on the
score for each measure (Table 1-2). Ammonia concentration in the water (as mg N/L),
habitat assessment, and LDI for the buffer are assigned scores of 0, 1, or 2; whereas the
hydrologic index is assigned scores of 0, 1, 2, or 3. Overall HDG values potentially
range from 0 (no detectable human induced disturbance) to 9 (extreme anthropogenic
disturbance) (Fore 2004). Original categorical scoring criteria for the HDG were
established from correlations of the HDG categories (ammonia, hydrologic condition,
habitat index, and LDI BF) with the macroinvertebrate metric EPT (Ephemeroptera,
Plecoptera, and Trichoptera) taxa richness. All HDG scores used in this analysis were
taken directly from HDG scores provided by FDEP.

Project Overview

The community composition of the macrophyte assemblage was sampled in
flowing wetland systems including forested strand (n = 10) and forested floodplain (n =
14) wetlands throughout Florida. Our primary goal was to develop a preliminary Florida
Wetland Condition Index (FWCI) for freshwater flowing water wetlands, building upon
the FWCI for isolated depressional herbaceous (Lane et al. 2003) and forested (Reiss and
Brown 2005) freshwater wetlands. Our secondary goal was to determine the appropriate
scale for LDI buffers for flowing water wetlands that would best capture the current
wetland condition, based on correlations with the FWCI. Our third goal was to correlate
measures of the HDG (Fore 2002) and macroinvertebrate community data from the
Florida SCI with our macrophyte FWCI and landscape LDI measures of ecosystem
condition.
Wetland study sites were sought in various landscape settings that included
natural, agricultural, and urban land uses. The elongated forested wetlands associated
with flowing waters were in many areas continuous for extended geographic distances,
and the forested edges bordered a great variety of land use activities. Therefore, it was









Table 1-2. Categorical scoring criteria used to calculate the Human Disturbance Gradient
(HDG). The HDG is the sum of the categorical scores for the four individual
measures ammonia concentration in the water (as mg N/L), hydrologic index, habitat
assessment, and LDI for the buffer. The HDG range is from 0 (no detectable human
induced disturbance) to 9 (extreme anthropogenic disturbance).

HDG Categorical Score
HDG Measure 0 1 2 3
Ammonia concentration (mg N/L) < 0.1 0.1 2.0 > 2.0
Hydrologic Condition < 6 6 7 8 9 10
Habitat Assessment > 65 50 65 < 50
LDI for the buffer < 2.0 2.0 3.5 > 3.5


difficult to ascertain the specific land use activity contributing most heavily or frequently
to perturbations in community composition. While a priori land use categories were
assigned (reference, agricultural, urban), wetlands were also divided into two groups
representing low and high landscape development intensities.











CHAPTER 2
METHODS

Twenty-four freshwater forested wetlands, including 10 strand and 14 floodplain
wetlands, were sampled during 2003. This chapter describes site selection, Landscape
Development Intensity (LDI) index calculations, field-data collection, and methods of
statistical analyses.

Study Area

Florida has been dissected into smaller geographic zones using a number of
different methods, including bioregions (Griffith et al. 1994) with 13 subecoregions
(grouped into four ecoregions for small wadeable streams of panhandle, northeast,
peninsula, and Everglades); lake ecoregions of Florida (Griffith et al. 1997) with 47
subecoregions; and Florida wetland regions determined with a hydrologic model by Lane
(2000), which included physical surficiall geology, soils, digital elevation model, slope)
and climatic (precipitation, potential evapotranspiration, runoff, annual days of freezing)
variables with four ecoregions (panhandle, north, central, and south). When boundaries
for the bioregions of Griffith et al. (1994) and wetland regions of Lane (2000) were
compared, there were similarities among the panhandle and south/Everglades boundaries;
however, the north/northeast and central/peninsula ecoregions differed in extent (Figure
2-1). We adopted the four wetland regions developed by Lane (2000) to partition the
state during site selection. However, Human Disturbance Gradient (HDG) and Stream
Condition Index (SCI) scores available for the floodplain forests were designated
according to the Florida bioregions (Griffith et al. 1994); as such we use both wetland
region and bioregion categories in our analyses.
The four wetland regions defined by Lane (2000) are categorized as panhandle,
north, central, and south. The panhandle wetland region is characterized by less human
development than the other ecoregions. Streams in the panhandle wetland region
typically run from north to south, discharging into the Gulf of Mexico. The primary
ecosystems are scrub and high pine, temperate hardwood forests, pine flatwoods and dry
prairies, and swamps (Fernald and Purdum 1992). The major population centers of
Tallahassee, Panama City, and Pensacola are included within the panhandle wetland
region (Davis 1967). The north wetland region has similar ecosystem types, primarily
with pine flatwoods and dry prairies and less scrub and high pine, temperate hardwood
forests, and swamps (Fernald and Purdum 1992). Two major drainage features in the
north wetland region include the Santa Fe and Suwannee River system and the lower St.
John's River, with Jacksonville being the major city situated along the northeastern
Atlantic coast. The central wetland region has a distinguishing central ridge feature with
a higher maximum elevation than the north and south wetland regions (Fernald and
Purdum 1992). The central wetland region is characterized by pine flatwoods and dry
prairies with scrub and high pine along the ridge. There is less area in temperate
hardwood forests, less area in swamps, and an increase in the area of inland lakes and
freshwater marshes (Fernald and Purdum 1992). The largest population centers include



























Legend
W Ecoregions(Lane 2000)
L Stream Ecoregions (Gnffith et al 1994)

0 75 150 300 Kilometers


Figure 2-1. Florida wetland regions defined by climatic and physical variables (solid
line; Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994).


Orlando and Tampa. The south wetland region is unique with the nearly flat terrain of
the Everglades (less than 2 m relief, Femald and Purdum 1992). Much of the south is
swamps and freshwater marsh, and few streams remain unaltered. Lake Okeechobee is in
this wetland region as are the urban centers of Miami and Ft. Myers located on the
densely populated east and west coasts, respectively.

Site Selection

During the 2003 growing season (May-July), 24 forested wetlands were sampled
throughout the state of Florida (Figure 2-2). Random site selection was not feasible
given the necessity of obtaining permission to access private lands and the non-random
pattern of land development in Florida. The forested strand wetlands (n=10) were
arranged spatially throughout Florida, so that the distribution was spread among three of
the four Florida wetland regions (Lane 2000), including north (n=2), central (n=5), and
south (n=3). No strands were sampled in the panhandle wetland region. The forested
floodplain wetlands (n=14) were selected as a subset of systems that have been sampled
using the SCI (Fore 2004). Many of the streams that had received a score of poor
according to the SCI no longer had floodplain forest vegetation along the banks, but
rather had mowed grass or paved banks. As such, these streams did not fit our criteria for

















S-



Legend
o Forested Strands
Forested Floodplains
Ecoregions (Lane 2000)
Panhandle
North
| |Central
| South


0 75 150 300 Kilometers
I I I I I I I


Figure 2-2. Study site location of 24 forested wetlands in Florida (+strands n=10;
floodplains n=14). Wetland region boundaries follow Lane (2000).


site selection, and where therefore excluded from sampling. Forested floodplain
wetlands sampled were located in the panhandle (n=2), north (n=8), central (n=3), and
south (n=1) wetland regions.
A priori categories were used to define sample wetlands as reference, agricultural,
and urban. The suite of reference wetlands included wetlands surrounded by intact native
ecosystems. These wetlands were thought to represent the best possible wetland
condition currently in Florida. These wetlands could be categorized as reference standard
wetlands by the USEPA (1998a), and have been described as wetlands considered to be
the least altered that are reflective of characteristic levels of wetland function.
Agricultural wetlands were defined as those currently surrounded by cattle pasture,
rangeland, row crops, citrus, and silvicultural land uses. Urban wetlands included those
embedded within commercial, industrial, and residential land uses. Many of the urban
wetlands sampled were suspected to previously have been embedded in agricultural land
uses due to the development patterns throughout Florida. Hereafter wetlands embedded
in primarily undeveloped landscapes are called reference; wetlands embedded in
primarily agricultural land uses are called agricultural; and wetlands embedded in
primarily urban land uses are called urban.
Table 2-1 provides general information about each sample wetland, including site
code, sample date, wetland type, wetland region (Lane 2000), county, bioregion (Griffith











Table 2-1. Site characterization for 24 freshwater forested wetlands.
Wetland


Site Code*
FF1
FF2
FF3
FF4
FF5
FF6
FF7
FF8
FF9
FF10
FF11
FF12
FF13
FF14
FS1
FS2
FS3
FS4
FS5
FS6
FS7
FS8
FS9
FS10


*Site Codes correspond to the wetland type (FF


floodplain forest; FS = forested strand) and were numbered in the order sampling occurred.


Wetland region from Lane (2000); Bioregion from Griffith et al. (1994).
Associated Stream = Stream Condition Index (SCI) stream data; NA = Not Applicable, forested strands were not associated with SCI streams.


Sample Date
May 19 2003
May 20 2003
May 22 2003
May 23 2003
May 25 2003
May 25 2003
May 25 2003
May 26 2003
May 29 2003
June 8 2003
June 13 2003
June 14 2003
June 24 2003
July 1 2003
May 24 2003
May 26 2003
May 27 2003
May 312003
June 1 2003
June 16 2003
June 24 2003
July 7 2003
July 8 2003


Wetland Type
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Floodplain
Strand
Strand
Strand
Strand
Strand
Strand
Strand
Strand
Strand


Region
North
North
North
Central
North
Central
Central
North
North
South
Panhandle
North
North
Panhandle
Central
Central
Central
Central
Central
South
South
North
North
South


County
Putnam
Marion
Volusia
Lake
Nassau
Polk
Polk
Nassau
Clay
Lee
Hamilton
Baker
Martin
Walton
Osceola
Osceola
Osceola
Osceola
Osceola
Palm Beach
Palm Beach
Alachua
Alachua
Martin


Bioregion
Peninsula
Peninsula
Peninsula
Peninsula
Northeast
Peninsula
Peninsula
Northeast
Northeast
Everglades
Northeast
Northeast
Peninsula
Panhandle
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula
Peninsula


A Priori Category
Urban
Reference
Urban
Agricultural
Urban
Reference
Reference
Urban
Urban
Urban
Agricultural
Agricultural/Urban
Reference
Agricultural
Reference
Agricultural
Agricultural
Urban
Urban
Urban
Reference
Agricultural
Reference
Reference


Associated Stream'
Orange Creek
Juniper Creek
Groover Branch
Blackwater Creek
Pigeon Creek
Livingston Creek
Tiger Creek
Alligator Creek
Green Creek
Leitner Creek
Rock Creek
Turkey Creek
Kitchen Creek
Limestone Creek
NA
NA
NA
NA
NA
NA
NA
NA
NA


July 14 2003 Strand









et al. 1994), apriori category of surrounding land use, and associated stream. Site codes
were assigned to preserve the anonymity of individual land owners.

Gradients of Landscape Development Intensity

Digital orthophoto imagery, available from Labins, The Land Boundary
Information System from the Florida Department of Environmental Protection (FDEP)
(available at http://www.labins.org/2003/index.cfm), were used to georectify transect
locations from GPS coordinates using ArcGIS 8.3 from Environmental Systems Research
Institute, Inc. Polygons were established for wetlands delineated from digital orthophoto
imagery. LDI scores were calculated using 1995 and 2000 land use coverages available
for download through the Florida Geographic Data Library (available at
http://www.fgdl.org/). Coverages for 1995 land use were available separately for each
Florida Water Management District (Northwest Florida, NWFWMD; Suwannee River,
SRWMD; St. Johns River, SJRWMD; Southwest Florida, SWFWMD; and South Florida,
SFWMD). More recent coverages (2000 land use) were available for a limited number of
sample wetlands including only those within the boundaries of the SJRWMD.
LDI index scores were calculated at the transect (LDI_T), wetland feature
(LDI F), and watershed (LDI WS) scale. At the transect scale, a 100 m buffer was
created around the downstream sample transect. Two LDI_T scores were then calculated
based on the area of the land uses within the 100 m buffer (Figure 2-3), one including the
wetland area and the other excluding the wetland area. The LDI_T score including the
wetland area will generally be lower than the LDI T score excluding the wetland area,
based on the increase in the percent of land assigned a 1.0 Natural Land/Open Space LDI
coefficient (because of the inclusion of the wetland area, which is assigned an LDI
coefficient of 1.0). Including the wetland area in the calculation was thought to weight
the LDI_T based on the size of the wetland sampled, which could be an important factor
influencing wetland condition. LDI_T calculations were repeated using a 200 m buffer
(Figure 2-3). LDI_T 1995 identification of land uses within the 100 m buffer were taken
from land use/land cover coverages and checked with 1995 digital orthophoto imagery.
LDI T 2000 identification of land uses were taken from 2000 land use/land cover
coverages and checked with 1999 digital orthophoto imagery. Land uses that had
changed since the photos were taken were updated based on land use maps drawn during
field site visits. This step was not possible for larger buffer areas where visibility of
surrounding land uses on site was hindered due to distance and barriers, and so
discrepancies may be apparent in 100 m buffer calculations (hand corrected for each
buffer) and larger 200 m buffers (not corrected, land use/land cover straight from
available GIS coverages).
LDI_F and LDIWS were also calculated once including and once excluding the
wetland area. The boundary for the LDI_F was established as the area making up the 200
m buffer upstream of the downstream transect (Figure 2-4). The boundary of the feature
was established when a distance of at least 30 m showed a break in wetland vegetation
established from photo-interpretation of the digital orthophoto imagery. The upstream
boundary for the LDIWS was established using the Better Assessment Science
Integrating Point and Nonpoint Sources 3.0 (BASINS 3.0) environmental analysis system
(USEPA 2001) (Figure 2-5). BASINS is a multi-purpose environmental assessment tool




















I : .. .




















17.1
...... ...K..
.... .......








.. .. ,



.. .. .. .


.. ... .

. ... ... "".. .
-. ... .. .." ... .. .. . .: .
. .. .. .. :. ... .. '. : '
;-. -.. -. -. : : : :... .


"'" '" ':' :" .. '.. '..'.. '... : .
".. .. .. .".. .. .. ... : :: : : '!
. ., .: : : ..... .. .,. ..
- -. .- --... .. ,

... ........... ... .......


Legend

I 1 FF3 LDI_T 100m buffer
FF3 LDI_T 200m buffer 0 50 100 200 Meters
Land Use
SPine Flatwoods
7 Residential Low Density Less than two dwelling units per acre
SResidential Med. Density Two to five dwelling units per acre

Wetland Forested Mixed


Figure 2-3. Boundaries of the 100 and 200 m buffers drawn around the downstream
transect for the LDI_T calculation at Forested Floodplain 3 (FF3). Two LDI_T values
were calculated for each sample wetland, one including the wetland area and the other
excluding the wetland area.


.................... ...

................ ......
. .. .

...... ............. ...

........... .... ......

............... .... ...

................

.... ..............
. .. .
.. .... .
...... ......... ......

........ ........... ...
. .

....... ........ ...
. .. .

....... ............

...............
.. .. ... .. .


............



......... ....














































I--I
I--I
Land Us
I I

I I








rTFFA


FF3 LDIF 100m b111
FF3 LDI_F 200m blD r
e
Coniferous Pine
Disturbed Land
Field Crops
Herbaceous
Improved Pastures
Pine Flatwoods
Reservoirs less than 10 acres (4 hectares) which are dominant features
Residential Low Density Less than two dwelling units per acre
Residential Med. Density Two to five dwelling units per acre
Roads and Highways
Shrub and Brushland
wetlandd Coniferous Forest
Wetland Forested Mixed
Vbodland Pastures


0 185 370
I 1 l l 1


740 Meters
I


Figure 2-4. Boundaries of 100 and 200 m buffers around a wetland feature and
designated land use. The 200 m buffer was used for the LDI_F calculation at Forested
Floodplain 3 (FF3).











































Legend
I I FF3 LDI_WS boundary
/ FF3 Feature
Land Use
I Natural Land
Agricultural Land
Urban Land


0 750 1,500 3,000 Meters
I I I I I I I I I


Figure 2-5. Boundary of the watershed buffer around Forested Floodplain 3 (FF3) for the
LDI_WS calculation. Four LDIWS values were calculated for each sample wetland:
two LDI_WSs were calculated based on an equal weighting by area of all land uses
within the upstream watershed including a calculation inclusive of the wetland area
(LDIWSED/w) and exclusive of the wetland area (LDIWS_ED/wo); and two
distance weighted LDIs were calculated in which land uses nearest to the sample
wetland were more highly weighted, including linear weighting (LDI WSDWlin)
and weighting based on an exponential decay function (LDI WS DWexp).









designed for watershed and water quality studies that integrates a GIS (ArcView 3.2),
national watershed database, and environmental assessment and modeling tools (USEPA
2001). The digital terrain model from the National Elevation Dataset, which is a 30 m
raster-based dataset produced by the United States Geological Survey (USGS), was used
for watershed delineation.
Two separate types of LDI_WSs were calculated based on an equal weighting by
area of all land uses within the upstream watershed (LDIWSED) and a distance
weighted calculation, where land uses occurring closer to the sampling point were
weighted more than those occurring farther away (WSLDIDW). LDI_WSED scores
were calculated both including (LDIWSED/w) and excluding (LDIWS ED/wo) the
wetland area. The distance weighted calculations were done in two ways. First, a
calculation was made using a linear weighting of land uses (LDI WS DWlin); second,
a calculation was made using an exponential decay function, where land uses nearest the
sampling point were weighted most significantly (LDI WS DW_exp).
GIS analyses were performed in ArcGIS 8.3 (Environmental Systems Research
Institute, Inc. 2002). The following equation was used to calculate LDI:

LDITotal = % LUi LDIi (2-1)

where %LUi is the percent of a land use within the buffer of interest and LDIi is the LDI
coefficient for a particular land use based on the amount of nonrenewable energy use per
unit area in the surrounding landscape (Table 1-1). The LDI coefficient values and LDI
equation were based on work by Brown and Vivas (2005). Potential LDI coefficients
ranged from a minimum of 1.0 (Natural Land/Open Space) to a maximum of 10.0
(Central Business District).

Field-data Collection

A concise summary of field-data collection procedures follows. Appendix A
provides more detailed descriptions of field-data collection techniques in the format of
Standard Operating Procedures (SOPs) for field use.

Transect Sampling Design

Transects were established perpendicular to the hydrologic flow of the system.
The wetland/upland boundary was determined based on the Florida Unified Wetland
Delineation Methodology (Chapter 62-340, F.A.C.), using a combination of wetland
plant presence according to wetland plant status (e.g. obligate, facultative wetland,
facultative, or upland) and wetland hydrologic indicators (e.g. lichen lines, moss collars).
Modifications of transect establishment were implemented for strand and floodplain
wetlands types, described below.
Sampling in forested strands consisted of establishing four transects that extended
from the wetland/upland boundary to the middle of the channelized flow in the center of
the strand (Figure 2-6a). Transects were established at 25 m intervals. While it is
recognized that strands were not perfectly symmetrical in nature, an effort was made to
establish transects that represented a cross-section of the strand along the gradient of







































Figure 2-6. Idealized transect layout for forested strand wetlands. (A) Standard transect
layout included four transects that represented a cross-section of the strand along the
gradient of wetland/upland boundary to the water channel, with transect establishment
along both sides of the water course. (B) Alternative transect placement included four
transects established on the same side of the water channel due to limited access or
permission restrictions.


















-hanneI --




Flo. L


F:r~e~,Iecj Siran.1
V'bl I a nd


T4 El









upland/wetland boundary to the water channel. In some sampling efforts wetland forests
were sampled along both sides of the water course. However, on some occasions, limited
access or lack of permission restricted sampling to the same side of the water course. In
these instances, four transects were established on the same side of the water channel
(Figure 2-6b).
Forested floodplain wetlands varied in size, with many wetlands spanning an area
broader than 100 m across. To encourage comparable sampling efforts among the
different width strand and floodplain systems, a maximum 50 m transect length was
established for floodplain systems. For forested floodplains narrower than 50 m wide,
four transects were established that spanned the area from the upland/wetland boundary,
as determined by wetland plant presence and hydrologic indicators, to the edge of the
stream channel. The initial transect was established downstream at a point closest to the
SCI sampling location. Consecutive transects were established upstream at 25 m
intervals. In forested floodplain wetlands wider than 50 m, the initial transect was
established at the upland/wetland boundary and extending 50 m inward perpendicular to
the stream. The second transect was established 25 m upstream, starting at the waters
edge and extending a full 50 m towards the upland/wetland boundary, perpendicular to
the water course. The pattern was repeated for the third and fourth transects (Figure 2-7).
While four transects of 50 m length were considered the optimal sampling effort, this was
not always achieved due to limited access and development pressures resulting in limited
remaining areas of forested floodplain.
Along each transect, a series of 1 m wide by 5 m long quadrats was established
back to back. Living macrophytes rooted within each quadrat were identified to the
lowest taxonomic level possible. When field identification was limited, a sample
specimen was collected, pressed, and identified in the laboratory. An expert Florida
botanist (Dr. David Hall) was consulted for identification of unknown specimens.
Taxonomic information including species, genus, and family were compiled for all of the
macrophytes identified. Additional characteristics were collected for use in metric
development, including category (annual or perennial, evergreen or deciduous,
indigenous or exotic) and growth form (aquatic, fern, grass, herb, sedge, shrub, tree, or
vine). References specific to Florida were consulted first (Tobe et al. 1998; Wunderlin
and Hansen 2003), and additional information was supplemented from other sources (in
the following order: Godfrey and Wooten 1981, Wunderlin 1998, and USDA NRCS
2002). Wetland indicator status (i.e. obligate, facultative wetland, etc.) for each species
was obtained from FAC Ch. 62-340. Wetland indicator statuses for species not listed in
FAC Ch. 62-340 were obtained from Tobe et al. (1998), Wunderlin and Hansen (2003),
and USDA NRCS (2002), in that order. When species were not assigned a Florida
specific indicator status, the National Wetlands Inventory indicator status was used
(Wunderlin and Hansen 2003; USDA NRCS 2002, in that order). When information was
still unavailable for plant characteristics in published literature, Florida botanists (who
also participated in the Floristic Quality Assessment Index) were consulted.

Floristic Quality Assessment Index
A Floristic Quality Assessment Index (FQAI) has been included in many of the
multimetric biotic indices created for the macrophyte assemblage. The concept of FQAI
was developed by Wilhelm and Ladd (1988) for vegetation around Chicago, Illinois.













elnd Siran
\VIelland


Main Water
Channel --


T4 .41 SO-... 1 S. 20 5-10


Flow Direction 4


'.11,I


15-20


Figure 2-7. Idealized transect layout for forested floodplain wetlands. Transect 1 began
at the upland/wetland boundary and extended 50 m inward, perpendicular to the
channelized flow. Transect 2 began at the edge of the channel and extended 50 m
towards the wetland/upland boundary. This pattern repeated for transects 3 and 4.









This method of scoring plant species based on expert botanist opinion has been used in
Michigan (Herman et al. 1997), Ohio (Andreas and Lichvar 1995; Fennessy et al. 1998;
Mack 2001), Ontario (Francis et al. 2000), North Dakota (Mushet et al. 2002), and
Florida (Lane et al. 2003; Cohen et al. 2004; Reiss and Brown 2005). The FQAI
provides a quantitative means of assessing the fidelity of a plant to a particular
environment through the Delphi technique (Kent 2000), where individual botanists
assigned coefficients (termed coefficients of conservatism) to individual species. This
technique assumes that the collective decision by a group of expert botanists is more
accurate than the professional judgment of one individual (Kent 2000).
The Florida FQAI enlisted regional expert botanists to provide quantitative scores
for vegetation in the form of a coefficient of conservatism assigned individually to each
species. The FQAI score for an individual wetland was calculated as:

FQAI = [ Z (C + C2 +... Cn) ]/N (2-2)

where C = species specific coefficient of conservatism score (CC), and N = species
richness. This equation was considered a modified FQAI because previous studies used
the square root of native species richness as the denominator. Theories on the importance
of species richness suggest that higher species richness signifies a more valuable
ecosystem, which can be quantified by using the square root function (Fennessy et al
1998). However, a recent study by Cohen et al. (2004) found a stronger relationship
along a disturbance gradient for the mean CC (Eq. 2-2) than with the traditional FQAI
equation (using the square root of native species richness in the denominator). They also
reported that using total species richness (i.e. including exotic species in the calculation
of species richness) improved the relationship of mean CC with the human disturbance
gradient (measured with LDI). The sum of the species CC scores was divided by total
species richness (Eq. 2-2) in this study to account for potential differences in species
richness due to differences in wetland regions (Lane 2000), bioregions (Griffith et al.
1994), apriori land use categories, or other unspecified differences.
Coefficient of conservatism scores were obtained from Florida botanist surveys to
compliment previous FQAI efforts for Florida. FQAI scores were adopted first from
depressional forested wetlands (Reiss and Brown 2005) and second from depressional
herbaceous wetlands (Lane et al. 2003; Cohen et al. 2004). Correlations between CC
scores for species from the depressional forested and herbaceous wetlands were strong
(unpublished data). In an effort to streamline the acquisition of species specific CC
scores for the FQAI for strand and floodplain wetlands, CC scores were only obtained for
those species not included in previous Florida FQAI surveys. As such, each botanist was
sent a list of species identified in the forested strand and floodplain wetlands that did not
already have a CC score for the previous Florida FQAI efforts (n = 86). Botanists
participating in the 2003 FQAI effort for forested strand and floodplain wetlands included
Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and Bruce Tatje. Botanists









scored each species based on its faithfulness to Florida forested strand and floodplain
wetlands. Potential CC scores ranged from 0 to 10:
0 Exotic and native species that act as opportunistic invaders,
included species that commonly occur in disturbed ecosystems
1-3 Species that were widely distributed and occurred in disturbed
ecosystems
4-6 Species with a faithfulness to a particular ecosystem, but also
tolerant of moderate levels of disturbance
7-8 Species typical of well-established ecosystems that sustain only
minor disturbances
9-10 Species occurring within a narrow set of ecological conditions
Species with low CC scores were considered tolerant of many disturbances, whereas
species with high CC scores were considered to occur within a narrow set of stable
ecological condition. Appendix B lists the CC scores for the macrophyte species
identified in this study.

Data Analysis

Summary Statistics
Summary statistics for the macrophyte assemblage were calculated at the species
level, including richness (R), evenness (E), Shannon diversity (H'), Simpson diversity
(D), and Whittaker's beta diversity (1w). Richness (R) was defined as the total number of
distinct taxa identified within the sample wetland. Evenness (E), described as the
fraction of maximum possible diversity in a wetland (McCune and Grace 2002), was
calculated as the Shannon diversity (H') value divided by the natural log of richness:
E = H'/ In (R) (2-3)
The Shannon diversity index has been described as measuring the "information content"
of a sample unit where maximum diversity yields maximum uncertainty (McCune and
Grace 2002). For Shannon diversity calculations (H'), the sample unit was an individual
forested wetland:
H'= -Y pi log(pi) (2-4)
pi= ni / N (2-5)
where ni was the number of occurrences of taxon i, and N was the total number of
occurrences of all taxa at a wetland. The number of occurrences represented the number
of quadrats a species occurred in, and the total number of occurrences of all taxa at a
wetland represented the sum of the total number of quadrats of all of the species
identified.
Simpson diversity (the compliment of Simpson's original index, which was a
measure of dominance) was a measure of the likelihood species that are randomly chosen
in sampling will be different. The equation used to calculate Simpson diversity was:


D = I pi'


(2-6)









Whittaker's beta diversity (1w) was computed as a calculation of overall beta
diversity, or the compositional change represented in a sample. Whittaker's beta
diversity was calculated as the number of species at a particular forested wetland (Sc)
divided by the average species richness per quadrat (S), minus one:
pw = [Sc / S] 1 (2-7)
The resulting value for Whitaker's beta diversity was described as the "number of distinct
communities" (McCune and Grace 2002). When Ow equaled zero, all of the sample units
contained all of the species. Some multivariate methods strongly depend on beta
diversity, and as a general rule beta diversity greater than 5 was considered high
(McCune and Grace 2002).
Summary statistic means within a priori land use categories were compared with
Fisher's Least Significant Difference (LSD) pair wise comparison test using Minitab
(Version 13.1, Minitab Statistical Software). The strength of using Fisher's LSD was in
the comparison of unequal group sizes (Ott and Longnecker 2001; Minitab 2000).
Sample wetlands were divided into two groups based on 1995 LDI_F/wo scores
including low (LDI < 2.0) and high (LDI > 2.0) LDI groups. Comparisons were made for
summary statistics between low and high LDI groups using the non-parametric Mann-
Whitney U-Test in Minitab (Ott and Longnecker 2001).
Overall calculations of beta and gamma diversity were calculated for sample
wetlands in the three apriori land use categories. Gamma diversity was calculated as the
overall number of taxa encountered at all sample wetlands per apriori land use category.
Higher gamma diversity for an a priori land use category suggested a greater difference
among the species composition of wetlands within that a priori land use category,
assuming a similar number of wetlands were sampled within each a priori land use
category. Beta diversity was calculated as a priori category gamma diversity divided by
the mean site taxa richness.

Regional Compositional Analysis
The Multi-Response Permutation Procedure (MRPP) was used to test the
similarity of macrophyte community composition among Lane's (2000) four Florida
wetland regions (see further application of MRPP in Zimmerman et al. 1985; McCune et
al. 2000; McCune and Grace 2002). MRPP is a nonparametric technique which tests for
no difference between groups (the null hypothesis) and is available in PCORD (Version
4.1 from MJM Software, Gleneden Beach, Oregon). It is an appropriate procedure for
ecological community data as it does not require distributional assumptions of normality
and homogeneity of variances. The Sorensen distance measure was used to calculate the
average weighted within-group distance.
MRPP analysis provided a test statistic (T), p-value, and chance-corrected within-
group agreement (A), which described within-group similarity. When A equaled one, all
items were identical within groups, and when A equaled zero, differences within-groups
equaled that expected by chance. McCune and Grace (2002) found that most values of A
were less than 0.1 in community ecology. MRPP was calculated across all wetland
region groups (panhandle versus north versus central versus south; wetland regions
according to Lane 2000) as well as for multiple pair wise comparisons (panhandle versus
north, panhandle versus central, panhandle versus south, north versus central, north
versus south, and central versus south) when sample size was appropriate. An appropriate









sample size was defined by a minimum of two wetlands within each category for
comparison. For example, if there were seven wetlands being compared with six in one
region and only one in a different region, the MRPP comparison analysis could not be
executed. Additional MRPP tests were calculated across all bioregions (panhandle versus
northeast versus peninsula versus Everglades; bioregions according to Griffith et al.
1994) as well as for multiple pair wise comparisons (panhandle versus northeast,
panhandle versus peninsula, panhandle versus Everglades, northeast versus peninsula,
northeast versus Everglades, and peninsula versus Everglades) when sample size was
appropriate.

Community Composition
Community composition of the macrophyte assemblage was summarized in a
non-metric multidimensional scaling (NMDS) ordination. NMDS, an ordination
technique designed to compress multi-dimensional space, is particularly agreeable with
ecological data because it does not rely on linear relationships among variables, which
generally do not accurately characterize ecological trends. This has been described as a
compensation for the "zero-truncation problem" through the use of ranked distances (a
common characteristic of non-parametric analyses) and the use of an appropriate distance
measure (McCune and Grace 2002). The "zero-truncation problem" refers to the
extraordinary number of zeros (denoting a species is absent from a sampling location)
typical in community ecology data sets. For example, in the species by site matrix used
in our data analysis (283 species by 24 sites), 85% of the cells within the matrix were
assigned a zero (i.e. the species does not occur at the site). Other ordination techniques
depend upon a value for each measured variable for each sample unit. However NMDS
is useful for the presence/absence data sets, where many species are not present, or
receive a zero in the species by site matrix. By ranking the variables, NMDS caters to the
non-parametric community ecology data set.
NMDS was used to explore the dissimilarities of the community composition of
the macrophyte assemblage among sample wetlands. The Sorensen (Bray-Curtis)
distance measure was used for ordination. The dimensionality was chosen based on an
initial 6-dimensional run in autopilot mode, which suggested an optimal 2-dimensional
solution. To find the optimal 2 dimensional solution, 50 runs with real data and 50
randomized runs were performed with the instability criterion set at 0.00001 and the
maximum number of iterations to reach a stable solution set at 500. This procedure was
repeated 20 times to insure stability and reproducibility in results. The final run was
completed with the starting point set as the results from the best experimental 2-
dimensional run, with the lowest stress and best overall fit. LDIF/wo, species richness,
Shannon diversity, Simpson diversity, Whittaker's beta diversity, and sampling date (as
Julian date) were tested for correlation with the NMDS ordination axes with Pearson
correlation coefficients.

Metric Development

In the context of this study, metrics are defined as biological attributes which have
a consistent and predictable response to anthropogenic activities (Karr and Chu 1997).
Metrics were summarized in three main categories, including tolerance indicators,









community structure, and community balance. Metrics were as the proportion (P) by
dividing the number of species (n) fitting a particular metric category (ex. exotic species)
divided by the total species richness (N) for each sample wetland:
Pi = ni /Ni (2-8)
where i represents an individual wetland.
Scatter plots were constructed for each candidate metric versus 1995
LDIF/wo_200m to ensure correlations were visually distinguishable. Candidate metrics
were accepted if they successfully distinguished between low (LDI < 2.0) and high LDI
(LDI > 2.0) groups using the Mann-Whitney U-test (p < 0.10) calculated with Version
13.1, Minitab Statistical Software, and showed consistent and predictable trends along the
gradient of anthropogenic land use activity according to the strength and significance of
the Spearman's correlation coefficient (p < 0.05) calculated with Analyze-It software v.
1.67 for Microsoft Excel. The Spearman rank correlation tested for an association
between two related variables, and was a non-parametric alternative to the Pearson
correlation.
Macrophyte metrics were adopted from previous FWCI studies (Lane et al. 2003;
Reiss and Brown 2005), included tolerance (tolerant species; sensitive species), exotic,
FQAI, and longevity and plant growth form (native perennial). Additional metrics were
explored for summary statistics (richness, evenness, and diversity), as well as dominance,
annual to perennial ratio, wetland indicator status, native evergreen, native deciduous,
and basal area (total, deciduous, or evergreen).
Tolerance metrics were calculated with Indicator Species Analysis (ISA) in
PCORD (Version 4.1 from MJM Software, Gleneden Beach, Oregon). Sample wetlands
were categorized into low and high LDI groups and analyzed with ISA, which evaluates
the abundance and faithfulness of species in a defined group (McCune and Grace 2002).
ISA is used to detect and describe the value of species indicative of environmental
conditions. It requires a priori groups and data on the presence of taxa in each group.
Groups are commonly defined by categorical environmental variables, levels of
disturbance, experimental treatments, presence and absence of a target species, or habitat
types (McCune and Grace 2002). Calculated indicator values are based on two standards,
faithfulness and exclusion. Faithfulness is defined mathematically by a particular species
always being present in a particular group. Additionally, the perfect indicator species is
exclusive to that group, meaning it never occurs in other groups. Indicator values range
from 0 (no indication) to 100 (a perfect indication of a particular group).
Multiple ISA iterations were conducted to determine sensitive and tolerant
indicator species. Sample wetlands were categorized based on 1995 LDIF/wo_200m
scores, and ISA calculations were completed for consecutive LDI breaks from 1.0
through 3.5 at each 0.5 increment using the presence/absence of taxa at each wetland. In
total ISA was run six times. Species significantly associated (p < 0.10) with the higher
LDI group (LDI > 1.0; LDI > 1.5; LDI > 2.0; LDI > 2.5; LDI > 3.0; LDI > 3.5) for each
iteration were included in the overall list of tolerant indicator species; species
significantly associated (p < 0.10) with the lower LDI group (LDI = 1.0; LDI < 1.5; LDI
< 2.0; LDI < 2.5; LDI < 3.0; LDI < 3.5) for each iteration were included in the overall list
of sensitive indicator species. Exotic species were excluded from the sensitive indicator
species list, as this list was constructed to represent reference biological condition, which









should exclude the presence of nonnative species. A random number seed was used for
each ISA iteration.
Calculated indicator values were tested for statistical significance using a Monte
Carlo randomization technique with 1000 randomized runs. Indicator species categorized
as tolerant species were associated with a higher LDI group; indicator species categorized
as sensitive species were associated with a lower LDI group. Tolerant and sensitive
indicator species lists were compiled based on a collection of significant indicator species
at each incremental LDI break. This method was employed due to the small sample size
of wetlands available to construct indicator species lists. The proportion indicator species
metrics were calculated for each wetland as the number of indicator species (either
tolerant or sensitive) divided by species richness.
The proportion exotic species metric was calculated as the number of species that
were exotic to Florida divided by species richness. The timeline for determining the
exotic status of a species was set near the beginning of European settlement in North
America. Sources consulted to determine the exotic status of a species included Godfrey
and Wooten (1981), Tobe et al. (1998), Wunderlin (1998), USDA NRCS (2002), and
Wunderlin and Hansen (2003).
FQAI scores for each wetland were calculated based on the sum of CC scores for
each species identified divided by total species richness at each wetland (Eq. 2-2). FQAI
scores had the potential to range from 0.0-10.0. Wetlands with higher FQAI scores
represented wetlands with species indicative of more stable environmental conditions.
The proportion native perennial species was calculated as the number of native
perennial species encountered divided by wetland species richness. Each species was
categorized as native or exotic and annual or perennial according to Godfrey and Wooten
(1981), Tobe et al. (1998), Wunderlin (1998), USDA NRCS (2002), and Wunderlin and
Hansen (2003). Wienhold and Van der Valk (1989) and Ehrenfeld and Schneider (1991)
determined that disturbance often favors annual species over perennial species and
promotes the invasion of nonnative perennials in wetlands. Galatowitsch et al. (2000)
found that while native perennial cover was reduced in wetlands impacted by cultivation,
the occurrence of introduced perennials rather than annuals increased in stormwater
impacted wetlands.

Florida Wetland Condition Index

The Florida Wetland Condition Index (FWCI) consists of individual metrics,
which were scaled and added together. Metric scoring was based on an approach from
the Stream Condition Index (SCI), a Florida based biological index of the
macroinvertebrate assemblage used to discern stream condition (Fore 2004) and also used
in the development of the FWCI for depressional forested wetlands (Reiss 2004).
Metrics were natural log transformed to improve the normality of distribution. The 5th to
95th percentile values of each metric were normalized from 0 to 10, with 10 always
representing the reference biological wetland condition.
An agglomerative cluster analysis in PCORD was used to determine wetland
groups (McCune and Grace 2002), for wetlands with similar macrophyte community
composition. The dissimilarity matrix was constructed using the Sorensen distance
measure and the flexible beta linkage method (3 = -0.25), which is a flexible clustering









setting designed to reduce chaining in the dendrogram. The resulting dendrogram was
pruned using ISA, which provided an objective, quantitative means of selecting the
optimum number of clusters representing the most ecologically meaningful point of
pruning (McCune and Grace 2002). Group membership determined in each step of the
agglomerative cluster analysis was used as the grouping variable for ISA. Using the
species by site presence/absence matrix, indicator values were calculated for each species
occurring in three or more sites for each step of the dendrogram. This is an entirely
separate application of ISA as that used to construct the metrics for tolerant and sensitive
indicator species. This method of using ISA to determine the most appropriate and
meaningful point on a dendrogram for understanding a significant number of groups for a
data set is described more completely by McCune and Grace (2002). The resulting p-
values from the randomized Monte Carlo tests were averaged among all species at each
step of the dendrogram, and the cluster step with the lowest average p-value was used to
determine the level in the dendrogram yielding the greatest ecologically meaningful
information (McCune and Grace 2002). FWCI scores and LDI index values among
clusters were then compared using Fisher's LSD pair wise comparison (p < 0.05)
available in Minitab.
To determine how the biological data compared to the intensity of landscape
development, Spearmans rank correlation was used to compare individual metrics and the
FWCI with the multiple LDI calculations. For forested floodplain wetlands, comparisons
of the FWCI and LDIF/wo_200m were made with the HDG and SCI scores using the
non-parametric Spearmans rank correlation coefficient (HDG and SCI scores were not
available for forested strands). The LDIF/wo_200m and HDG represented measures of
anthropogenic influence; whereas the FWCI and SCI represented biological indices based
on the macrophyte community composition in the forested wetland and
macroinvertebrate community composition in the channelized stream flow, respectively.











CHAPTER 3
RESULTS


Gradients of Anthropogenic Activity

Two measures of anthropogenic activity in the landscape were correlated to
physical, chemical, and biological measures of wetland condition, including the
Landscape Development Intensity (LDI) index and the Human Disturbance Gradient
(HDG).

Landscape Development Intensity
LDI index calculations were completed for the forested strand (Table 3-1) and
floodplain (Table 3-2) wetlands. Transect (LDI_T) and feature (LDI F) scale
calculations were calculated for all 24 wetlands using 1995 land use coverages and 2000
land use coverages for wetlands located within the boundary of the St Johns River Water
Management District (SJRWMD). Land uses in the 100 m buffers for LDI_T
calculations were delineated from digital orthophoto imagery and assigned land use
categories based on ground truthed land use maps drawn during the field site visit.
Watershed (WS) scale LDI calculations were completed for the 14 forested floodplain
wetlands.
A comparison between LDI calculations including (1995_LDI_T/w_100m) and
excluding (1995 LDIT/wo_100m) wetland area (Figure 3-1) show that calculations
excluding the wetland area (1995 LDI_T/wo_100m) for agricultural and urban wetlands
were consistently higher than those including the wetland area (1995 LDIT/w_100m)
for the forested wetlands, as can be seen by the values generally occurring above the 1:1
line. Four reference wetlands received a score of 1.0 using both calculation methods. In
effect the calculation of LDI_T/w compensates for differences in wetland area between
study wetlands. However, this difference in area may or may not correlate to differences
in biological response such as changes in macrophyte community composition to
anthropogenic activities in the landscape.
The correlation between 1995 LDI calculations at the transect and feature scales
with 200 m buffers (Figure 3-2) was low (R2 = 0.32). Because the forested strand and
floodplain wetlands involve channelized flowing water, it is anticipated that the feature
scale LDI would better reflect the biological response signal of organisms that are
affected by upstream anthropogenic activities. Three wetlands, one in each a priori land
use category, received identical scores for the transect (1995 LDI_T/wo_200m) and
feature scales (1995 LDIF/wo_200m). Ten wetlands received higher 1995
LDI F/wo 200m than 1995 LDI_T/wo_200m scores, including six reference, one
agricultural, and three urban wetlands; whereas the remaining 11 wetlands received
higher 1995 LDI_T/wo_200m scores, including five agricultural and six urban wetlands.
Watershed scale LDI calculations using equal weighting for all area within a
watershed, excluding the wetland area, (1995 LDIWSED/wo) and a distance weighted












Table 3-1. Landscape Development Intensity (LDI) index scores for 10 forested strands using 1995 and 2000 land use coverages.
LDI calculations were completed at the transect (T) and wetland feature (F) scale, and including (/w) and excluding (/wo) the
wetland area. See text for further details of individual calculations.


FS1 FS2 FS3 FS4 FS5 FS6 FS7 FS8 FS9 FS10
A Priori Group R A A U U U U A R R
1995 LDI T/w 100m 1.2 2.4 2.3 1.2 1.9 3.7 1.0 2.6 1.3 1.2
1995 LDI T/wo 100m 1.3 3.5 3.3 1.4 2.3 4.6 1.0 3.5 1.4 1.2
1995 LDI T/w 200m 1.3 2.6 2.4 1.2 2.1 3.5 1.0 2.9 1.4 1.0
1995 LDI T/wo 200m 1.3 3.2 3.5 1.3 3.5 3.9 1.0 3.2 1.5 1.0
1995 LDI F/w 200m 2.6 2.5 3.1 1.9 3.0 3.0 1.0 3.0 1.7 1.0
1995 LDI F/wo 200m 3.0 2.9 3.1 1.9 3.5 3.2 1.0 3.1 2.1 1.0
2000 LDI T/w 100m 1.2 2.6 2.3 4.3 5.4 5.8 1.0 6.1 1.0 1.0
2000 LDI T/wo 100m 1.2 3.8 2.5 8.0 7.0 7.5 1.0 8.0 1.0 1.2
2000 LDI T/w 200m 1.5 2.6 2.3 na na na na 6.9 1.4 na
2000 LDI T/wo 200m 1.5 3.2 2.3 na na na na 7.4 1.6 na
2000 LDI F/w 200m 2.5 2.5 2.6 na na na na 6.9 1.7 na
2000 LDI F/wo 200m 2.9 2.9 3.0 na na na na 7.2 2.1 na
na 2000 land use coverages were not available for these wetlands.












Table 3-2. Landscape Development Intensity (LDI) index scores for 14 forested floodplain wetlands using 1995 and 2000 land use
coverages. LDI calculations were completed at the transect (T), wetland feature (F), and watershed (WS) scale, and including (/w)
and excluding (/wo) the wetland area. See text for further details of individual calculations.

Site Code FF1 FF2 FF3 FF4 FF5 FF6 FF7 FF8 FF9 FF10 FF11 FF12 FF13 FF14
A Priori Land Use U R U A U R R U U U A A R A
1995 LDI T/w 100m 1.9 1.0 5.2 1.5 1.9 1.1 1.6 3.2 1.0 4.8 1.2 4.2 1.0 1.2
1995 LDI T/wo 100m 3.2 1.0 7.6 2.2 3.3 1.1 2.1 4.4 1.0 6.9 1.6 6.3 1.0 1.5
1995 LDI T/w 200m 2.2 1.0 5.9 1.2 2.3 1.0 1.0 3.3 1.1 4.0 1.3 4.3 1.0 1.7
1995 LDI T/wo 200m 3.1 1.0 7.1 1.3 3.1 1.0 1.0 3.9 1.1 4.7 1.6 6.5 1.0 2.1
1995 LDI F/w 200m 1.5 1.1 3.2 1.6 1.7 1.4 1.9 2.0 1.5 5.7 1.6 2.4 1.9 1.7
1995 LDI F/wo 200m 2.0 1.2 3.7 2.3 1.9 1.4 2.5 2.5 1.8 7.1 1.6 2.8 2.1 1.8
1995 LDI WS ED/w 2.4 1.5 3.8 2.2 1.8 2.7 3.1 2.2 1.7 2.1 1.4 1.9 1.8 2.0
1995 LDI WS ED/wo 2.5 1.5 2.9 2.3 1.9 2.7 3.4 2.5 1.7 2.1 1.4 1.9 1.8 2.1
1995 LDI WS DW lin 2.2 1.4 2.2 2.2 2.1 2.7 2.8 2.3 1.5 2.2 1.4 2.0 1.7 2.0
1995 LDIWSDWexp 2.3 1.4 2.7 2.2 2.1 2.6 2.6 2.6 1.5 2.8 1.4 2.8 1.6 1.9
2000 LDI T/w 100m 1.9 1.0 5.2 1.5 1.9 1.1 1.6 3.3 1.0 5.6 1.2 2.9 1.0 1.2
2000 LDI T/wo 100m 3.2 1.0 7.6 2.2 3.3 1.1 2.1 4.5 1.0 8.1 1.6 5.2 1.0 1.5
2000 LDI T/w 200m 2.2 1.0 5.9 1.0 2.2 na na 3.4 1.3 na na 4.1 na na
2000 LDI T/wo 200m 3.1 1.0 7.1 1.1 3.1 na na 4.8 1.4 na na 5.6 na na
2000 LDI F/w 200m 1.5 1.1 3.0 1.6 1.8 na na 2.0 1.5 na na 2.3 na na
2000 LDI F/wo 200m 2.0 1.3 3.5 2.2 2.0 na na 2.5 1.7 na na 2.7 na na
2000 LDI WS ED/w 2.3 1.5 2.3 2.2 1.7 na na 2.3 1.6 na na 1.9 na na
2000 LDI WS ED/wo 2.4 1.5 2.3 2.3 1.8 na na 2.6 1.8 na na 1.9 na na
2000 LDI WS DW lin 2.2 1.5 2.4 2.2 1.9 na na 2.3 1.7 na na 2.0 na na
2000 LDI WS DW exp 2.2 1.4 2.8 2.2 1.9 na na 2.6 1.6 na na 2.8 na na
na 2000 land use coverages were not available for these wetlands.












R2 0.98



1:1 Line


1/ Strands
2 Floodplains
1:1 Line
0
0 /"-^----------------

0 2 4 6 8 10
1995 LDI T/w 100m



Figure 3-1. Comparison between the LDI calculations including (1995 LDI_T/w_100m)
and excluding (1995 LDI_T/wo_100m) wetland area (R2 = 0.98). Calculations
excluding the wetland area (1995 LDI_T/wo_100m) for agricultural and urban
wetlands were consistently higher than those including the wetland area (1995
LDI_T/w_100m) for the forested wetlands. Four reference wetlands received a score
of 1.0 using both calculation methods.


R2 0.32


Strands
Floodplains
1:1 Line


0 2 4 6 8 10
1995 LDI F/wo 200m



Figure 3-2. Comparison between LDI calculations at the transect
(1995 LDIT/wo_200m) and feature (1995 LDIF/wo_200m) scale for 24 forested
wetlands (R2 = 0.32).


13
13 a P










approach using linear weighting (1995 LDIWS DWlin) showed a stronger correlation
(R2 = 0.82) for the 14 forested floodplain wetlands (Figure 3-3).
LDI calculations were also completed for multiple years, using land use/land
cover data from 1995 and 2000. Comparison between 1995 and 2000 LDI calculations at
the feature scale (without the wetland area) for 13 forested wetlands showed a moderate
correlation (R2 = 0.41) (Figure 3-4). However, 12 of the wetlands had nearly equal LDI
scores for the 1995 and 2000 LDIF/wo_200m calculations with one clear disagreement,
FS8, which had a significant area of the 200m buffer categorized as Low Intensity
Pasture (with livestock) in 1995 and Institutional in 2000. The land use was not actually
changed, only the classification on the GIS coverage was changed, as the property is part
of the University of Florida's (thus the Institutional classification) Beef Research Unit
(thus the Low Intensity Pasture (with livestock) classification). Removing FS8 because
of erroneous differences in the land use/land cover coverages, the remaining 12 wetlands
had strongly correlated 1995 and 2000 LDIF/wo_200m calculations (R2 = 0.99 without
FS8).
LDI F/wo 200m LDI calculations for 1995 land use/land cover were selected for
further analysis for four reasons. First, 1995 land use/land cover data were available for
all of the 24 forested wetlands sampled. Second, calculations for 13 forested wetlands
having both 1995 and 2000 LDIF/wo_200m calculations showed similar values as
described above (Figure 3-4). Using the older land use/land cover data from 1995 may
have limited implications in changes to LDI scores as changes in land use are localized to
areas with recent development. In these areas it may be argued that more recent
coverages are needed, however at the statewide scale the 1995 land use coverage
provides a baseline for LDI calculations. We are not suggesting that the most recent and
complete coverage should not be used, but simply that the LDI calculation should not be
disregarded because the coverage is ten years old. Third, because the 24 wetlands
surround channelized flowing water, the feature scale LDI calculations were used to
capture the human activity influencing the upstream wetlands. Fourth, the LDI T_100m
calculations included hand delineation of land use according to interpretation of digital
orthophoto imagery and ground truthing; however, the application of LDI as a remote
(GIS computer based) tool for assessing human activity in the landscape, using widely
available coverages, was an important consideration for future application.

Human Disturbance Gradient
Thirteen of the forested floodplain wetlands sampled corresponded to storet
stations established by FDEP for water quality monitoring. Data were available for the
four components of the Human Disturbance Gradient (HDG), including ammonia
concentration, hydrologic condition, habitat assessment (habitat condition index), and
LDI for the buffer (Table 3-3). Using the scoring criteria established by FDEP (Table 1-
2), each of the HDG components were scored individually and summed to get the HDG
for each storet station. HDG scores were adopted directly from previous FDEP
assessments, and no scoring was done directly for this research. HDG scores potentially
range from 0-9, with 0 representing the reference condition and 9 representing a highly
impaired condition. The maximum HDG score for the 13 forested floodplain wetlands
was a 4 at the urban wetland FF10 (storet 28020234) for the most recent sampling period















R2 = 0.82



*t*


.*


* Floodplains
1:1 Line


1995 LDI WS ED/wo


Figure 3-3. Comparison between LDI calculations at the watershed scale using equal
weighting for all area within a watershed (1995 LDIWSED/wo) and a distance
weighted approach using linear weighting (1995 LDIWSDWlin) (R2 = 0.82).


R2 0.41










o Strands
* Floodplains
1:1 Line


1995 LDI F/wo 200m


Figure 3-4. Comparison between 1995 and 2000 LDI calculations
(without the wetland area) for 13 forested wetlands (R2 = 0.41).


at the feature scale


2
1
21













Table 3-3. Human Disturbance Gradient (HDG) for 13 storet stations, which correspond with the forested floodplains sampled. Some
forested floodplain wetlands have multiple HDG calculations for different sampling dates (1992-2000). The four components of
the HDG were also provided, including ammonia concentration, Hydrologic Condition, Habitat Index, and LDI buffer. HDG
scoring criteria were presented in Table 1-2.

Ammonia (NH3) NH3 Hydrologic Hydro Habitat Habitat LDI LDI BF
Code STORET HDG Date HDG (mg N/L) SCORE Condition SCORE Assessment SCORE Buffer SCORE
FF1 20020404 8/15/2000 0 0 4 0 84 0 1.38 0
FF1 20020404 8/15/2000 0 0.04 0 4 0 84 0 1.38 0
FF2 20010454 7/19/1999 0 0.01 0 1 0 88 0 1.06 0
FF3 27010580 8/22/2000 2 0.10 1 4 0 58 1 1.79 0
FF4 20010455 2/23/1999 0 -- -- 2 0 84 0 1.02 0
FF5 19010099 8/17/1993 1 0 3 0 65 1 1.27 0
FF5 19010099 3/5/1998 0 -- -- 3 0 88 0 1.27 0
FF5 19010099 2/1/2000 0 0.04 0 3 0 91 0 1.27 0
FF5 19010099 7/9/1996 0 0.02 0 3 0 90 0 1.27 0
FF5 19010099 8/18/1992 0 -- -- 3 0 91 0 1.27 0
FF6 26011019 3/8/1999 0 -- -- 3 0 86 0 1.51 0
FF7 26011020 7/20/1999 0 0 2 0 89 0 1.04 0
FF8 19020027 7/15/1996 1 0.08 0 7 1 88 0 1.40 0
FF9 20030246 4/17/1997 0 0 2 0 99 0 1.21 0
FF10 28020234 10/9/2000 4 0.12 1 7 1 81 0 4.73 2
FF10 28020234 9/11/1996 3 0.06 0 7 1 86 0 4.73 2
FF11 21010032 1/17/1995 0 -- -- 3 0 97 0 1.16 0
FF12 19010072 3/22/1999 0 0 2 0 86 0 1.56 0
FF14 32010024 7/14/1998 0 0.02 0 5 0 75 0 1.33 0
no data available
*below detection limits









(10/9/2000). [Multiple data points were available for many of the SCI points used in this
research project, and the most complete data from the most recent sampling period was
used for correlation purposes.] The moderate to low range of scores suggests that the
wetlands sampled represented wetlands on the lower scale of impairment. This may be
an accurate statement when considering that many stream sampling locations with SCI
data were visited during site reconnaissance, but found unacceptable for this study
because the floodplain wetland had been completed altered, often because the area had
been sodded or paved as part of adjacent human activity in the surrounding landscape.
These sites, considered highly impaired on the SCI, were not considered for macrophyte
sampling and FWCI development, as they no longer hosted wetland vegetation.

Water Quality
Water quality data (chemical and physical parameters) were obtained for 13 of the
forested floodplain wetlands sampled that corresponded to storet stations. Water
parameters available included temperature (C), pH, specific conductivity (umhos/cm),
dissolved oxygen (mg 02/L), turbidity (NTU), Total Kjeldahl Nitrogen (mg/L), ammonia
(as mg N/L), and total phosphorus (mg P/L) (Table 3-4). FF5 an urban wetland had the
lowest values for temperature (7.1 C), specific conductivity (49 umhos/cm), and the
highest value for dissolved oxygen (10.2 mg 02/L). In contrast FF8, another urban
wetland, had the highest values for turbidity (5.7 NTU), Total Kjeldahl Nitrogen (1.50
mg N/L), and total phosphorus (0.28 mg P/L). A third urban wetland FF10 had the
lowest value for dissolved oxygen (2.6 mg 02/L), and the highest values for temperature
(25.8 C) and ammonia nitrogen (0.120 mg N/L).

Data Analysis

Statewide, 24 wetlands were sampled with 283 macrophyte species, representing
190 genera and 97 families identified. The most abundant species was the vine Vitis
rotundifolia (muscadine), which was found rooted within vegetation quadrats at 23 of 24
study wetlands. The second most abundant species was the tree Acer rubrum (red maple)
found at 75% of the study wetlands. The most common fern was Woodwardia virginica
(Virginia chain fern) found at 54% of the wetlands; the most common graminoid was
Panicum commutatum (variable witchgrass) found at 46% of the wetlands. Boehmeria
cylindrica (false nettle) was the most common herbaceous species, found at 63% of the
wetlands; the two most common shrubs were Cephalanthus occidentalis (buttonbush) and
Myrica cerifera (wax myrtle), each found at 71% of the wetlands; the second most
common tree was Sabalpalmetto (cabbage palm) found at 71% of the wetlands; and the
second most common vine was Toxicodendron radicans (Eastern poison ivy) found
at71% of the wetlands. Of the species encountered, 110 species (39%) occurred at a
minimum of three sample wetlands (12.5%). Nearly one-fifth of the species identified
(48 species or 17%) were identified at only two sample wetlands; approximately two-
fifths of the species identified (126 species or 44%) were found at only one wetland. In
total the forested strand wetlands hosted 161 species, and the forested floodplain
wetlands hosted 214 species. The forested strand and floodplain wetlands had 91 species
in common.














Table 3-4. Water quality (chemical and physical parameters) for 13 storet stations, which correspond with the forested floodplains
sampled. Three forested floodplain wetlands (FF1, FF5, and FF10) have multiple data from different sampling dates (1993-2000).

Specific Dissolved Total Kjeldahl Ammonia Total
Code Date STORET Temperature pH Conductivity Oxygen Turbidity Nitrogen Nitrogen Phosphorus
C umhos/cm mg /L NTU mg/L mg/L mg/L


FF1 8/15/2000
FF1 8/15/2000
FF2 7/19/1999
FF3 8/22/2000
FF4 2/23/1999
FF5 8/17/1993
FF5 3/5/1998
FF5 2/1/2000
FF5 7/9/1996
FF6 3/8/1999
FF7 7/20/1999
FF8 7/15/1996
FF9 4/17/1997
FF10 10/9/2000
FF10 9/11/1996
FF11 1/17/1995
FF12 3/22/1999
FF14 7/14/1998
- no data available


20020404
20020404
20010454
27010580
20010455
19010099
19010099
19010099
19010099
26011019
26011020
19020027
20030246
28020234
28020234
21010032
19010072
32010024


23.7
23.7
23.2
25.0
12.8
25.0


121
121
1580
746
230
70


6.2
6.2
6.8
2.7
7.2


1.2
1.8


0.75
0.75
0.27
0.62


0.004
0.038
0.012
0.100


0.000

0.039
0.024

0.000
0.080
0.000
0.120
0.059

0.000
0.017


0.08
0.08
0.04
0.13


0.04
0.02
0.02
0.02

0.00
0.28
0.00
0.18
0.11

0.05
0.02









Summary Statistics
Species richness (R), species evenness (E), Shannon diversity (H'), Simpson
diversity (D), and Whittaker's beta diversity (PW) were calculated based on the
macrophyte assemblage for each sample wetland. Species richness ranged from 21
species at FF2 (a reference floodplain forested wetland), to 77 species at FF5 (a wetland
embedded in an urbanized landscape). Species evenness was consistent among all a
priori land use groups, and species evenness only showed differences at the hundred
thousandth decimal place, explaining why reported values were identical for two decimal
places. Shannon diversity ranged from 3.05 to 4.34 at FF2 and FF5, respectively; these
were the same wetlands with the lowest (21) and highest (77) species richness,
respectively. Similarly, FF2 had the lowest Simpson diversity index value (0.95), and
FF5 and FF8 (another urban wetland) shared the highest Simpson diversity index value
(0.99). Whittaker's beta diversity ranged from a low of 1.8 at FF6 (a reference wetland),
to a high of 7.4 at FF4 (an agricultural wetland). A complete table of summary statistics
for each wetland is presented in Appendix C.
Table 3-5 summarizes comparisons of mean richness, evenness, and diversity
calculations by apriori land use category. Urban wetlands had the mean greatest species
richness (50 + 17) followed by agricultural wetlands (41 + 10). This trend was evident
for both Shannon diversity, with urban wetlands having greater mean values (3.86
0.35). Whittaker's beta diversity was greatest in agricultural wetlands (4.94 1.07).
Beta and gamma diversity were calculated for overall a priori land use categories, with
urban wetlands having the highest beta and gamma diversity (4.0 and 201, respectively).
None of the summary statistics were significantly different among a priori land use
categories (Fisher's LSD pair wise comparison, a = 0.05) as the standard deviations
overlap for most of the summary statistics.
Similarly, summary statistics (including species richness, species evenness,
Shannon diversity, and Whittaker's beta diversity) were not significantly different
between low (1995 LDIF/wo_200m < 2.0) and high (1995 LDI_F/wo_200m > 2.0) LDI
groups (Mann-Whitney U-Test; Table 3-6). However, wetlands in the high LDI group
had higher (though not statistically significant) mean species richness (47 17), Shannon
diversity (3.80 + 0.34), Whittaker's beta diversity (4.56 1.44), and overall beta (4.9)
and gamma (233) diversity.

Regional Compositional Analysis
MRPP was calculated across all wetlands grouped according to wetland regions
(Lane 2000), as well as for multiple pair wise comparisons. Sample size limitations
prevented all possible wetland region pair wise comparisons when a minimum of two
wetlands per wetland region and land use category were not available. Results for the
MRPP tests included the test statistic (T), chance-corrected within-group agreement (A),
and significance value (p) (Table 3-7). Only the MRPP pair wise comparison of the
panhandle and north wetland regions for the compiled pool of all wetlands (including a
priori reference, agricultural, and urban) was not significant at the a = 0.05 level,
suggesting that there were regionally significant differences among species composition
across all wetland regions. Of the possible pair wise comparisons within a priori land
use categories, the MRPP analysis did not suggest a difference among species
composition in north and central reference wetlands or agricultural wetlands for all









Table 3-5. Richness, evenness, and diversity of
priori land use categories.


the macrophyte assemblage among a


Reference Agricultural Urban
Species richness (R) 36 + 11a 41 + 10a 50 17
Species evenness (E) 1.00 + 0.00a 1.00 + 0.00a 1.00 + 0.00a
Shannon diversity (H') 3.55 + 0.32a 3.69 + 0.25a 3.86 + 0.35a
Simpson's diversity (D) 0.97 0.01a 0.97 0.01a 0.98 0.01a
Whittaker's Beta diversity (PW) 3.53 1.03 4.94 1.07T 4.57 1.38a
Beta diversity 3.7 3.7 4.0
Gamma diversity 134 153 201
Categories with similar letters were not significantly different (Fisher's LSD, a=0.05)


wetland regions. However, all MRPP tests among urban wetlands suggested a difference
in the species composition of wetlands among the wetland regions (multiple comparison
panhandle versus north versus central versus south; pair wise north versus central; and
pair wise north versus south). While limitations to interpretation are apparent due to
small sample sizes within each wetland region, the differences in species composition
among urban wetland showed the greatest variability.
The MRPP analysis was repeated using wetlands grouped according to bioregions
(Griffith et al. 1994). Of the forested wetland sampled, only one was located in the
panhandle bioregion and one in the Everglades bioregion. All of the reference wetlands
(n=7) were located in the peninsula bioregion. Thus the only comparisons possible were
between a priori agricultural and urban wetlands in the northeast and peninsula
ecoregions, limiting the MRPP analysis to two comparisons. A significant difference in
species composition was not found among agricultural or urban wetlands in the northeast
and peninsula bioregions (Table 3-7).


Table 3-6. Richness, evenness, and diversity of the macrophyte assemblage between low
(LDI < 2.0) and high (LDI > 2.0) LDI groups.

Low LDI High LDI W p
Species richness (R) 39 + 11 47 + 17 120 0.32
Species evenness (E) 1.00 + 0.00 1.00 + 0.00 153 0.38
Shannon diversity (H') 3.62 0.31 3.80 + 0.34 120 0.32
Simpson's diversity (D) 0.97 + 0.01 0.98 + 0.01 120 0.32
Whittaker's Beta diversity (PW) 4.16 1.44 4.56 + 1.44 123 0.42
Beta diversity 4.3 4.9
Gamma diversity 169 233


W' = Mann-Whitney U-test statistic


p` signi icance value










Table 3-7. Macrophyte community composition similarity among wetland regions (Lane
2000) and bioregions (Griffith et al. 1994) with MRPP.

Sites (n) T^ A' p#


Florida Ecoregions (Lane 2000)
All wetlands
All regions (P vs N vs C vs S)
Panhandle vs north
Panhandle vs central
Panhandle vs south
North vs central
North vs south
Central vs south
Reference wetlands
North vs central
Agricultural wetlands
All regions (P vs N vs C vs S)
Urban wetlands
All regions (P vs N vs C vs S)
North vs central
North vs south
Bioregions (Griffith et al. 1994)
Agricultural wetlands
Northeast vs peninsula


-4.2
0.2
-1.9
-2.4
-3.2
-3.7
-4.0


0.07 0.00*
0.56
0.04 0.04*
0.12 0.02*
0.04 0.01*
0.07 0.01*
0.07 0.00*


6 -1.2 0.04 0.10


7 0.4


-4.4
-2.8
-3.9


0.18
0.13
0.15


0.65

0.00*
0.01*
0.01*


6 -1.5 0.12 0.07


Urban Wetlands
Northeast vs peninsula 9 -1.6 0.05 0.07
A high ITI value and significant p-value (p<0.05) suggests a difference in species
composition.
ST = the MRPP test statistic
SA = the chance corrected within-group agreement
#p = the significance value.


Community Composition
Macrophyte community composition was summarized in a 2-dimensional
nonmetric multidimensional scaling (NMDS) ordination to explore gradients in
macrophyte community composition (Figure 3-5). The final solution had an overall
stress of 17.8 with a final stability of 0.0002, which is considered a fair stress value
useful for ordinations with community data sets (Kruskal 1964; Clarke 1993; McCune
and Grace 2002). Axis 1 explained 53.3% variance, and axis 2 explained 29.2%
additional variance. No measured variables were correlated with the ordination axes,
including 1995 LDI F/wo_200m, species richness, Shannon diversity, Simpson diversity,
Whittaker's beta diversity, and Julian date.
Wetlands appeared to be broadly grouped on axis 1 according to wetland region,
and perhaps the gradients of latitude and longitude would have been appropriate














FF3
FF14 FF8
59


FF12
FF9
0


FF4
FF5 F
Am


FF2
0



Ecoregions (Lane 2000)
V South
Central
North
Panhandle


FF10
v


FS1 FF13
*-0-


Axis 1: 53.3%


FS6


FS10
V


Figure 3-5. NMDS ordination bi-plot of 24 sample wetlands in macrophyte species
space; wetlands are labeled according to site code; symbols correlate to wetland
regions (Lane 2000). Axis 1 explained 53.3% variance; axis 2 explained an additional
29.2% variance.


explanatory variables for the ordination axes. For purposes of speculation, the second
axis may be correlated to wetland type, as only two of the forested strands (FS5 and FS7)
fell above the horizontal line of axis 2 at the center of the plot; whereas the remaining
eight strands fell on or below the horizontal line. In contrast, only three of the forested
floodplain wetlands (FF2, FF11, and FF13) fell below the horizontal line; while, the
remaining 11 floodplain wetlands fell on or above the horizontal line. Perhaps there were
distinct differences in the macrophyte community composition between wetland types
(strand versus floodplain) detected with the NMDS, which could be supported by the
observation that the forested strand and floodplain wetlands in this study shared just 91
species. This correlates to 43% of the species identified at any of the 14 forested
floodplain wetlands also occurring at any of the 10 forested strand wetlands, which had
fewer species identified (161 species) meaning that 57% of the species identified within
the forested strands also occurred in at least one floodplain wetland. However,









considering that 126 species (45% of the entire 283 species identified for this study) were
only sampled at one of the 24 wetlands included in this study, perhaps the dissimilarity of
species composition among wetland type (strand versus floodplain) is not as significant
as some other unmeasured environmental variable. A meaningful grouping of wetlands
based on apriori land use classification was not apparent.

Metric Selection
Five metrics that were significantly correlated with LDI (Spearmans correlation
coefficient |r|>0.50, p<0.05) were selected for inclusion in the preliminary FWCI for
forested strand and floodplain wetlands (Table 3-8). Metrics selected for inclusion were
the proportion tolerant indicator species (TOL); proportion sensitive indicator species
(SEN); Floristic Quality Assessment Index (FQAI); proportion exotic species (EX); and
proportion native perennial species (NP). The proportion of tolerant indicator species
and proportion exotic species increased with increasing development intensity; whereas,
the proportion sensitive indicator species, FQAI, and proportion native perennial species
decreased with increased landscape development intensity. All metrics significantly (p <
0.05) differentiated between low (LDI<2.0) and high (LDI>2.0) 1995 LDIF/wo_200m
groups (Table 3-9).

Tolerance metrics
Tolerant and sensitive indicator species were determined statewide using
Indicator Species Analysis (PCORD). Table 3-10 provides a list of 19 tolerant indicator
species. Figure 3-6 shows a scatter plot of proportion tolerant indicator species versus
1995 LDIF/wo_200m. The proportion tolerant indicator species increased with
increasing development intensity. FF10 (an urban floodplain forest in the south wetland
region (Lane 2000) or Everglades bioregion (Griffith et al. 1994) had the highest
proportion tolerant indicator species (0.29), followed by FS5 (0.28) and FF7 (0.25),
which were both in the central wetland region (Lane 2000) or peninsula bioregion
(Griffith et al. 1994). Two wetlands in the peninsula bioregion (Griffith et al. 1994) had
less than 0.05 proportion tolerant indicator species, including FS10 a reference strand in
the south wetland region (Lane 2000) and FF2 a reference floodplain in the north wetland
region (Lane 2000).


Table 3-8. Spearmans correlation coefficients for macrophyte metrics and FWCI with
1995 LDI F/wo 200m.

Metric Spearman's r p-value
Proportion tolerant indicator species 0.65 0.001
Proportion sensitive indicator species -0.83 <0.001
FQAI score -0.50 0.012
Proportion exotic species 0.54 0.006
Proportion native perennial species -0.56 0.004
FWCI -0.75 <0.001









Table 3-9. Comparisons among five macrophyte metrics and the FWCI between low
(LDI < 2.0) and high (LDI > 2.0) LDI groups (LDI F/wo_200m).

Metric Low LDI High LDI W p


Tolerant indicator species 0.11 + 0.06 0.19 + 0.06 115.5 0.007
Sensitive indicator species 0.15 + 0.08 0.05 0.05 14 0.001
FQAI score 4.72 0.68 4.10 + 0.64 34 0.036
Exotic species 0.03 0.04 0.10 + 0.08 109 0.022
Native perennial species 0.95 0.06 0.86 + 0.11 34 0.036


FWCI 38.33 + 7.91
Values represent the mean + standard deviation
W = the Mann-Whitney U-Test statistic
p' = the significance value


21.54 12.00


Table 3-11 provides a list of the 16 sensitive indicator species. Figure 3-7 shows
that the proportion sensitive indicator species decreased with increasing development
intensity. The wetland with the highest proportion sensitive indicator was FF2 (0.38; a
reference wetland), followed by and FF9 (0.18; an urban wetland) and FS7 (0.17; a
reference wetland). Three wetlands had no sensitive indicator species including two in
the central wetland region (Lane 2000) or peninsula bioregion (Griffith et al. 1994) (the
urban strand FS5 and the agricultural strand FS3) and one wetland in the south wetland
region (Lane 2000) or Everglades bioregion (Griffith et al. 1994) (FF10).
Shrub and tree species were included in the ISA for both tolerant and sensitive
metrics. Metrics developed based on the macrophyte community composition included
woody species rooted within the sampling quadrats, as structure was thought to play an
important role in the biological condition of flowing water forested wetlands. Excluding
the tree and shrub layers would seemingly underscore the importance of these woody
species. In fact, tree and shrub species comprised 52.6% of the tolerant and 62.5% of the
sensitive indicator species lists (Tables 3-10 and 3-11). The five tolerant indicator tree
species were Acer rubrum (red maple), Ilex cassine (dahoon holly), Liquidambar
styraciflua (sweetgum), Nyssa sylvatica var. biflora (swamp tupelo), and Prunus
caroliniana (Carolina laurelcherry) (Table 3-10); the five tolerant indicator shrub species
were Callicarpa americana (American beautyberry), Cephalanthus occidentalis
(buttonbush), Ludwigia peruviana (water-primrose), Sambucus canadensis (elderberry),
and Viburnum obovatum (Walter viburnum). Vines also comprised a high percentage of
the tolerant indicator species list, including Berchemia scandens (rattan vine), Smilax
auriculata (earleaf greenbrier), Smilax tamnoides (bristly greenbrier), and Toxicodendron
radicans (Eastern poison ivy).
Of the sensitive indicator species, 43.8% were shrubs, 18.8% trees, 18.8% herbs,
12.5% graminoids, 6.3% ferns, and 0% vines (Table 3-11). The three sensitive indicator
tree species included Fraxinus caroliniana (Carolina ash), Persea palustris (swamp bay)
and Pinus elliottii (slash pine). Seven shrub species were categorized as sensitive
indicator species, including Agarista populifolia (Florida hobble-bush), Hypericum
hypericoides (St. Andrew's cross), Ilex coriacea (bay-gall holly), Lyonia lucida (fetter-


0.001


x












Table 3-10. Tolerant indicator species for forested strand and floodplain wetlands.


Botanical Name Common Name LDI Break (Indicator Value, p-value) Growth Form
Acer rubrum Red Maple 1.0 (81.8, 0.057) Tree
1.5 (90, 0.005)
Berchemia scandens Rattan Vine 3.5 (58.3, 0.062) Vine
Bidens alba Hairy Beggar-Ticks 3.5 (54.9, 0.097) Herb
Callicarpa americana American Beautyberry 3.5 (75, 0.053) Shrub
Cephalanthus occidentalis Buttonbush 1.0 (77.3, 0.071) Shrub
1.5 (61, 0.072)
Commelina diffusa Dayflower 2.0 (50.6, 0.034) Herb
3.0 (67.3, 0.015)


Hydrocotyle verticillata
Ilex cassine
Liquidambar styraciflua
Ludwigia peruviana
Nyssa sylvatica var. biflora
Panicum rigidulum
Prunus caroliniana
Rhynchospora miliacea
Sambucus canadensis
Smilax auriculata
Smilax tamnoides
Toxicodendron radicans


Pennywort
Dahoon Holly
Sweetgum
Water-Primrose
Swamp Tupelo
Red-Top Panicum
Carolina Laurelcherry
Millet Beakrush
Elderberry
Earleaf Greenbrier
Bristly Greenbrier
Eastern Poison Ivy


3.0
1.5
1.5
3.5
1.5


(43.7, 0.068)
(55, 0.095)
(60, 0.092)
(54.9, 0.087)
(60, 0.092)
(47.4, 0.031)
(27.3, 0.092)
(43.7, 0.072)
(27.3, 0.09)
(72.4, 0.086)
(37.7, 0.051)
(77.3, 0.07)
(61, 0.068)
(57.3, 0.069)
(58.3, 0.052)


Viburnum obovatum Walter viburnum Shrub


Herb
Tree
Tree
Shrub
Tree
Graminoid
Tree
Graminoid
Shrub
Vine
Vine
Vine


Viburnum obovatum


Walter viburnum


Shrub











0.3
A*)
10


0.2

**

[


Floodplains
o 0
S0o Strands
0.0
0 2 4 6 8 10
1995 LDI F/wo 200m

Figure 3-6. The proportion of tolerant indicator species at wetlands increased with
increasing development intensity (LDI).


bush), Rhododendron viscosum (swamp azalea), Vaccinium arboreum sparkleberryy), and
Vaccinium corymbosum highbushh blueberry). The two sensitive indicator graminoid
species were Cladiumjamaicense (saw-grass) and Panicum hemitomon (maidencane).

Floristic Quality Assessment Index metric
Wetland FQAI scores decreased with increasing 1995 LDI_F/wo (Figure 3-8). Of
the 283 species identified in the flowing water wetlands, 17 were assigned coefficient of
conservatism (CC) scores of zero, and 11 additional species received scores less than 1.0.
Of the species receiving a zero CC score, nine (53%) were listed as Category I invasive
exotics, and two (12%) were listed as Category II invasive exotics (EPPC 2003). [The
rankings of Category I or II invasive exotics are from the Florida Exotic Pest Plant
Council (EPPC), which focuses on identifying exotic pest species. Category I species
include those invasive exotic species considered responsible for changes to native plant
communities through the displacement of natives, changes in community structure or
ecological functions, and hybridizing with natives, based on documented ecological
damage (EPPC 2003). Category II species have not yet altered native plant communities,
but have increased in abundance or frequency and may become ranked as Category I with
confirmed ecological damage (EPPC 2003).]
The species with the highest CC score was Pieris phyllyreifolia (climbing fetter-
bush) (9.5), followed by Panicum abscissum (cut-throat grass) (9.22), Taxodium
ascendens (pond cypress) (8.8), Asplenium heterochroum (bicolored spleenwort) (CC =
8.5), and Pinckneya bracteata (fever-tree) (8.3). A complete list of CC scores is
available in Appendix B. Wetland FQAI scores were significantly correlated with LDI












Table 3-11. Sensitive indicator species for forested strand and floodplain wetlands.


Botanical Name
Agarista populifolia
Centella asiatica
Cladium jamaicense
Eupatorium capillifolium
Fraxinus caroliniana
Hypericum hypericoides
Ilex coriacea
Lyonia lucida


Panicum hemitomon
Persea palustris
Phlebodium aureum


Common Name
Florida Hobble-Bush
Coinwort
Saw-Grass
Dog Fennel
Carolina Ash
St. Andrew's Cross
Bay-Gall Holly
Fetter-Bush


Maidencane
Swamp Bay
Golden Polypody


LDI Break
2.5
2.5
1.5
1.0
2.5
2.5
2.5
2.5
3.0
2.0
2.5
1.0
1.5


(Indicator Value, p-value)
(30.8, 0.092)
(38.6, 0.082)
(45.5, 0.059)
(88, 0.034)
(38.6, 0.064)
(30.8, 0.095)
(38.5, 0.037)
(53.6, 0.017)
(52.9, 0.054)
(38.6, 0.042)
(61.2, 0.006)
(95.7, 0.008)
(45.5, 0.065)


Growth Form
Shrub
Herb
Graminoid
Herb
Tree
Shrub
Shrub
Shrub


Graminoid
Tree
Fern


Pinus elliottii Slash Pine 1.0 (95.7, 0.008) Tree
1.5 (45.5, 0.065)
Rhododendron viscosum Swamp Azalea 2.5 (38.6, 0.081) Shrub
Saururus cernuus Lizard's Tail 3.0 (64.7, 0.01) Herb
Vaccinium arboreum Sparkleberry 2.5 (30.8, 0.098) Shrub
Vaccinium corymbosum Highbush Blueberry 3.0 (52.9, 0.038) Shrub











* Floodplains
o Strands


0 2 4 6 8 10

1995 LDI F/wo 200m


Figure 3-7. The proportion sensitive indicator species at
increasing development intensity (LDI).


4


3


2


S8

I OH

0


wetlands decreased with


* Floodplains
o Strands


0 2 4 6 8 10

1995 LDI F/wo 200m


Figure 3-8. FQAI scores decreased with increasing landscape development intensity.


D3*
of


0.2



0.1



0.0


I I -I a I


- I


13









gradient (Irl = 0.50; p = 0.012; Table 3-8). When wetlands were divided into two groups
based on 1995 LDI_F/wo_200m (LDI < 2.0 and LDI > 2.0), there was a significant
difference between FQAI scores (U = 34; p < 0.05) (Table 3-9).
The range of wetland FQAI scores was 3.0, though the scale of species CC scores
ranged from 0.0-9.5. The wetland with the highest FQAI was FF2 (5.6), a reference
floodplain forest in the north (Lane 2000) or peninsula stream (Griffith et al. 1994)
ecoregions. The wetland receiving the lowest FQAI score was FF10 (2.7), an urban
floodplain forest in southwest Florida in the south wetland region (Lane 2000) or
Everglades bioregion (Griffith et al. 1994). Three wetlands in the low LDI group with
low FQAI scores included FS7 (4.3), FS10 (3.9), and FF6 (3.4). Sixty-seven percent of
the wetlands in the low LDI group had an FQAI score greater than 4.5; whereas 73% of
wetlands in the high LDI group had an FQAI score less than 4.5.

Exotic species metric
The proportion of exotic species at a wetland was significantly correlated with the
gradient of development intensity in the surrounding landscape (Irl = 0.54, p = 0.006;
Table 3-8). Figure 3-9 shows that the proportion of exotic species increased with
increasing LDI. The south wetland region (Lane 2000; or Everglades bioregion from
Griffith et al. 1994) hosted the wetland with the greatest proportion exotic species, FF10
(0.29), which also received the lowest FQAI score (2.7). The wetland with the second
highest proportion exotic species was FS3 (0.18), an agricultural strand surrounded by
cattle pasture. Six floodplain wetlands had no exotic species present, including two
reference wetlands (FF2 and FF7), two agricultural wetlands (FF4 and FF11), and two
urban wetlands (FF5 and FF9).
The proportion of exotic species at wetlands was significantly different between
low and high LDI groups (W = 109, p = 0.022; Table 3-9). Table 3-12 lists the 35 exotic
species encountered throughout the forested strand and floodplain wetlands. The most
common exotic species was the herbaceous species Commelina diffusa (dayflower) found
at 10 (42%) of the 24 forested wetlands, which was also categorized as a tolerant
indicator species (Table 3-10). Sixty-nine percent of the exotic species (n=24) were
found at only one forested wetland. Thirty-seven percent of the species (n=13) were
listed as Category I invasive exotic species, and 14% (n=5) were listed as Category II
invasive exotic species (EPPC 2003). Forty-three percent of the exotic species were
herbs, 20% vines, 14% shrubs, 11% trees, 9% graminoids, and 3% ferns.

Native perennial species metric
Of the 283 macrophyte species identified, 234 (83%) were classified as native
perennials. Figure 3-10 shows that the proportion of native perennial species at a wetland
decreased with increasing development intensity. The native perennial species metric
was significantly correlated with LDI (Spearman Irl = 0.56, p = 0.004; Table 3-8); and
there was a significant difference between the proportion native perennial species at low
and high LDI group wetlands (W = 34, p = 0.036; Table 3-9). FF10 had the lowest
proportion native perennial species (0.60). Four wetlands hosted entirely native perennial
species, including one reference wetland (FF2), two agricultural wetlands (FF4 and
FF 11), and one urban wetland (FF9).











0.3
*


'13
0.2
o B
0



0 *0
S* Floodplains
139 e Strands
E 0







0.0 ,-

0 2 4 6 8 10

1995 LDI F/wo 200m


Figure 3-9. The proportion of exotic species at a wetland increased with increasing
development intensity.


Florida Wetland Condition Index
Five metrics were included in the preliminary FWCI for forested strand and
floodplain wetlands, including proportion tolerant indicator species, proportion sensitive
indicator species, FQAI score, proportion exotic species, and proportion native perennial
species. Metrics were natural log transformed to improve normality. Metric scoring was
done on a continuous scale spreading between the 5th to 95th percentiles of the values for
the sampled wetlands. Scores for each metric were then added together to create the
preliminary forested strand and floodplain FWCI with a scale of 0-50, with 50
representing the reference condition of high biological integrity. Appendix D provides
metric scoring criteria.
Figure 3-11 shows that FWCI scores decreased with increasing development
intensity. Correlations between macrophyte metrics and FWCI with LDI were significant
(p < 0.05) for all of the metrics and the FWCI (Irl = 0.75, p < 0.001) (Table 3-8). One
wetland, FF2 a reference forested floodplain wetland, received a perfect 50 on the FWCI
scale, also receiving a perfect 10 score for all five metrics. One wetland scored the
lowest potential score of zero, FF10 which was an urban forested floodplain wetland in
the south wetland region (Lane 2000) or Everglades bioregion (Griffith et al. 1994). All
of the wetlands in the low LDI group scored above the midpoint of 25 on the FWCI
scale; while 64% of wetlands in the high LDI group scored below 25.









Table 3-12. Exotic species identified at 24 forested strand and floodplain wetlands.

Botanical Name Common Name EPPC Growth Form
Abrus precatorius Rosary Pea I Vine
Alternanthera philoxeroides Alligator Weed II Herb
Alternanthera sessilis Sessile Alligator Weed Herb
Begonia cucullata Wax Begonia II Herb
Cinnamomum camphora Camphor Tree I Tree
Colocasia esculenta Elephant's Ear I Herb
Commelina communis Asiatic Dayflower Herb
Commelina diffusa Dayflower Herb
Cuphea carthagenensis Columbia waxweed Herb
Cynodon dactylon Bermudagrass Graminoid
Cyperus difformis Variable Flatsedge Graminoid
Emiliafosbergii Florida Tasselflower Herb
Eugenia uniflora Surinam Cherry I Shrub
Koelreuteria elegans Flamegold II Tree
Ligustrum sinense Chinese Privet I Shrub
Lonicerajaponica Japanese Honeysuckle I Vine
Ludwigia peruviana Water-Primrose Shrub
Lygodium japonicum Japanese Climbing Fern I Vine
Lygodium microphyllum Small-leaf Climbing Fern I Vine
Melaleuca quinquenervia Punk Tree; Melaleuca I Tree
Merremia dissecta Alamo Vine Vine
Mimosa pigra Black Mimosa I Shrub
Momordica charantia Balsampear Vine
Oeceoclades maculata Monk Orchid Herb
Paspalum notatum Bahiagrass Graminoid
Phyllanthus urinaria Water Leafflower Herb
Sapium sebiferum Chinese Tallowtree I Tree
Schefflera actinophylla Australian Umbrella Tree I Tree
Schinus terebinthifolius Brazilian Pepper I Shrub
Thelypteris dentata Downy Maiden Fern Fern
Trifolium repens White Clover Herb
Urena lobata Caesarweed II Herb
Wisteria sinensis Chinese Wisteria II Vine
Xyrisjupicai Richard's Yellow-Eyed-Grass Herb
Youngia japonica Oriental False Hawksbeard Herb
Exotic Pest Plant Council (EEPC) categories from EEPC (2003).












[]
O *


* Floodplains
o Strands


* *OD
]


0 2 4 6 8 10

1995 LDI F/wo 200m


Figure 3-10. The proportion of native perennial species
development intensity (LDI).


decreased with increasing


50 *
*


O3
40 o

s*
300


* Floodplains
o Strands


10


0 to LDI 200m
0 2 4 6 8 10

1995 LDI F/wo 200m


Figure 3-11. Forested Wetland Condition Index (FWCI)
increasing development intensity (LDI).


scores decreased with


1.0


a 0.9


0.8


" 0.7


e 0.6









Cluster Analysis
Cluster analysis determined wetland groupings based on macrophyte community
composition using a species by site presence/absence matrix. Based on the lowest
average p-value for all species from the randomized Monte Carlo tests used in the ISA
analysis, the most ecologically meaningful cluster for dendrogram pruning was at cluster
step 5 (Figure 3-12) with the lowest average species p-value of 0.27. The highest number
of significant (p < 0.05) indicator species was found at cluster step 3, which had 37
significant indicator species (Figure 3-13). Cluster steps 4 and 5 had the second highest
number of significant indicator species at 31. Exploration of the sample sites assigned to
the groups at cluster step 3 suggested an association between group membership and
spatial distribution of the wetlands throughout Florida. [Note: ISA used to determine the
most ecologically meaningful clusters for dendrogram pruning is separate from ISA used
in determining tolerant and sensitive indicator species. For ISA used to determine
meaningful clusters for dendrogram pruning, group membership was assigned based on
the wetland groupings established through cluster analysis. For ISA used to determine
sensitive and tolerant indicator species, group membership was assigned based on the
calculated LDI score for each wetland.] The groups defined through the agglomerative
cluster analysis at cluster step 5 were roughly defined by wetland regions (Lane 2000),
bioregions (Griffith et al. 1994), and apriori land use categories, including:
Cluster 1: FF2, a north wetland region, peninsula bioregion, reference
floodplain;
Cluster 2: Nine floodplains, including two in the panhandle, six in the
north, and one in the central wetland regions; one in the panhandle,
five in the northeast, and three in the peninsula bioregions; and four
agricultural and five urban wetlands;
Cluster 3: Five strand and two floodplain wetlands; one in the north and
six in the central wetland regions; all seven in the peninsula bioregion;
four reference, one agricultural, and two urban wetlands;
Cluster 4: Three strand and two floodplains, including one in the north
and four in the south wetland regions; four in the peninsula and one in
the Everglades bioregions; three reference and two urban wetlands;
Cluster 5: Two floodplain wetlands with one in the north and one in the
central wetland regions; both in the peninsula bioregion; and both
agricultural wetlands.
Wetland groups based on the third cluster step combined the aforementioned groups from
the cluster step 5, so that:
Cluster 1: Combines Clusters 1 & 2
Cluster 2: Composed of Cluster 3
Cluster 3: Combines Clusters 4 & 5
It appear that the initial grouping for cluster step three was based on spatial location and
that further cluster steps separate out wetlands based on impairment. Figure 3-14 shows
that based on wetland groups from cluster step 3, FWCI scores for wetlands in Cluster 1
(37.5 10.5) were significantly different from wetlands in Cluster 2 (24.4 10.0) and
Cluster 3 (19.8 13.1) (p < 0.05). Wetland FWCI scores for Cluster 2 and Cluster 3
were not significantly different from one another. Regionalization was apparent in the
wetland groupings of Cluster 3 (at cluster step 3); as Cluster 3 was comprised of the all of













0.6

0.5 -

0.4

0.3

S0.2

0.1

0.0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Number of Clusters

Figure 3-12. Change in average species p-value from the randomized Monte Carlo tests
at each step in clustering. The minimum average p-value (0.27) was found at cluster
step 5.





40


20
o 30


20

.a
10
-o

0


2 3 4 5


6 7 8 9 10 11 12 13 14 15 16 17 18 19 20


Number of Clusters

Figure 3-13. Change in the number of significant indicator species from the indicator
species analysis performed at each step in clustering. The maximum number of
significant indicator species (37) was found at cluster step 3 (p < 0.05), followed by 31
significant indicator species at cluster steps 4 and 5.












50 -
a
40 -

30 b

20-

10 -

0_
I I I
Cluster 1 Cluster 2 Cluster 3


Figure 3-14. FWCI scores for three wetland clusters based on macrophyte community
composition. Boxes represent 25th and 75th quartiles, lines represent median FWCI
scores per cluster, dots represent the mean, and vertical lines represent the range.
Clusters with similar letters were not significantly different (p<0.05).


the south wetland region wetlands (Lane 2000), including the only wetland sampled in
the Everglades bioregion (Griffith et al. 1994). Table 3-13 provides means and standard
deviations for cluster FWCI and LDI F/wo 200m scores. LDI F/wo 200m scores were
not significantly different among clusters.

Landscape Development Intensity Index
and the Florida Wetland Condition Index

Nearly all of the correlation (110 of 120, or 92%) between the five macrophyte
metrics and the FWCI with the 20 variations of LDI were significantly correlated at the
flexible p < 0.10 level (Table 3-14). Additionally, nearly half of the comparisons (58 of
120, or 48.3%) were significantly correlated at the strictest significance level (p < 0.01).
Three of the LDI calculations (1995 LDI_F/w_200m, 2000 LDI_F/w_200m, and 2000
LDI_F/wo_200m) were significant at the strictest significance level (p < 0.01) for all five
metrics and the FWCI, which should not be surprising given that metrics were selected
for inclusion in the FWCI based on correlations with the 1995 LDIF/wo_200m (using
Spearman correlation with LDI, visually distinguishable and ecologically meaningful
pattern when graphed with LDI, and differentiation among LDI groups with the Mann
Whitney U-test).
The proportion tolerant indicator species metric was significantly correlated with
1995 and 2000 LDI_F (wetland feature scale) calculations (p < 0.01). However, the
proportion tolerant indicator species metric was not significantly correlated at the strictest
significance level (p < 0.01) for any of the 1995 or 2000 LDI_T transectt scale)









Table 3-13. FWCI scores and LDI values for wetland clusters (at the third cluster step)
based on macrophyte community composition.


Wetland Clusters FWCI LDI F/wo 200m
1 37.5 + 10.5a 2.2 0.7a
2 24.4 + 10.0b 2.5 0.7a
3 19.8 13.1b 2.9 2.1a
Clusters with similar letters within columns were not significantly different (p < 0.05).
LDIF/wo_200m represents the LDI calculated at the feature scale excluding the wetland
area within a 200 m buffer.


calculations. The proportion sensitive indicator species metric was more strongly
correlated with the LDI calculations at the transect and feature scales (11 of 12
correlations with |r| > 0.63, p < 0.01, remaining one at Irl = 0.57, p < 0.05) than the
watershed scales (1995 and 2000 LDI WS DW_exp at p < 0.01; 2000 LDI WS DW lin
at p < 0.05; 2000 LDIWS_ED/wo not significant; remaining four LDIWS correlations
at p < 0.10).
The FQAI score metric was not significantly correlated with three of the 1995
transect level LDI calculations, but was strongly (I|r > 0.66) and significantly (p < 0.01)
correlated with all of the 1995 and 2000 watershed scale LDI calculations. Watershed
scale LDI calculations were completed for floodplain forested wetlands only, leaving
speculation as to whether these results suggest that FQAI scoring reflect landscape level
anthropogenic activity (e.g. exotic species with low CC scores entering a system due to
anthropogenic activities) or whether FQAI scoring was biased towards larger flowing
water systems such as floodplain forests. The proportion exotic species and proportion
native perennial species metrics were more strongly correlated with the feature scale LDI
calculations than transect or watershed scales. In fact, three watershed 1995 LDI
calculations were not significant with the proportion exotic species or the proportion
native perennial species metrics, including 1995 LDIWSED/w, 1995
LDIWSED/wo, and 1995 LDIWSDWlin.
The strongest correlation between the FWCI and LDI was with the 2000
LDIWS_DW_exp (I|r = 0.95; p < 0.01), though the correlation with 2000
LDIF/wo_200m (I|r = 0.94; p < 0.01) was also remarkably strong. The two correlations
with the highest Spearmans correlation coefficients were between the proportion tolerant
indicator species metric and the 2000 LDIWS_DW_exp and between the FQAI score
metric and the 2000 LDIWSDWlin (I|r = 0.96; p < 0.01).













Table 3-14. Correlations of metrics and FWCI scores with 20 variations of the LDI index. Differences in LDI calculations include
1995 or 2000 land use; transect (T), feature (F), or watershed (WS) scale buffers; including (/w) or excluding (/wo) the wetland
area; 100 (100m) or 200 meter (200m) buffers; equal distance (ED) or distance weighted (DW); and linear (lin) or exponential
(exp) weighting. Column headings refer to the five metrics and the FWCI: proportion tolerant indicator species (TOL), proportion sensitive indicator
species (SEN), Floristic Quality Assessment Index (FQAI), proportion exotic species (EX), and proportion native perennial species (NP).


LDI r
1995 LDI T/w 100m
1995 LDI T/wo 100m
1995 LDI T/w 200m
1995 LDI T/wo 200m
1995 LDI F/w 200m
1995 LDI F/wo 200m
1995 LDI WS ED/w
1995 LDI WS ED/wo
1995 LDI WS DW lin
1995 LDIWSDW_exp
2000 LDI T/w 100m
2000 LDI T/wo 100m
2000 LDI T/w 200m
2000 LDI T/wo 200m
2000 LDI F/w 200m
2000 LDI F/wo 200m
2000 LDI WS ED/w
2000 LDI WS ED/wo
2000 LDI WS DW lin
2000 LDI WS DW exp
*p<0.01 **p<0.05 ***p<0.10 ns


1= TOL
24 0.44
24 0.41
24 0.38
24 0.45
24 0.72
24 0.66
14 0.60
14 0.59
14 0.58
14 0.80
24 0.51
24 0.51
13 0.54
13 ns
13 0.78
13 0.82
8 0.73
8 0.65
8 0.83
8 0.96
- not significant


SEN
-0.68
-0.65
-0.63
-0.69
-0.90
-0.83
-0.49
-0.49
-0.49
-0.72
-0.72
-0.72
-0.69
-0.57
-0.93
-0.92
-0.68
ns
-0.80
-0.88


FQAI
-0.39
ns
ns
ns
-0.57
-0.50
-0.71
-0.67
-0.69
-0.75
-0.49
-0.42
-0.74
-0.66
-0.88
-0.91
-0.95
-0.86
-0.96
-0.88


EX
0.51
0.46
0.49
0.51
0.61
0.54
ns
ns
ns
0.60
0.48
0.44
0.83
0.73
0.79
0.82
0.77
0.73
0.77
0.83


NP
-0.48
-0.43
-0.44
-0.44
-0.63
-0.55
-0.50
ns
ns
-0.65
-0.54
-0.49
-0.85
-0.77
-0.85
-0.84
-0.83
-0.69
-0.81
-0.78


FWCI
-0.58
-0.52
-0.49
-0.52
-0.79
-0.75
-0.57
-0.53
-0.53
-0.70
-0.63
-0.59
-0.81
-0.72
-0.91
-0.94
-0.88
-0.78
-0.92
-0.95









Human Disturbance Gradient and Stream Condition Index

Thirteen of the forested floodplain wetlands had Human Disturbance Gradient
(HDG) and Stream Condition Index (SCI) scores for sampling events taken within the
channelized flow of the forested floodplain wetlands (Table 3-15). Sample dates for the
FWCI and LDI were during 2003 (Table 2-1); sample dates for the HDG and SCI ranged
from 1995-2000 (Table 3-15). Table 3-15 provides storet identification numbers, HDG
and SCI sample dates, a priori land use categories, bioregions (Griffith et al. 1994),
wetland regions (Lane 2000), FWCI, LDIF/wo_200m scores, HDG scores, and SCI
scores for the 13 forested floodplain wetlands. Data for the most recent and complete
sampling event for each wetland, as presented in Table 3-15, was selected for correlation
purposes.
The HDG was significantly correlated with the SCI (Irl = 0.58, p<0.05),
LDIF/wo_200m (Irl = 0.66, p<0.05), and FWCI (Irl = 0.74, p<0.01) (Table 3-16).
However, the SCI was not significantly correlated with either the LDIF/wo_200m or the
FWCI. The correlation between the LDIF/wo_200m and FWCI (Irl = 0.68, p<0.05)
(Table 3-16) for the 13 forested floodplain wetlands (only those with HDG and SCI
scores), was weaker than the correlation between them for the entire data set of 24
forested flowing water wetlands (Irl = 0.75, p<0.001) (Table 3-8).













Table 3-15. Forested Wetland Condition Index (FWCI), Landscape Development Intensity Index (LDI F/wo_200m), Human
Disturbance Gradient (HDG), and Stream Condition Index (SCI) data for 13 forested floodplain wetlands.


Wetland
Region
North
North
North
Central
North
Central
Central
North
North
South
Panhandle
North
Panhandle


FWCI
39.1
50.0
16.4
43.6
44.0
26.2
28.2
23.6
46.6
0.0
42.0
33.5
36.5


LDI F/wo 200m
2.0
1.2
3.7
2.3
1.9
1.4
2.5
2.5
1.8
7.1
1.6
2.8
1.8


HDG
0
0
2
0
0
0
0
1
0
4
0
0
0


SCI
80
70
20
75
55
80
90
70
95
50
25
85
80


STORET ID refers to Florida Department of Environmental Protection (FDEP) database ID.
HDG/SCI Date refers to original sample date corresponding to HDG and SCI data. FWCI and LDI were calculated for the 2003 site visit.
Land Use refers to a priori land use category (R-reference, A-agricultural, U-urban)
Bioregion from Griffith et al. (1994)
Wetland Region from Lane (2000)
LDI F/wo 200m refers to the LDI calculated at the feature scale, excluding the wetland area, within a 200 m buffer.


Site
FF1
FF2
FF3
FF4
FF5
FF6
FF7
FF8
FF9
FF10
FF11
FF12
FF14


STORET
ID
20020404
20010454
27010580
20010455
19010099
26011019
26011020
19020027
20030246
28020234
21010032
19010072
32010024


HDG/SCI
Date
8/15/2000
7/19/1999
8/22/2000
2/23/1999
2/1/2000
3/8/1999
7/20/1999
7/15/1996
4/17/1997
10/9/2000
1/17/1995
3/22/1999
7/14/1998


Land Use
U
R
U
A
U
R
R
U
U
U
A
A
A


Bioregion
Peninsula
Peninsula
Peninsula
Peninsula
Northeast
Peninsula
Peninsula
Northeast
Northeast
Everglades
Northeast
Northeast
Panhandle






63


Table 3-16. Correlations among four measures of ecosystem condition or anthropogenic
activity, including the Human Disturbance Gradient (HDG), Stream Condition Index
(SCI), Landscape Development Intensity Index (1995 LDIF/wo_200m), and the
Florida Wetland Condition Index (FWCI) for freshwater forested floodplain wetlands.
Values are Spearmans rank correlation coefficients.


HDG SCI LDI F
SCI -0.58*
LDI F 0.66 ns
FWCI -0.74 ns -0.68*
*p<0.01, **p<0.05, ns=not significant











CHAPTER 4
DISCUSSION


The primary objective of this research was to develop a preliminary Florida
Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland
study sites were sought in various a priori designated land use categories that included
undeveloped, agricultural, and urban land uses. The preliminary FWCI provides a
quantitative measure of the biological integrity of forested strand and floodplain wetlands
in Florida. Comprised of five metrics, the preliminary FWCI was developed based on the
community composition of the macrophyte species assemblages (Table 4-1). Metrics
were selected for inclusion in the FWCI based on the correlation (nonparametric
Spearman correlation coefficient) of each metric with the Landscape Development
Intensity (LDI), an independent measure of anthropogenic activity in the landscape
calculated for each wetland (Brown and Vivas 2005); based on a metrics visually
distinguishable correlation with LDI in a scatter plot; and based on a statistical difference
of metric values between low and high LDI groups (Mann-Whitney U-test). The FWCI
was composed of individual metrics, which were scaled and added together, creating the
preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest
score of 50 reflecting the highest biological integrity and the lowest score of zero
reflecting a lack of biological integrity or no similarity to the reference wetland condition.
The contribution of this research to our understanding of changes in the
macrophyte community composition of forested strand and floodplain wetlands in
relation to differing anthropogenic activities in the surrounding landscape can be
summarized in five main points:
1. Five macrophyte based metrics including proportion tolerant indicator
species, proportion sensitive indicator species, Floristic Quality
Assessment Index (FQAI) score, proportion exotic species, and proportion
native perennial species, were useful biological indicators for defining
biological integrity for forested strand and floodplain wetland vegetation;
2. Vegetation richness, evenness, and diversity were not sensitive to apriori
land use categories or development intensities in the surrounding
landscape for forested strand and floodplain wetlands;
3. The Landscape Development Intensity (LDI) index was a useful tool
correlating with the measured biological condition of vegetation for
forested strand and floodplain wetlands;
4. Regional species lists for metrics would enhance the forested strand and
floodplain Florida Wetland Condition Index (FWCI);
5. An FWCI with a set of core metrics could be developed for Florida
freshwater wetlands, which includes separate species lists for indicator
species by wetland type and ecoregions and separate Floristic Quality
Assessment Index (FQAI) scores for species by wetland type.









Table 4-1. The five metrics of the preliminary Florida Wetland Condition Index for
freshwater forested strand and floodplain wetlands based on the macrophyte species
assemblage.

FWCI Metrics
1. Proportion Tolerant Indicator Species
2. Proportion Sensitive Indicator Species
3. Floristic Quality Assessment Index Score
4. Proportion Exotic Species
5. Proportion Native Perennial Species


Describing Biological Integrity

Biological indicators were useful in determining the biological integrity of
freshwater forested strand and floodplain wetlands. For the purposes of this study,
biological integrity has been defined quantitatively with the FWCI. The FWCI
incorporated five metrics from the macrophyte species assemblages (Table 4-1). Strong
correlations between the FWCI and the intensity of development in the surrounding
landscape (based on the use of nonrenewable energy and calculated with the LDI)
suggest that changes in macrophyte community composition quantified as metrics were
captured by the LDI. It has been proposed that organisms respond to environmental
gradients by colonizing a range of feasible conditions beyond which the organisms fail to
persist (ter Braak 1987). By selecting species that occur throughout the range of
measurable environmental parameters, the FWCI defined and detected deviations from
the condition of reference wetlands based on macrophyte community composition. Each
of the FWCI metrics addressed some disparity from the assumed range of feasible
conditions. The proportion sensitive indicator species metric showed the strongest
correlation with LDI, suggesting that the presence of a suite of taxa characteristic of
wetlands with high biological integrity may be the most effective means of identifying
changes in macrophyte community composition in freshwater forested strand and
floodplain wetlands associated with changes in anthropogenic activities.

Richness, Evenness, and Diversity

Measures of richness, evenness, and diversity of the macrophyte assemblage were
not sensitive to differences in land use or development intensity in the surrounding
landscape within the forested strand and floodplain wetlands. Perhaps due to a limited
sample size and high variability inherent in the landscape (e.g. regional differences, land
use differences, etc.) no statistical test on summary statistics produced statistically
significant results. Nevertheless, wetlands in the high LDI group (LDI > 2.0) had higher
(though not statistically significant) mean species richness and diversity (Shannon
diversity, Whittaker's beta diversity, and overall beta and gamma diversity). Species
evenness and Simpson's diversity were remarkably similar among the sample wetlands;
in effect no differences were detectable.









General ecological theory predicts a decrease in plant diversity resulting from an
increase in anthropogenic assaults such as grazing (Blanch and Brock 1994; Grace and
Jutila 1999) and nutrient enrichment (Bedford et al. 1999), though the forested strand and
floodplain wetlands displayed the opposite trend. Mitsch and Gosselink (1993) report
that freshwater forested wetlands have low species diversity as compared to other
ecosystems, so perchance macrophyte species that entered wetlands in developed
landscapes were merely taking advantage of available habitat and in fact increased the
overall species diversity. However, Ewel (1990) notes that due to high topographic and
soil variability, river swamps may be the most diverse type swamp in Florida. Clearly
there is some uncertainty within the published literature as to anticipated and abnormal
diversity for Florida forested strand and floodplain wetlands.
Many of the species entering wetlands in developed landscapes were categorized
as exotic species, and the increased incidence of exotic species has long been associated
with disturbed ecosystems (Galatowitsch 1999b; Cronk and Fennessy 2001). An increase
in the frequency of exotic species has been attributed to drainage and hydrologic
alterations (Hobbs and Heunneke 1992; David 1999; Galatowitsch et al. 1999b),
increased human development (Cronk and Fennessy 2001), and ecosystem scale
alterations such as clear-cut harvests (Devine 1998). Within the study wetlands, the
proportion of exotic species increased with increasing development intensity in the
surrounding landscape. It appears that the influx of exotic species added to rather than
diminished the species richness and diversity within the freshwater forested strand and
floodplain wetlands.
As such, the presence of exotic species alone may not be an ideal indicator of
biological integrity. While many ecological theories have been established suggesting
that the presence or occurrence of exotic species increases with anthropogenic
disturbance (Cronk and Fennessy 2001; Galatowitsch 1999b), there was a inconsistent
pattern of occurrence of exotic species in the forested strand and floodplain wetlands
sampled. For example, six of the freshwater forested floodplain wetlands, including two
reference, two agricultural, and two urban wetlands, hosted zero exotic species. That
four floodplain wetlands in developed landscapes (including two agricultural and two
urban wetlands) hosted no exotic species was somewhat contradictory to the theories of
increased exotic species occurrence in disturbed ecosystems. However, for the complete
dataset, the trend of an increase in the proportion of exotic species with increasing
landscape development intensity held.
Some concerns arise considering the discrepancies on the exotic status of some
species. For example, not all of the species listed by the Exotic Pest Plant Council
(EPPC) as invasive exotics altering native plant communities (Category I) or invasive
exotics increasing in frequency or abundance with the potential to alter native plant
communities (Category II) (EPPC 2003) received an FQAI score of 0.0 corresponding to
species that act as opportunistic invaders, including species that commonly occur in
disturbed ecosystems. The disagreement on the status of a species as an invasive exotic
translates into disagreement as to the meaning of a particular exotic species occurring
within an ecosystem. As a case in point, two exotic species were included as tolerant
indicator species including Commelina diffusa (dayflower) and Ludwigia peruviana
(water-primrose). Neither of the two tolerant indicator species that are also exotic species
was listed as an EPPC Category I or II species. In fact, none of the 13 Category I









invasive exotic species identified in the forested strand and floodplain wetlands was
categorized as a tolerant indicator species.
The intent of this research was to assess the current biological integrity of Florida
freshwater forested flowing water wetlands in order to develop a quantitative wetland
condition index. As such the comparison was made against the current day reference
standard of biological integrity apparent in the reference strand and floodplain wetlands
sampled. While we may innately believe that reference wetlands host no exotic species,
we found that five of nine wetlands surrounded by low development intensity (56%)
hosted at least one exotic species. Increasingly, it may become apparent that even
reference wetlands with the highest current standard of biological integrity may host
some proportion of exotic species. This may be even more apparent in the southern half
of the state, where drainage and development have altered nearly all of the Florida
landscape. In fact, the south wetland region reference wetland (FS10) had the highest
proportion exotic species of all wetlands in the low LDI group by nearly 10%.
Consequently, establishing a baseline for the reference condition of biological integrity is
crucial for the application of indices of biotic integrity. Consensus should be reached as
to whether scientists proceed by establishing a present day baseline for future
assessments or by maintaining and updating a moving baseline for future application of
the FWCI. Implications for both methods are complex.

Measuring Anthropogenic Activity

The variable sensitivities of three different independently derived indices
compared to the forested strand and floodplain FWCI, including the Landscape
Development Intensity index (LDI; Lane et al. 2003; Brown and Vivas 2005), the Human
Disturbance Gradient (HDG; Fore 2004), and the Stream Condition Index (SCI; Fore
2004), suggest that multiple measures of biological integrity may be more effective at
describing ecosystem wide biological integrity than any single measure based on an
individual species assemblage or surrounding land use activity. Measurements of
anthropogenic activity such as the LDI and HDG seek to describe ecosystem biological
integrity from the perspective of outside anthropogenic influences which act to alter an
ecosystem. While the LDI was used as a remote based measure of human development
intensity, the HDG used both remote and local conditions integrating four components
ranging from the in-stream chemical water quality, habitat structure, and hydrologic
alteration, to a remote landscape assessment (using a form of LDI). For the HDG each of
the four components was given equal weighting in determining the influence of
anthropogenic activities on the biological integrity of an ecosystem. The LDI and HDG
were strongly correlated, suggesting that the LDI alone may capture the influence from
human development intensity on ecosystem biological integrity without the need for
additional sampling and sample processing associated with in-stream chemical water
quality, habitat structure, and hydrologic alteration.
The FWCI was strongly correlated with both the HDG and the LDI, though the
significance of correlation was slightly stronger with the HDG than the LDI. The caveat
here was that only three of the thirteen floodplain wetlands had an HDG greater than
zero, and because these scores were tested using the non-parametric Spearmans rank
correlation coefficient test the strong correlations may simply be a factor of zero HDG









values. In fact, our findings of strong correlations of LDI and HDG must be interpreted
with caution because so few of the wetlands received HDG values greater than zero (an
HDG score of zero represents no detectable human induced disturbance). This was likely
due to the fact that the HDG and SCI were calculated in the channelized water course,
whereas the LDI and FWCI were calculated for the surrounding floodplain forest. Many
of the streams with low SCI and high HDG scores (considered those with low biological
integrity) were not sampled for the FWCI and LDI as no floodplain forest remained, with
the banks of the channelized water course being either sodded and mowed or paved.
Therefore we should cautiously interpret correlations and discrepancies between the SCI
(no significant correlations with FWCI or LDI) and HDG with the FWCI and LDI
because the range of impaired conditions used to develop the FWCI was much narrower
than that used to develop the SCI.
The lack of correlations of SCI with both FWCI and LDI, suggest that in-stream
macroinvertebrate based measures of biological condition and surrounding forested
wetland macrophyte based measures of biological condition did not respond in a
consistent manner to changes in anthropogenic activity. Using both the in-stream
macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte
FWCI biological assessment methods may provide a more complete picture of the overall
condition of a wetland and associated stream at a particular spatial location. While
agreement in the ranking of the biological condition of study wetlands using the FWCI
and SCI was anticipated, discrepancies among the ranking from the different assemblages
may provide great insight into biological condition as different species assemblages
respond to changes in anthropogenic activities and the associated changes in inflows (e.g.
nutrient enrichment) over different time scales. Additionally, use of the forested strand
and floodplain FWCI may lead to specific conclusions as to the biological condition of
local or nearby anthropogenic activity, while use of the SCI may enhance understanding
of larger watershed scale influences from anthropogenic activity (i.e. due to the
convergence of surface water within the watershed associated with stream flow).
By correlating the community composition based FWCI and the landscape based
LDI, we were better able to understand anthropogenic influence on biological integrity on
forested wetland ecosystems. We found that the LDI index was a useful tool in
approximating biological integrity. Its primary power was in the reproducible, objective,
and quantitative methods employed to obtain a score based on the use on non-renewable
energy in the surrounding landscape. A second strength of the LDI index was apparent in
the practical application of a remote GIS based method of describing ecosystem condition
as a starting point to identify potential areas for further biological sampling.

Regionalization of the Florida Wetland Condition Index

Pronounced differences in the local climate across Florida, such as differences in
the amount of annual rainfall, seasonal maximum temperatures, and number of freeze
days (Fernald and Purdum 1992; Lane 2000) and the broad latitudinal and longitudinal
ranges of sample wetlands (26.3N -30.8N latitude, 80.1W-82.1W longitude), suggest
differences would be found among macrophyte community composition in Florida
wetlands. In fact, compositional differences of the macrophyte species assemblages were









found among Lane's (2000) Florida wetland regions. However, the study sample size
limited the development of regional indicator species and metric scoring criteria.
Wetland clusters based on macrophyte community composition (species presence
by site) roughly correlated with wetland spatial distribution throughout the state. The
distribution of wetlands in the cluster groups suggests that wetlands located in the
northern area of Florida may have higher biological integrity than wetlands in the central
or southern peninsula given a statewide scoring approach of the preliminary forested
strand and floodplain FWCI. Accordingly, most of the human development in Florida
has occurred along the east and west coastal areas of peninsular Florida (Femald and
Purdum 1992), suggesting that while the reference wetlands selected in the south and
central wetland regions were selected as the best possible examples of reference type
conditions, they may be more affected by development in the surrounding landscape
(such as compounded secondary effects) than their panhandle and north wetland region
reference counterparts. While the ease and utility of a single statewide FWCI would
seemingly prevail over regional indicator species and metric scoring criteria, the
necessity of scoring each wetland region based on the best possible reference conditions
cannot be overlooked (as suggested by Karr and Chu 1999).
Regionalization of biological indices has been suggested throughout the literature.
One of the main premises behind indices of biological integrity is a comparison of "like
to like" (Gerristen et al. 2000), that is, to reduce the noise in background variability in
biological data, which could be accomplished through regionally based indices. The
lower FWCI scores for wetlands in the central and southern peninsula may also be a
factor of the smaller sample size for wetlands located in those wetland regions. Since the
macrophyte community composition was found to be different between wetland regions,
wetlands in the south and central wetland regions were underrepresented and therefore
had less influence in the indicator species analysis and the metric scoring criteria. A
larger sample size would improve uncertainty related to questions surrounding the
biological integrity of wetlands in the south and central wetland regions.

Florida Wetland Condition Index Independent of Wetland Type

Recent works by Lane et al. (2003) and Reiss and Brown (2005) presented a five
metric FWCI for isolated depressional herbaceous wetlands and a six metric FWCI for
isolated depressional forested wetlands in Florida based on the community composition
of the macrophyte assemblage. The depressional herbaceous FWCI was created based on
a sample size of 75 freshwater marshes surrounded by reference (n=34) and agricultural
(n=40) land uses throughout peninsular Florida. The depressional forested FWCI was
created based on a sample size of 118 freshwater wetlands surrounded by reference
(n=37), agricultural (n=40) and urban (n=41) land uses throughout the extent of Florida.
The five metrics of the preliminary FWCI for forested strand and floodplain wetlands
were nearly identical to the five metrics of the depressional herbaceous FWCI (with an
adaptation from annual to perennial ratio from the depressional herbaceous FWCI
changed to the proportion native perennial species for the depressional forested and
forested strand and floodplain FWCIs) and similar to five of the depressional forested
FWCI metrics.









The five macrophyte metrics included in the three wetland type FWCIs
(depressional herbaceous, depressional forested, forested strand and floodplain) were
tolerant and sensitive indicator species, FQAI score, exotic species, and native perennial
species (annual to perennial ratio in the depressional herbaceous FWCI). Tolerant and
sensitive indicator species lists were constructed independently for each wetland type. Of
the 16 tolerant indicator species for flowing water wetlands, five also occurred as tolerant
indicator species for depressional forested wetlands (Reiss and Brown 2005). Two of
those species, Commelina diffusa (dayflower) and Ludwigia peruviana (water-primrose)
also occurred as tolerant indicator species for depressional herbaceous wetland (Lane et
al. 2003). The additional three overlapping tolerant indicator species between the
flowing water and depressional forested wetlands were Acer rubrum (red maple),
Sambucus canadensis (elderberry), and Toxicodendron radicans (Eastern poison ivy).
One tolerant indicator species for flowing water wetlands, Nyssa sylvatica var. biflora
(swamp tupelo) was found to be a sensitive indicator species for depressional herbaceous
wetlands (Lane et al. 2003); and the tolerant indicator species for forested strand and
floodplain wetlands Panicum rigidulum (red-top panicum) was found to be a sensitive
indicator species for both depressional herbaceous and forested wetlands. Clearly a
larger sample size and refinement of indicator species analysis for the forested strand and
floodplain wetlands FWCI could reduce the inconsistencies among indicator species.
Similarly, four sensitive indicator species were common among depressional
herbaceous, depressional forested, and forested strand and floodplain wetlands, including
Cladium jamaicense (saw-grass), Lyonia lucida (fetter-bush), Panicum hemitomon
(maidencane), and Pinus elliottii (slash pine). Vaccinium corymbosum highbushh
blueberry) was also considered a sensitive indicator species for both depressional
herbaceous and forested strand and floodplain wetlands. However, three sensitive
indicator species for forested strand and floodplain wetlands were listed as tolerant
indicator species for depressional forested (Saururus cernuus, lizard's tail) or
depressional herbaceous and forested (Centella asiatica, coinwort; Eupatorium
capillifolium, dog fennel) wetlands.
The three additional metrics included in the FWCIs were Floristic Quality
Assessment Index score, exotic species, and native perennial species (modified from the
depressional herbaceous wetland FWCI that used annual to perennial ratio). The variant
of the annual to perennial ratio (as native perennial species) was used to account for
variable conditions at urban wetlands, which were not included in the study of
depressional herbaceous wetlands but were studied in both the depressional forested and
forested strand and floodplain wetland research. Wienhold and Van der Valk (1989) and
Ehrenfeld and Schneider (1991) determined that disturbance often favors annual species
over perennial species and promotes the invasion of nonnative perennials in wetlands.
Galatowitsch et al. (2000) found that while native perennial cover was reduced in
wetlands impacted by cultivation, the occurrence of introduced perennials rather than
annuals increased in stormwater impacted wetlands. The sixth depressional forested
FWCI metric was the wetland status species (including both obligate and facultative
wetland species), which was not included for the preliminary forested strand and
floodplain wetland FWCI as it did not meet selection criteria for inclusion.
Overall, the depressional herbaceous, depressional forested, and flowing water
forested wetland FWCIs included five similar metrics. Perhaps the strong similarity of









metrics selected for inclusion in the FWCIs suggests that a universal assessment index
with core metrics could be constructed regardless of wetland type. However, it would
likely be necessary to maintain independent indicator species lists and metric scoring
criteria for different wetland types within each wetland region.

Limitations and Further Research

Generally wetlands were visited only once, with a complete sample effort lasting
just one day, which provided a mere snapshot of wetland condition. Visiting these
wetlands only once did not allow insight into seasonal or yearly variations in the
macrophyte assemblage; and the preliminary forested strand and floodplain FWCI would
benefit from inter-seasonal validation. The FWCI would also benefit from validation
based on a new set of wetlands to test the repeatability of this index. A larger sample size
of wetlands will improve the scoring criteria of the FWCI based on wetland regions for
metrics such as indicator species analysis. Funding for additional wetland sampling and
FWCI refinement is in the application phase. Regionalization appears to be an important
next step in refining the FWCI for all wetland types, as this study was limited to a
statewide approach due to small sample sizes within each wetland region.

Conclusions

The use of the macrophyte assemblage for a biological assessment of Florida
freshwater flowing water forested wetlands provided a useful tool for detecting changes
in biological integrity associated with changes in anthropogenic activity. While richness
and diversity measures of macrophyte community composition were not particularly
sensitive to changes in landscape development intensity, metrics used as biological
indicators of changes in macrophyte community composition were. In fact, the strong
correlation between the landscape scale human disturbance gradient (LDI) and the local
scale wetland condition index of biological integrity (FWCI) demonstrated the potential
value of using the LDI index as an initial indication of biological integrity, which can be
further tested with chemical and physical parameters and compared against assemblage
specific biological indices.
Due to similarities with metrics from the FWCIs for depressional herbaceous
(Lane et al. 2003) and depressional forested (Reiss and Brown 2005) wetlands, a multi-
metric multi-assemblage FWCI could be constructed for all freshwater wetlands
throughout the state of Florida, with a set of core metrics and specific indicator species
and metric scores based on wetland type within the wetland regions. While the forested
strand and floodplain FWCI for flowing water systems can not be used to predict changes
in the physical and chemical parameters of a wetland, its strength lies in providing an
overview of biological integrity through the integration of changes in macrophyte
community composition from cumulative effects. The quantitative score of biological
integrity established through the FWCI should not be used as a surrogate for wetland
value, but as an objective, quantitative means of comparing changes in community
composition along gradients of human development intensity, which can be used
objectively to assess the biological integrity of Florida's wetlands.









APPENDIX A
STANDARD OPERATING PROCEDURES

CHECKLIST OF MATERIALS/FIELD EQUIPMENT

Miscellaneous
/ SOPs
/ Large cooler with frozen ice bottles for unknown vegetation
/ Waders
/ Garmin III GPS unit
/ Florida Gazetteer
/ Machete
/ Aerial photo & FLUCCS codes of site

Vegetation Transects
/ 2-3 100m transect tapes
/ 1 m PVC with distance marks (cm)
/ 2-3 compasses
/ Clipboards
/ Field data sheets a minimum of 10 per site
/ Site Characterization & WRAP sheets 1 per person per site
/ Pencils
/ Sharpie
/ Bag for unknown plants
/ Masking tape
/ Field ID manuals
/ Prism for basal area
/ Hand lens
/ Index cards









SOPS FOR FORESTED STRAND WETLANDS: VEGETATION

1. Note the direction of flow through the landscape.
2. Locate a line running through the center of the strand along the flow gradient. This is
the center-line.
3. Randomly select a starting point for the initial transect. Each consecutive transect
will begin approximately 25 m upstream of the initial transect, so that a stretch of
approximately 100 m will be sampled along the length of the strand. Run transects
perpendicular to the main channelized flow.
4. At the beginning of each transect, delineate the edge of the wetland using a
combination of wetland plants and hydrologic indicators. Be conservative on the side
of the wetland.
5. Establish the transect using a meter tape and a compass. Each transect will start with
0 meters at the wetland edge and run into the center-line (established in step 2).
6. Use a separate field data sheet for each transect. If the number of species located on a
transect exceeds the number of columns on the data sheet, start a new data sheet. Be
thorough in completing field data sheets including information on site, transect
direction, date, and data recorder. Specify if there are multiple field data sheets for a
single transect.
7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long,
record all species rooted within these elongated quadrats.
8. Plant species names are recorded on the data sheets using the full genus and species
names. Each unknown species is given a unique ID code using the transect number
(ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.).
9. Collect voucher specimens for all unknown species being sure to get plant
inflorescence and roots, tag samples with properly labeled masking tape, and put into
a labeled collection bag. Note the color of the inflorescence on the label, as the
flowers often do not preserve well. Index cards can be used to protect especially
sensitive parts. When vegetation sampling is complete, store the collection bag in a
cooler on ice until identification can be completed.
10. Voucher specimens are identified in the field on the day of sampling. Unidentified
plants will be placed in a plant press for further clarification and identification. Plant
nomenclature follows FDEP's Florida Wetland Plant Identification Manual (Tobe et
al. 1998). If time prohibits immediate pressing, unknown plants should be stored in
the cooler.
11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be
recorded. Use the data sheet for basal area, and record basal area per species using
variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent
elbow. Looking through the prism count the number of trees per species that fall
within the variable area plot. The prism shall be centered over the sampling point at
all times, with the field person rotating around the prism so that the entire circular
area (3600) around the point of sampling is included.









Forested Strand Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands


Site: Transect Number:


Date: Data Recorder:

















I---------------------------- ------------------ --- -- -------
I I
I 0 I







[- -- --- --_ -_ --.- _- _- -----------_ -_ -_ _------_ -. _- _- -
i 0-5 m
5-10
S10-15
15-20
20-25
S25-30
S30-35
S35-40
S40-45
45-50
50-55
S55-60
60-65
I 65-70
70-75
75-80
I 80-85
85-90
S90-95
S95-100 -L. -








Forested Strand Wetlands Field Data Sheet Basal Area UF Center for Wetlands

Site: Transect Direction:


Date: Data Recorder:

r ---------------------------------------------------------------------------- -----------------



I IC










- - - ---- --- - - -
10l
I I







I I


I 0 m
50I I








6 0 m

70 m

80 m

90 m
I Im_______
I 10mI

! 3 0 m_______________________________________!

. 4 0 m_______________________________________j

! 5 0 m_______________________________________!

j 6 0 m_______________________________________j

! 7 0 m_______________________________________!

j 8 0 m_______________________________________j

! 9 0 m ____ ______ ______ ______ ______ ______ ____!
I 10 mI









SOPS FOR FORESTED FLOODPLAIN WETLANDS: VEGETATION

1. Note the direction of flow through the landscape.
2. Locate a line running through the center of the strand along the flow gradient. This is
the center-line.
3. Randomly select a starting point for the initial transect, preferably this point with
coincide with a Stream Condition Index (SCI) sample point. Each consecutive
transect will begin approximately 25 m upstream of the initial transect, so that a
stretch of approximately 100 m will be sampled along the length of the strand. Run
transects perpendicular to the main channelized flow.
4. Delineate the wetland line using a combination of wetland plants and hydrologic
indicators. Be conservative on the side of the wetland.
5. Establish the transect using a meter tape and a compass. Transects will be limited to
a maximum length of 50 m. The first transect will begin with 0 meters at the wetland
edge and run towards the center-line (established in step 2). The second transect will
begin 25 m upstream of the first transect. Transect 2 begins at the edge of the
channelized flow and runs through the wetland perpendicular to the flow for a
maximum of 50 m. Repeat placement for transects 3 and 4.
6. Use a separate field data sheet for each transect. If the number of species located on a
transect exceeds the number of columns on the data sheet, start a new data sheet. Be
thorough in completing field data sheets including information on site, transect
direction, date, and data recorder. Specify if there are multiple field data sheets for a
single transect.
7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long,
record all species rooted within these elongated quadrats.
8. Plant species names are recorded on the data sheets using the full genus and species
names. Each unknown species is given a unique ID code using the transect number
(ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.).
9. Collect voucher specimens for all unknown species being sure to get plant
inflorescence and roots, tag samples with properly labeled masking tape, and put into
a labeled collection bag. Note the color of the inflorescence on the label, as the
flowers often do not preserve well. Index cards can be used to protect especially
sensitive parts. When vegetation sampling is complete, store the collection bag in a
cooler on ice until identification can be completed.
10. Voucher specimens are identified in the field on the day of sampling. Unidentified
plants will be placed in a plant press for further clarification and identification. Plant
nomenclature follows FDEP's Florida Wetland Plant Identification Manual (Tobe et
al. 1998). If time prohibits immediate pressing, unknown plants should be stored in
the cooler.
11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be
recorded. Use the data sheet for basal area, and record basal area per species using
variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent






77


elbow. Looking through the prism count the number of trees per species that fall
within the variable area plot. The prism shall be centered over the sampling point at
all times, with the field person rotating around the prism so that the entire circular
area (3600) around the point of sampling is included.











Forested Floodplain Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands


Site: Transect Number:

Date: Data Recorder:


















0-5 m

5-10
I I











I -1









15-20

20-25
I I























I 25-30
I I

























35-40

I 40-45
S5-50 _




! 35-40 _

I 40-45 I

| 45-50











Forested Floodplain Wetlands Field Data Sheet Basal Area UF Center for Wetlands


Site: Transect Direction:


Date: Data Recorder:





I I)





0I








I30
I I


















40 m











APPENDIX B
COEFFICIENT OF CONSERVATISM SCORES

Table B-1. Coefficient of Conservatism (CC) scores for macrophyte species identified in
forested strand and floodplain wetlands in Florida. CC scores were assigned from FQAI
surveys from isolated depressional forested wetlands (Reiss and Brown 2005) first,
followed by isolated depressional herbaceous wetlands (Lane et al. 2003). Species
without scores from previous FQAI studies were scores according to their faithfulness
and fidelity to freshwater flowing water wetlands (strands and floodplains) in 2003 by
five expert Florida botanists (Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and
Bruce Tatje).


Species
Abrus precatorius
Acer rubrum
Acer saccharum
Agarista populifolia
Alnus serrulata
Alternanthera philoxeroides
Alternanthera sessilis
Amaranthus australis
Ambrosia artemisiifolia
Ampelopsis arborea
Amphicarpum muhlenbergianum
Andropogon virginicus
Annona glabra
Apios americana
Ardisia escallonioides
Arisaema triphyllum
Aronia arbutifolia
Arundinaria gigantea
Asimina parviflora
Asimina reticulata
Asplenium heterochroum
Aster elliottii
Bacopa caroliniana
Bacopa monnieri
Begonia cucullata
Berchemia scandens
Bidens alba
Bidens mitis
Bignonia capreolata
Blechnum serrulatum
Boehmeria cylindrica
Bumelia lycioides
Callicarpa americana
Campsis radicans
Carex albolutescens


CC Score
0.0
5.2
6.0
6.5
6.0
0.0
0.7
2.6
0.7
3.3
5.0
2.6
6.8
3.1
7.0
6.5
5.7
5.3
5.0
4.4
8.5
4.2
6.0
4.3
1.5
5.1
1.0
3.8
4.4
5.5
4.5
6.0
2.4
3.3
3.6


1999/2000 2001/2002 2003
Herbaceous Forested Flowing
1











1999/2000 2001/2002 2003
Species CC Score Herbaceous Forested Flowing
Carex crus-corvi 6.0 1
Carex leptalea 6.8 1
Carex lupulina 6.3 1
Carex stipata 4.5 1
Carpinus caroliniana 6.8 1
Carya glabra 5.8 1
Celtis laevigata 5.0 1
Centella asiatica 1.9 1
Cephalanthus occidentalis 6.0 1
Chasmanthium laxum 6.0 1
Chasmanthium nitidum 6.3 1
( l '.',t it,, icaco 6.3 1
Cinnamomum camphora 0.2 1
Cladium jamaicense 5.5 1
Clematis crispa 5.5 1
Colocasia esculenta 0.0 1
Commelina communis 2.5 1
Commelina dillttm. 1.7 1
Conyza canadensis 0.3 1
Cornusfoemina 6.6 1
Crataegus marshallii 6.3 1
Crinum americanum 7.6 1
Cuphea c ,,i i,,. oi, ,i, 1.4 1
Cynodon dactylon 0.0 1
Cyperus difformis 2.0 1
Cyperus globulosus 1.8 1
Cyperus haspan 2.6 1
Cyperus ligularis 3.2 1
Cyperus odoratus 3.6 1
Cyrilla racemiflora 4.5 1
Decumaria barbara 6.3 1
Dichondra carolinensis 1.9 1
Dichromena colorata 5.5 1
Digitaria serotina 1.8 1
Diodia virginiana 2.4 1
Dioscoreafloridana 5.5 1
Diospyros virginiana 4.0 1
Eclipta alba 1.7 1
Emiliafosbergii 0.5 1
Erechtites hieraciifolius 2.1 1
Erianthus giganteus 6.0 1
Eryngium baldwini 4.4 1
Eryngium prostratum 4.0 1
Erythrina herbacea 4.0 1
Eugenia uniflora 1.3 1
Euonymus americanus 5.5 1
Eupatorium capillifolium 0.5 1
Eupatorium mikanioides 5.5 1












Species
Eupatorium perfoliatum


CC Score
5.9


1999/2000 2001/2002 2003
Herbaceous Forested Flowing
1


Eustachys petraea 2.0 1
Fagus grandifolia 7.5 1
Fraxinus caroliniana 7.1 1
Galium tinctorium 3.1 1
Gaylussacia dumosa 5.4 1
Gaylussaciafrondosa 6.7 1
Gelsemium sempervirens 4.0 1
Gordonia lasianthus 6.7 1
Gratiola hispida 6.0 1
Hamamelis virginiana 6.5 1
Hibiscus coccineus 6.6 1
Hydrocotyle bonariensis 3.3 1
Hydrocotyle ranunculoides 3.1 1
Hydrocotyle umbellata 2.9 1
Hydrocotyle verticillata 3.1 1
Hypericum brachyphyllum 6.8 1
Hypericum hypericoides 4.0 1
Hypericum myrtifolium 5.5 1
Hypericum tetrapetalum 5.0 1
Hypoxis curtissii 5.7 1
Hyptis alata 4.3 1
Ilex cassine 8.1 1
flex coriacea 6.0 1
Ilex glabra 4.3 1
flex opaca var. opaca 6.0 1
flex vomitoria 4.8 1
Ipomoea hederifolia 2.0 1
Ipomoea pandurata 4.8 1
Ipomoea sagittata 5.4 1
Itea virginica 7.9 1
Juncus effusus 1.9 1
Juncus megacephalus 3.3 1
Juncus polycephalos 3.3 1
Juniperus virginiana 5.2 1
Justicia ovata 5.5 1
Koelreuteria elegans 2.0 1
Lachnanthes caroliana 3.1 1
Lemna minor 1.0 1
Leucothoe axillaris 7.0 1
Leucothoe racemosa 6.2 1
Ligustrum sinense 0.0 1
Lindernia granidilora 3.6 1
Liquidambar styraciflua 3.3 1
Liriodendron tulipifera 6.8 1
Lobelia cardinalis 6.8 1
Lonicerajaponica 0.0 1
Ludwigia maritima 3.3 1











1999/2000 2001/2002 2003


Herbaceous Forested Flowing


Ludwigia palustris 4.0 1
Ludwigia peruviana 1.2 1
Ludwigia pilosa 5.8 1
Ludwigia repens 2.9 1
Lycopus rubellus 5.2 1
Lygodium japonicum 0.0 1
Lygodium microphyllum 0.0 1
Lyoniafruticosa 6.0 1
Lyonia lucida 6.0 1
Lyonia mariana 6.8 1
Magnolia giraditllra 6.2 1
Magnolia virginiana var. australis 8.1 1
Mateleafloridana 6.7 1
Melaleuca quinquenervia 0.0 1
Merremia dissecta 0.3 1
Micranthemum glomeratum 4.0 1
Micranthemum umbrosum 4.3 1
Micromeria brownei 4.8 1
Mikania scandens 2.4 1
Mimosa pigra 0.7 1
Mitchella repens 6.7 1
Momordica charantia 0.0 1
Morus rubra 4.4 1
Myrica cerifera 3.1 1
Myrsine guianensis 5.2 1
Nephrolepis exaltata 3.8 1
Nyssa ogeche 7.0 1
Nyssa sylvatica var. biflora 7.4 1
Oeceoclades maculata 0.4 1
Oplismenus setarius 3.3 1
Orontium aquaticum 7.6 1
Osmunda cinnamomea 5.5 1
Osmunda regalis 6.9 1
Oxypolisfiliformis 6.7 1
Panicum abscissum 9.2 1
Panicum anceps 4.3 1
Panicum commutatum 4.5 1
Panicum dichotomum 4.0 1
Panicum ensifolium 5.0 1
Panicum erectifolium 5.7 1
Panicum hemitomon 5.0 1
Panicum rigidulum 4.5 1
Panicum spretum 5.4 1
Panicum tenue 4.2 1
Parietaria praetermissa 3.0 1
Parthenocissus quinquefolia 3.0 1
Paspalum conjugatum 3.1 1
Paspalum notatum 0.0 1


Species


CC Score












Species
Passiflora incarnata


CC Score
3.0


1999/2000 2001/2002 2003
Herbaceous Forested Flowing
1


Peltandra virginica 5.8 1
Persea borbonia 6.3 1
Persea palustris 7.4 1
Phlebodium aureum 6.8 1
Phyla nodiflora 1.4 1
Phyllanthus urinaria 0.0 1
Phytolacca americana 1.2 1
Pieris phyllyreifolia 9.5 1
Pinckneya bracteata 8.3 1
Pinus clausa 5.6 1
Pinus elliottii 4.0 1
Pinus taeda 3.3 1
Plucheafoetida 3.8 1
Pluchea rosea 3.6 1
Polygala rugelii 8.2 1
Polygonum densiflorum 5.3 1
Polygonum hirsutum 8.2 1
Polygonum hydropiperoides 2.6 1
Pontederia cordata 5.0 1
Proserpinaca palustris 3.8 1
Prunus caroliniana 3.0 1
Prunus serotina 3.6 1
Psychotria nervosa 5.2 1
Psychotria sulzneri 5.5 1
Pteridium aquilinum 3.6 1
Ptilimnium capillaceum 3.1 1
Quercus laurifolia 3.6 1
Quercus michauxii 5.7 1
Quercus nigra 2.1 1
Quercus virginiana 4.2 1
Rhododendron canescens 6.8 1
Rhododendron viscosum 7.6 1
Rhus copallinum 2.4 1
Rhynchospora baldwinii 5.7 1
Rhynchospora inundata 6.0 1
Rhynchospora microcephala 4.8 1
Rhynchospora miliacea 7.1 1
Rubus argutus 2.1 1
Rubus trivialis 1.9 1
Ruellia caroliniensis 4.3 1
Rumex verticillatus 4.8 1
Sabal minor 6.2 1
Sabalpalmetto 4.5 1
Sabatia calycina 6.2 1
Sagittariafiliformis 6.0 1
Sagittaria lancifolia 4.5 1
Sagittaria latifolia 5.0 1











1999/2000 2001/2002 2003
Herbaceous Forested Flowing
1


Sambucus canadensis 1.7 1
Samolus valerandi 5.6 1
Sanicula canadensis 5.7 1
Sapium sebiferum 0.0 1
Saururus cernuus 5.5 1
,. Ii, l actinophylla 0.5 1
Schinus terebinthifolius 0.0 1
Scleria triglomerata 4.8 1
Serenoa repens 4.5 1
Setaria geniculata 3.1 1
Sida rhombifolia 1.0 1
Smilax auriculata 3.8 1
Smilax bona-nox 2.6 1
Smilax glauca 3.3 1
Smilax laurifolia 5.2 1
Smilaxpumila 6.0 1
Smilax smallii 4.5 1
Smilax tamnoides 3.6 1
Smilax walteri 6.0 1
Solidagofistulosa 3.6 1
Sparganium americanum 6.7 1
Sporobolus floridanus 7.1 1
Stenotaphrum secundatum 0.8 1
\, ii ,, i aquatica 7.4 1
Symplocos tinctoria 6.0 1
Taxodium ascendens 8.8 1
Taxodium distichum 7.2 1
Thelypteris dentata 3.4 1
Thelypteris hispidula 4.5 1
Thelypteris palustris 5.8 1
Tilia americana 5.5 1
Toxicodendron radicans 1.9 1
Triadenum virginicum 5.0 1
Trichostema dichotomum 4.5 1
Trifolium repens 0.0 1
Tripsacum dactyloides 4.0 1
Ulmus americana 7.4 1
Urena lobata 0.0 1
Vaccinium arboreum 6.4 1
Vaccinium corymbosum 5.7 1
Vaccinium stamineum 5.8 1
Vaccinium tenellum 5.5 1
Viburnum dentatum 6.0 1
Viburnum nudum 5.0 1
Viburnum obovatum 4.7 1
Viola affinis 5.5 1
Vitis aestivalis 2.9 1


Species
Salvia lyrata


CC Score
3.2











1999/2000 2001/2002 2003
Species CC Score Herbaceous Forested Flowing
Vitis cinerea 2.0 1
Vitis rotundifolia 2.1 1
Vitis shuttleworthii 3.5 1
Wisteria sinensis 1.0 1
Woodwardia areolata 5.7 1
Woodwardia virginica 4.8 1
Xyrisjupicai 1.7 1
Youngia japonica 0.0 1









APPENDIX C
SUMMARY STATISTICS

Table C-1. Summary statistics of richness (R), evenness (E), Shannon diversity (H'),
Simpson diversity (D), and Whittaker's beta diversity (3w) for the macrophyte
assemblage (species level).

Site R E H' D pw
FF1 59 1.0001 4.08 0.98 5.88
FF2 21 1.0002 3.05 0.95 3.73
FF3 56 0.9999 4.03 0.98 4.86
FF4 46 1.0001 3.83 0.98 7.36
FF5 77 1.0000 4.34 0.99 6.82
FF6 25 1.0000 3.22 0.96 1.78
FF7 32 1.0001 3.47 0.97 3.49
FF8 75 0.9999 4.32 0.99 6.11
FF9 45 1.0001 3.81 0.98 3.90
FF10 35 0.9999 3.56 0.97 3.18
FF11 29 0.9999 3.37 0.97 4.73
FF12 46 1.0001 3.83 0.98 5.15
FF13 48 0.9999 3.87 0.98 4.93
FF14 60 0.9999 4.09 0.98 5.94
FS1 39 1.0001 3.66 0.97 2.68
FS2 39 1.0001 3.66 0.97 4.90
FS3 33 1.0001 3.50 0.97 4.45
FS4 55 0.9999 4.01 0.98 5.03
FS5 39 1.0001 3.66 0.97 3.91
FS6 29 0.9999 3.37 0.97 3.02
FS7 30 0.9999 3.40 0.97 2.96
FS8 35 0.9999 3.56 0.97 2.12
FS9 47 1.0000 3.85 0.98 3.73
FS10 42 1.0001 3.74 0.98 4.43









APPENDIX D
METRIC SCORING FOR THE MACROPHYTE FLORIDA WETLAND CONDITION
INDEX FOR FLOWING WATER SYSTEMS


1. Calculate values for the 5 metrics:
1 Proportion tolerant indicator species
2 Proportion sensitive indicator species
3 FQAI score
4 Proportion exotic species
5 Proportion native perennial species

2. Take the natural log of metrics to improve distribution.
= In (metric value + 1)
1 is added to avoid errors related to taking the natural log of a zero value

3. Use the scoring equations to normalize scores between 0 and 10.
Metrics that increase with increasing LDI tolerant, exotic metrics:
= 10 (( metric 5th percentile) ( 10 / ( 95th percentile 5th percentile)))

Metrics that decrease with increasing LDI sensitive, FQAI, native perennial
metrics:
= (( metric 5th percentile) (10 / ( 95th percentile 5th percentile)))

Below are the 5th and 95th percentiles for each metric (transformed values are presented,
see step 2 above):

5th Percentile 95th Percentile
Proportion tolerant indicator species 0.05 0.24
Proportion sensitive indicator species 0.00 0.16
FQAI score 1.44 1.84
Proportion exotic species 0.00 0.17
Proportion native perennial species 0.56 0.69


4. Rescore, so that the metrics in the outer 5th percentiles receive scores of 0 or 10.

= IF ( score < 0, 0, (IF ( score >= 10, score, 10)))









LIST OF REFERENCES


Adams, S.M. 2002. Biological indicators of aquatic ecosystem stress: introduction and
overview. Pages 1-11 in S.M. Adams, editor. Biological indicators of aquatic
ecosystem stress. American Fisheries Society. Bethesda, Maryland, USA.
Analyse-it Software, Ltd. 1997-2003. version 1.67. Leeds, England, United Kingdom.
Andreas, B.K. and R.W. Lichvar. 1995. A floristic assessment system for northern Ohio.
Wetlands Research Program Technical Report WRP-DE-8. U.S. Army Corps of
Engineers Waterways Experiment Station, Vicksburg, Mississippi, USA.
Apfelbeck, R. 2000. Developing preliminary bioassessment protocols for Montana
wetlands, State of Montana Department of Environmental Quality. Helena, Montana,
USA.
Arcview GIS 3.2 Environmental Systems Research Institute, Inc. 1999. Neuron Data,
Inc. 1991-1996. Portions copyright 1991-1995 Arthur D. Applegate. Found at:
http://www.esri.com/. Redlands, California, USA.
ArcGIS 8.3 Environmental Systems Research Institute, Inc. 1999-2002. Found at:
http://www.esri.com/arcgis. Redlands, California, USA.
Barbour, M.T., J. Gerristen, G.E. Griffith, R. Frydenborg, E. McCarron, J.S. White, and
M.L. Bastian. 1996a. A framework for biological criteria for Florida streams using
benthic macroinvertebrates. Journal of the North American Benthological Society
15(2): 185-211.
Barbour, M.T., J. Gerristen, and J.S. White. 1996b. Development of the Stream
Condition Index (SCI) for Florida. A Report to the Florida Department of
Environmental Protection, Stormwater and Nonpoint Source Management Section.
Tetra Tech, Inc. Owing Mills, Maryland, USA.
Bedford, B.L., M.R. Wabridge, and A. Aldous. 1999. Patterns in nutrient availability
and plant diversity of temperate North American wetlands. Ecology 80(7): 2151-
2169.
Blanch, S.J. and M.A. Brock. 1994. Effects of grazing and depth on two wetland plant
species. Australian Journal of Marine and Freshwater Research 45: 1387-1394.
Brown, M.T. and M.B. Vivas. 2005. Landscape Development Intensity Index.
Environmental Monitoring and Assessment 101: 289-309.
Brown, M.T. and S. Ulgiati. 2005. Emergy, transformity, and ecosystem health. Pages
333-352 in S.E. Jorgensen, R. Costanza, and F. Xu, editors. Handbook of ecological
indicators for assessment of ecosystem health. Taylor and Francis, Boca Raton,
Florida, USA.
Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community
structure. Australian Journal of Ecology 18: 117-143.
Cohen, M.J., S.M. Carstenn, and C.R. Lane. 2004. Floristic quality indices for biotic
assessment of depressional marsh condition in Florida. Ecological Applications 14(3):
784-794.
Cronk, J. K. and M.S. Fennessy. 2001. Wetland plants: biology and ecology. Lewis
Publishers. Boca Raton, Florida, USA.
Crowder, A. and D.S. Painter. 1991. Submerged macrophytes in Lake Ontario: current
knowledge, importance, threats to stability, and needed studies. Canadian Journal of
Fisheries and Aquatic Sciences 48:1539-1545.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EXCYGBYUG_3840U0 INGEST_TIME 2012-02-07T17:21:25Z PACKAGE AA00004283_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

Pilot Study The Florida Wetland Condition Index (F WCI): Preliminary Development of Biological Indicators for Fore sted Strand and Floodplain Wetlands Report Submitted to the Florida Department of Environmental Protection Under Contract #WM-683 Kelly Chinners Reiss and Mark T. Brown Howard T. Odum Center for Wetlands University of Florida Gainesville, Florida 32611-6350 June 2005

PAGE 2

ACKNOWLEDGMENTS Research on biological indicators was supported by a grant to Mark T. Brown, principle investigator, from the Florida Department of Environmental Protection (FDEP). FDEP staff provided support for this research, particularly Russ Frydenborg, Ellen McCarron, Ashley ONeal, Erica Hernandez, Julie Espy, Tom Frick, Joy Jackson, Liz Miller, Johnny Richardson, and Lori Wolfe. FDEP staff served as reviewers for the draft report, including Russ Frydenborg, Connie Bersock, Nia Wellendorf, Julie Espy, and Joy Jackson. Additionally, acknowledgement is due to the systems ecology research group at the Howard T. Odum Center for Wetlands. In particular, Chuck Lane (who developed biological indicators for Florida marshes) provided a framework for this analysis. Assistance in field-data collection, laboratory analysis, data entry, and/or feedback on statistical analyses from Eliana Bardi, Matt Cohen, Tony Davanzo, Melissa Friedman, Kristina Jackson, Joanna Reilly-Brown, Vanessa Rumancik, Kris Sullivan, Jim Surdick, and Casey Chinners Virata, was particularly valuable. Acknowledgement is due to the Florida botanists who participated in the Floristic Quality Assessment index surveys from 1999-2004, including: Guy Anglin, Anthony Arcuri, Dan Austin, Keith Bradley, Kathy Burks, David Hall, Ashley ONeal, Jim Poppleton, Nina Raymond, Bruce Tatje, John Tobe, and Wendy Zomlefer. This project and the preparation of this report were funded in part by a Section 319 Nonpoint Source Management grant from the U.S. Environmental Protection Agency through a contract with the Florida Department of Environmental Protection. ii

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS................................................................................................ii LIST OF TABLES...............................................................................................................v LIST OF FIGURES..........................................................................................................vii EXECUTIVE SUMMARY...............................................................................................ix CHAPTER 1 INTRODUCTION AND OVERVIEW...................................................................1 Forested Strand and Floodplain Wetlands...............................................................1 Historical Perspective.............................................................................................. 5 Biological Indicators of Ecosystem Integrity..........................................................6 Quantifying Anthropogenic Influence..................................................................... 8 Landscape Development Intensity Index..................................................... 8 Human Disturbance Gradient...................................................................... 9 Project Overview...................................................................................................11 2 METHODS............................................................................................................13 Study Area.............................................................................................................13 Site Selection......................................................................................................... 14 Gradients of Landscape Development Intensity....................................................17 Field-data Collection..............................................................................................21 Transect Sampling Design.........................................................................21 Floristic Quality Assessment.....................................................................24 Data Analysis ........................................................................................................27 Summary Statistics.....................................................................................27 Regional Compositional Analysis..............................................................28 Community Composition...........................................................................29 Metric Development..............................................................................................30 Florida Wetland Condition Index..........................................................................31 3 RESULTS..............................................................................................................33 Gradients of Anthropogenic Activity.....................................................................33 Landscape Development Intensity.............................................................33 Human Disturbance Gradient....................................................................37 iii

PAGE 4

Water Quality.............................................................................................40 Data Analysis.........................................................................................................40 Summary Statistics.....................................................................................42 Regional Compositional Analysis..............................................................42 Community Composition...........................................................................44 Metric Selection.........................................................................................46 Tolerance metrics...........................................................................46 Floristic Quality Assessment Index metric....................................49 Exotic species metric.....................................................................52 Native perennial species metric.....................................................52 Florida Wetland Condition Index..............................................................53 Cluster Analysis.........................................................................................56 Landscape Development Intensity Index and the Florida Wetland Condition Index..................................................................................................58 Human Disturbance Gradient and the Stream Condition Index............................61 4 DISCUSSION ........................................................................................................64 Describing Biological Integrity..............................................................................65 Richness, Evenness, and Diversity........................................................................65 Measuring Anthropogenic Activity.......................................................................67 Regionalization of the Florida Wetland Condition Index......................................69 Florida Wetland Condition Index Independent of Wetland Type.........................69 Limitations and Further Research..........................................................................71 Conclusions............................................................................................................71 APPENDIX A Standard Operating Procedures..............................................................................72 B Coefficient of Conservatism Scores.......................................................................80 C Summary Statistics.................................................................................................87 D Metric Scoring Criteria..........................................................................................88 REFERENCES..................................................................................................................89 iv

PAGE 5

LIST OF TABLES Table Page 1-1 Non-renewable energy use and LDI coefficients per land use used in the calculation of the LDI index..................................................................................10 1-2 Categorical scoring criteria used to calculate the Human Disturbance Gradient (HDG).....................................................................................................12 2-1 Site characterization for 24 freshwater forested wetlands.....................................16 3-1 Landscape Development Intensity (LDI) index scores for 10 forested strands using 1995 and 2000 land use coverages...................................................34 3-2 Landscape Development Intensity (LDI) index scores for 14 forested floodplain wetlands using 1995 and 2000 land use coverages..............................35 3-3 Human Disturbance Gradient (HDG) for 13 storet stations, which correspond with the forested floodplains sampled................................................39 3-4 Water quality (chemical and physical parameters) for 13 storet stations, which correspond with the forested floodplains sampled......................................41 3-5 Richness, evenness, and diversity of the macrophyte assemblage among a priori land use categories....................................................................................43 3-6 Richness, evenness, and diversity of the macrophyte assemblage between low (LDI < 2.0) and high (LDI 2.0) LDI groups................................................43 3-7 Macrophyte community composition similarity among Florida wetland regions (Lane 2000) and bioregions (Griffith et al. 1994) with MRPP.................44 3-8 Spearmans correlation coefficients for the macrophyte metrics and FWCI with LDI_F/wo_200m................................................................................46 3-9 Comparisons among five macrophyte metrics and the FWCI between low (LDI < 2.0) and high (LDI 2.0) LDI groups (LDI_F/wo_200m)................47 3-10 Tolerant indicator species for forested strand and floodplain wetlands................48 3-11 Sensitive indicator species for forested strand and floodplain wetlands...............50 3-12 Exotic species identified at 24 forested strand and floodplain wetlands...............54 3-13 FWCI scores and LDI values for wetland clusters based on macrophyte community composition.........................................................................................59 3-14 Correlations of metrics and FWCI scores with 20 variations of the LDI index....60 v

PAGE 6

3-15 Forested Wetland Condition Index (FWCI), Landscape Development Intensity Index (LDI_F/wo_200m), Human Disturbance Gradient (HDG), and Stream Condition Index (SCI) data available for 13 forested floodplain wetlands................................................................................................62 3-16 Correlations among four measures of ecosystem condition or anthropogenic activity, including the Human Disturbance Gradient (HDG), Stream Condition Index (SCI), Landscape Development Intensity Index (LDI_F/wo_200m), and the Florida Wetland Condition Index (FWCI).....63 4-1 The five metrics of the preliminary Florida Wetland Condition Index for freshwater forested strand and floodplain wetlands based on the macrophyte species assemblage.................................................................................................65 vi

PAGE 7

LIST OF FIGURES Figure Page 1-1 Photograph showing the interior of a freshwater forested strand wetland in Osceola County, Florida......................................................................................3 1-2 Photograph showing the interior of a freshwater forested floodplain wetland in Polk County, Florida..............................................................................4 2-1 Florida wetland regions defined by climatic and physical variables (solid line; Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994)..................14 2-2 Study site location of 24 forested wetlands in Florida...........................................15 2-3 Boundaries of the 100 and 200 m buffers drawn around the downstream transect for the LDI_T calculation at Forested Floodplain 3 (FF3).......................18 2-4 Boundaries of 100 and 200 m buffers around a wetland feature and designated land use................................................................................................19 2-5 Boundary of the watershed buffer around Forested Floodplain 3 (FF3) for the LDI_WS calculation...................................................................................20 2-6 Idealized transect layout for forested strand wetlands...........................................22 2-7 Idealized transect layout for forested floodplain wetlands....................................25 3-1 Comparison between the LDI calculations including (1995_LDI_T/w_100m) and excluding (1995 LDI_T/wo_100m) wetland area...........................................36 3-2 Comparison between LDI calculations at the transect (1995_LDI_T/wo_200m) and feature (1995 LDI_F/wo_200m) scale..................3 6 3-3 Comparison between LDI calculations at the watershed scale using equal weighting for all area within a watershed (1995 LDI_WS_ED/wo) and a distance weighted approach using linear weighting (1995 LDI_WS_DW_lin)....38 3-4 Comparison between 1995 and 2000 LDI calculations at the feature scale (without the wetland area) for 13 forested wetlands..............................................38 3-5 NMDS ordination bi-plot of 24 sample wetlands in macrophyte species space....45 3-6 The proportion of tolerant indicator species at wetlands increased with increasing development intensity...........................................................................49 3-7 The proportion sensitive indicator species at wetlands decreased with increasing development intensity...........................................................................5 1 3-8 FQAI scores decreased with increasing landscape development intensity............51 3-9 The proportion of exotic species at a wetland increased with increasing development intensity............................................................................................52 vii

PAGE 8

3-10 The proportion of native perennial species decreased with increasing development intensity (LDI)..................................................................................55 3-11 Forested Wetland Condition Index (FWCI) scores decreased with increasing development intensity (LDI)................................................................55 3-12 Change in average species p-value from the randomized Monte Carlo tests at each step in clustering........................................................................................57 3-13 Change in the number of significant indicator species from the indicator species analysis performed at each step in clustering............................................57 3-14 FWCI scores for three wetland clusters based on macrophyte community composition............................................................................................................58 viii

PAGE 9

EXECUTIVE SUMMARY PILOT STUDY THE FLORIDA WETLAND CONDITION INDEX (FWCI): PRELIMINARY DEVELOPMENT OF BIOLOGICAL INDICATORS FOR FORESTED STRAND AND FLOODPLAIN WETLANDS Over 30 years ago, the federal Water Pollution and Control Act obliged states to protect and restore the chemical, physical, and biological integrity of waters, and charged states with establishing water quality standards for all waters within state boundaries including wetlands. Criteria for defining water quality could be narrative or numeric, and it could be addressed through chemical, physical, or biological standards. Initially, states used chemical and physical criteria (testing waters for chemical concentrations or physical conditions that exceeded criteria), assuming losses in ecosystem integrity if the criteria were exceeded (Danielson 1998). The United States Environmental Protection Agency (USEPA) recognized the potential of biological criteria to assess water quality standards and in the late 1980s required states to use biological indicators to accomplish the goals of the Clean Water Act (USEPA 1990). In effect, biological assessment has evolved into one of the standard monitoring tools of water resource-protection agencies over the past 2 decades (Gerristen et al. 2000). Such biological assessment programs have been created for lakes and streams throughout the United States ( Barbour et al. 1996a ; Karr and Chu 1999 ; Gerristen et al. 2000), and more recently efforts to assess wetland condition have been initiated (Mack 2001; USEPA 2002). Within Florida, biological indices have been created based on macroinvertebrate community composition for streams (Barbour et al. 1996a; Fore 2004), lakes (Gerristen and White 1997), and isolated depressional freshwater herbaceous (Lane et al. 2003) and forested (Reiss and Brown 2005) wetlands. Biological indices have also been created based on the community composition of the diatom and macrophyte assemblages for Florida freshwater herbaceous and forested wetlands (Lane et al. 2003; Reiss and Brown 2005). The primary objective of this research was to develop a preliminary Florida Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland study sites were sought in various a priori designated land use categories that included natural, agricultural, and urban land uses. An independent measure of anthropogenic activity in the landscape was calculated for each wetland using the Landscape Development Intensity index (LDI) (Brown and Vivas 2005). The contribution of this research to our understanding of changes in the macrophyte community composition of forested strand and floodplain wetlands in relation to different anthropogenic activities in the surrounding landscape can be summarized in five main points: 1. Five macrophyte based metrics including proportion tolerant indicator species, proportion sensitive indicator species, Floristic Quality Assessment Index (FQAI) score, proportion exotic species, and proportion native perennial species, were useful biological indicators for defining biological integrity for forested strand and floodplain wetland vegetation; ix

PAGE 10

2. Vegetation richness, evenness, and diversity were not sensitive to a priori land use categories or development intensities in the surrounding landscape for forested strand and floodplain wetlands; 3. The Landscape Development Intensity (LDI) index was a useful tool correlating with the measured biological condition of vegetation for forested strand and floodplain wetlands; 4. Regional species lists for metrics would enhance the forested strand and floodplain Florida Wetland Condition Index (FWCI); 5. An FWCI with a set of core metrics could be developed for Florida freshwater wetlands, which includes separate species lists for indicator species by wetland type and ecoregions and separate Floristic Quality Assessment Index (FQAI) scores for species by wetland type. The FWCI provided a quantitative measure of the biological integrity of forested strand and floodplain wetlands in Florida. Comprised of five metrics, the FWCI was developed based on the community composition of the macrophyte species assemblages. Metrics were selected for inclusion in the FWCI based on the correlation (nonparametric Spearmans correlation coefficient) of each metric with a quantitative gradient of Landscape Development Intensity (LDI); based on a metrics visually distinguishable correlation with LDI in a scatter plot; and based on a statistical difference of metric values between low and high LDI groups (Mann-Whitney U-test). The FWCI was composed of individual metrics, which were scaled and added together, creating the preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest score of 50 reflecting the highest biological integrity and the lowest score of zero reflecting a lack of biological integrity or no similarity to the reference wetland condition. The five macrophyte metrics that met the three selection criteria (Spearmans correlation coefficient (|r| 0.50, p < 0.05), visually distinguishable scatter plots, and Mann-Whitney U-test between LDI groups (p<0.10)) were proportion tolerant indicator species; proportion sensitive indicator species; Floristic Quality Assessment Index (FQAI) score; proportion exotic species; and proportion native perennial species. Forested Strand and Floodplain FWCI Metrics for Wetland Vegetation 1. Proportion Tolerant Indicator Species 2. Proportion Sensitive Indicator Species 3. Floristic Quality Assessment Index (FQAI) Score 4. Proportion Exotic Species 5. Proportion Native Perennial Species The variable sensitivities of three different independently derived indices compared to the forested strand and floodplain FWCI, including the Landscape Development Intensity index (LDI; Lane et al. 2003; Brown and Vivas 2005), the Human Disturbance Gradient (HDG; Fore 2004), and the Stream Condition Index (SCI; Fore 2004), suggest that multiple measures of biological integrity may be more effective at describing ecosystem wide biological integrity than any single measure based on an individual species assemblage or surrounding land use activity. However, the strong x

PAGE 11

correlations among FWCI, LDI, and HDG (Spearmans correlation coefficient |r| 0.58, p < 0.05), and lack of correlations of SCI with both FWCI and LDI, suggest that in-stream macroinvertebrate based measures of biological condition and surrounding forested wetland macrophyte based measures of biological condition did not respond in a consistent manner to changes in anthropogenic activity. Using both the in-stream macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte FWCI biological assessment methods may provide a more complete picture of the overall condition of a wetland and associated stream at a particular spatial location. While agreement in the ranking of the biological condition of study wetlands using the FWCI and SCI was anticipated, discrepancies among the ranking from the different assemblages may provide great insight into biological condition as different species assemblages respond to changes in anthropogenic activities and the associated changes in inflows (e.g. nutrient enrichment) over different time scales. Additionally, use of the forested strand and floodplain FWCI may lead to specific conclusions as to the biological condition of local or nearby anthropogenic activity, while use of the SCI may enhance understanding of larger watershed scale influences from anthropogenic activity (i.e. due to the convergence of water within the watershed associated with stream flow). The quantitative score of biological integrity established through the FWCI can be used as an objective, quantitative means of comparing changes in macrophyte community composition for wetlands, including those impacted by varying degrees of anthropogenic influence. While the forested strand and floodplain FWCI for flowing water systems can not be used to predict changes in the physical and chemical parameters of a wetland, its strength lies in providing an overview of biological integrity through the integration of changes in macrophyte community composition from cumulative effects. xi

PAGE 12

CHAPTER 1 INTRODUCTION AND OVERVIEW Assessment techniques for categorizing ecosystem condition have been established for many Florida ecosystems, including freshwater lakes (Lake Condition Index, LCI: Gerristen and White 1997), streams (BioRecon and Stream Condition Index, SCI: Fore 2004), and depressional freshwater wetlands (Florida Wetland Condition Index, Lane et al. 2003; Reiss and Brown 2005). This research furthers the development of the Florida Wetland Condition Index (FWCI), establishing preliminary metrics with species lists specific to forested strand and floodplain wetlands, referred to as flowing water systems. The overall goal of the FWCI is to use changes in community composition to characterize the biological condition of wetland ecosystems. Wetland condition is scored on a numeric scale developed from the addition of scores from individual metrics. A metric is defined as a biological attribute having a consistent and predictable response to anthropogenic activities (Karr and Chu 1997). Each metric represents an indication of biological integrity, or a signal of ecosystem condition, based on a change in community composition from the reference standard condition. The reference standard condition is defined as the condition of wetlands surrounded by undeveloped landscapes and without apparent human induced alterations. By designating a measure of ecosystem condition we refer to what others have described as ecosystem integrity, defined by Karr and Dudley (1981) as the ability of an aquatic ecosystem to support and maintain a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of the natural habitats of the region. Forested Strand and Floodplain Wetlands Wetlands have been categorized in many different ways based on any number of community attributes including, but not limited to, dominant vegetation, hydrology, soil type, and location in the landscape (Mitsch and Gosselink 1993; Keddy 2000; Kent 2000). One of the most widely recognized classification systems in North America is that by Cowardin et al. (1979). Our study focused on what Cowardin et al. (1979) categorize palustrine wetlands, commonly described as freshwater marshes and swamps. More specifically, palustrine wetlands are defined as nontidal wetland ecosystems with trees, shrubs, persistent emergents, or emergent mosses or lichens as the dominant vegetation type, and tidal wetlands with these vegetation types with ocean-derived salinity levels below 0.5 (Cowardin et al. 1979). Palustrine wetlands occur throughout the landscape as small, shallow, permanent or intermittent water bodies; shoreward of lakes, river channels, or estuaries; on river floodplains; in isolated catchments; on slopes; or as islands in lakes or rivers (Cowardin et al. 1979). The category of palustrine wetlands includes eight classes: aquatic bed, emergent, forested, moss-lichen, rock bottom, scrub-shrub, unconsolidated bottom, and unconsolidated shore. Wetlands for this study were primarily in the forested wetland class, which includes wetlands in all water regimes, except subtidal wetlands, that are characterized by 1

PAGE 13

2 woody vegetation 6m tall or greater. The structure of palustrine forested wetlands typically includes an overstory of trees with an understory of young trees and shrubs and an understory of herbaceous species (Cowardin et al. 1979). The class of palustrine forested wetlands includes six subclasses and dominance types, including broad-leaved deciduous, needle-leaved deciduous, broad-leaved evergreen, needle-leaved evergreen, dead, and indeterminate deciduous. Four categories of water modifiers are used to describe the palustrine forested wetlands in this study (as categorized on the National Wetlands Inventory GIS coverage available from the Florida Geographic Data Library at http://www.fgdl.org ) including temporarily flooded, seasonally flooded, semipermanently flooded, and permanently flooded. During the growing season, temporarily flooded wetlands have surface water present for brief periods with a water table typically well below the soil surface. Vegetation in temporarily flooded wetlands consists of facultative species, including those that grow in both uplands and wetlands. Early in the growing season seasonally flooded wetlands have surface water standing for extended periods most often without surface water late in the season but with a water table near the soil surface. Similarly, semipermanently flooded forested wetlands generally have standing surface water throughout the growing season and a water table at or near the soil surface when not flooded. At the flood extreme, permanently flooded wetlands have standing surface water throughout the year with a vegetation community of obligate wetland species (Cowardin et al. 1979). Within the text Ecosystems of Florida (Myers and Ewel, eds. 1990), Ewel (1990) describes approximately 10 distinctive types of swamps. The freshwater forested strands in this study most closely resemble wetlands in the cypress pond and strand category. Wharton et al. (1976) described cypress strands as a diffuse freshwater stream flowing through a shallow depression on a greatly sloping plain. While Mitsch and Gosselink (1993) suggest that cypress strands are found primarily in south Florida, Ewel (1990) notes that cypress strands are common throughout Florida and are found where water flow is sufficient to create a depression channel in areas with little slope but where actual flow is seldom observed. The definition of forested strands for this study broadly encompasses all of these definitions with the primary distinction of forested strands including evidence of channelized flow (Figure 1-1), though actual flow was rarely observed at any of the strands during the 2003 growing season during the period of sampling. The forested floodplain wetlands in this study most closely resemble the river swamp category in Ecosystems of Florida (Myers and Ewel, eds. 1990); however the floodplain forests in this study were associated with smaller river systems than those typically characterized by Ewel (1990). Forested floodplain wetlands in this study were associated with low order streams and rivers and were not associated with the main channels of the largest river systems in Florida (ex. Apalachicola, Suwannee, etc.). The forested floodplain wetlands in this study can also be categorized as riparian wetlands such named for the influence on the wetland environment by the adjacent stream or river system. Mitsch and Gosselink (1993) note vast differences among riparian wetlands, with the common link being the interconnection between the riparian zone, the river or stream, and the adjacent upland environment. Riparian wetlands in the southeastern United States are characterized by low-lying, low slope, and broad floodplain areas with seasonally pulsing hydrologic influences on well developed soils (Mitsch and Gosselink

PAGE 14

3 Figure 1-1. Photograph showing the interior of a freshwater forested strand wetland in Osceola County, Florida. The dark organic layer in the center of the photo shows evidence of flowing water during times of high water. 1993). The floodplain wetlands in this study are similar to the strands, however standing water was always observed in the channelized stream (Figure 1-2). The average hydroperiod (the seasonal pattern and length of saturated soils or standing water level during a year (Ewel 1990; Mitsch and Gosselink 1993)) varies among strand and floodplain wetlands, with strands having a moderate length hydroperiod with saturated soils or standing water for six to nine months a year and floodplain forests having a short length hydroperiod with generally less than six months of saturated soils or standing water during a year (Ewel 1990). Fire frequency in forested strand and floodplain wetlands ranges from moderate frequency (approximately one per 20 years) for strands to low frequency (approximately one per 100 years) for floodplain wetlands (Ewel 1990). Strands and floodplains also differ in their organic matter accumulation depths, with strands having high organic matter accumulation with an organic soil layer greater than 1 m deep and floodplain wetlands having low organic matter accumulation with an organic soil layer less than 1 m deep (Ewel 1990). Additionally, strand and floodplain wetlands have different primary water sources with strands receiving most water from shallow groundwater sources and the main source of water for floodplain wetlands from surface water originating from the associated stream or river (Ewel 1990). Despite these differences in fire frequency, organic matter accumulation, and water source, the species composition is similar among forested strand and floodplain wetlands. Common shared tree species in strand and floodplain wetlands include Acer

PAGE 15

4 Figure 1-2. Photograph showing the interior of a freshwater forested floodplain wetland in Polk County, Florida. The channelized stream is visible in the center of the photo showing the presence of a permanently flooded and flowing stream adjacent to the floodplain wetland. rubrum (red maple), Fraxinus caroliniana (water ash), Gordonia lasianthus (loblolly bay), Liquidambar styraciflua (sweetgum), Magnolia virginiana (sweet bay), Nyssa sylvatica (black gum), Persea palustris (swamp bay), Quercus laurifolia (swamp laurel oak), Sabal palmetto (cabbage palm), Salix caroliniana (coastal plain willow), and Taxodium distichum (baldcypress). Additional tree species common to strands include Annona glabra (pond apple), Pinus elliottii (slash pine), and Pinus palustris (longleaf pine). The tree stratum of floodplain wetlands has greater species richness with additional common tree species including Alnus serrulata (hazel alder), Betula nigra (river birch), Carpinus caroliniana (American hornbeam), Carya aquatica (water hickory), Carya glabra (pignut hickory), Celtis laevigata (hackberry), Chamaecyparis thyoides (Atlantic white cedar), Diospyros virginiana (persimmon), Fraxinus pennsylvanica (green ash, red ash), Fraxinus profunda (pumpkin ash), Gleditsia aquatica (water locust), Magnolia grandiflora (southern magnolia), Nyssa aquatica (water tupelo), Pinus glabra (spruce pine), Pinus taeda (loblolly pine), Planera aquatica (planer tree), Platanus occidentalis (American sycamore), Quercus lyrata (overcup oak), Quercus michauxii (basket oak, swamp chestnut oak), Quercus nigra (water oak), Quercus virginiana (live oak), Rhapidophyllum hystrix (needle palm), Sabal minor (bluestem, dwarf palmetto), Salix nigra (black willow), and Ulmus americana (American elm).

PAGE 16

5 Forested strand and floodplain wetlands also share a number of species in the shrub stratum including Cephalantus occidentalis (buttonbush), Clethra alnifolia (sweet pepperbush), Cliftonia monophylla (black titi), Cyrilla racemiflora (titi), Ilex cassine (dahoon holly), Itea virginica (Virginia willow), Lyonia lucida (fetterbush), Myrica cerifera (wax myrtle), and Rubus argutus (blackberry). Additional common shrub species in forested strands include Chrysobalanus icaco (coco plum), Ilex glabra (gallberry), Leucothoe racemosa (fetterbush), Myrica heterophylla (northern bayberry), Myrsine guianensis (myrsine), Psychotria sulzneri (wild coffee), Psychotria undata (wild coffee), and Vaccinium arboretum (sparkleberry). Other common shrub species in river swamps include Aronia arbutifolia (red chokeberry), Crataegus marshallii (parsley haw), Ilex decidua (possum haw), Ilex vomitoria (yaupon), Leucothoe axillaries (dog-hobble), Rhododendron viscosum (swamp honeysuckle), Rubus betulifolius (blackberry), Sambucus canadensis (elderberry), Sebastiana fruticosa (Sebastian bush), Viburnum nudum (swamp haw), and Viburnum obovatum (small viburnum, black haw). The species composition of woody vines in strand and floodplain wetlands vary a great deal according to Ewel (1990), as there is only one shared species, Smilax laurifolia (bamboo-vine, catbrier). Common woody vine species in strands include Ampelopsis arborea (pepper vine), Ficus aurea (strangler fig), Ficus citrifolia (wild banyan tree), Vitis aestivalis (summer grape), and Vitis shuttleworthii (calusa grape); and common woody vine species in floodplain wetlands include Ampelopsis arborea (pepper vine), Aster carolinianus (climbing aster), Smilax walteri (coral greenbrier), Toxicodendron radicans (poison ivy), and Vitis rotundifolia (muscadine grape). Species lists for common tree, shrub, and woody vine species were adopted from Ewel (1990). Forested strand and floodplain wetlands provide important habitat for wildlife such as invertebrates, amphibians, reptiles, birds, and mammals. Benthic invertebrates form the base of the forested wetland food chain, and water quality is strongly related to the diversity of the benthic macroinvertebrate community (Ewel 1990). Strands and floodplain wetlands provide valuable habitat for bird and mammal species characterized by low vegetation density and high cavity density. Though these wetlands differ somewhat in their relative contributions to bird and mammal habitat as strands have low canopy insect production, low production of edible fruits and seeds, and high presence of water, whereas forested floodplain wetlands have high canopy insect production, high production of edible fruits and seeds, and low presence of water (Ewel 1990). Historical Perspective Over 30 years ago, the Water Pollution and Control Act (later referred to as the Clean Water Act, 1972) required states to restore and maintain the chemical, physical, and biological integrity of the Nations waters (USEPA 1990). This legislation included establishing water quality standards for all waters within state boundaries, including wetlands. Such water quality criteria could be qualitative or quantitative, and it could be addressed through chemical, physical, or biological standards. Initially, states used chemical and physical criteria (testing waters for chemical concentrations or physical conditions that exceeded known standards), assuming losses in ecosystem integrity if these standards were exceeded (Danielson 1998).

PAGE 17

6 Several shortcomings have been noted when deriving ecosystem integrity based on exceeding established limits for chemical and physical parameters. Such criteria have been considered incomplete in their ability to reflect more than the temporal concentration of substances within a water body (Karr 1993). For instance, the use of toxicity parameters for determining ecosystem integrity may falsely indicate high ecosystem integrity when a single toxicity parameter went overlooked. This same water body could have elevated levels of other toxins or metals that went untested, or be physically altered so that it has lost functions typically associated with a fully functioning water body (Karr and Chu 1997). Furthermore, chemical and physical sampling may not occur during specific loading events and may therefore incompletely describe the ecological condition of the system. Adams (2002) points out that other environmental factors such as sedimentation, alterations to habitat, varying temperature and oxygen levels, and changes in ecological aspects like food availability and predator-prey relationships are not reflected with chemical criteria alone. James and Kleinow (1994) note that different organisms respond in different ways to the amount, persistence, and exposure of chemical compounds otherwise foreign to an organism; and single-valued chemical and physical criteria of water quality may overlook important biological implications. Alternatively, biological indicators integrate the spatial and temporal effects of the environment on resident organisms, and are suitable for assessing the possible effects of multifaceted changes in ecosystems (Adams 2002). Karr and Chu (1997) and Adams (2002) note that biological indicators signal changes in the environment that might otherwise be overlooked or underestimated by methods that depend on chemical criteria alone. Organisms have an intricate relationship with their environment, reflecting current and cumulative ecosystem condition (Karr 1981). The presence of biological organisms reveals chemical exposure, expressing changes in the physical, chemical, and biological components of the ecosystem through changes in community composition (Adams 2002). The United States Environmental Protection Agency (USEPA) recognized the potential of biological criteria to assess water quality standards and in the late 1980s required states to use biological indicators to accomplish the goals of the Clean Water Act (USEPA 1990). In effect, biological assessment has evolved into one of the standard monitoring tools of water resource protection agencies over the last two decades (Gerristen et al. 2000). Biological criteria and monitoring programs through the USEPA have been created for lakes and streams throughout the United States (Barbour et al. 1996a; Karr and Chu 1999; Gerristen et al. 2000), and more recently efforts to assess wetland condition have been initiated (USEPA 2002). Indicators of Biological Integrity Biological monitoring to assess ecosystem condition has been applied widely in ecological research. The primary aim of biological monitoring is to detect changes in abundance, structure, and diversity of target species assemblages. One trend in biological monitoring has led to the development of indices of biotic integrity (referred to as IBIs), for different species assemblages including diatoms (Fore and Grafe 2002; Fore 2004); macrophytes (Galatowitsch et al. 1999a; Gernes and Helgen 1999; Mack 2001; Lane

PAGE 18

7 2003); macroinvertebrates (Kerans and Karr 1994; Barbour et al. 1996b); amphibians (Micacchion 2004); fish (Schulz et al. 1999); and birds (OConnell et al. 1998). Perhaps the most common species assemblage chosen for use in the development of IBIs is the macroinvertebrate assemblage, because many of the macroinvertebrate species rely entirely on the conditions of their aquatic environment for habitat, food, and reproductive activities. In Florida there are currently three biological indices that use the community composition of the macroinvertebrate assemblage to detect changes in biological integrity including the Lake Condition Index (LCI, Gerristen and White 1997), Stream Condition Index (SCI, Fore 2004), and the Florida Wetland Condition Index for depressional freshwater wetlands (FWCI, Lane et al. 2003; Reiss and Brown 2005). However, the forested strand and floodplain systems targeted in this study have varying hydrologic regimes from temporarily to permanently flooded, complicating macroinvertebrate collection due to variable hydrologic conditions. As such, we have chosen the macrophyte species assemblage for use in the preliminary forested strand and floodplain FWCI. Wetland macrophytes are defined as aquatic emergent, submergent, or floating plants growing in or near water (USEPA 1998); and are described as distinguishing landscape features. The spatial distribution of macrophytes in the landscape occurs according to a multitude of factors, including hydroperiod, water chemistry, and substrate type, as well as other factors such as available seed source and climate. Fennessy et al. ( 2001 ) state that the community composition of wetland macrophytes typifies the physical, chemical, and biological wetland dynamic in time and space. Macrophytes play a vital role in supporting the structure and function of wetlands by providing food and habitat for other assemblages including algae, macroinvertebrates, fish, amphibians, reptiles, birds, and mammals; and macrophyte populations can be used as a diagnostic tool to assess other aspects of the wetland environment. Crowder and Painter (1991) state that a lack of macrophytes where they are otherwise expected to grow suggests reduced wildlife populations from lack of food or cover and/or water quality concerns such as toxic chemical constituents, increased turbidity, or increased salinity. In contrast, an overgrowth of particular macrophytes may signify increased nutrient loading (USEPA 1998). Many advantages of studying macrophytes as indicators of wetland condition have been noted, including their large, obvious size; ease of identification, to at least some useful taxonomic level; known response to toxicity tests; and general lack of ability to move to avoid unfavorable conditions (Danielson 1998; Cronk and Fennessy 2001). Additionally, macrophytes readily respond to changes in nutrient, light, toxic contaminant, metal, herbicide, turbidity, water, and salt levels. They can also be sampled in the field with transects, or remotely from aerial photography; and well-established field methods of sampling macrophytes exist (USEPA 2003). Furthermore, the USEPA (2003) states additional advantages of using the macrophyte assemblage, including that they do not require laboratory analysis, can easily be used for calculating simple abundance metrics, and are superb integrators of environmental condition. In general, macrophytes represent a useful assemblage for describing wetland condition (Mack 2001). Schindler (1987) alleges that macrophytes can provide a more integrated picture of wetland function than measures such as nutrient cycling, productivity, decomposition, or chemical and physical composition.

PAGE 19

8 There are however some noted shortcomings of using macrophytes as biological indicators. These include the potential delay in response time for perennial shrub and tree species, difficulty identifying taxa to the species level in certain seasons and for some genera, uneven herbivory patterns, and varied pest-management practices (Cronk and Fennessy 2001). Despite these limitations, macrophytes have provided strong signals of anthropogenic influence (USEPA 2003). In fact, many states have begun using macrophytes in their wetland biological assessment programs, including, Minnesota (Galatowitsch et al. 1999a; Gernes and Helgen 1999), Montana (Apfelbeck 2000), North Dakota (Mushet et al. 2002), Ohio (Mack 2001), and Florida (Lane et al. 2003; Reiss and Brown 2005). Quantifying Anthropogenic Influence Wetlands occupy a large portion of the Florida landscape. An estimate from the 1780s reported 8,225,000 ha of wetlands in Florida (Dahl 2000). By the mid-1980s, the National Wetlands Inventory estimated Florida had 4,467,000 ha of wetlands remaining, translating into a loss in Florida of roughly 45% of the pre-1780s wetland area (Mitsch and Gosselink 1993; Dahl 2000). Throughout the continental United States, similar trends were apparent, with a drastic decline in the surface area of wetlands. More specifically, Dahl ( 2000 ) reported that 98% of all wetland losses throughout the continental United States from 1986 to 1997 were losses to freshwater wetlands. Of the remaining freshwater wetlands, 40% were adjacent to agricultural lands and therefore potentially affected by land use practices such as herbicide and pesticide application, irrigation, livestock watering and wastes, soil erosion, and deposition. An additional 17% were adjacent to urban or rural development. Freshwater non-tidal wetlands experienced the greatest development pressure just inland from coastlines as the demand for housing, transportation infrastructure, and commercial and recreational facilities increased ( Dahl 2000 ). These changes in land use were proportionally more widespread in Florida than much of the continental United States due to the remarkable length of coastline along both the Atlantic Ocean and Gulf of Mexico coasts of Florida. Agricultural and urban development activities influence an array of changes to the physical, chemical, and biological characteristics in nearby ecosystems. There have been numerous attempts at quantifying anthropogenic influence based on quantitative indices, for example, the Wetland Rapid Assessment Procedure (WRAP; Miller and Boyd 1999), the Minnesota disturbance index (Gernes and Helgen 1999), the Human Disturbance Gradient (HDG; Fore 2004), and the Landscape Development Intensity (LDI) index (Brown and Vivas 2005). Our study incorporates the LDI index as a measure of anthropogenic influence. Additionally, for thirteen floodplain wetlands, HDG scores have been obtained from previous studies (Fore 2004; Florida Department of Environmental Protection Geographic Information Systems map layers). Landscape Development Intensity Index The LDI index has been used as a gauge of human activity based on a development intensity measure derived from nonrenewable energy use in the surrounding landscape. The underlying concept behind calculating the LDI (quantifying the nonrenewable energy use per unit area in the surrounding landscape) stems from earlier

PAGE 20

9 works by Odum (1995), who pioneered emergy analysis for environmental accounting. [Emergy is an environmental accounting term referring to expressing energy use in solar equivalents (Odum 1995).] Brown and Ulgiati (2005) suggest that landscape condition, or ecosystem health, is strongly related to the surrounding intensity of human activity, and that ecological communities are affected by the direct, secondary, and cumulative impacts of activities in the surrounding landscape. Healthy ecosystems are defined as those with integrity and sustainability, which correlate to limited development in the surrounding landscape and the maintenance of ecosystem structure and function, even when stressors (e.g. flooding, drought, etc.) are present (Brown and Ulgiati 2005). The LDI scale encompasses a gradient from completely natural to highly developed land use intensity, and is calculated based on the percent of the area in a particular land use within the designated area surrounding the wetland (ex. 100 meters, 200 meters, etc.) multiplied times the LDI coefficient (Table 1-1), which is defined by the amount of nonrenewable energy use for a given land use (Brown and Vivas 2005). The LDI coefficient does not account for any individual causal agent directly, but instead, may represent the combined effects of air and water pollutants, physical damage, changes in the suite of environmental conditions (ex. groundwater levels, increased flooding), or a combination of such factors, all of which enter the natural ecological system from the surrounding developed landscape. Wetlands surrounded by more intense activities such as highways and multi-family residential land uses receive higher LDI index values, as the highest LDI coefficient of 10.0 is assigned to the urban land use category of Central Business District. Undeveloped land uses such as wetlands, lakes, and upland forests are assigned an LDI coefficient of 1.0, the lowest possible value, based on no use of nonrenewable energy in these ecosystems. Human Disturbance Gradient The Human Disturbance Gradient (HDG) is a quantitative measure used to assess the level of human disturbance to an ecosystem based on the cumulative score of four independent measures of the environment including ammonia concentration in the water (as mg N/L), hydrologic index, habitat assessment, and LDI for the buffer which included an area of 100 m on each side of the stream and 10 km upstream of the sampling point ( Fore 2004 ). The first measure included in the HDG, ammonia concentration, is included as a summary variable defining water quality because of its consistent correlation with other water quality parameters (i.e. total phosphate) and its availability in the dataset ( Fore 2004 ). Additionally, increases in ammonia concentration are thought to be evidenced from both agricultural (e.g., fertilizer and farming practices) and urban activities, whereas changes in total phosphorus concentrations are thought to be solely associated with agricultural operations ( Fore 2004 ), so using ammonia concentration accounts for changes in both agricultural and urban land use activities. The second measure included in the HDG is an estimate of hydrologic condition of the stream site, which is scored on-site at the discretion of the biologist conducting the sampling effort. Scores for the hydrologic condition range from 1-10, with a score of 1-2 Excellent representing a natural, undisturbed system with few impervious surfaces, high connectivity with ground water and surface features, and a natural flow regime ( Fore 2004 ). Intermediate categories include 3-4 Good, 5-6 Moderate, and 7-8 Poor. The final category, 9-10 Very Poor, is reserved for those systems with a flow regime entirely

PAGE 21

10 Table 1-1. Non-renewable energy use and LDI coefficients per land use used in the calculation of the LDI index. Land Use Nonrenewable Energy Use (E14 solar equivalent joules/ha/yr) LDI Coefficient Natural Land / Open Water 0.0 1.00 Pine Plantation 5.1 1.58 Low Intensity Open Space / Recreational 6.7 1.85 Unimproved Pastureland (with livestock) 8.3 2.06 Improved pasture (no livestock) 19.5 2.89 Low Intensity Pasture (with livestock) 36.9 3.51 High Intensity Pasture (with livestock) 51.5 3.83 Citrus 65.4 4.06 Medium Intensity Open Space / Recreational 67.3 4.09 Row crops 117.1 4.63 High Intensity Agriculture (dairy farm) 201.0 5.15 Single Family Residential (Low-density) 1077.0 6.79 Recreational / Open Space (High-intensity) 1230.0 6.92 Single Family Residential (Med-density) 2461.5 7.59 Low Intensity Transportation 3080.0 7.81 Single Family Residential (High-density) 3729.5 7.99 Low Intensity commercial (Comm Strip) 3758.0 8.00 Institutional 4042.2 8.07 Highway (4 lane) 5020.0 8.28 Industrial 5210.6 8.32 Multi-family residential (Low rise) 7391.5 8.66 High intensity commercial (Mall) 12661.0 9.18 Multi-family residential (High rise) 1285.0 9.19 Central Business District (Avg 2 stories) 16150.3 9.42 Electric Power Facility 29401.3 10.00 controlled by human modification, with a flashy hydrograph and extreme alteration of the natural ecosystem ( Fore 2004 ). Habitat assessment (or habitat condition index as termed in Fore 2004 ) is the third measure included in the HDG. The Florida Department of Environmental Protection (FDEP) has developed Standard Operating Procedures (SOPs) for river and stream habitat assessments (FDEP-SOP-001/01: Form FD 9000-5 June 1, 2001 available from the Bureau of Laboratories at http://www.dep.state.fl.us/labs/library/forms.htm ). The habitat assessment is separated into four primary habitat components: substrate diversity (number of diverse, productive habitats), substrate availability (percent of stream reach composed of productive habitat), water velocity (based on the maximum observed

PAGE 22

11 velocity, where higher velocities receive higher scores), and habitat smothering (as the percent of the stream reach covered by sand or silt accumulation). The four secondary habitat parameters include artificial channelization (as an assessment of anthropogenic modification of the stream reach), bank stability (evidence of bank stability/instability from erosion or bank failure), riparian buffer zone width (estimate of the width of the vegetation on the least buffered side), and riparian zone vegetation quality (estimate of community composition and structure). These eight habitat parameters are scored by the field biologist on a scale of 1-20, with 20 representing the highest habitat quality. Three of the secondary habitat components (bank stability, riparian buffer zone width, and riparian zone vegetation quality) are scored separately on a scale of 1-10 (with 10 representing the highest habitat quality) for both the right and left banks (when added together these three categories are still given equal weighting, with half of the score reflecting the condition of each bank). Scores for each of the eight parameters are summed, and a stream is then assigned a categorical score of optimal (134-160), suboptimal (91-123), marginal (54-80), or poor (11-43) based on the total score ( Fore 2004 ). The fourth measure of the HDG is an LDI score calculated within a 100 m buffer on each side of the stream for 10 km upstream from the sampling point (LDI_BF) ( Fore 2004 ). Each measure included in the HDG is categorically assigned a score based on the score for each measure ( Table 1-2 ). Ammonia concentration in the water (as mg N/L), habitat assessment, and LDI for the buffer are assigned scores of 0, 1, or 2; whereas the hydrologic index is assigned scores of 0, 1, 2, or 3. Overall HDG values potentially range from 0 (no detectable human induced disturbance) to 9 (extreme anthropogenic disturbance) ( Fore 2004 ). Original categorical scoring criteria for the HDG were established from correlations of the HDG categories (ammonia, hydrologic condition, habitat index, and LDI_BF) with the macroinvertebrate metric EPT (Ephemeroptera, Plecoptera, and Trichoptera) taxa richness. All HDG scores used in this analysis were taken directly from HDG scores provided by FDEP. Project Overview The community composition of the macrophyte assemblage was sampled in flowing wetland systems including forested strand (n = 10) and forested floodplain (n = 14) wetlands throughout Florida. Our primary goal was to develop a preliminary Florida Wetland Condition Index (FWCI) for freshwater flowing water wetlands, building upon the FWCI for isolated depressional herbaceous ( Lane et al. 2003 ) and forested ( Reiss and Brown 2005 ) freshwater wetlands. Our secondary goal was to determine the appropriate scale for LDI buffers for flowing water wetlands that would best capture the current wetland condition, based on correlations with the FWCI. Our third goal was to correlate measures of the HDG ( Fore 2002 ) and macroinvertebrate community data from the Florida SCI with our macrophyte FWCI and landscape LDI measures of ecosystem condition. Wetland study sites were sought in various landscape settings that included natural, agricultural, and urban land uses. The elongated forested wetlands associated with flowing waters were in many areas continuous for extended geographic distances, and the forested edges bordered a great variety of land use activities. Therefore, it was

PAGE 23

12 Table 1-2. Categorical scoring criteria used to calculate the Human Disturbance Gradient (HDG). The HDG is the sum of the categorical scores for the four individual measures ammonia concentration in the water (as mg N/L), hydrologic index, habitat assessment, and LDI for the buffer. The HDG range is from 0 (no detectable human induced disturbance) to 9 (extreme anthropogenic disturbance). HDG Categorical Score HDG Measure 0 1 2 3 Ammonia concentration (mg N/L) < 0.1 0.1 2.0 > 2.0 Hydrologic Condition < 6 6 7 8 9 10 Habitat Assessment > 65 50 65 < 50 LDI for the buffer < 2.0 2.0 3.5 > 3.5 difficult to ascertain the specific land use activity contributing most heavily or frequently to perturbations in community composition. While a priori land use categories were assigned (reference, agricultural, urban), wetlands were also divided into two groups representing low and high landscape development intensities.

PAGE 24

CHAPTER 2 METHODS Twenty-four freshwater forested wetlands, including 10 strand and 14 floodplain wetlands, were sampled during 2003. This chapter describes site selection, Landscape Development Intensity (LDI) index calculations, field-data collection, and methods of statistical analyses. Study Area Florida has been dissected into smaller geographic zones using a number of different methods, including bioregions ( Griffith et al. 1994 ) with 13 subecoregions (grouped into four ecoregions for small wadeable streams of panhandle, northeast, peninsula, and Everglades); lake ecoregions of Florida ( Griffith et al. 1997 ) with 47 subecoregions; and Florida wetland regions determined with a hydrologic model by Lane ( 2000 ), which included physical (surficial geology, soils, digital elevation model, slope) and climatic (precipitation, potential evapotranspiration, runoff, annual days of freezing) variables with four ecoregions (panhandle, north, central, and south). When boundaries for the bioregions of Griffith et al ( 1994 ) and wetland regions of Lane ( 2000 ) were compared, there were similarities among the panhandle and south/Everglades boundaries; however, the north/northeast and central/peninsula ecoregions differed in extent ( Figure 2-1 ). We adopted the four wetland regions developed by Lane ( 2000 ) to partition the state during site selection. However, Human Disturbance Gradient (HDG) and Stream Condition Index (SCI) scores available for the floodplain forests were designated according to the Florida bioregions ( Griffith et al. 1994 ); as such we use both wetland region and bioregion categories in our analyses. The four wetland regions defined by Lane ( 2000 ) are categorized as panhandle, north, central, and south. The panhandle wetland region is characterized by less human development than the other ecoregions. Streams in the panhandle wetland region typically run from north to south, discharging into the Gulf of Mexico. The primary ecosystems are scrub and high pine, temperate hardwood forests, pine flatwoods and dry prairies, and swamps ( Fernald and Purdum 1992 ). The major population centers of Tallahassee, Panama City, and Pensacola are included within the panhandle wetland region ( Davis 1967 ). The north wetland region has similar ecosystem types, primarily with pine flatwoods and dry prairies and less scrub and high pine, temperate hardwood forests, and swamps ( Fernald and Purdum 1992 ). Two major drainage features in the north wetland region include the Santa Fe and Suwannee River system and the lower St. Johns River, with Jacksonville being the major city situated along the northeastern Atlantic coast. The central wetland region has a distinguishing central ridge feature with a higher maximum elevation than the north and south wetland regions ( Fernald and Purdum 1992 ). The central wetland region is characterized by pine flatwoods and dry prairies with scrub and high pine along the ridge. There is less area in temperate hardwood forests, less area in swamps, and an increase in the area of inland lakes and freshwater marshes ( Fernald and Purdum 1992 ). The largest population centers include 13

PAGE 25

14 Figure 2-1. Florida wetland regions defined by climatic and physical variables (solid line; Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994). Orlando and Tampa. The south wetland region is unique with the nearly flat terrain of the Everglades (less than 2 m relief; Fernald and Purdum 1992). Much of the south is swamps and freshwater marsh, and few streams remain unaltered. Lake Okeechobee is in this wetland region as are the urban centers of Miami and Ft. Myers located on the densely populated east and west coasts, respectively. Site Selection During the 2003 growing season (May-July), 24 forested wetlands were sampled throughout the state of Florida (Figure 2-2). Random site selection was not feasible given the necessity of obtaining permission to access private lands and the non-random pattern of land development in Florida. The forested strand wetlands (n=10) were arranged spatially throughout Florida, so that the distribution was spread among three of the four Florida wetland regions (Lane 2000), including north (n=2), central (n=5), and south (n=3). No strands were sampled in the panhandle wetland region. The forested floodplain wetlands (n=14) were selected as a subset of systems that have been sampled using the SCI (Fore 2004). Many of the streams that had received a score of poor according to the SCI no longer had floodplain forest vegetation along the banks, but rather had mowed grass or paved banks. As such, these streams did not fit our criteria for

PAGE 26

15 Figure 2-2. Study site location of 24 forested wetlands in Florida (strands n=10; floodplains n=14). Wetland region boundaries follow Lane ( 2000 ). site selection, and where therefore excluded from sampling. Forested floodplain wetlands sampled were located in the panhandle (n=2), north (n=8), central (n=3), and south (n=1) wetland regions. A priori categories were used to define sample wetlands as reference, agricultural, and urban. The suite of reference wetlands included wetlands surrounded by intact native ecosystems. These wetlands were thought to represent the best possible wetland condition currently in Florida. These wetlands could be categorized as reference standard wetlands by the USEPA ( 1998a ), and have been described as wetlands considered to be the least altered that are reflective of characteristic levels of wetland function. Agricultural wetlands were defined as those currently surrounded by cattle pasture, rangeland, row crops, citrus, and silvicultural land uses. Urban wetlands included those embedded within commercial, industrial, and residential land uses. Many of the urban wetlands sampled were suspected to previously have been embedded in agricultural land uses due to the development patterns throughout Florida. Hereafter wetlands embedded in primarily undeveloped landscapes are called reference; wetlands embedded in primarily agricultural land uses are called agricultural; and wetlands embedded in primarily urban land uses are called urban. Table 2-1 provides general information about each sample wetland, including site code, sample date, wetland type, wetland region ( Lane 2000 ), county, bioregion ( Griffith

PAGE 27

Table 2-1. Site characterization for 24 freshwater forested wetlands. Site Code* Sample Date Wetland Type Wetland Region ^ County Bioregion ^ A Priori Category Associated Stream FF1 May 19 2003 Floodplain North Putnam Peninsula Urban Orange Creek FF2 May 20 2003 Floodplain North Marion Peninsula Reference Juniper Creek FF3 May 22 2003 Floodplain North Volusia Peninsula Urban Groover Branch FF4 May 23 2003 Floodplain Central Lake Peninsula Agricultural Blackwater Creek FF5 May 25 2003 Floodplain North Nassau Northeast Urban Pigeon Creek FF6 May 25 2003 Floodplain Central Polk Peninsula Reference Livingston Creek FF7 May 25 2003 Floodplain Central Polk Peninsula Reference Tiger Creek FF8 May 26 2003 Floodplain North Nassau Northeast Urban Alligator Creek FF9 May 29 2003 Floodplain North Clay Northeast Urban Green Creek FF10 June 8 2003 Floodplain South Lee Everglades Urban Leitner Creek FF11 June 13 2003 Floodplain Panhandle Hamilton Northeast Agricultural Rock Creek FF12 June 14 2003 Floodplain North Baker Northeast Agricultural/Urban Turkey Creek FF13 June 24 2003 Floodplain North Martin Peninsula Reference Kitchen Creek FF14 July 1 2003 Floodplain Panhandle Walton Panhandle Agricultural Limestone Creek FS1 May 24 2003 Strand Central Osceola Peninsula Reference NA FS2 May 26 2003 Strand Central Osceola Peninsula Agricultural NA FS3 May 27 2003 Strand Central Osceola Peninsula Agricultural NA FS4 May 31 2003 Strand Central Osceola Peninsula Urban NA FS5 June 1 2003 Strand Central Osceola Peninsula Urban NA FS6 June 16 2003 Strand South Palm Beach Peninsula Urban NA FS7 June 24 2003 Strand South Palm Beach Peninsula Reference NA FS8 July 7 2003 Strand North Alachua Peninsula Agricultural NA FS9 July 8 2003 Strand North Alachua Peninsula Reference NA FS10 July 14 2003 Strand South Martin Peninsula Reference NA *Site Codes correspond to the wetland type (FF = floodplain forest; FS = forested strand) and were numbered in the order sampling occurred. ^ Wetland region from Lane ( 2000 ); Bioregion from Griffith et al. ( 1994 ). Associated Stream = Stream Condition Index (SCI) stream data; NA = Not Applicable, forested strands were not associated with SCI streams.

PAGE 28

17 et al. 1994 ), a priori category of surrounding land use, and associated stream. Site codes were assigned to preserve the anonymity of individual land owners. Gradients of Landscape Development Intensity Digital orthophoto imagery, available from Labins The Land Boundary Information System from the Florida Department of Environmental Protection (FDEP) (available at http://www.labins.org/2003/index.cfm ), were used to georectify transect locations from GPS coordinates using ArcGIS 8.3 from Environmental Systems Research Institute, Inc. Polygons were established for wetlands delineated from digital orthophoto imagery. LDI scores were calculated using 1995 and 2000 land use coverages available for download through the Florida Geographic Data Library (available at http://www.fgdl.org/ ). Coverages for 1995 land use were available separately for each Florida Water Management District (Northwest Florida, NWFWMD; Suwannee River, SRWMD; St. Johns River, SJRWMD; Southwest Florida, SWFWMD; and South Florida, SFWMD). More recent coverages (2000 land use) were available for a limited number of sample wetlands including only those within the boundaries of the SJRWMD. LDI index scores were calculated at the transect (LDI_T), wetland feature (LDI_F), and watershed (LDI_WS) scale. At the transect scale, a 100 m buffer was created around the downstream sample transect. Two LDI_T scores were then calculated based on the area of the land uses within the 100 m buffer ( Figure 2-3 ), one including the wetland area and the other excluding the wetland area. The LDI_T score including the wetland area will generally be lower than the LDI_T score excluding the wetland area, based on the increase in the percent of land assigned a 1.0 Natural Land/Open Space LDI coefficient (because of the inclusion of the wetland area, which is assigned an LDI coefficient of 1.0). Including the wetland area in the calculation was thought to weight the LDI_T based on the size of the wetland sampled, which could be an important factor influencing wetland condition. LDI_T calculations were repeated using a 200 m buffer ( Figure 2-3 ). LDI_T 1995 identification of land uses within the 100 m buffer were taken from land use/land cover coverages and checked with 1995 digital orthophoto imagery. LDI_T 2000 identification of land uses were taken from 2000 land use/land cover coverages and checked with 1999 digital orthophoto imagery. Land uses that had changed since the photos were taken were updated based on land use maps drawn during field site visits. This step was not possible for larger buffer areas where visibility of surrounding land uses on site was hindered due to distance and barriers, and so discrepancies may be apparent in 100 m buffer calculations (hand corrected for each buffer) and larger 200 m buffers (not corrected, land use/land cover straight from available GIS coverages). LDI_F and LDI_WS were also calculated once including and once excluding the wetland area. The boundary for the LDI_F was established as the area making up the 200 m buffer upstream of the downstream transect ( Figure 2-4 ). The boundary of the feature was established when a distance of at least 30 m showed a break in wetland vegetation established from photo-interpretation of the digital orthophoto imagery. The upstream boundary for the LDI_WS was established using the Better Assessment Science Integrating Point and Nonpoint Sources 3.0 (BASINS 3.0) environmental analysis system ( USEPA 2001 ) ( Figure 2-5 ). BASINS is a multi-purpose environmental assessment tool

PAGE 29

18 Figure 2-3. Boundaries of the 100 and 200 m buffers drawn around the downstream transect for the LDI_T calculation at Forested Floodplain 3 (FF3). Two LDI_T values were calculated for each sample wetland, one including the wetland area and the other excluding the wetland area.

PAGE 30

19 Figure 2-4. Boundaries of 100 and 200 m buffers around a wetland feature and designated land use. The 200 m buffer was used for the LDI_F calculation at Forested Floodplain 3 (FF3).

PAGE 31

20 Figure 2-5. Boundary of the watershed buffer around Forested Floodplain 3 (FF3) for the LDI_WS calculation. Four LDI_WS values were calculated for each sample wetland: two LDI_WSs were calculated based on an equal weighting by area of all land uses rshed including a calculation inclusive of the wetland area e of the wetland area (LDI_WS_ED/wo); and two dista within the upstream wate(LDI_WS_ED/w) and exclusiv nce weighted LDIs were calculated in which land uses nearest to the sample wetland were more highly weighted, including linear weighting (LDI_WS_DW_lin) and weighting based on an exponential decay function (LDI_WS_DW_exp).

PAGE 32

21 designed for watershed and water quality studies that integrates a GIS (ArcView 3.2), national watershed database, and environmental assessment and modeling tools (USEPA 2001). The digital terrain model from the National Elevation Dataset, which is a 30 m raster-bere weighted more than those occurring farther away (WS_LDI_DW). LDI_WS_ED scores were calculated both including (LDI_WS_ED/w) and excluding (LDI_WS_ED/wo) the wetland area. The distance weighted calculations were done in two ways. First, a calculation was made using a linear weighting of land uses (LDI_WS_DW_lin); second, a calculation was made using an exponential decay function, where land uses nearest the sampling point were weighted most significantly (LDI_WS_DW_exp). GIS analyses were performed in ArcGIS 8.3 (Environmental Systems Research Institute, Inc. 2002). The following equation was used to calculate LDI: LDITotal = % LUi LDIi (2-1) where %LUi is the percent of a land use within the buffer of interest and LDIi is the LDI coefficient for a particular land use based on the amount of nonrenewable energy use per unit area in the surrounding landscape (Table 1-1). The LDI coefficient values and LDI equation were based on work by Brown and Vivas (2005). Potential LDI coefficients ranged from a minimum of 1.0 (Natural Land/Open Space) to a maximum of 10.0 (Central Business District). Field-data Collection A concise summary of field-data collection procedures follows. Appendix A provides more detailed descriptions of field-data collection techniques in the format of Standard Operating Procedures (SOPs) for field use. Transect Sampling Design Transects were established perpendicular to the hydrologic flow of the system. he wetland/upland boundary was determined based on the Florida Unified Wetland plafacMowe strands consisted of establishing four transects that extended from the wetland/upland boundary to the middle of the channelized flow in the center of the strand (Figure 2-6a). Transects were established at 25 m intervals. While it is recognized that strands were not perfectly symmetrical in nature, an effort was made to establish transects that represented a cross-section of the strand along the gradient of ased dataset produced by the United States Geological Survey (USGS), was used for watershed delineation. Two separate types of LDI_WSs were calculated based on an equal weighting by area of all land uses within the upstream watershed (LDI_WS_ED) and a distance weighted calculation, where land uses occurring closer to the sampling point w T Delineation Methodology (Chapter 62-340, F.A.C.), using a combination of wetland nt presence according to wetland plant status (e.g. obligate, facultative wetland, ultative, or upland) and wetland hydrologic indicators (e.g. lichen lines, moss collars). difications of transect establishment were implemented for strand and floodplain tlands types, described below. Sampling in forested

PAGE 33

22 (A) Figure 2-6. Idealized transect layout for forested strand wetlands. (A) Standard transect layout included four transects that represented a cross-section of the strand along the gradient of wetland/upland boundary to the water channel, with transect establishment along both sides of the water course. (B) Alternative transect placement included four transects established on the same side of the water channel due to limited access or permission restrictions. 0-5m 0-5m 5-10 5-10 10-15 10-15 15-2015-2020-2520-2525-3025-3030-35T1 T2Upland/Wetland Boundary Main Water Channel 0-5m 5-10 10-15 15-20 20-25T3 0-5m 5-10 10-15 15-2020-2525-30T425-3030-3535-40 25m Flow Direction Forested Strand Wetland Forested Strand Wetland Upland Upland

PAGE 34

23 (B) Upland/Wetland Boundary Main Water Channel Flow Direction Forested Strand Wetland Upland 0-5m5-1010-15T1 15-2020-25 T3 0-5m5-1010-1515-2020-2525-30 T4 25m 0-5m5-1010-1515-2020-25 25-30 30-35 35-40 25-30 25m 30-35T2 30-35 25m 0-5m5-1010-1515-2020-25 25-30 30-35 35-40

PAGE 35

24 upland/wetland boundary to the water channel. In some sampling efforts wetland forests were sampled along both sides of the water course. However, on some occasions, limited access or lack of permission restricted sampling to the same side of the water course. In these instances, four transects were established on the same side of the water channel (Figure 2-6b). Forested floodplain wetlands varied in size, with many wetlands spanning an area broader than 100 m across. To encourage comparable sampling efforts among the different width strand and floodplain systems, a maximum 50 m transect length was established for floodplain systems. For forested floodplains narrower than 50 m wide, four transects were established that spanned the area from the upland/wetland boundary, as determined by wetland plant presence and hydrologic indicators, to the edge of the stream channel. The initial transect was established downstream at a point closest to the SCI sampling location. Consecutive transects were established upstream at 25 m intervals. In forested floodplain wetlands wider than 50 m, the initial transect was established at the upland/wetland boundary and extending 50 m inward perpendicular to the stream. The second transect was established 25 m upstream, starting at the waters edge and extending a full 50 m towards the upland/wetland boundary, perpendicular to the water course. The pattern was repeated for the third and fourth transects (Figure 2-7). While four transects of 50 m length were considered the optimal sampling effort, this was not always achieved due to limited access and development pressures resulting in limited remaining areas of forested floodplain. Along each transect, a series of 1 m wide by 5 m long quadrats was established back to back. Living macrophytes rooted within each quadrat were identified to the lowest taxonomic level possible. When field identification was limited, a sample specimen was collected, pressed, and identified in the laboratory. An expert Florida botanist (Dr. David Hall) was consulted for identification of unknown specimens. Taxonomic information including species, genus, and family were compiled for all of the macrophytes identified. Additional characteristics were collected for use in metric development, including category (annual or perennial, evergreen or deciduous, indigenous or exotic) and growth form (aquatic, fern, grass, herb, sedge, shrub, tree, or vine). References specific to Florida were consulted first (Tobe et al. 1998; Wunderlin and Hansen 2003), and additional information was supplemented from other sources (in the following order: Godfrey and Wooten 1981, Wunderlin 1998, and USDA NRCS 2002). Wetland indicator status (i.e. obligate, facultative wetland, etc.) for each species was obtained from FAC Ch. 62-340. Wetland indicator statuses for species not listed in FAC Ch. 62-340 were obtained from Tobe et al. (1998), Wunderlin and Hansen (2003), and USDA NRCS (2002), in that order. When species were not assigned a Florida specific indicator status, the National Wetlands Inventory indicator status was used (Wunderlin and Hansen 2003; USDA NRCS 2002, in that order). When information was still unavailable for plant characteristics in published literature, Florida botanists (who also participated in the Floristic Quality Assessment Index) were consulted. muwaois. Floristic Quality Assessment Index A Floristic Quality Assessment Index (FQAI) has been included in many of the ltimetric biotic indices created for the macrophyte assemblage. The concept of FQAI s developed by Wilhelm and Ladd (1988) for vegetation around Chicago, Illin

PAGE 36

25 Figure 2-7. Idealized transect layout for forested floodplain wetlands. Transect 1 began at the upland/wetland boundary and extended 50 m inward, perpendicular to the channelized flow. Transect 2 began at the edge of the channel and extended 50 m towards the wetland/upland boundary. This pattern repeated for transects 3 and 4. Upland/Wetland Boundary Main Water Channel Forested Strand Wetland Upland 0-5m-10 510-15 15-2020-2525-3030-35 35-40 40-4545-50 T4 25m 0-5m 5-10 10-15 15-2020-25T1 0-5m 5-10 10-15 15-20 20-2525-30 25m Flow Direction 25-30 30-35T3 35-4040-4545-50 25m 30-35 35-4040-4545-50T2 0-5m5-1010-15 15-20 20-25 25-3030-35 35-40 40-4545-50

PAGE 37

26 This method of scoring plant species based on expert botanist opinion has been used in Michigan (Herman et al. 1997), Ohio (Andreas and Lichvar 1995; Fennessy et al. 1998; Mack 2001). 2002), and Floridaanehe FQAI provides a quaparticular environbotanists assigned coefficd coefficients of conservatism) to individual species. This techniqsmts is more accurate thnefor vegin to each species. The FQidual wetland was calculated as: (2-2) species richness as the denominator. Theories on the importance f species richness suggest that higher species richness signifies a more valuable ecosystem, which can be quantifieduare root function (Fennessy et al 998). However, a recent study by Cohen et al. (2004) found a stronger relationship radient for the mean CC (Eq. 2-2) than with the traditional FQAI equatiot al. 1994), categories, or other unspecified differences. onservatism scores were obtained from Florida botanist survefrom the depressional forested and herbaceous wetlands were strong ncluded in previous Florida FQAI surveys. As such, each botanis Ontario (Francis et al. 2000), North Dakota (Mushet et al (Le t al. 2003; Cohen et al. 2004; Reiss and Brown 2005). Tntitative means of assessing the fidelity of a plant to a ment through the Delphi technique (Kent 2000), where individual ients (terme ue asues that the collective decision by a group of expert botanisa th professional judgment of one individual (Kent 2000). The Florida FQAI enlisted regional expert botanists to provide quantitative scores etation the form of a coefficient of conservatism assigned individuallyAI score for an indiv FQAI = [ (C + C + C) ] / N 1 2 n where C = species specific coefficient of conservatism score (CC), and N = species richness. This equation was considered a modified FQAI because previous studies used the square root of native o by using the sq 1 along a disturbance g n (using the square root of native species richness in the denominator). They also reported that using total species richness (i.e. including exotic species in the calculation of species richness) improved the relationship of mean CC with the human disturbance gradient (measured with LDI). The sum of the species CC scores was divided by total species richness ( Eq. 2-2 ) in this study to account for potential differences in species richness due to differences in wetland regions ( Lane 2000 ), bioregions ( Griffith e a priori land useCoefficient of cys to compliment previous FQAI efforts for Florida. FQAI scores were adopted first from depressional forested wetlands ( Reiss and Brown 2005 ) and second from depressional herbaceous wetlands ( Lane et al. 2003 ; Cohen et al. 2004 ). Correlations between CC scores for species (unpublished data). In an effort to streamline the acquisition of species specific CC scores for the FQAI for strand and floodplain wetlands, CC scores were only obtained for those species not it was sent a list of species identified in the forested strand and floodplain wetlands that did not already have a CC score for the previous Florida FQAI efforts (n = 86). Botanists participating in the 2003 FQAI effort for forested strand and floodplain wetlands included Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and Bruce Tatje. Botanists

PAGE 38

27 scored cosystems hat were widely distributed and occurred in disturbed cal of well-established ecosystems that sustain only le diversity in a wetland (McCune and Grace 2002), was alculated as the Shannon diversity (H') value divided by the natural log of richness: (2-3) s of taxon i, and N was the total number of occurre each species based on its faithfulness to Florida forested strand and floodplain wetlands. Potential CC scores ranged from 0 to 10: 0 Exotic and native species that act as opportunistic invaders, included species that commonly occur in disturbed e 1-3 Species t ecosystems 4-6 Species with a faithfulness to a particular ecosystem, but also tolerant of moderate levels of disturbance 7-8 Species typi minor disturbances 9-10 Species occurring within a narrow set of ecological conditions Species with low CC scores were considered tolerant of many disturbances, whereas species with high CC scores were considered to occur within a narrow set of stable ecological condition. Appendix B lists the CC scores for the macrophyte species identified in this study. Data Analysis Summary Statistics Summary statistics for the macrophyte assemblage were calculated at the species level, including richness (R), evenness (E), Shannon diversity (H'), Simpson diversity (D), and Whittakers beta diversity ( W ). Richness (R) was defined as the total number of distinct taxa identified within the sample wetland. Evenness (E), described as the fraction of maximum possib c E = H' / ln (R) The Shannon diversity index has been described as measuring the information content of a sample unit where maximum diversity yields maximum uncertainty ( McCune and Grace 2002 ). For Shannon diversity calculations (H'), the sample unit was an individual forested wetland: H' = p i log(p i ) (2-4) p i = n i / N (2-5) where n i was the number of occurrence nces of all taxa at a wetland. The number of occurrences represented the number of quadrats a species occurred in, and the total number of occurrences of all taxa at a wetland represented the sum of the total number of quadrats of all of the species identified. Simpson diversity (the compliment of Simpsons original index, which was a measure of dominance) was a measure of the likelihood species that are randomly chosen in sampling will be different. The equation used to calculate Simpson diversity was: D = 1 p i 2 (2-6)

PAGE 39

28 Whittakers beta diversity ( W ) was computed as a calculation of overall beta diversity, or the compositional change represented in a sample. Whittakers beta diversity was calculated as the number of species at a particular forested wetland (S c ) divided by the average species richness per quadrat (S), minus one: w = [S c / S] 1 (2-7) The resulting value for Whitakers beta diversity was described as the number of distinct communities ( McCune and Grace 2002 ). When w equaled zero, all of the sample units contained all of the species. Some multivariate methods strongly depend on beta diversity, and as a general rule beta diversity greater than 5 was considered high ich described within-group similarity. When A equaled one, all items were identical within groups, and when A equaled zero, differences within-groups equaled that expected by chance. 02) found that most values of A ere less than 0.1 in community ecology. MRPP was calculated across all wetland region versus south, and central versus south) when sample size was appropriate. An appropriate (McCune and Grace 2002). Summary statistic means within a priori land use categories were compared with Fishers Least Significant Difference (LSD) pair wise comparison test using Minitab ( Version 13.1, Minitab Statistical Software ). The strength of using Fishers LSD was in the comparison of unequal group sizes ( Ott and Longnecker 2001 ; Minitab 2000 ). Sample wetlands were divided into two groups based on 1995 LDI_F/wo scores including low (LDI < 2.0) and high (LDI 2.0) LDI groups. Comparisons were made for summary statistics between low and high LDI groups using the non-parametric Mann-Whitney U-Test in Minitab ( Ott and Longnecker 2001 ). Overall calculations of beta and gamma diversity were calculated for sample wetlands in the three a priori land use categories. Gamma diversity was calculated as the overall number of taxa encountered at all sample wetlands per a priori land use category. Higher gamma diversity for an a priori land use category suggested a greater difference among the species composition of wetlands within that a priori land use category, assuming a similar number of wetlands were sampled within each a priori land use category. Beta diversity was calculated as a priori category gamma diversity divided by the mean site taxa richness. Regional Compositional Analysis The Multi-Response Permutation Procedure (MRPP) was used to test the similarity of macrophyte community composition among Lanes (2000) four Florida wetland regions (see further application of MRPP in Zimmerman et al. 1985; McCune et al. 2000; McCune and Grace 2002). MRPP is a nonparametric technique which tests for no difference between groups (the null hypothesis) and is available in PCORD (Version 4.1 from MJM Software, Gleneden Beach, Oregon). It is an appropriate procedure for ecological community data as it does not require distributional assumptions of normality and homogeneity of variances. The Srensen distance measure was used to calculate the average weighted within-group distance. MRPP analysis provided a test statistic (T), p-value, and chance-corrected within-group agreement (A), wh McCune and Grace (20 w groups (panhandle versus north versus central versus south; wetland regions according to Lane 2000 ) as well as for multiple pair wise comparisons (panhandle versus north, panhandle versus central, panhandle versus south, north versus central, north

PAGE 40

29 sample size was defined by a minimum of two wetlands within each category for comparison. For example, if there were seven wetlands being compared with six in one region and only one in a different region, the MRPP comparison analysis could not be executenorthease comparisons (panhandle versus northeast, panhanot accurately characterize ecological trends. This has been described as a compen85% of the cells within the matrix were assigne final run was comple d. Additional MRPP tests were calculated across all bioregions (panhandle versus st versus peninsula versus Everglades; bioregions according to Griffith et al. 1994) as well as for multiple pair wi dle versus peninsula, panhandle versus Everglades, northeast versus peninsula, northeast versus Everglades, and peninsula versus Everglades) when sample size was appropriate. Community Composition Community composition of the macrophyte assemblage was summarized in a non-metric multidimensional scaling (NMDS) ordination. NMDS, an ordination technique designed to compress multi-dimensional space, is particularly agreeable with ecological data because it does not rely on linear relationships among variables, which generally do n sation for the zero-truncation problem through the use of ranked distances (a common characteristic of non-parametric analyses) and the use of an appropriate distance measure ( McCune and Grace 2002 ). The zero-truncation problem refers to the extraordinary number of zeros (denoting a species is absent from a sampling location) typical in community ecology data sets. For example, in the species by site matrix used in our data analysis (283 species by 24 sites), d a zero (i.e. the species does not occur at the site). Other ordination techniques depend upon a value for each measured variable for each sample unit. However NMDS is useful for the presence/absence data sets, where many species are not present, or receive a zero in the species by site matrix. By ranking the variables, NMDS caters to the non-parametric community ecology data set. NMDS was used to explore the dissimilarities of the community composition of the macrophyte assemblage among sample wetlands. The Srensen (Bray-Curtis) distance measure was used for ordination. The dimensionality was chosen based on an initial 6-dimensional run in autopilot mode, which suggested an optimal 2-dimensional solution. To find the optimal 2 dimensional solution, 50 runs with real data and 50 randomized runs were performed with the instability criterion set at 0.00001 and the maximum number of iterations to reach a stable solution set at 500. This procedure was repeated 20 times to insure stability and reproducibility in results. The ted with the starting point set as the results from the best experimental 2-dimensional run, with the lowest stress and best overall fit. LDI_F/wo, species richness, Shannon diversity, Simpson diversity, Whittakers beta diversity, and sampling date (as Julian date) were tested for correlation with the NMDS ordination axes with Pearson correlation coefficients. Metric Development In the context of this study, metrics are defined as biological attributes which have a consistent and predictable response to anthropogenic activities (Karr and Chu 1997). Metrics were summarized in three main categories, including tolerance indicators,

PAGE 41

30 community structure, and community balance. Metrics were as the proportion (P) by dividing the number of species (n) fitting a particular metric category (ex. exotic species) dividedistent and predictable trends along the gradiene adopted from previous FWCI studies (Lane et al. 2003; Reiss aus, and basGrace 2002). Calculated indicator values are based on two standards, faithfulness and exclusion. Faithfulness is defined mathematically by a particular species always being present in a erfect indicator species is xclusive to that group, meaning it never occurs in other groups. Indicator values range from 0nt indicator species; species signific by the total species richness (N) for each sample wetland: P i = n i / N i (2-8) where i represents an individual wetland. Scatter plots were constructed for each candidate metric versus 1995 LDI_F/wo_200m to ensure correlations were visually distinguishable. Candidate metrics were accepted if they successfully distinguished between low (LDI < 2.0) and high LDI (LDI 2.0) groups using the Mann-Whitney U-test (p < 0.10) calculated with Version 13.1, Minitab Statistical Software, and showed cons t of anthropogenic land use activity according to the strength and significance of the Spearmans correlation coefficient (p < 0.05) calculated with Analyze-It software v. 1.67 for Microsoft Excel. The Spearman rank correlation tested for an association between two related variables, and was a non-parametric alternative to the Pearson correlation. Macrophyte metrics wer nd Brown 2005 ), included tolerance (tolerant species; sensitive species), exotic, FQAI, and longevity and plant growth form (native perennial). Additional metrics were explored for summary statistics (richness, evenness, and diversity), as well as dominance, annual to perennial ratio, wetland indicator status, native evergreen, native deciduo al area (total, deciduous, or evergreen). Tolerance metrics were calculated with Indicator Species Analysis (ISA) in PCORD ( Version 4.1 from MJM Software, Gleneden Beach, Oregon ). Sample wetlands were categorized into low and high LDI groups and analyzed with ISA, which evaluates the abundance and faithfulness of species in a defined group ( McCune and Grace 2002 ). ISA is used to detect and describe the value of species indicative of environmental conditions. It requires a priori groups and data on the presence of taxa in each group. Groups are commonly defined by categorical environmental variables, levels of disturbance, experimental treatments, presence and absence of a target species, or habitat types ( McCune and particular group. Additionally, the p e (no indication) to 100 (a perfect indication of a particular group). Multiple ISA iterations were conducted to determine sensitive and tolerant indicator species. Sample wetlands were categorized based on 1995 LDI_F/wo_200m scores, and ISA calculations were completed for consecutive LDI breaks from 1.0 through 3.5 at each 0.5 increment using the presence/absence of taxa at each wetland. In total ISA was run six times. Species significantly associated (p < 0.10) with the higher LDI group (LDI > 1.0; LDI > 1.5; LDI > 2.0; LDI > 2.5; LDI > 3.0; LDI > 3.5) for each iteration were included in the overall list of tolera antly associated (p < 0.10) with the lower LDI group (LDI = 1.0; LDI 1.5; LDI 2.0; LDI 2.5; LDI 3.0; LDI 3.5) for each iteration were included in the overall list of sensitive indicator species. Exotic species were excluded from the sensitive indicator species list, as this list was constructed to represent reference biological condition, which

PAGE 42

31 should exclude the presence of nonnative species. A random number seed was used for each ISA iteration. Calculated indicator values were tested for statistical significance using a Monte Carlo randomization technique with 1000 randomized runs. Indicator species categorized as tolerant species were associated with a higher LDI group; indicator species categorized as sensitive species were associated with a lower LDI group. Tolerant and sensitive indicator species lists were compiled based on a collection of significant indicator species at each incremental LDI break. This method was employed due to the small sample size of wetlands available to construct indicator species lists. The proportion indicator species metrics were calculated for each wetland as the number of indicator species (either tolerant or sensitive) divided by species richness. The proportion exotic species metric was calculated as the number of species that were exotic to Florida divided by species richness. The timeline for determining the exotic status of a species was set near the beginning of European settlement in North America. Sources consulted to determine the exotic status of a species included Godfrey and Wooten ( 1981 ), Tobe et al. ( 1998 ), Wunderlin ( 1998 ), USDA NRCS ( 2002 ), and Wunderlin and Hanse n (2003). determined that disturbance often favors annual species over perennial species and promotes the invasion of nonnative perennials in wetlands. Galatowitsch et al. (2000) found that while native perennial cover was reduced in wetlands impacted by cultivation, the occurrence of introduced perennials rather than annuals increased in stormwater impacted wetlands. Florida Wetland Condition Index The Florida Wetland Condition Index (FWCI) consists of individual metrics, which were scaled and added together. Metric scoring was based on an approach from the Stream Condition Index (SCI), a Florida based biological index of the macroinvertebrate assemblage used to discern stream condition (Fore 2004) and also used in the development of the FWCI for depressional forested wetlands (Reiss 2004). Metrics were natural log transformed to improve the normality of distribution. The 5th to 95th percentile values of each metric were normalized from 0 to 10, with 10 always representing the reference biological wetland condition. An agglomerative cluster analysis in PCORD was used to determine wetland groups (McCune and Grace 2002), for wetlands with similar macrophyte community composition. The dissimilarity matrix was constructed using the Srensen distance measure and the flexible beta linkage method ( = -0.25), which is a flexible clustering FQAI scores for each wetland were calculated based on the sum of CC scores for each species identified divided by total species richness at each wetland (Eq. 2-2). FQAI scores had the potential to range from 0.0-10.0. Wetlands with higher FQAI scores represented wetlands with species indicative of more stable environmental conditions. The proportion native perennial species was calculated as the number of native perennial species encountered divided by wetland species richness. Each species was categorized as native or exotic and annual or perennial according to Godfrey and Wooten (1981), Tobe et al. (1998), Wunderlin (1998), USDA NRCS (2002), and Wunderlin and Hansen (2003). Wienhold and Van der Valk (1989) and Ehrenfeld and Schneider (1991)

PAGE 43

32 setting designed to reduce chaining in the dendrogram. The resulting dendrogram was pruned using ISA, which provided an objective, quantitative means of selecting the optimum number of clusters represest ecologically meaningful point of pruning (McCune and Grace 2002). Gership determined in each step of the agglomerative cluster analysis was used ashe grouping variable for ISA. Using the ecies by site presence/absence matrix, indicator values were calculated for each species occurring in three or mom. This is an entirely separate application of ISA as that used to construct the metrics for tolerant and sensitive indicatoom the randomized Monte Carlo tests were averaged among all species at each step of the dendrogram, and the cluster step with the lowest average p-value was used to ram yielding the greatest ecologically meaningful informaor forested strands). The LDI_F/wo_200m and HDG represented measures of anthrop nting the mo roup memb t sp re sites for each step of the dendrogra r species. This method of using ISA to determine the most appropriate and meaningful point on a dendrogram for understanding a significant number of groups for a data set is described more completely by McCune and Grace (2002). The resulting p-values fr determine the level in the dendrog tion ( McCune and Grace 2002 ). FWCI scores and LDI index values among clusters were then compared using Fishers LSD pair wise comparison (p < 0.05) available in Minitab To determine how the biological data compared to the intensity of landscape development, Spearmans rank correlation was used to compare individual metrics and the FWCI with the multiple LDI calculations. For forested floodplain wetlands, comparisons of the FWCI and LDI_F/wo_200m were made with the HDG and SCI scores using the non-parametric Spearmans rank correlation coefficient (HDG and SCI scores were not available f ogenic influence; whereas the FWCI and SCI represented biological indices based on the macrophyte community composition in the forested wetland and macroinvertebrate community composition in the channelized stream flow, respectively.

PAGE 44

33 CHAPTER 3 RESULTS nic Activity Two measures of anthropoge nic activity in the lands cape were correlated to d the Hu man Disturbance Gradient cape Development Intensity LD e forested strand ( Table 3-1 ) and tions were calculated for all 24 we tlands using 1995 land use coverages and 2000 se co eDistrict (SJRWMD). Land uses in the 100 m buffers for LDI_T tioner a l orthophoto imagery and assigned land use ries hed (WS) scale LDI calculations were com a ulations including (1995_LDI_T/w_100m) and 5 LDI_T/wo_100m) wetland area ) show that calculations ltural and urban wetlands sistently higher than those incl uding the wetland area (1995 LDI_T/w_100m) re curring above the 1:1 r ethods. In calculation of LD I_T/w compensates foferences in wetland area between la ay not corr elate to differences ice unity composition to enic activ i lsc at the transect and feature scales buffers ( ause the forested strand and ticipated that the feature gnal of organisms that are a priori land ct (1995 LDI_T/wo_200m) and and three urba n wetlands; whereas the rema ining 11 wetlands received scores, including five agricultural and six urban wetlands. equal weighting for all area within a nd area, (1995 LDI_WS_ED/wo) and a distance weighted physical, chem Landscape Developm (HDG). Lands floodplain ( calcula land u Managem calcula catego Waters wetlands. A com excluding excluding the wetland area ( were con for the fo line. Fou effect the study wet in biolog anthropog The correlation between 1995 LDI calculations with 200 m floodplain wetlands involve channe scale LDI would better reflec affected by upstream use category, received identi feature scales (1995 LDI_F/wo_200m LDI_F/wo_200m agricultural, higher 1995 LDI_T/wo_200m Watershed scale LDI calculations using watershed, excluding the wetla G r adi en ts o f A nt hr opo ge ic al, a e nd nt b In iol ten ogi s ity cal ( m LD ea I) su ind r es of wetland condition, including the ex an I in Ta de b x c le alc 3-2 ula ) wetlands. Transect (LDI_T) and feature (LDI_F) scale tio ns we re co mp let ed for th ver nt ag es f or we tla nd s lo ca ted wi thi n the boundary of the St Johns River Water s w bas e d on eli g nea rou te nd d f tr rom uth d ed igit la ed nd us e m a ps drawn during the field site visit. pleted for the 14 forested floodplain ( Figure 3-1 ) for agricu r dif ay or m cores, including six reference, one p riso n bet we en LD I calc (199 19 95 LD I_ T/w o_ 10 0m sted refe w re et nce lan w ds, etla as nd ca s r n b ece e s ive een d b a s y the values generally oc c ore of 1.0 using both calculation m nd al r s. Ho spo we nse ver s th uc is h dif as fer ch enc ang e i e s in macrophyte comm n a r ea m ities Figure 3-2 anthropogenic activities. Three wetlands, one in each an n the 99 and ) wa t the biological response si LD ape. /w s low (R lized flowing water, it is an I_T 2 = 0m 0.32). Bec s cal scores for the transe 5 ) Ten wetlands received higher 1995 20 th 1 o_

PAGE 45

Table 3-1. Landscape Development Intensity (LDI) index scores for 10 forested strands using 1995 and 2000 land use coverages. LDI calculations were completed at the transect (T) and wetland feature (F) scale, and including (/w) and excluding (/wo) the wetland area. See text for further details of individual calculations. FS1 FS2 FS3 FS4 FS5 FS6 FS7 FS8 FS9 FS10 A Priori Group R A A U U U U A R R 1995 LDI_T/w_100m1.22.42.31.21.93.71.02.61.3 DI_T/w0m1.33.53.31.42.34.61.03.51.4 LDI_T/w_200m1.32.62.41.22.13.51.02.91.4 DI_T/w0m1.33.23.51.33.53.91.03.21.5 LDI_F/w_200m2.62.53.11.93.03.01.03.01.7 DI_F/w0m3.02.93.11.93.53.21.03.12.1 LDI_T/w_100m1.22.62.34.35.45.81.06.11.0 DI_T/w0m1.23.82.58.07.07.51.08.01.0DI_T/wm 1.52.62.3na na na na 6.91.4na I_T/w0m1.53.22.3na na na na 7.41.6na I_F/wm 2.52.52.6na na na na 6.91.7na DI_F/w0m2.92.93.0na na na na 7.22.1na 1.2 1995L o_10 1.2 1995 1.0 1995L o_20 1.0 1995 1.0 1995L o_20 1.0 2000 1.0 2000L o_10 1.2 2000 L _200 2000 LDLD o_20 2000 _200 2000 L o_20 na 20 00 land userage naila thtla cove es wer ot av ble for ese we n ds.

PAGE 46

ent Intensity (LDI) index scores for 14 forested floodplain wetlands using 1995 and 2000 land usecoverages. LDI calculations were completed at the transect (T), w and watershed (WS) scale, and including (/w) and excluding (/wo) the wetland area. See text for further details ations. Site CoFF1 FF2 FF3 FF4 FF5 8 FF9 FF10 FF11FF12FF13FF14 etland feature (F),of individual calculFF6 FF7 FF de A Prioe U R U A U U U A A R A ri Land Us R R U 1995 1.9 1.0 5.2 1.5 4.8 1.2 4.2 1.0 1.2 1995 3.2 1.0 7.6 2.2 6.9 1.6 6.3 1.0 1.5 1995 2.2 1.0 5.9 1.2 4.0 1.3 4.3 1.0 1.7 1995 3.1 1.0 7.1 1.3 4.7 1.6 6.5 1.0 2.1 1995 1.5 1.1 3.2 1.6 5.7 1.6 2.4 1.9 1.7 1995 2.0 1.2 3.7 2.3 7.1 1.6 2.8 2.1 1.8 1995 2.4 1.5 3.8 2.2 2.1 1.4 1.9 1.8 2.0 2.5 1.5 2.9 2.1 1.4 1.9 1.8 2.1 1995 2.2 1.4 2.2 2.2 2.1 2.7 2.8 2.3 1.5 2.2 1.4 2.0 1.7 2.0 2.3 1.4 2.7 2.8 1.4 2.8 1.6 1.9 2000 1.9 1.0 5.2 1.5 5.6 1.2 2.9 1.0 1.2 2000 3.2 1.0 7.6 2.2 8.1 1.6 5.2 1.0 1.5 2000 2.2 1.0 5.9 1.0 2.2 na 1.3 na na 4.1 na na 2000 3.1 1.0 7.1 1.1 3.1 na 1.4 na na 5.6 na na 2000 1.5 1.1 3.0 1.6 1.8 na 1.5 na na 2.3 na na 2000 2.0 1.3 3.5 2.2 2.0 na 1.7 na na 2.7 na na 2000 2.3 1.5 2.3 2.2 1.7 na 1.6 na na 1.9 na na 2000 2.4 1.5 2.3 2.3 1.8 na 1.8 na na 1.9 na na 2000 2.2 1.5 2.4 2.2 1.9 na 1.7 na na 2.0 na na 2000 LDI_WS_DW_exp 2.2 1.4 2.8 2.2 1.9 na1.6 na na 2.8 na na 1.9 1.1 1.6 3.2 1.0 3.3 1.1 2.1 4.4 1.0 2.3 1.0 1.0 3.3 1.1 3.1 1.0 1.0 3.9 1.1 1.7 1.4 1.9 2.0 1.5 1.9 1.4 2.5 2.5 1.8 1.8 2.7 3.1 2.2 1.7 2.3 1.9 2.7 3.4 2.5 1.7 2.2 2.1 2.6 2.6 2.6 1.5 1.9 1.1 1.6 3.3 1.0 3.3 1.1 2.1 4.5 1.0 na 3.4 na 4.8 na 2.0 na 2.5 na 2.3 na 2.6 na 2.3 na 2.6 Table 3-2. Landscape Developm na 2000 laoverages were not available for these wetlands. LDI_T/w_100m LDI_T/wo_100m LDI_T/w_200mLDI_T/wo_200mLDI_F/w_200mLDI_F/wo_200mLDI_WS_ED/w 1995 LDI_WS_ED/wo LDI_WS_DW_lin 1995 LDI_WS_DW_exp LDI_T/w_100mLDI_T/wo_100mLDI_T/w_200mLDI_T/wo_200mLDI_F/w_200mLDI_F/wo_200mLDI_WS_ED/w LDI_WS_ED/wo LDI_WS_DW_lin nd use c

PAGE 47

36 R2 = 0.9810 0246802468101995 LDI_T/w_100m1995 LDI_T/wo_100m Strands Floodplains 1:1 Line 1:1 Line Figure 3-1. Comparison between the LDI calculations including (1995_LDI_T/w_100m) and excluding (1995 LDI_T/wo_100m) wetland area (R 2 = 0.98). Calculations excluding the wetland area (1995 LDI_T/wo_100m) for agricultural and urban wetlands were consistently higher than those including the wetland area (1995 LDI_T/w_100m) for the forested wetlands. Four reference wetlands received a score of 1.0 using both calculation methods. R2 = 0.3246810LDI_T/wo_200m 1995 0202468101995 LDI_F/wo_200m Strands Floodplains 1:1 Line Figure 3-2. Comparison between LDI calculations at the transect (1995_LDI_T/wo_200m) and feature (1995 LDI_F/wo_200m) scale for 24 forested wetlands (R 2 = 0.32).

PAGE 48

37 approach using linear weighting (1995 LDI_WS_DW_lin) showed a stronger correlation (R2 = 0.82) for the 14 forested floodplain wetlands (Figure 3-3). LDI calculations were also completed for multiple years, using land use/land cover data from 1995 and 2000. Comparison between 1995 and 2000 LDI calculations at the feature scale (without the wetland area) for 13 forested wetlands showed a moderate correlation (R2 = 0.41) (Figure 3-4). However, 12 of the wetlands had nearly equal LDI scores for the 1995 and 2000 LDI_F/wo_200m calculations with one clear disagreement, FS8, which had a significant area of the 200m buffer categorized as Low Intensity Pasture (with livestock) in 1995 and Institutional in 2000. The land use was not actually changed, only the classification on the GIS coverage was changed, as the property is part of the University of Floridas (thus the Institutional classification) Beef Research Unit (thus the Low Intensity Pasture (with livestock) classification). Removing FS8 because of erroneous differences in the land use/land cover coverages, the remaining 12 wetlands had strongly correlated 1995 and 2000 LDI_F/wo_200m calculations (R2 = 0.99 without FS8). allhas as described above (Figure 3-4). Using the older land use/land cover data from 1995 may have limited implications in changes to LDI scores as changes in land use are localized to t development. In these areas it may be argued that more recent overages are needed, however at the statewide scale the 1995 land use coverage rovides a baseline for LDI calculations. We are not suggesting that the most recent and complete coverage should not be used, but simply that the LDI calculation should not be disregarded because the coverage is ten years old. Third, because the 24 wetlands surround channelized flowing water, the feature scale LDI calculations were used to capture the human activity influencing the upstream wetlands. Fourth, the LDI_T_100m calculations included hand delineation of land use according to interpretation of digital orthophoto imagery and ground truthing; however, the application of LDI as a remote (GIS computer based) tool for assessing human activity in the landscape, using widely available coverages, was an important consideration for future application. Human Disturbance Gradient Thirteen of the forested floodplain wetlands sampled corresponded to storet stations established by FDEP for water quality monitoring. Data were available for the four components of the Human Disturbance Gradient (HDG), including ammonia concentration, hydrologic condition, habitat assessment (habitat condition index), and LDI for the buffer (Table 3-3). Using the scoring criteria established by FDEP (Table 1-), each of the HDG components were scored individually and summrange fromimpaired condition. The maximum HDG score for the 13 forested floodplain wetlands was a 4 at the urban wetland FF10 (storet 28020234) for the most recent sampling period LDI_F/wo_200m LDI calculations for 1995 land use/land cover were selected for further analysis for four reasons. First, 1995 land use/land cover data were available for of the 24 forested wetlands sampled. Second, calculations for 13 forested wetlands ving both 1995 and 2000 LDI_F/wo_200m calculations showed similar value areas with recenc p 2 ed to get the HDG for each storet station. HDG scores were adopted directly from previous FDEP assessments, and no scoring was done directly for this research. HDG scores potentially 0-9, with 0 representing the reference condition and 9 representing a highly

PAGE 49

38 R2 = 0.8201231995 LDI_WS_ED/wo19LDI_WSin 4 0 1 2 3 4 95 _DW_l Floodplains 1:1 Line ll area within a watershed (1995 LDI_WS_ED/wo) and a distance weighting (1995 LDI_WS_DW_lin) (R2 = 0.82). Figure 3-3. Comweightingweighted approach using linear Figure 3-4. Com(without the wetland area) fo p for a arison between LDI calculati ons at the watershed scale using equal parison between 1995 and 20 00 LDI calculations at the feature scale r 13 forested wetlands (R 2 = 0.41). R 2 = 0.41 0 262000 LDI_F/200m 4 8 02 460m 1 995 LD I_F /wo _20 wo_ Strands 1:1 Line F lood pla ins

PAGE 50

man Disturbance Gradient (HDG) for 13 storet stations, which correspond with the forested floodplains sampled. Some foulns fo9thncentratiobinsc Code 3 RE rested floodplain wetlands have multiple HDG calce HDG were also provided, including ammonia cooring criteria were presented in Table 1-2. STORET HDG Date HDG Ammonia (NH atio r different sampling dates (19n, Hydrologic Condition, HaHydrologic Condition Hydro SCORE Habitat Assessmen 2-2tat It 00 0). The four components of dex, and LDI buffer. HDG Habitat SCORE LDI Buffer LDI_BF SCORE 3 ) (mg N/L) NHSCO FF1 20020404 8/15/2000 0 0 4 0 84 0 1.38 0 FF1 FF2 FF3 FF4 FF5 FF5 FF5 FF5 FF5 FF6 FF7 FF8 FF9 FF10 FF10 FF11 FF12 FF14 0 1.38 00 1.06 01 1.79 00 1.02 0 1 1.27 00 1.27 0 0 1.27 00 1.27 00 1.27 0 0 1.51 0 0 1.04 00 1.40 00 1.21 00 4.73 20 4.73 20 1.16 0 0 1.56 00 1.33 0 4 0 841 0 884 0 582 0 84 3 0 653 0 88 3 0 913 0 903 0 91 3 0 86 2 0 897 1 882 0 997 1 817 1 863 0 97 2 0 865 0 75 20020404 8/15/2000 0 0.04 0 20010454 7/19/1999 0 0.01 0 27010580 8/22/2000 2 0.10 120010455 2/23/1999 0 19010099 8/17/1993 1 019010099 3/5/1998 0 19010099 2/1/2000 0 0.04 0 19010099 7/9/1996 0 0.02 019010099 8/18/1992 0 -26011019 3/8/1999 0 26011020 7/20/1999 0 0 19020027 7/15/1996 1 0.08 0 20030246 4/17/1997 0 0 28020234 10/9/2000 4 0.12 1 28020234 9/11/1996 3 0.06 021010032 1/17/1995 0 19010072 3/22/1999 0 0 32010024 7/14/1998 0 0.02 0 no*bel data available ow detection limits Table 3-3. Hu

PAGE 51

40 ints were available for many of the SCI points used in this ject, and the most complete da ta from the most recent sampling period was oderate to low range of scores suggests that the ds sled represented wetlands on the lower scale of impairment. This may be u pling locations with SCI e si ss ncceptable for this study been comple ted altered, often because the area had or paved as part of adjacent human activity in the surrounding landscape. s nsidered for macrophyte Wquality data (chemical and physical parameters) were obtained for 13 of the d ater table included temperature ( C), pH, specific conductivity (umhos/cm), eygen (mg O2/L), turbidity (NTU), Total Kjeldahl Nitrogen (mg/L), ammonia N/L), and total phosphorus (mg P/L) ( Table 3-4 ). FF5 an urban wetland had the alur ratur7.oC), snductivi(49 umhos/cm), and the t va rast FF8, another urban (5.7 NTU), Total Kjeldahl Nitrogen (1.50 nd total phosphorus (0.28 mg P/ L). A third urban wetland FF10 had the v st values for temperature C Data Analysis S milies identified. The most abundant species was the vine Vitis li uscadine), which was found rooted within vegetation quadrats at 23 of 24 Acer rubrum (red maple) at 7o s Woodwardia virginica ia chain fern) found at 54% of the wetlands; the most common graminoid was (variable witchgrass) found at 46% of the wetlands. Boehmeria ica (f shru dentalis (buttonbush) and (wax myrtle), each found at 71% of the wetlands; the second most of the wetlands; and the o 39%) occurred at a um of three sam of the species identified ies or 1 nds; approximately twoonly one wetland. In f forested floodplain ds hosted e forested on. (10/9/2000). [Multiple data po research pro used for correlation purposes.] The m wetlan an acc data w because the floodplain wetland had been sodded These sampling and FW Water Quality foreste parame dissolv (as mg lowest v highes wetland, had the highest values for turbidity mg N/L), a lowest (25.8 amp rate s re vi tate site m d en du t w rin he g n con te si rec der on ing nai t hat an m ce, an b y s ut tre fou am s d u am na ites, c on sid C ere I d d ev hig elo hl pm y im en p t, air as ed the o y n th no e lon SC ge I, r h wer ost e ed no w t co etland vegetation. ater floo ers av d ox dp ail lain we tla nds s am pled that corresponded to storet stations. Wo es lue fo fo tem dis pe so e ( ox 1 pe m cif g ic O co2 ty ont r lve d yg en (1 0.2 /L ). In c alue ) and for am di m sso on lv ia ed nitr ox og yg en en (0 (2. .12 6 0 mg mg O N 2 /L /L) ), an d t he hig he o 190 genera and 97 fa rotundifo study wetlands. The second m found (Virgin Panicum commutatum cylindr wetlands; the two m Myrica cerifera common tree was second m at71% of the wetlands. Of the species en minim (48 spec fifths of the species total the wetlan in comm tatew a (m id e, 24 we tla nd s w e re sam p led w ith 28 3 ma crophyte species, representing os tla t nd abu s nd Th an e t s mo pec st ies co w mm as o th n f e tr ern ee wa 5% f th e s tu dy we a lse ne ttl o e) st c wa om s t m he on m os t c bs om we mo re n Ce he ph rba ala ce nt ou hu s species, found at 63% of the s o cci S ab al p pa le lm we et tla to nd (ca Toxicodendron radicans s bb (12 ag .5 e p countered, 110 species ( %) alm ) N e fou arl nd y o a ne t 7 (Eastern poison ivy) found -fi 1% fth st common vine was 7% stran 21 ) identified (126 species 4 s we pe re d wetlands cie id s. en Th tifi ed at ho on ste ly or 44%) were found at d strand and floodplain wetlands had 91 species tw 161 species, and the o sa mp le we tla orested

PAGE 52

T al s pl a lae n 0 C ur able 3-4. Water quality (chemic sampled. Three forested floodpode Date STORET Temp an in werat d p hys ical parameter tlands (FF1, FF5, ae pH Specific Conductivit ) for 13 storet stations which d FF10) have multiple data fry Dissolved Oxygen Turbidity To correspond with the forested flood om different sampling dates (1993-20tal Kjeldahl Nitrogen Ammonia Nitrogen Total Phosphorus ins 0). oF1 8/15/2000 20020404 2 C 3.7 umhos/cm 6.60 121 mg /L NTU 6.2 mg/L mg/L mg/L 0.75 0.004 0.08 F F F F F F F F 1 F .7 F .5 F .6 F .7 F .7 F .6 F .8 F .6 F .4 F .3 F1 8/15/2000 20020404 2 F2 7/19/1999 20010454 2 F3 8/22/2000 27010580 2 F4 2/23/1999 20010455 1 F5 8/17/1993 19010099 2 F5 3/5/1998 19010099 F5 2/1/2000 19010099 7 F5 7/9/1996 19010099 24 F6 3/8/1999 26011019 18 F7 7/20/1999 26011020 25 F8 7/15/1996 19020027 25 F9 4/17/1997 20030246 16 F10 10/9/2000 28020234 24 F10 9/11/1996 28020234 25 F11 1/17/1995 21010032 11 F12 3/22/1999 19010072 16 F14 7/14/1998 32010024 23 3.7 3.2 5.0 2.8 5.0 6.61 121 7.10 1580 7.40 746 6.90 230 6.36 70 5.10 75 5.25 49 6.90 219 6.72 172 5.90 101 5.60 68 7.20 399 7.40 451 3.97 79 6.10 77 7.70 189 6.2 1.2 6.8 1.8 2.7 7.2 3.2 10.2 1.9 5.3 7.0 6.6 1.5 4.2 5.7 7.7 3.5 2.6 2.0 3.1 8.3 6.8 7.1 3.0 0.75 0.038 0.08 0.27 0.012 0.04 0.62 0.100 0.13 0.34 0.000 0.04 0.47 0.02 0.25 0.039 0.02 0.66 0.024 0.02 0.26 0.000 0.00 1.50 0.080 0.28 0.46 0.000 0.00 0.88 0.120 0.18 0.75 0.059 0.11 0.46 0.000 0.05 0.12 0.017 0.02 no data available

PAGE 53

42 Summary Statistics Species richness (R), species evenness (E), Shannon diversity (H'), Simpson iversity (D), and Whittakers beta diversity (W) were calculated based on the mblage for each samd. ness rarom 21 sence floodplain fotland),cies at FF5 (a wetland eized landscape). evenneonsisng all a p, and speciesly showed differences at the hundred t, explainingd vaendecimal places. Shannon diversity ranged fromy; these w with the) a(7ichness, r thpsod5), and FF5 and FF8 (another urban wetland) shared the highest Simpson diversity index value (ers beta diversity ranged fromow of 1.8 at (a reference wetland), to a high of 7.4 at FF4 (an agricultural wetland). A completee of summtatistics fable 3-5 summarizes comparisons of mean richness, evenness, and diversity alculations by a priori land use category. Urban wetlands had the mean greatest species ifferent among a priori land use categorta diversity (4.56 1.44), and overall beta (4.9) a (233) diversity. grouped according to wetland regions ), as well as for multiple pair wise comparisons. Sample size limitations prevented all possible wetland region pair par a umf two wetlands per wetland region and land use caere nblsor the MRPP statistic (T),ecroret (A), and significance value (p) (). R crof the panhandle and north wetland regions for et (ing a priori and urbanigthe level, suggesignresp csition across all wetland regwia priori land use categories, the MRPP analysis did not suggest a difference among species composition in north and central reference wetlands or agricultural wetlands for all d acrophyte assem ple wetlan Species rich nged f pecies at FF2 (a refermbedded in an urban rested we Species to 77 spess was c tent amo riori land use groups evenness on housandth decimal place why reporte lues were id tical for two 3.05 to 4.34 at FF2 and FF5, respectiv lowest (21 el7) species r ere the same wetlands nd highest espectively. Similarly, FF2 had e lowest Sim n diversity in ex value (0.9 0.99). Whittak a l FF6 tabl ary s or each wetland is presented in Appendix C. T c richness (50 17) followed by agricultural wetlands (41 10). This trend was evident for both Shannon diversity, with urban wetlands having greater mean values (3.86 0.35). Whittakers beta diversity was greatest in agricultural wetlands (4.94 1.07). Beta and gamma diversity were calculated for overall a priori land use categories, with urban wetlands having the highest beta and gamma diversity (4.0 and 201, respectively). None of the summary statistics were significantly d ies (Fishers LSD pair wise comparison, = 0.05) as the standard deviations overlap for most of the summary statistics. Similarly, summary statistics (including species richness, species evenness, Shannon diversity, and Whittakers beta diversity) were not significantly different between low (1995 LDI_F/wo_200m < 2.0) and high (1995 LDI_F/wo_200m 2.0) LDI groups (Mann-Whitney U-Test; Table 3-6 ). However, wetlands in the high LDI group had higher (though not statistically significant) mean species richness (47 17), Shannon diversity (3.80 0.34), Whittakers be and gamm Regional Compositional Analysis MRPP was calculated across all wetlands (Lane 2000 wise com isons when minim o tegory w ot availa e. Re ults f tests included the test chance-corrOnly the M ted within-gPP pair wise up agompa emenison Table 3-7 the compiled pool of all w lands includ reference, agricultural, ) was not s nificant at = 0.05 ting that there were regionally sions. Of the possible pair wise com ificant diffe nces among paris eciesthin ompo ons

PAGE 54

43 Table 3-5. Richness, evenness, and diversity of the macrophyte assemblage among a priori land use categories. Referenceculturrban Agri al U Spec36 11a41 10a50 17a ies richness (R) Species evenness (E) 1.00 0.00a1.00 0.00a1.00 0.00aShanon div.32a3 0 0Simpons d0.97 0.01a0 0 0Whittaker's 3.53 1.03a4.9 1 1Betaiversi3.7Gamma dive134 n ersity (H') 3.55 0 .69 .2a 5 3.86 .35 a s iversity (D) .97 .01 a 0.98 .01 a Beta diversity (W) 4 .07 a 4.57 .38 a d ty 3.7 4.0 rsity 153 201 Categnot significantly different (Fisher's LSD, 0.05) wetlanmong urban wetlands suggested a difference in thethe wetland regions (multiple comarison panhandle versus north versus central versus south; pair wise north versus central; and pair wse norations to interpre pue to small sampleh wetland region, the direnc s sition among urban wetland showed the greatest variability. using wetlands grouped according to bioregions (Griffe forested wetland sampled, only one wapanhandle bioregion and one in the Everglades bioregion. All of the reference wetlands (n=7) were located in the peninsula bioregion. Thus the only comarisons possible were between a purban wetlands thehaninsula ecoregiofference in species coms not found among agricultural or urban wetlands in the northeast and peni Table 3-6. Richness, evenness, and diversity of the macrophyte assemblage between low I 2.0) LDI groups. ories with similar letters were = d regio ns. However, all MRPP tests aion of wetlands among species composit p i th versus south)c While limit tation are ap arent d sizes within ea ffe es in specie compo The MRPP analysis was repeated ith et al. 1994). Of th s located in the p riori agricultural and in nort east d pen ns, limiting the MRPP analysis to two comparisons. A significant diposition wa nsula bioregions (Table 37). (LDI < 2.0) and high (LD Low LDIHigh LDI W ^ p` Species richness (R) 39 1147 17 120 0.32 Species evenness (E) 1.00 0.001.00 0.00 153 0.38Shannon diversity (H') 3.62 0.313.80 0.34 120 0.32Simpsons diversity (D) 0.97 0.010.98 0.01 120 0.32Whittaker's Beta diversity (W) 4.16 1.444.56 1.44 123 0.42B eta diversity 4.34.9 Gamma diversity 169233 W = Mann-Whitney U-test statistic p` = significance value ^

PAGE 55

44 Table 3-7. Macrophyte community composition similarity among wetland regions (Lane 2000) and bioregions (Griffith et al. 1994) with MRPP. Sites (n)T^A` p# Florida Ecoregions (Lane 2000) All wetlands All regions (P vs N vs C vs S) 24-4.2 0.07 0.00* Panhandle vs north 120.2 -0.56* Panhandle vs central 10-1.9 0.04 0.04* Panhandle vs south 6-2.4 0.12 0.02* North vs central 18-3.2 0.04 0.01* North vs south 14-3.7 0.07 0.01* Central vs south 12-4.0 0.07 0.00*Reference wetlands North vs central 6-1.2 0.04 0.10*Agricultural wetlands All regions (P vs N vs C vs S) 70.4 -0.65*Urban wetlands All regions (P vs N vs C vs S) 10-4.4 0.18 0.00* North vs central 7-2.8 0.13 0.01* North vs south 8-3.9 0.15 0.01*Bioregions (Griffith et al. 1994) Agricultural wetlands Northeast vs peninsula 6-1.5 0.12 0.07*Urban Wetlands Northeast vs peninsula 9-1.6 0.05 0.07 A high |T| value and significant p-value (p<0.05) suggests a difference in species composition. ^ T = the MRPP test statistic ` A = the chance corrected within-group agreement # p = the signi ficance value. Community Composition Macrophyte community composition was summarized in a 2-dimensional nonmetric multidimensional scaling (NMDS) ordination to explore gradients in macrophyte community composition (Figure 3-5). The final solution had an overall stress of 17.8 with a final stability of 0.0002, which is considered a fair stress value useful for ordinations with community data sets (Kruskal 1964; Clarke 1993; McCune and Grace 2002). Axis 1 explained 53.3% variance, and axis 2 explained 29.2% additional variance. No measured variables were correlated with the ordination axes, including 1995 LDI_F/wo_200m, species richness, Shannon diversity, Simpson diversity, Whittakers beta diversity, and Julian date. Wetlands appeared to be broadly grouped on axis 1 according to wetland region, and perhaps the gradients of latitude and longitude would have been appropriate

PAGE 56

45 Figure 3-5. NMDS ordination bi-plot of 24 sample wetlands in macrophyte species space; wetlands are labeled according to site code; symbols correlate to wetland regions ( Lane 2000 ). Axis 1 explained 53.3% variance; axis 2 explained an additional 29.2% variance. xplanatory variables for the ordination axes. For purposes of speculation, the second fell above the horizontal line of axis 2 at the center of the plot; whereas the remaining ight strands fell on or below the horizontal line. In contrast, only three of the forested floodplain wetlands (FF2, FF11, and FF13) fell below the horizontal line; while, the remaining 11 floodplain wetlands fell on or above the horizontal line. Perhaps there were distinct dnity(strand vMDS, which could be supported by the observation that the forested strand and floodplain wetlands inthis stred just 91 species. of the species ied aany o4 forested floodplain wetlands also occurring at any of the 10 forested strand wetlands, which had fewer speg that 57% of the species identified within the forested strands also occurred in at least one floodplain wetland. However, e axis may be correlated to wetland type, as only two of the forested strands (FS5 and FS7) e ifferences in the macrophyte commu composition between wetland types ersus floodplain) detected with the N udy sha This correlates to 43% identif t f the 1 cies identified (161 species) meanin FF1 FF10 FF11 FF12 FF13 FF14 FF2 FF3 FF4 FF5 FF6 FF7 FS5 FF8 FF9 FS1 FS10 FS2 FS3 FS4 FS6 FS7 FS8 FS9 Axis 1: 53.3% A xis 2: 29.2 % Ecoregions (Lane 2000)SouthCentralNorthPanhandle

PAGE 57

46 c onsidering that 126 species (45% of the entire 283 species identified for this study) were speis not as significant s some other unmeasured environmental variable. A meaningful grouping of wetlands based ori land use classificationparen Metricifilatedeation coeffic p<0.05) were senclusrelimCI for foresteodplain wetlan8). Medn were the proportion tolerant indicator spe propivepecies (SEN); Floristic Quality Assessment ); pric species (EX); and proporive perennial species roantipecies and procreasing development intensity; whereas, the pro FQAI, and proportion native perennial species decreascape development intensity. All metrics significantly (p < .05) differentiated between low (LDI<2.0) and high (LDI2.0) 1995 LDI_F/wo_200m roups (Table 3-9). ad the highest proport only sampled at one of the 24 wetlands included in this study, perhaps the dissimilarity of cies composition among wetland type (strand versus floodplain) a n a prio was not ap t. Selection Five metrics that were sign cantly corre with LDI (Sp rmans correla ient |r|0.50, lected for i ion in the p ina ry FW d strand and flo ds (Table 3-cies (TOL); etrics selectortion sensit for in indi clusiocator s Index (FQAI oportion exot tion nat (NP). The p portion of toler ind cator s portion exotic species increased with in portion sensitive indicator species, ed with increased lands 0 g Tolerance metrics Tolerant and sensitive indicator species were determined statewide using Indicator Species Analysis ( PCORD ). Table 3-10 provides a list of 19 tolerant indicator species. Figure 3-6 shows a scatter plot of proportion tolerant indicator species versus 1995 LDI_F/wo_200m. The proportion tolerant indicator species increased with increasing development intensity. FF10 (an urban floodplain forest in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ) h ion tolerant indicator species (0.29), followed by FS5 (0.28) and FF7 (0.25), which were both in the central wetland region ( Lane 2000 ) or peninsula bioregion ( Griffith et al. 1994 ). Two wetlands in the peninsula bioregion ( Griffith et al. 1994 ) had less than 0.05 proportion tolerant indicator species, including FS10 a reference strand in the south wetland region ( Lane 2000 ) and FF2 a reference floodplain in the north wetland region ( Lane 2000 ). Table 3-8. Spearmans correlation coefficients for macrophyte metrics and FWCI with 1995 LDI_F/wo_200m. Metric Spearmans r p-value Proportion tolerant indicator species 0.65 0.001 Proportion sensitive indicator species -0.83 <0.001 FQAI score -0.50 0.012 Proportion exotic species 0.54 0.006 Proportion native perennial species -0.56 0.004 FWCI -0.75 <0.001

PAGE 58

47 Table 3-9. Comparisons among five macrophyte metrics and the FWCI between low (LDI < 2.0) and high (LDI 2.0) LDI groups (LDI_F/wo_200m). Metric Low LDIHigh LDI W^p` Tolerant indicator species 0.11 0.060.19 0.06 115.5 0.007 Sensitive indicator species 0.15 0.080.05 0.05 14 0.001FQAI score 4.72 0.684.10 0.64 34 0.036Exotic species 0.03 0.040.10 0.08 109 0.022Native perennial species 0.95 0.06 0.86 0.11 34 0.036FWCI 38.33 7.9121.54 12.00 16 0.001 Values represent the mean standard deviation W^ = the Mann-Whitney U-Test statistic p` = the significance value Table 3-11 provides a list of the 16 sensitive indicator species. Figure 3-7 shows that the proportion sensitive indicator species decreased with increasing development intensity. The wetland with the highest proportion sensitive indicator was FF2 (0.38; a reference wetland), followed by and FF9 (0.18; an urban wetland) and FS7 (0.17; a reference wetland). Three wetlands had no sensitive indicator species including two in the central wetland region (Lane 2000) or peninsula bioregion (Griffith et al. 1994) (the urban strand FS5 and the agricultural strand FS3) and one wetland in the south wetland region (Lane 2000) or Everglades bioregion (Griffith et al. 1994) (FF10). Shrub and tree species were included in the ISA for both tolerant and sensitive metrics. Metrics developed based on the macrophyte community composition included woody species rooted within the sampling quadrats, as structure was thought to play an important role in the biological condition of flowing water forested wetlands. Excluding the tree and shrub layers would seemingly underscore the importance of these woody species. In fact, tree and shrub species comprised 52.6% of the tolerant and 62.5% of the sensitive indicator species lists (Tables 3-10 and 3-11). The five tolerant indicator tree species were Acer rubrum (red maple), Ilex cassine (dahoon holly), Liquidambar styraciflua (sweetgum), Nyssa sylvatica var. biflora (swamp tupelo), and Prunus caroliniana (Carolina laurelcherry) (Table 3-10); the five tolerant indicator shrub species were Callicarpa americana (American beautyberry), Cephalanthus occidentalis (buttonbush), Ludwigia peruviana (water-primrose), Sambucus canadensis (elderberry), and Viburnum obovatum (Walter viburnum). Vines also comprised a high percentage of the tolerant indicator species list, including Berchemia scandens (rattan vine), Smilax auriculata (earleaf greenbrier), Smilax tamnoides (bristly greenbrier), and Toxicodendron radicans (Eastern poison ivy). Of the sensitive indicator species, 43.8% were shrubs, 18.8% trees, 18.8% herbs, 12.5% graminoids, 6.3% ferns, and 0% vines (Table 3-11). The three sensitive indicator tree species included Fraxinus caroliniana (Carolina ash), Persea palustris (swamp bay) and Pinus elliottii (slash pine). Seven shrub species were categorized as sensitive indicator species, including Agarista populifolia (Florida hobble-bush), Hypericum hypericoides (St. Andrews cross), Ilex coriacea (bay-gall holly), Lyonia lucida (fetter

PAGE 59

Table 3-10. Tolerant indicator species for forested strand and floodplain wetlands. Botanical Name Common Name LDI Break (Indicator Value, p-value) Growth Form Acer rubrum Red Maple 1.0 (81.8, 0.057) Tree 1.5 (90,0.005) Berchemia scandens Rattan Vine 3.5 (58.3, 0.062) Vine Bidens alba Hairy Beggar-Ticks 3.5 (54.9, 0.097) Herb Callicarpa americana American Beautyberry 3.5 (75, 0.053) Shrub Cephalanthus occidentalis Buttonbush 1.0 (77.3,0.071) Shrub 1.5 (61,0.072) Commelina diffusa Dayflower 2.0 (50.6,0.034) Herb 3.0 (67.3,0.015) Hydrocotyle verticillata Pennywort 3.0 (43.7,0.068) Herb Ilex cassine Dahoon Holly 1.5 (55, 0.095) Tree Liquidambar styraciflua Sweetgum 1.5 (60,0.092) Tree Ludwigia peruviana Water-Primrose 3.5 (54.9, 0.087) Shrub Nyssa sylvatica var. biflora Swamp Tupelo 1.5 (60, 0.092) Tree Panicum rigidulum Red-Top Panicum 3.0 (47.4, 0.031) Graminoid Prunus caroliniana Carolina Laurelcherry 2.5 (27.3, 0.092) Tree Rhynchospora miliacea Millet Beakrush 3.0 (43.7, 0.072) Graminoid Sambucus canadensis Elderberry 2.5 (27.3,0.09) Shrub Smilax auriculata Earleaf Greenbrier 3.5 (72.4, 0.086) Vine Smilax tamnoides Bristly Greenbrier 3.0 (37.7, 0.051) Vine Toxicodendron radicans Eastern Poison Ivy 1.0 (77.3, 0.07) Vine 1.5 (61,0.068) 2.0 (57.3,0.069) Viburnum obovatum Walter viburnum 3.5 (58.3, 0.052) Shrub

PAGE 60

49 0.00.10.20.302468101995 LDI_F/wo_200mProportion Tolerant Indicator Species Floodplains Strands Figure 3-6. The proportion of tolerant indicator species at wetlands increased with increasing development intensity (LDI). bush), Rhododendron viscosum (swamp azalea), Vaccinium arboreum (sparkleberry), and Vaccinium corymbosum (highbush blueberry). The two sensitive indicator graminoid species were Cladium jamaicense (saw-grass) and Panicum hemitomon (maidencane). Floristic Quality Assessment Index metric Wetland FQAI scores decreased with increasing 1995 LDI_F/wo ( Figure 3-8 ). Of the 283 species identified in the flowing water wetlands, 17 were assigned coefficient of conservatism (CC) scores of zero, and 11 additional species received scores less than 1.0. Of the species receiving a zero CC score, nine (53%) were listed as Category I invasive exotics, and two (12%) were listed as Category II invasive exotics ( EPPC 2003 ). [The rankings of Category I or II invasive exotics are from the Florida Exotic Pest Plant Council (EPPC), which focuses on identifying exotic pest species. Category I species include those invasive exotic species considered responsible for changes to native plant communities through the displacement of natives, changes in community structure or ecological functions, and hybridizing with natives, based on documented ecological damage ( EPPC 2003 ). Category II species have not yet altered native plant communities, but have increased in abundance or frequency and may become ranked as Category I with confirmed ecological damage ( EPPC 2003 ).] The species with the highest CC score was Pieris phyllyreifolia (climbing fetter-bush) (9.5), followed by Panicum abscissum (cut-throat grass) (9.22), Taxodium ascendens (pond cypress) (8.8), Asplenium heterochroum (bicolored spleenwort) (CC = 8.5), and Pinckneya bracteata (fever-tree) (8.3). A complete list of CC scores is available in Appendix B Wetland FQAI scores were significantly correlated with LDI

PAGE 61

Table 3-11. Sensitive indicator species for forested strand and floodplain wetlands. Botanical Name Common Name LDI Break (Indicator Value, p-value) Growth Form Agarista populifolia Florida Hobble-Bush 2.5 (30.8, 0.092) Shrub Centella asiatica Coinwort 2.5 (38.6,0.082) Herb Cladium jamaicense Saw-Grass 1.5 (45.5,0.059) Graminoid Eupatorium capillifolium Dog Fennel 1.0 (88, 0.034) Herb Fraxinus caroliniana Carolina Ash 2.5 (38.6, 0.064) Tree Hypericum hypericoides St. Andrew's Cross 2.5 (30.8, 0.095) Shrub Ilex coriacea Bay-Gall Holly 2.5 (38.5, 0.037) Shrub Lyonia lucida Fetter-Bush 2.5 (53.6,0.017) Shrub 3.0 (52.9,0.054) Panicum hemitomon Maidencane 2.0 (38.6,0.042) Graminoid Persea palustris Swamp Bay 2.5 (61.2, 0.006) Tree Phlebodium aureum Golden Polypody 1.0 (95.7, 0.008) Fern 1.5 (45.5,0.065) Pinus elliottii Slash Pine 1.0 (95.7, 0.008) Tree 1.5 (45.5,0.065) Rhododendron viscosum Swamp Azalea 2.5 (38.6, 0.081) Shrub Saururus cernuus Lizard's Tail 3.0 (64.7, 0.01) Herb Vaccinium arboreum Sparkleberry 2.5 (30.8,0.098) Shrub Vaccinium corymbosum Highbush Blueberry 3.0 (52.9, 0.038) Shrub

PAGE 62

51 0.00.10.20.30.402468101995 LDI_F/wo_200mProportion Sensitive Indicator Species Floodplains Strands Figure 3-7. The proportion sensitive indicator species at wetlands decreased with increasing development intensity (LDI). 012345602468101995 LDI_F/wo_200mFQAI Score Floodplains Strands Figure 3-8. FQAI scores decreased with increasing landscape development intensity.

PAGE 63

52 gradient (|r| = 0.50; p = 0.012; Table 3-8 ). When wetlands were divided into two groups based on 1995 LDI_F/wo_200m (LDI < 2.0 and LDI 2.0), there was a significant difference between FQAI scores (U = 34; p < 0.05) ( Table 3-9 ). The range of wetland FQAI scores was 3.0, though the scale of species CC scores ranged from 0.0-9.5. The wetland with the highest FQAI was FF2 (5.6), a reference floodplain forest in the north ( Lane 2000 ) or peninsula stream ( Griffith et al. 1994 ) ecoregions. The wetland receiving the lowest FQAI score was FF10 (2.7), an urban floodplain forest in southwest Florida in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ). Three wetlands in the low LDI group with low FQAI scores included FS7 (4.3), FS10 (3.9), and FF6 (3.4). Sixty-seven percent of the wetlands in the low LDI group had an FQAI score greater than 4.5; whereas 73% of wetlands in the high LDI group had an FQAI score less than 4.5. Exotic species metric The proportion of exotic species at a wetland was significantly correlated with the gradient of development intensity in the surrounding landscape (|r| = 0.54, p = 0.006; Table 3-8 ). Figure 3-9 shows that the proportion of exotic species increased with increasing LDI. The south wetland region ( Lane 2000 ; or Everglades bioregion from Griffith et al. 1994 ) hosted the wetland with the greatest proportion exotic species, FF10 (0.29), which also received the lowest FQAI score (2.7). The wetland with the second highest proportion exotic species was FS3 (0.18), an agricultural strand surrounded by cattle pasture. Six floodplain wetlands had no exotic species present, including two reference wetlands (FF2 and FF7), two agricultural wetlands (FF4 and FF11), and two urban wetlands (FF5 and FF9). The proportion of exotic species at wetlands was significantly different between low and high LDI groups (W = 109, p = 0.022; Table 3-9 ). Table 3-12 lists the 35 exotic species encountered throughout the forested strand and floodplain wetlands. The most common exotic species was the herbaceous species Commelina diffusa (dayflower) found at 10 (42%) of the 24 forested wetlands, which was also categorized as a tolerant indicator species ( Table 3-10 ). Sixty-nine percent of the exotic species (n=24) were found at only one forested wetland. Thirty-seven percent of the species (n=13) were listed as Category I invasive exotic species, and 14% (n=5) were listed as Category II invasive exotic species ( EPPC 2003 ). Forty-three percent of the exotic species were herbs, 20% vines, 14% shrubs, 11% trees, 9% graminoids, and 3% ferns. Native perennial species metric Of the 283 macrophyte species identified, 234 (83%) were classified as native perennials. Figure 3-10 shows that the proportion of native perennial species at a wetland decreased with increasing development intensity. The native perennial species metric was significantly correlated with LDI (Spearman |r| = 0.56, p = 0.004; Table 3-8 ); and there was a significant difference between the proportion native perennial species at low and high LDI group wetlands (W = 34, p = 0.036; Table 3-9 ). FF10 had the lowest proportion native perennial species (0.60). Four wetlands hosted entirely native perennial species, including one reference wetland (FF2), two agricultural wetlands (FF4 and FF11), and one urban wetland (FF9).

PAGE 64

53 0.00.10.20.302468101995 LDI_F/wo_200mProportion Exotic Species Floodplains Strands Figure 3-9. The proportion of exotic species at a wetland increased with increasing development intensity. Florida Wetland Condition Index Five metrics were included in the preliminary FWCI for forested strand and floodplain wetlands, including proportion tolerant indicator species, proportion sensitive indicator species, FQAI score, proportion exotic species, and proportion native perennial species. Metrics were natural log transformed to improve normality. Metric scoring was done on a continuous scale spreading between the 5 th to 95 th percentiles of the values for the sampled wetlands. Scores for each metric were then added together to create the preliminary forested strand and floodplain FWCI with a scale of 0-50, with 50 representing the reference condition of high biological integrity. Appendix D provides metric scoring criteria. Figure 3-11 shows that FWCI scores decreased with increasing development intensity. Correlations between macrophyte metrics and FWCI with LDI were significant (p < 0.05) for all of the metrics and the FWCI (|r| = 0.75, p < 0.001) ( Table 3-8 ). One wetland, FF2 a reference forested floodplain wetland, received a perfect 50 on the FWCI scale, also receiving a perfect 10 score for all five metrics. One wetland scored the lowest potential score of zero, FF10 which was an urban forested floodplain wetland in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ). All of the wetlands in the low LDI group scored above the midpoint of 25 on the FWCI scale; while 64% of wetlands in the high LDI group scored below 25.

PAGE 65

54 Table 3-12. Exotic species identified at 24 forested strand and floodplain wetlands. Botanical Name Common Name EPPC Growth Form Abrus precatorius Rosary Pea I Vine Alternanthera philoxeroides Alligator Weed II Herb Alternanthera sessilis Sessile Alligator Weed Herb Begonia cucullata Wax Begonia II Herb Cinnamomum camphora Camphor Tree I Tree Colocasia esculenta Elephant's Ear I Herb Commelina communis Asiatic Dayflower Herb Commelina diffusa Dayflower Herb Cuphea carthagenensis Columbia waxweed Herb Cynodon dactylon Bermudagrass Graminoid Cyperus difformis Variable Flatsedge Graminoid Emilia fosbergii Florida Tasselflower Herb Eugenia uniflora Surinam Cherry I Shrub Koelreuteria elegans Flamegold II Tree Ligustrum sinense Chinese Privet I Shrub Lonicera japonica Japanese Honeysuckle I Vine Ludwigia peruviana Water-Primrose Shrub Lygodium japonicum Japanese Climbing Fern I Vine Lygodium microphyllum Small-leaf Climbing Fern I Vine Melaleuca quinquenervia Punk Tree; Melaleuca I Tree Merremia dissecta Alamo Vine Vine Mimosa pigra Black Mimosa I Shrub Momordica charantia Balsampear Vine Oeceoclades maculata Monk Orchid Herb Paspalum notatum Bahiagrass Graminoid Phyllanthus urinaria Water Leafflower Herb Sapium sebiferum Chinese Tallowtree I Tree Schefflera actinophylla Australian Umbrella Tree I Tree Schinus terebinthifolius Brazilian Pepper I Shrub Thelypteris dentata Downy Maiden Fern Fern Trifolium repens White Clover Herb Urena lobata Caesarweed II Herb Wisteria sinensis Chinese Wisteria II Vine Xyris jupicai Richard's Yellow-Eyed-Grass Herb Youngia japonica Oriental False Hawksbeard Herb Exotic Pest Plant Council (EEPC) categories from EEPC ( 2003 ).

PAGE 66

55 0.50.60.70.80.91.002468101995 LDI_F/wo_200mProportion Native Perennial Species Floodplains Strands Figure 3-10. The proportion of native perennial species decreased with increasing development intensity (LDI). 0102030405002468101995 LDI_F/wo_200mFWCI Floodplains Strands Figure 3-11. Forested Wetland Condition Index (FWCI) scores decreased with increasing development intensity (LDI).

PAGE 67

56 Cluster Analysis Cluster analysis determined wetland groupings based on macrophyte community composition using a species by site presence/absence matrix. Based on the lowest average p-value for all species from the randomized Monte Carlo tests used in the ISA analysis, the most ecologically meaningful cluster for dendrogram pruning was at cluster step 5 ( Figure 3-12 ) with the lowest average species p-value of 0.27. The highest number of significant (p < 0.05) indicator species was found at cluster step 3, which had 37 significant indicator species ( Figure 3-13 ). Cluster steps 4 and 5 had the second highest number of significant indicator species at 31. Exploration of the sample sites assigned to the groups at cluster step 3 suggested an association between group membership and spatial distribution of the wetlands throughout Florida. [Note: ISA used to determine the most ecologically meaningful clusters for dendrogram pruning is separate from ISA used in determining tolerant and sensitive indicator species. For ISA used to determine meaningful clusters for dendrogram pruning, group membership was assigned based on the wetland groupings established through cluster analysis. For ISA used to determine sensitive and tolerant indicator species, group membership was assigned based on the calculated LDI score for each wetland.] The groups defined through the agglomerative cluster analysis at cluster step 5 were roughly defined by wetland regions ( Lane 2000 ), bioregions ( Griffith et al. 1994 ), and a priori land use categories, including: Cluster 1: FF2, a north wetland region, peninsula bioregion, reference floodplain; Cluster 2: Nine floodplains, including two in the panhandle, six in the north, and one in the central wetland regions; one in the panhandle, five in the northeast, and three in the peninsula bioregions; and four agricultural and five urban wetlands; Cluster 3: Five strand and two floodplain wetlands; one in the north and six in the central wetland regions; all seven in the peninsula bioregion; four reference, one agricultural, and two urban wetlands; Cluster 4: Three strand and two floodplains, including one in the north and four in the south wetland regions; four in the peninsula and one in the Everglades bioregions; three reference and two urban wetlands; Cluster 5: Two floodplain wetlands with one in the north and one in the central wetland regions; both in the peninsula bioregion; and both agricultural wetlands. Wetland groups based on the third cluster step combined the aforementioned groups from the cluster step 5, so that: Cluster 1: Combines Clusters 1 & 2 Cluster 2: Composed of Cluster 3 Cluster 3: Combines Clusters 4 & 5 It appear that the initial grouping for cluster step three was based on spatial location and that further cluster steps separate out wetlands based on impairment. Figure 3-14 shows that based on wetland groups from cluster step 3, FWCI scores for wetlands in Cluster 1 (37.5 10.5) were significantly different from wetlands in Cluster 2 (24.4 10.0) and Cluster 3 (19.8 13.1) (p < 0.05). Wetland FWCI scores for Cluster 2 and Cluster 3 were not significantly different from one another. Regionalization was apparent in the wetland groupings of Cluster 3 (at cluster step 3); as Cluster 3 was comprised of the all of

PAGE 68

57 0.00.10.20.30.40.50.60.7234567891011121314151617181920Number of ClustersAverage p-value Figure 3-12. Change in average species p-value from the randomized Monte Carlo tests at each step in clustering. The minimum average p-value (0.27) was found at cluster step 5. 010203040234567891011121314151617181920Number of ClustersNumber Significant Indicators Figure 3-13. Change in the number of significant indicator species from the indicator species analysis performed at each step in clustering. The maximum number of significant indicator species (37) was found at cluster step 3 (p < 0.05), followed by 31 significant indicator species at cluster steps 4 and 5.

PAGE 69

58 Cluster 3Cluster 2Cluster 1 50403020100 FWCI a b b Figure 3-14. FWCI scores for three wetland clusters based on macrophyte community composition. Boxes represent 25 th and 75 th quartiles, lines represent median FWCI scores per cluster, dots represent the mean, and vertical lines represent the range. Clusters with similar letters were not significantly different (p<0.05). the south wetland region wetlands ( Lane 2000 ), including the only wetland sampled in the Everglades bioregion ( Griffith et al. 1994 ). Table 3-13 provides means and standard deviations for cluster FWCI and LDI_F/wo_200m scores. LDI_F/wo_200m scores were not significantly different among clusters. Landscape Development Intensity Index and the Florida Wetland Condition Index Nearly all of the correlation (110 of 120, or 92%) between the five macrophyte metrics and the FWCI with the 20 variations of LDI were significantly correlated at the flexible p < 0.10 level ( Table 3-14 ). Additionally, nearly half of the comparisons (58 of 120, or 48.3%) were significantly correlated at the strictest significance level (p < 0.01). Three of the LDI calculations (1995 LDI_F/w_200m, 2000 LDI_F/w_200m, and 2000 LDI_F/wo_200m) were significant at the strictest significance level (p < 0.01) for all five metrics and the FWCI, which should not be surprising given that metrics were selected for inclusion in the FWCI based on correlations with the 1995 LDI_F/wo_200m (using Spearman correlation with LDI, visually distinguishable and ecologically meaningful pattern when graphed with LDI, and differentiation among LDI groups with the Mann Whitney U-test). The proportion tolerant indicator species metric was significantly correlated with 1995 and 2000 LDI_F (wetland feature scale) calculations (p < 0.01). However, the proportion tolerant indicator species metric was not significantly correlated at the strictest significance level (p < 0.01) for any of the 1995 or 2000 LDI_T (transect scale)

PAGE 70

59 Table 3-13. FWCI scores and LDI values for wetland clusters (at the third cluster step) based on macrophyte community composition. Wetland Clusters FWCI LDI_F/wo_200m 1 37.5 10.5 a 2.2 0.7 a 2 24.4 10.0 b 2.5 0.7 a 3 19.8 13.1 b 2.9 2.1 a Clusters with similar letters within columns were not significantly different (p < 0.05). LDI_F/wo_200m represents the LDI calculated at the feature scale excluding the wetland area within a 200 m buffer. calculations. The proportion sensitive indicator species metric was more strongly correlated with the LDI calculations at the transect and feature scales (11 of 12 correlations with |r| 0.63, p < 0.01, remaining one at |r| = 0.57, p < 0.05) than the watershed scales (1995 and 2000 LDI_WS_DW_exp at p < 0.01; 2000 LDI_WS_DW_lin at p < 0.05; 2000 LDI_WS_ED/wo not significant; remaining four LDI_WS correlations at p < 0.10). The FQAI score metric was not significantly correlated with three of the 1995 transect level LDI calculations, but was strongly (|r| > 0.66) and significantly (p < 0.01) correlated with all of the 1995 and 2000 watershed scale LDI calculations. Watershed scale LDI calculations were completed for floodplain forested wetlands only, leaving speculation as to whether these results suggest that FQAI scoring reflect landscape level anthropogenic activity (e.g. exotic species with low CC scores entering a system due to anthropogenic activities) or whether FQAI scoring was biased towards larger flowing water systems such as floodplain forests. The proportion exotic species and proportion native perennial species metrics were more strongly correlated with the feature scale LDI calculations than transect or watershed scales. In fact, three watershed 1995 LDI calculations were not significant with the proportion exotic species or the proportion native perennial species metrics, including 1995 LDI_WS_ED/w, 1995 LDI_WS_ED/wo, and 1995 LDI_WS_DW_lin. The strongest correlation between the FWCI and LDI was with the 2000 LDI_WS_DW_exp (|r| = 0.95; p < 0.01), though the correlation with 2000 LDI_F/wo_200m (|r| = 0.94; p < 0.01) was also remarkably strong. The two correlations with the highest Spearmans correlation coefficients were between the proportion tolerant indicator species metric and the 2000 LDI_WS_DW_exp and between the FQAI score metric and the 2000 LDI_WS_DW_lin (|r| = 0.96; p < 0.01).

PAGE 71

Table 3-14. Correlations of metrics and FWCI scores with 20 variations of the LDI index. Differences in LDI calculations include 1995 or 2000 land use; transect (T), feature (F), or watershed (WS) scale buffers; including (/w) or excluding (/wo) the wetland area; 100 (100m) or 200 meter (200m) buffers; equal distance (ED) or distance weighted (DW); and linear (lin) or exponential (exp) weighting. Column headings refer to the five metrics and the FWCI: proportion tolerant indicator species (TOL), proportion sensitive indicator species (SEN), Floristic Quality Assessment Index (FQAI), proportion exotic species (EX), and proportion native perennial species (NP). LDI n= TOL SEN FQAI EX NP FWCI 1995 LDI_T/w_100m 24 0.44 ** -0.68 -0.39 *** 0.51 ** -0.48 ** -0.58 1995 LDI_T/wo_100m 24 0.41 ** -0.65 ns 0.46 ** -0.43 ** -0.52 1995 LDI_T/w_200m 24 0.38 *** -0.63 ns 0.49 ** -0.44 ** -0.49 ** 1995 LDI_T/wo_200m 24 0.45 ** -0.69 ns 0.51 ** -0.44 ** -0.52 ** 1995 LDI_F/w_200m 24 0.72 -0.90 -0.57 0.61 -0.63 -0.79 1995 LDI_F/wo_200m 24 0.66 -0.83 -0.50 ** 0.54 -0.55 -0.75 1995 LDI_WS_ED/w 14 0.60 ** -0.49 *** -0.71 ns -0.50 *** -0.57 ** 1995 LDI_WS_ED/wo 14 0.59 ** -0.49 *** -0.67 ns ns -0.53 ** 1995 LDI_WS_DW_lin 14 0.58 ** -0.49 *** -0.69 ns ns -0.53 *** 1995 LDI_WS_DW_exp 14 0.80 -0.72 -0.75 0.60 ** -0.65 ** -0.70 2000 LDI_T/w_100m 24 0.51 ** -0.72 -0.49 ** 0.48 ** -0.54 -0.63 2000 LDI_T/wo_100m 24 0.51 ** -0.72 -0.42 ** 0.44 ** -0.49 ** -0.59 2000 LDI_T/w_200m 13 0.54 *** -0.69 -0.74 0.83 -0.85 -0.81 2000 LDI_T/wo_200m 13 ns -0.57 ** -0.66 ** 0.73 -0.77 -0.72 2000 LDI_F/w_200m 13 0.78 -0.93 -0.88 0.79 -0.85 -0.91 2000 LDI_F/wo_200m 13 0.82 -0.92 -0.91 0.82 -0.84 -0.94 2000 LDI_WS_ED/w 8 0.73 ** -0.68 *** -0.95 0.77 ** -0.83 ** -0.88 2000 LDI_WS_ED/wo 8 0.65 *** ns -0.86 0.73 ** -0.69 *** -0.78 ** 2000 LDI_WS_DW_lin 8 0.83 ** -0.80 ** -0.96 0.77 ** -0.81 ** -0.92 2000 LDI_WS_DW_exp 8 0.96 -0.88 -0.88 0.83 ** -0.78 ** -0.95 *p<0.01 **p<0.05 ***p<0.10 ns not significant

PAGE 72

61 Human Disturbance Gradient and Stream Condition Index Thirteen of the forested floodplain wetlands had Human Disturbance Gradient (HDG) and Stream Condition Index (SCI) scores for sampling events taken within the channelized flow of the forested floodplain wetlands ( Table 3-15 ). Sample dates for the FWCI and LDI were during 2003 ( Table 2-1 ); sample dates for the HDG and SCI ranged from 1995-2000 ( Table 3-15 ). Table 3-15 provides storet identification numbers, HDG and SCI sample dates, a priori land use categories, bioregions ( Griffith et al. 1994 ), wetland regions ( Lane 2000 ), FWCI, LDI_F/wo_200m scores, HDG scores, and SCI scores for the 13 forested floodplain wetlands. Data for the most recent and complete sampling event for each wetland, as presented in Table 3-15 was selected for correlation purposes. The HDG was significantly correlated with the SCI (|r| = 0.58, p<0.05), LDI_F/wo_200m (|r| = 0.66, p<0.05), and FWCI (|r| = 0.74, p<0.01) ( Table 3-16 ). However, the SCI was not significantly correlated with either the LDI_F/wo_200m or the FWCI. The correlation between the LDI_F/wo_200m and FWCI (|r| = 0.68, p<0.05) ( Table 3-16 ) for the 13 forested floodplain wetlands (only those with HDG and SCI scores), was weaker than the correlation between them for the entire data set of 24 forested flowing water wetlands (|r| = 0.75, p<0.001) ( Table 3-8 ).

PAGE 73

Table 3-15. Forested Wetland Condition Index (FWCI), Landscape Development Intensity Index (LDI_F/wo_200m), Human Disturbance Gradient (HDG), and Stream Condition Index (SCI) data for 13 forested floodplain wetlands. Site STORET ID HDG/SCI Date Land Use Bioregion Wetland Region FWCI LDI_F/wo_200m HDG SCI FF1 20020404 8/15/2000 U Peninsula North 39.1 2.0 0 80 FF2 20010454 7/19/1999 R Peninsula North 50.0 1.2 0 70 FF3 27010580 8/22/2000 U Peninsula North 16.4 3.7 2 20 FF4 20010455 2/23/1999 A Peninsula Central 43.6 2.3 0 75 FF5 19010099 2/1/2000 U Northeast North 44.0 1.9 0 55 FF6 26011019 3/8/1999 R Peninsula Central 26.2 1.4 0 80 FF7 26011020 7/20/1999 R Peninsula Central 28.2 2.5 0 90 FF8 19020027 7/15/1996 U Northeast North 23.6 2.5 1 70 FF9 20030246 4/17/1997 U Northeast North 46.6 1.8 0 95 FF10 28020234 10/9/2000 U Everglades South 0.0 7.1 4 50 FF11 21010032 1/17/1995 A Northeast Panhandle 42.0 1.6 0 25 FF12 19010072 3/22/1999 A Northeast North 33.5 2.8 0 85 FF14 32010024 7/14/1998 A Panhandle Panhandle 36.5 1.8 0 80 STORET ID refers to Florida Department of Environmental Protection (FDEP) database ID. HDG/SCI Date refers to original sample date corresponding to HDG and SCI data. FWCI and LDI were calculated for the 2003 site visit. Land Use refers to a priori land use category (R-reference, A-agricultural, U-urban) Bioregion from Griffith et al ( 1994 ) Wetland Region from Lane ( 2000 ) LDI_F/wo_200m refers to the LDI calculated at the feature scale, excluding the wetland area, within a 200 m buffer.

PAGE 74

63 Table 3-16. Correlations among four measures of ecosystem condition or anthropogenic activity, including the Human Disturbance Gradient (HDG), Stream Condition Index (SCI), Landscape Development Intensity Index (1995 LDI_F/wo_200m), and the Florida Wetland Condition Index (FWCI) for freshwater forested floodplain wetlands. Values are Spearmans rank correlation coefficients. HDG SCI LDI_F SCI -0.58 ** LDI_F 0.66 ** ns FWCI -0.74 ns -0.68 ** *p<0.01, **p<0.05, ns=not significant

PAGE 75

CHAPTER 4 DISCUSSION The primary objective of this research was to develop a preliminary Florida Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland study sites were sought in various a priori designated land use categories that included undeveloped, agricultural, and urban land uses. The preliminary FWCI provides a quantitative measure of the biological integrity of forested strand and floodplain wetlands in Florida. Comprised of five metrics, the preliminary FWCI was developed based on the community composition of the macrophyte species assemblages ( Table 4-1 ). Metrics were selected for inclusion in the FWCI based on the correlation (nonparametric Spearman correlation coefficient) of each metric with the Landscape Development Intensity (LDI), an independent measure of anthropogenic activity in the landscape calculated for each wetland ( Brown and Vivas 2005 ); based on a metrics visually distinguishable correlation with LDI in a scatter plot; and based on a statistical difference of metric values between low and high LDI groups (Mann-Whitney U-test). The FWCI was composed of individual metrics, which were scaled and added together, creating the preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest score of 50 reflecting the highest biological integrity and the lowest score of zero reflecting a lack of biological integrity or no similarity to the reference wetland condition. The contribution of this research to our understanding of changes in the macrophyte community composition of forested strand and floodplain wetlands in relation to differing anthropogenic activities in the surrounding landscape can be summarized in five main points: 1. Five macrophyte based metrics including proportion tolerant indicator species, proportion sensitive indicator species, Floristic Quality Assessment Index (FQAI) score, proportion exotic species, and proportion native perennial species, were useful biological indicators for defining biological integrity for forested strand and floodplain wetland vegetation; 2. Vegetation richness, evenness, and diversity were not sensitive to a priori land use categories or development intensities in the surrounding landscape for forested strand and floodplain wetlands; 3. The Landscape Development Intensity (LDI) index was a useful tool correlating with the measured biological condition of vegetation for forested strand and floodplain wetlands; 4. Regional species lists for metrics would enhance the forested strand and floodplain Florida Wetland Condition Index (FWCI); 5. An FWCI with a set of core metrics could be developed for Florida freshwater wetlands, which includes separate species lists for indicator species by wetland type and ecoregions and separate Floristic Quality Assessment Index (FQAI) scores for species by wetland type. 64

PAGE 76

65 Table 4-1. The five metrics of the preliminary Florida Wetland Condition Index for freshwater forested strand and floodplain wetlands based on the macrophyte species assemblage. FWCI Metrics 1. Proportion Tolerant Indicator Species 2. Proportion Sensitive Indicator Species 3. Floristic Quality Assessment Index Score 4. Proportion Exotic Species 5. Proportion Native Perennial Species Describing Biological Integrity Biological indicators were useful in determining the biological integrity of freshwater forested strand and floodplain wetlands. For the purposes of this study, biological integrity has been defined quantitatively with the FWCI. The FWCI incorporated five metrics from the macrophyte species assemblages ( Table 4-1 ). Strong correlations between the FWCI and the intensity of development in the surrounding landscape (based on the use of nonrenewable energy and calculated with the LDI) suggest that changes in macrophyte community composition quantified as metrics were captured by the LDI. It has been proposed that organisms respond to environmental gradients by colonizing a range of feasible conditions beyond which the organisms fail to persist ( ter Braak 1987 ). By selecting species that occur throughout the range of measurable environmental parameters, the FWCI defined and detected deviations from the condition of reference wetlands based on macrophyte community composition. Each of the FWCI metrics addressed some disparity from the assumed range of feasible conditions. The proportion sensitive indicator species metric showed the strongest correlation with LDI, suggesting that the presence of a suite of taxa characteristic of wetlands with high biological integrity may be the most effective means of identifying changes in macrophyte community composition in freshwater forested strand and floodplain wetlands associated with changes in anthropogenic activities. Richness, Evenness, and Diversity Measures of richness, evenness, and diversity of the macrophyte assemblage were not sensitive to differences in land use or development intensity in the surrounding landscape within the forested strand and floodplain wetlands. Perhaps due to a limited sample size and high variability inherent in the landscape (e.g. regional differences, land use differences, etc.) no statistical test on summary statistics produced statistically significant results. Nevertheless, wetlands in the high LDI group (LDI 2.0) had higher (though not statistically significant) mean species richness and diversity (Shannon diversity, Whittakers beta diversity, and overall beta and gamma diversity). Species evenness and Simpsons diversity were remarkably similar among the sample wetlands; in effect no differences were detectable.

PAGE 77

66 General ecological theory predicts a decrease in plant diversity resulting from an increase in anthropogenic assaults such as grazing ( Blanch and Brock 1994 ; Grace and Jutila 1999 ) and nutrient enrichment ( Bedford et al. 1999 ), though the forested strand and floodplain wetlands displayed the opposite trend. Mitsch and Gosselink ( 1993 ) report that freshwater forested wetlands have low species diversity as compared to other ecosystems, so perchance macrophyte species that entered wetlands in developed landscapes were merely taking advantage of available habitat and in fact increased the overall species diversity. However, Ewel ( 1990 ) notes that due to high topographic and soil variability, river swamps may be the most diverse type swamp in Florida. Clearly there is some uncertainty within the published literature as to anticipated and abnormal diversity for Florida forested strand and floodplain wetlands. Many of the species entering wetlands in developed landscapes were categorized as exotic species, and the increased incidence of exotic species has long been associated with disturbed ecosystems ( Galatowitsch 1999b ; Cronk and Fennessy 2001 ). An increase in the frequency of exotic species has been attributed to drainage and hydrologic alterations ( Hobbs and Heunneke 1992 ; David 1999 ; Galatowitsch et al. 1999b ), increased human development ( Cronk and Fennessy 2001 ), and ecosystem scale alterations such as clear-cut harvests ( Devine 1998 ). Within the study wetlands, the proportion of exotic species increased with increasing development intensity in the surrounding landscape. It appears that the influx of exotic species added to rather than diminished the species richness and diversity within the freshwater forested strand and floodplain wetlands. As such, the presence of exotic species alone may not be an ideal indicator of biological integrity. While many ecological theories have been established suggesting that the presence or occurrence of exotic species increases with anthropogenic disturbance ( Cronk and Fennessy 2001 ; Galatowitsch 1999b ), there was a inconsistent pattern of occurrence of exotic species in the forested strand and floodplain wetlands sampled. For example, six of the freshwater forested floodplain wetlands, including two reference, two agricultural, and two urban wetlands, hosted zero exotic species. That four floodplain wetlands in developed landscapes (including two agricultural and two urban wetlands) hosted no exotic species was somewhat contradictory to the theories of increased exotic species occurrence in disturbed ecosystems. However, for the complete dataset, the trend of an increase in the proportion of exotic species with increasing landscape development intensity held. Some concerns arise considering the discrepancies on the exotic status of some species. For example, not all of the species listed by the Exotic Pest Plant Council (EPPC) as invasive exotics altering native plant communities (Category I) or invasive exotics increasing in frequency or abundance with the potential to alter native plant communities (Category II) ( EPPC 2003 ) received an FQAI score of 0.0 corresponding to species that act as opportunistic invaders, including species that commonly occur in disturbed ecosystems. The disagreement on the status of a species as an invasive exotic translates into disagreement as to the meaning of a particular exotic species occurring within an ecosystem. As a case in point, two exotic species were included as tolerant indicator species including Commelina diffusa (dayflower) and Ludwigia peruviana (water-primrose). Neither of the two tolerant indicator species that are also exotic species was listed as an EPPC Category I or II species. In fact, none of the 13 Category I

PAGE 78

67 invasive exotic species identified in the forested strand and floodplain wetlands was categorized as a tolerant indicator species. The intent of this research was to assess the current biological integrity of Florida freshwater forested flowing water wetlands in order to develop a quantitative wetland condition index. As such the comparison was made against the current day reference standard of biological integrity apparent in the reference strand and floodplain wetlands sampled. While we may innately believe that reference wetlands host no exotic species, we found that five of nine wetlands surrounded by low development intensity (56%) hosted at least one exotic species. Increasingly, it may become apparent that even reference wetlands with the highest current standard of biological integrity may host some proportion of exotic species. This may be even more apparent in the southern half of the state, where drainage and development have altered nearly all of the Florida landscape. In fact, the south wetland region reference wetland (FS10) had the highest proportion exotic species of all wetlands in the low LDI group by nearly 10%. Consequently, establishing a baseline for the reference condition of biological integrity is crucial for the application of indices of biotic integrity. Consensus should be reached as to whether scientists proceed by establishing a present day baseline for future assessments or by maintaining and updating a moving baseline for future application of the FWCI. Implications for both methods are complex. Measuring Anthropogenic Activity The variable sensitivities of three different independently derived indices compared to the forested strand and floodplain FWCI, including the Landscape Development Intensity index (LDI; Lane et al. 2003 ; Brown and Vivas 2005 ), the Human Disturbance Gradient (HDG; Fore 2004 ), and the Stream Condition Index (SCI; Fore 2004 ), suggest that multiple measures of biological integrity may be more effective at describing ecosystem wide biological integrity than any single measure based on an individual species assemblage or surrounding land use activity. Measurements of anthropogenic activity such as the LDI and HDG seek to describe ecosystem biological integrity from the perspective of outside anthropogenic influences which act to alter an ecosystem. While the LDI was used as a remote based measure of human development intensity, the HDG used both remote and local conditions integrating four components ranging from the in-stream chemical water quality, habitat structure, and hydrologic alteration, to a remote landscape assessment (using a form of LDI). For the HDG each of the four components was given equal weighting in determining the influence of anthropogenic activities on the biological integrity of an ecosystem. The LDI and HDG were strongly correlated, suggesting that the LDI alone may capture the influence from human development intensity on ecosystem biological integrity without the need for additional sampling and sample processing associated with in-stream chemical water quality, habitat structure, and hydrologic alteration. The FWCI was strongly correlated with both the HDG and the LDI, though the significance of correlation was slightly stronger with the HDG than the LDI. The caveat here was that only three of the thirteen floodplain wetlands had an HDG greater than zero, and because these scores were tested using the non-parametric Spearmans rank correlation coefficient test the strong correlations may simply be a factor of zero HDG

PAGE 79

68 values. In fact, our findings of strong correlations of LDI and HDG must be interpreted with caution because so few of the wetlands received HDG values greater than zero (an HDG score of zero represents no detectable human induced disturbance). This was likely due to the fact that the HDG and SCI were calculated in the channelized water course, whereas the LDI and FWCI were calculated for the surrounding floodplain forest. Many of the streams with low SCI and high HDG scores (considered those with low biological integrity) were not sampled for the FWCI and LDI as no floodplain forest remained, with the banks of the channelized water course being either sodded and mowed or paved. Therefore we should cautiously interpret correlations and discrepancies between the SCI (no significant correlations with FWCI or LDI) and HDG with the FWCI and LDI because the range of impaired conditions used to develop the FWCI was much narrower than that used to develop the SCI. The lack of correlations of SCI with both FWCI and LDI, suggest that in-stream macroinvertebrate based measures of biological condition and surrounding forested wetland macrophyte based measures of biological condition did not respond in a consistent manner to changes in anthropogenic activity. Using both the in-stream macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte FWCI biological assessment methods may provide a more complete picture of the overall condition of a wetland and associated stream at a particular spatial location. While agreement in the ranking of the biological condition of study wetlands using the FWCI and SCI was anticipated, discrepancies among the ranking from the different assemblages may provide great insight into biological condition as different species assemblages respond to changes in anthropogenic activities and the associated changes in inflows (e.g. nutrient enrichment) over different time scales. Additionally, use of the forested strand and floodplain FWCI may lead to specific conclusions as to the biological condition of local or nearby anthropogenic activity, while use of the SCI may enhance understanding of larger watershed scale influences from anthropogenic activity (i.e. due to the convergence of surface water within the watershed associated with stream flow). By correlating the community composition based FWCI and the landscape based LDI, we were better able to understand anthropogenic influence on biological integrity on forested wetland ecosystems. We found that the LDI index was a useful tool in approximating biological integrity. Its primary power was in the reproducible, objective, and quantitative methods employed to obtain a score based on the use on non-renewable energy in the surrounding landscape. A second strength of the LDI index was apparent in the practical application of a remote GIS based method of describing ecosystem condition as a starting point to identify potential areas for further biological sampling. Regionalization of the Florida Wetland Condition Index Pronounced differences in the local climate across Florida, such as differences in the amount of annual rainfall, seasonal maximum temperatures, and number of freeze days ( Fernald and Purdum 1992 ; Lane 2000 ) and the broad latitudinal and longitudinal ranges of sample wetlands (26.3N -30.8N latitude, 80.1W-82.1W longitude), suggest differences would be found among macrophyte community composition in Florida wetlands. In fact, compositional differences of the macrophyte species assemblages were

PAGE 80

69 found among Lane s ( 2000 ) Florida wetland regions. However, the study sample size limited the development of regional indicator species and metric scoring criteria. Wetland clusters based on macrophyte community composition (species presence by site) roughly correlated with wetland spatial distribution throughout the state. The distribution of wetlands in the cluster groups suggests that wetlands located in the northern area of Florida may have higher biological integrity than wetlands in the central or southern peninsula given a statewide scoring approach of the preliminary forested strand and floodplain FWCI. Accordingly, most of the human development in Florida has occurred along the east and west coastal areas of peninsular Florida ( Fernald and Purdum 1992 ), suggesting that while the reference wetlands selected in the south and central wetland regions were selected as the best possible examples of reference type conditions, they may be more affected by development in the surrounding landscape (such as compounded secondary effects) than their panhandle and north wetland region reference counterparts. While the ease and utility of a single statewide FWCI would seemingly prevail over regional indicator species and metric scoring criteria, the necessity of scoring each wetland region based on the best possible reference conditions cannot be overlooked (as suggested by Karr and Chu 1999 ). Regionalization of biological indices has been suggested throughout the literature. One of the main premises behind indices of biological integrity is a comparison of like to like ( Gerristen et al. 2000 ), that is, to reduce the noise in background variability in biological data, which could be accomplished through regionally based indices. The lower FWCI scores for wetlands in the central and southern peninsula may also be a factor of the smaller sample size for wetlands located in those wetland regions. Since the macrophyte community composition was found to be different between wetland regions, wetlands in the south and central wetland regions were underrepresented and therefore had less influence in the indicator species analysis and the metric scoring criteria. A larger sample size would improve uncertainty related to questions surrounding the biological integrity of wetlands in the south and central wetland regions. Florida Wetland Condition Index Independent of Wetland Type Recent works by Lane et al. ( 2003 ) and Reiss and Brown ( 2005 ) presented a five metric FWCI for isolated depressional herbaceous wetlands and a six metric FWCI for isolated depressional forested wetlands in Florida based on the community composition of the macrophyte assemblage. The depressional herbaceous FWCI was created based on a sample size of 75 freshwater marshes surrounded by reference (n=34) and agricultural (n=40) land uses throughout peninsular Florida. The depressional forested FWCI was created based on a sample size of 118 freshwater wetlands surrounded by reference (n=37), agricultural (n=40) and urban (n=41) land uses throughout the extent of Florida. The five metrics of the preliminary FWCI for forested strand and floodplain wetlands were nearly identical to the five metrics of the depressional herbaceous FWCI (with an adaptation from annual to perennial ratio from the depressional herbaceous FWCI changed to the proportion native perennial species for the depressional forested and forested strand and floodplain FWCIs) and similar to five of the depressional forested FWCI metrics.

PAGE 81

70 The five macrophyte metrics included in the three wetland type FWCIs (depressional herbaceous, depressional forested, forested strand and floodplain) were tolerant and sensitive indicator species, FQAI score, exotic species, and native perennial species (annual to perennial ratio in the depressional herbaceous FWCI). Tolerant and sensitive indicator species lists were constructed independently for each wetland type. Of the 16 tolerant indicator species for flowing water wetlands, five also occurred as tolerant indicator species for depressional forested wetlands ( Reiss and Brown 2005 ). Two of those species, Commelina diffusa (dayflower) and Ludwigia peruviana (water-primrose) also occurred as tolerant indicator species for depressional herbaceous wetland ( Lane et al. 2003 ). The additional three overlapping tolerant indicator species between the flowing water and depressional forested wetlands were Acer rubrum (red maple), Sambucus canadensis (elderberry), and Toxicodendron radicans (Eastern poison ivy). One tolerant indicator species for flowing water wetlands, Nyssa sylvatica var. biflora (swamp tupelo) was found to be a sensitive indicator species for depressional herbaceous wetlands ( Lane et al. 2003 ); and the tolerant indicator species for forested strand and floodplain wetlands Panicum rigidulum (red-top panicum) was found to be a sensitive indicator species for both depressional herbaceous and forested wetlands. Clearly a larger sample size and refinement of indicator species analysis for the forested strand and floodplain wetlands FWCI could reduce the inconsistencies among indicator species. Similarly, four sensitive indicator species were common among depressional herbaceous, depressional forested, and forested strand and floodplain wetlands, including Cladium jamaicense (saw-grass), Lyonia lucida (fetter-bush), Panicum hemitomon (maidencane), and Pinus elliottii (slash pine). Vaccinium corymbosum (highbush blueberry) was also considered a sensitive indicator species for both depressional herbaceous and forested strand and floodplain wetlands. However, three sensitive indicator species for forested strand and floodplain wetlands were listed as tolerant indicator species for depressional forested (Saururus cernuus, lizards tail) or depressional herbaceous and forested (Centella asiatica, coinwort; Eupatorium capillifolium, dog fennel) wetlands. The three additional metrics included in the FWCIs were Floristic Quality Assessment Index score, exotic species, and native perennial species (modified from the depressional herbaceous wetland FWCI that used annual to perennial ratio). The variant of the annual to perennial ratio (as native perennial species) was used to account for variable conditions at urban wetlands, which were not included in the study of depressional herbaceous wetlands but were studied in both the depressional forested and forested strand and floodplain wetland research. Wienhold and Van der Valk ( 1989 ) and Ehrenfeld and Schneider ( 1991 ) determined that disturbance often favors annual species over perennial species and promotes the invasion of nonnative perennials in wetlands. Galatowitsch et al. ( 2000 ) found that while native perennial cover was reduced in wetlands impacted by cultivation, the occurrence of introduced perennials rather than annuals increased in stormwater impacted wetlands. The sixth depressional forested FWCI metric was the wetland status species (including both obligate and facultative wetland species), which was not included for the preliminary forested strand and floodplain wetland FWCI as it did not meet selection criteria for inclusion. Overall, the depressional herbaceous, depressional forested, and flowing water forested wetland FWCIs included five similar metrics. Perhaps the strong similarity of

PAGE 82

71 metrics selected for inclusion in the FWCIs suggests that a universal assessment index with core metrics could be constructed regardless of wetland type. However, it would likely be necessary to maintain independent indicator species lists and metric scoring criteria for different wetland types within each wetland region. Limitations and Further Research Generally wetlands were visited only once, with a complete sample effort lasting just one day, which provided a mere snapshot of wetland condition. Visiting these wetlands only once did not allow insight into seasonal or yearly variations in the macrophyte assemblage; and the preliminary forested strand and floodplain FWCI would benefit from inter-seasonal validation. The FWCI would also benefit from validation based on a new set of wetlands to test the repeatability of this index. A larger sample size of wetlands will improve the scoring criteria of the FWCI based on wetland regions for metrics such as indicator species analysis. Funding for additional wetland sampling and FWCI refinement is in the application phase. Regionalization appears to be an important next step in refining the FWCI for all wetland types, as this study was limited to a statewide approach due to small sample sizes within each wetland region. Conclusions The use of the macrophyte assemblage for a biological assessment of Florida freshwater flowing water forested wetlands provided a useful tool for detecting changes in biological integrity associated with changes in anthropogenic activity. While richness and diversity measures of macrophyte community composition were not particularly sensitive to changes in landscape development intensity, metrics used as biological indicators of changes in macrophyte community composition were. In fact, the strong correlation between the landscape scale human disturbance gradient (LDI) and the local scale wetland condition index of biological integrity (FWCI) demonstrated the potential value of using the LDI index as an initial indication of biological integrity, which can be further tested with chemical and physical parameters and compared against assemblage specific biological indices. Due to similarities with metrics from the FWCIs for depressional herbaceous ( Lane et al. 2003 ) and depressional forested ( Reiss and Brown 2005 ) wetlands, a multi-metric multi-assemblage FWCI could be constructed for all freshwater wetlands throughout the state of Florida, with a set of core metrics and specific indicator species and metric scores based on wetland type within the wetland regions. While the forested strand and floodplain FWCI for flowing water systems can not be used to predict changes in the physical and chemical parameters of a wetland, its strength lies in providing an overview of biological integrity through the integration of changes in macrophyte community composition from cumulative effects. The quantitative score of biological integrity established through the FWCI should not be used as a surrogate for wetland value, but as an objective, quantitative means of comparing changes in community composition along gradients of human development intensity, which can be used objectively to assess the biological integrity of Floridas wetlands.

PAGE 83

APPENDIX A STANDARD OPERATING PROCEDURES CHECKLIST OF MATERIALS/FIELD EQUIPMENT Miscellaneous SOPs Large cooler with frozen ice bottles for unknown vegetation Waders Garmin III GPS unit Florida Gazetteer Machete Aerial photo & FLUCCS codes of site Vegetation Transects 2-3 100m transect tapes 1 m PVC with distance marks (cm) 2-3 compasses Clipboards Field data sheets a minimum of 10 per site Site Characterization & WRAP sheets 1 per person per site Pencils Sharpie Bag for unknown plants Masking tape Field ID manuals Prism for basal area Hand lens Index cards 72

PAGE 84

73 SOPS FOR FORESTED STRAND WETLANDS: VEGETATION 1. Note the direction of flow through the landscape. 2. Locate a line running through the center of the strand along the flow gradient. This is the center-line. 3. Randomly select a starting point for the initial transect. Each consecutive transect will begin approximately 25 m upstream of the initial transect, so that a stretch of approximately 100 m will be sampled along the length of the strand. Run transects perpendicular to the main channelized flow. 4. At the beginning of each transect, delineate the edge of the wetland using a combination of wetland plants and hydrologic indicators. Be conservative on the side of the wetland. 5. Establish the transect using a meter tape and a compass. Each transect will start with 0 meters at the wetland edge and run into the center-line (established in step 2). 6. Use a separate field data sheet for each transect. If the number of species located on a transect exceeds the number of columns on the data sheet, start a new data sheet. Be thorough in completing field data sheets including information on site, transect direction, date, and data recorder. Specify if there are multiple field data sheets for a single transect. 7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long, record all species rooted within these elongated quadrats. 8. Plant species names are recorded on the data sheets using the full genus and species names. Each unknown species is given a unique ID code using the transect number (ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.). 9. Collect voucher specimens for all unknown species being sure to get plant inflorescence and roots, tag samples with properly labeled masking tape, and put into a labeled collection bag. Note the color of the inflorescence on the label, as the flowers often do not preserve well. Index cards can be used to protect especially sensitive parts. When vegetation sampling is complete, store the collection bag in a cooler on ice until identification can be completed. 10. Voucher specimens are identified in the field on the day of sampling. Unidentified plants will be placed in a plant press for further clarification and identification. Plant nomenclature follows FDEPs Florida Wetland Plant Identification Manual (Tobe et al. 1998). If time prohibits immediate pressing, unknown plants should be stored in the cooler. 11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be recorded. Use the data sheet for basal area, and record basal area per species using variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent elbow. Looking through the prism count the number of trees per species that fall within the variable area plot. The prism shall be centered over the sampling point at all times, with the field person rotating around the prism so that the entire circular area (360 o ) around the point of sampling is included.

PAGE 85

Forested Strand Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands Site: Transect Number: Date: Data Recorder: Species 0-5 m 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100

PAGE 86

Forested Strand Wetlands Field Data Sheet Basal Area UF Center for Wetlands Site: Transect Direction: Date: Data Recorder: Species 10 m 20 m 30 m 40 m 50 m 60 m 70 m 80 m 90 m 100 m

PAGE 87

76 SOPS FOR FORESTED FLOODPLAIN WETLANDS: VEGETATION 1. Note the direction of flow through the landscape. 2. Locate a line running through the center of the strand along the flow gradient. This is the center-line. 3. Randomly select a starting point for the initi al transect, preferab ly this point with coincide with a Stream Condition Index (SCI) sample point. Each consecutive transect will begin approximately 25 m upstr eam of the initial transect, so that a stretch of approximately 100 m will be samp led along the length of the strand. Run transects perpendicular to the main channelized flow. 4. Delineate the wetland line using a combin ation of wetland plants and hydrologic indicators. Be conservativ e on the side of the wetland. 5. Establish the transect using a meter tape and a compass. Transects will be limited to a maximum length of 50 m. The first transect will begin with 0 meters at the wetland edge and run towards the centerline (established in step 2) The second transect will begin 25 m upstream of the firs t transect. Transect 2 begins at the edge of the channelized flow and runs through the we tland perpendicular to the flow for a maximum of 50 m. Repeat placement for transects 3 and 4. 6. Use a separate field data sheet for each tr ansect. If the number of species located on a transect exceeds the number of columns on the data sheet, start a new data sheet. Be thorough in completing field data sheets including information on site, transect direction, date, and data recorder. Specify if there are multiple field data sheets for a single transect. 7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long, record all species rooted w ithin these elongated quadrats. 8. Plant species names are recorded on the data sheets using the full genus and species names. Each unknown species is given a uni que ID code using the transect number (ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.). 9. Collect voucher specimens for all unknown species being sure to get plant inflorescence and roots, tag samples with properly labeled masking tape, and put into a labeled collection bag. Note the color of the inflorescence on the label, as the flowers often do not preserve well. Index cards can be us ed to protect especially sensitive parts. When vegetation sampling is complete, store the collection bag in a cooler on ice until identi fication can be completed. 10. Voucher specimens are identified in the fi eld on the day of sampling. Unidentified plants will be placed in a plant press for fu rther clarification and identification. Plant nomenclature follows FDEPs Florida Wetland Plant Identification Manual (Tobe et al. 1998). If time prohibits immediate pres sing, unknown plants should be stored in the cooler. 11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be recorded. Use the data sheet for basal ar ea, and record basal area per species using variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent

PAGE 88

77 elbow. Looking through the prism count the number of trees per species that fall within the variable area plot. The prism shall be centered over the sampling point at all times, with the field person rotating around the prism so that the entire circular area (360 o ) around the point of sampling is included.

PAGE 89

Forested Floodplain Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands Site: Transect Number: Date: Data Recorder: Species 0-5 m 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50

PAGE 90

Forested Floodplain Wetlands Field Data Sheet Basal Area UF Center for Wetlands Site: Transect Direction: Date: Data Recorder: Species 10 m 20 m 30 m 40 m 50 m

PAGE 91

APPENDIX B COEFFICIENT OF CONSERVATISM SCORES Table B-1. Coefficient of Conservatism (CC) scores for macrophyte species identified in forested strand and floodplain wetlands in Florida. CC scores were assigned from FQAI surveys from isolated depressional forested wetlands (Reiss and Brown 2005) first, followed by isolated depressional herbaceous wetlands (Lane et al. 2003). Species without scores from previous FQAI studies were scores according to their faithfulness and fidelity to freshwater flowing water wetlands (strands and floodplains) in 2003 by five expert Florida botanists (Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and Bruce Tatje). Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Abrus precatorius 0.0 1 Acer rubrum 5.2 1 Acer saccharum 6.0 1 Agarista populifolia 6.5 1 Alnus serrulata 6.0 1 Alternanthera philoxeroides 0.0 1 Alternanthera sessilis 0.7 1 Amaranthus australis 2.6 1 Ambrosia artemisiifolia 0.7 1 Ampelopsis arborea 3.3 1 Amphicarpum muhlenbergianum 5.0 1 Andropogon virginicus 2.6 1 Annona glabra 6.8 1 Apios americana 3.1 1 Ardisia escallonioides 7.0 1 Arisaema triphyllum 6.5 1 Aronia arbutifolia 5.7 1 Arundinaria gigantea 5.3 1 Asimina parviflora 5.0 1 Asimina reticulata 4.4 1 Asplenium heterochroum 8.5 1 Aster elliottii 4.2 1 Bacopa caroliniana 6.0 1 Bacopa monnieri 4.3 1 Begonia cucullata 1.5 1 Berchemia scandens 5.1 1 Bidens alba 1.0 1 Bidens mitis 3.8 1 Bignonia capreolata 4.4 1 Blechnum serrulatum 5.5 1 Boehmeria cylindrica 4.5 1 Bumelia lycioides 6.0 1 Callicarpa americana 2.4 1 Campsis radicans 3.3 1 Carex albolutescens 3.6 1 80

PAGE 92

81 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Carex crus-corvi 6.0 1 Carex leptalea 6.8 1 Carex lupulina 6.3 1 Carex stipata 4.5 1 Carpinus caroliniana 6.8 1 Carya glabra 5.8 1 Celtis laevigata 5.0 1 Centella asiatica 1.9 1 Cephalanthus occidentalis 6.0 1 Chasmanthium laxum 6.0 1 Chasmanthium nitidum 6.3 1 Chrysobalanus icaco 6.3 1 Cinnamomum camphora 0.2 1 Cladium jamaicense 5.5 1 Clematis crispa 5.5 1 Colocasia esculenta 0.0 1 Commelina communis 2.5 1 Commelina diffusa 1.7 1 Conyza canadensis 0.3 1 Cornus foemina 6.6 1 Crataegus marshallii 6.3 1 Crinum americanum 7.6 1 Cuphea carthagenensis 1.4 1 Cynodon dactylon 0.0 1 Cyperus difformis 2.0 1 Cyperus globulosus 1.8 1 Cyperus haspan 2.6 1 Cyperus ligularis 3.2 1 Cyperus odoratus 3.6 1 Cyrilla racemiflora 4.5 1 Decumaria barbara 6.3 1 Dichondra carolinensis 1.9 1 Dichromena colorata 5.5 1 Digitaria serotina 1.8 1 Diodia virginiana 2.4 1 Dioscorea floridana 5.5 1 Diospyros virginiana 4.0 1 Eclipta alba 1.7 1 Emilia fosbergii 0.5 1 Erechtites hieraciifolius 2.1 1 Erianthus giganteus 6.0 1 Eryngium baldwini 4.4 1 Eryngium prostratum 4.0 1 Erythrina herbacea 4.0 1 Eugenia uniflora 1.3 1 Euonymus americanus 5.5 1 Eupatorium capillifolium 0.5 1 Eupatorium mikanioides 5.5 1

PAGE 93

82 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Eupatorium perfoliatum 5.9 1 Eustachys petraea 2.0 1 Fagus grandifolia 7.5 1 Fraxinus caroliniana 7.1 1 Galium tinctorium 3.1 1 Gaylussacia dumosa 5.4 1 Gaylussacia frondosa 6.7 1 Gelsemium sempervirens 4.0 1 Gordonia lasianthus 6.7 1 Gratiola hispida 6.0 1 Hamamelis virginiana 6.5 1 Hibiscus coccineus 6.6 1 Hydrocotyle bonariensis 3.3 1 Hydrocotyle ranunculoides 3.1 1 Hydrocotyle umbellata 2.9 1 Hydrocotyle verticillata 3.1 1 Hypericum brachyphyllum 6.8 1 Hypericum hypericoides 4.0 1 Hypericum myrtifolium 5.5 1 Hypericum tetrapetalum 5.0 1 Hypoxis curtissii 5.7 1 Hyptis alata 4.3 1 Ilex cassine 8.1 1 Ilex coriacea 6.0 1 Ilex glabra 4.3 1 Ilex opaca var. opaca 6.0 1 Ilex vomitoria 4.8 1 Ipomoea hederifolia 2.0 1 Ipomoea pandurata 4.8 1 Ipomoea sagittata 5.4 1 Itea virginica 7.9 1 Juncus effusus 1.9 1 Juncus megacephalus 3.3 1 Juncus polycephalos 3.3 1 Juniperus virginiana 5.2 1 Justicia ovata 5.5 1 Koelreuteria elegans 2.0 1 Lachnanthes caroliana 3.1 1 Lemna minor 1.0 1 Leucothoe axillaris 7.0 1 Leucothoe racemosa 6.2 1 Ligustrum sinense 0.0 1 Lindernia grandiflora 3.6 1 Liquidambar styraciflua 3.3 1 Liriodendron tulipifera 6.8 1 Lobelia cardinalis 6.8 1 Lonicera japonica 0.0 1 Ludwigia maritima 3.3 1

PAGE 94

83 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Ludwigia palustris 4.0 1 Ludwigia peruviana 1.2 1 Ludwigia pilosa 5.8 1 Ludwigia repens 2.9 1 Lycopus rubellus 5.2 1 Lygodium japonicum 0.0 1 Lygodium microphyllum 0.0 1 Lyonia fruticosa 6.0 1 Lyonia lucida 6.0 1 Lyonia mariana 6.8 1 Magnolia grandiflora 6.2 1 Magnolia virginiana var. australis 8.1 1 Matelea floridana 6.7 1 Melaleuca quinquenervia 0.0 1 Merremia dissecta 0.3 1 Micranthemum glomeratum 4.0 1 Micranthemum umbrosum 4.3 1 Micromeria brownei 4.8 1 Mikania scandens 2.4 1 Mimosa pigra 0.7 1 Mitchella repens 6.7 1 Momordica charantia 0.0 1 Morus rubra 4.4 1 Myrica cerifera 3.1 1 Myrsine guianensis 5.2 1 Nephrolepis exaltata 3.8 1 Nyssa ogeche 7.0 1 Nyssa sylvatica var. biflora 7.4 1 Oeceoclades maculata 0.4 1 Oplismenus setarius 3.3 1 Orontium aquaticum 7.6 1 Osmunda cinnamomea 5.5 1 Osmunda regalis 6.9 1 Oxypolis filiformis 6.7 1 Panicum abscissum 9.2 1 Panicum anceps 4.3 1 Panicum commutatum 4.5 1 Panicum dichotomum 4.0 1 Panicum ensifolium 5.0 1 Panicum erectifolium 5.7 1 Panicum hemitomon 5.0 1 Panicum rigidulum 4.5 1 Panicum spretum 5.4 1 Panicum tenue 4.2 1 Parietaria praetermissa 3.0 1 Parthenocissus quinquefolia 3.0 1 Paspalum conjugatum 3.1 1 Paspalum notatum 0.0 1

PAGE 95

84 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Passiflora incarnata 3.0 1 Peltandra virginica 5.8 1 Persea borbonia 6.3 1 Persea palustris 7.4 1 Phlebodium aureum 6.8 1 Phyla nodiflora 1.4 1 Phyllanthus urinaria 0.0 1 Phytolacca americana 1.2 1 Pieris phyllyreifolia 9.5 1 Pinckneya bracteata 8.3 1 Pinus clausa 5.6 1 Pinus elliottii 4.0 1 Pinus taeda 3.3 1 Pluchea foetida 3.8 1 Pluchea rosea 3.6 1 Polygala rugelii 8.2 1 Polygonum densiflorum 5.3 1 Polygonum hirsutum 8.2 1 Polygonum hydropiperoides 2.6 1 Pontederia cordata 5.0 1 Proserpinaca palustris 3.8 1 Prunus caroliniana 3.0 1 Prunus serotina 3.6 1 Psychotria nervosa 5.2 1 Psychotria sulzneri 5.5 1 Pteridium aquilinum 3.6 1 Ptilimnium capillaceum 3.1 1 Quercus laurifolia 3.6 1 Quercus michauxii 5.7 1 Quercus nigra 2.1 1 Quercus virginiana 4.2 1 Rhododendron canescens 6.8 1 Rhododendron viscosum 7.6 1 Rhus copallinum 2.4 1 Rhynchospora baldwinii 5.7 1 Rhynchospora inundata 6.0 1 Rhynchospora microcephala 4.8 1 Rhynchospora miliacea 7.1 1 Rubus argutus 2.1 1 Rubus trivialis 1.9 1 Ruellia caroliniensis 4.3 1 Rumex verticillatus 4.8 1 Sabal minor 6.2 1 Sabal palmetto 4.5 1 Sabatia calycina 6.2 1 Sagittaria filiformis 6.0 1 Sagittaria lancifolia 4.5 1 Sagittaria latifolia 5.0 1

PAGE 96

85 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Salvia lyrata 3.2 1 Sambucus canadensis 1.7 1 Samolus valerandi 5.6 1 Sanicula canadensis 5.7 1 Sapium sebiferum 0.0 1 Saururus cernuus 5.5 1 Schefflera actinophylla 0.5 1 Schinus terebinthifolius 0.0 1 Scleria triglomerata 4.8 1 Serenoa repens 4.5 1 Setaria geniculata 3.1 1 Sida rhombifolia 1.0 1 Smilax auriculata 3.8 1 Smilax bona-nox 2.6 1 Smilax glauca 3.3 1 Smilax laurifolia 5.2 1 Smilax pumila 6.0 1 Smilax smallii 4.5 1 Smilax tamnoides 3.6 1 Smilax walteri 6.0 1 Solidago fistulosa 3.6 1 Sparganium americanum 6.7 1 Sporobolus floridanus 7.1 1 Stenotaphrum secundatum 0.8 1 Stillingia aquatica 7.4 1 Symplocos tinctoria 6.0 1 Taxodium ascendens 8.8 1 Taxodium distichum 7.2 1 Thelypteris dentata 3.4 1 Thelypteris hispidula 4.5 1 Thelypteris palustris 5.8 1 Tilia americana 5.5 1 Toxicodendron radicans 1.9 1 Triadenum virginicum 5.0 1 Trichostema dichotomum 4.5 1 Trifolium repens 0.0 1 Tripsacum dactyloides 4.0 1 Ulmus americana 7.4 1 Urena lobata 0.0 1 Vaccinium arboreum 6.4 1 Vaccinium corymbosum 5.7 1 Vaccinium stamineum 5.8 1 Vaccinium tenellum 5.5 1 Viburnum dentatum 6.0 1 Viburnum nudum 5.0 1 Viburnum obovatum 4.7 1 Viola affinis 5.5 1 Vitis aestivalis 2.9 1

PAGE 97

86 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Vitis cinerea 2.0 1 Vitis rotundifolia 2.1 1 Vitis shuttleworthii 3.5 1 Wisteria sinensis 1.0 1 Woodwardia areolata 5.7 1 Woodwardia virginica 4.8 1 Xyris jupicai 1.7 1 Youngia japonica 0.0 1

PAGE 98

APPENDIX C SUMMARY STATISTICS Table C-1. Summary statistics of richness (R), evenness (E), Shannon diversity (H'), Simpson diversity (D), and Whittakers beta diversity (W) for the macrophyte assemblage (species level). Site R E H' D w FF1 59 1.0001 4.08 0.98 5.88 FF2 21 1.0002 3.05 0.95 3.73 FF3 56 0.9999 4.03 0.98 4.86 FF4 46 1.0001 3.83 0.98 7.36 FF5 77 1.0000 4.34 0.99 6.82 FF6 25 1.0000 3.22 0.96 1.78 FF7 32 1.0001 3.47 0.97 3.49 FF8 75 0.9999 4.32 0.99 6.11 FF9 45 1.0001 3.81 0.98 3.90 FF10 35 0.9999 3.56 0.97 3.18 FF11 29 0.9999 3.37 0.97 4.73 FF12 46 1.0001 3.83 0.98 5.15 FF13 48 0.9999 3.87 0.98 4.93 FF14 60 0.9999 4.09 0.98 5.94 FS1 39 1.0001 3.66 0.97 2.68 FS2 39 1.0001 3.66 0.97 4.90 FS3 33 1.0001 3.50 0.97 4.45 FS4 55 0.9999 4.01 0.98 5.03 FS5 39 1.0001 3.66 0.97 3.91 FS6 29 0.9999 3.37 0.97 3.02 FS7 30 0.9999 3.40 0.97 2.96 FS8 35 0.9999 3.56 0.97 2.12 FS9 47 1.0000 3.85 0.98 3.73 FS10 42 1.0001 3.74 0.98 4.43 87

PAGE 99

APPENDIX D METRIC SCORING FOR THE MACROPHYTE FLORIDA WETLAND CONDITION INDEX FOR FLOWING WATER SYSTEMS 1. Calculate values for the 5 metrics: 1 Proportion tolerant indicator species 2 Proportion sensitive indicator species 3 FQAI score 4 Proportion exotic species 5 Proportion native perennial species 2. Take the natural log of metrics to improve distribution. = ln (metric value + 1) 1 is added to avoid errors related to taking the natural log of a zero value 3. Use the scoring equations to normalize scores between 0 and 10. Metrics that increase with increasing LDI tolerant, exotic metrics: = 10 (( metric 5th percentile) ( 10 / ( 95th percentile 5th percentile))) Metrics that decrease with increasing LDI sensitive, FQAI, native perennial metrics: = (( metric 5th percentile) ( 10 / ( 95th percentile 5th percentile))) Below are the 5th and 95th percentiles for each metric (transformed values are presented, see step 2 above): 5th Percentile 95th Percentile Proportion tolerant indicator species 0.05 0.24 Proportion sensitive indicator species 0.00 0.16 FQAI score 1.44 1.84 Proportion exotic species 0.00 0.17 Proportion native perennial species 0.56 0.69 4. Rescore, so that the metrics in the outer 5th percentiles receive scores of 0 or 10. = IF ( score < 0, 0, ( IF ( score >= 10, score, 10))) 88

PAGE 100

LIST OF REFERENCES Adams, S.M. 2002. Biological indicators of aquatic ecosystem stress: introduction and overview. Pages 1-11 in S.M. Adams, editor. Biological indicators of aquatic ecosystem stress. American Fisheries Society. Bethesda, Maryland, USA. Analyse-it Software, Ltd. 1997-2003. version 1.67. Leeds, England, United Kingdom. Andreas, B.K. and R.W. Lichvar. 1995. A floristic assessment system for northern Ohio. Wetlands Research Program Technical Report WRP-DE-8. U.S. Army Corps of Engineers Waterways Experiment Station, Vicksburg, Mississippi, USA. Apfelbeck, R. 2000. Developing preliminary bioassessment protocols for Montana wetlands, State of Montana Department of Environmental Quality. Helena, Montana, USA. Arcview GIS 3.2 Environmental Systems Research Institute, Inc. 1999. Neuron Data, Inc. 1991-1996. Portions copyright 1991-1995 Arthur D. Applegate. Found at: http://www.esri.com/. Redlands, California, USA. ArcGIS 8.3 Environmental Systems Research Institute, Inc. 1999-2002. Found at: http://www.esri.com/arcgis. Redlands, California, USA. Barbour, M.T., J. Gerristen, G.E. Griffith, R. Frydenborg, E. McCarron, J.S. White, and M.L. Bastian. 1996a. A framework for biological criteria for Florida streams using benthic macroinvertebrates. Journal of the North American Benthological Society 15(2): 185-211. Barbour, M.T., J. Gerristen, and J.S. White. 1996b. Development of the Stream Condition Index (SCI) for Florida. A Report to the Florida Department of Environmental Protection, Stormwater and Nonpoint Source Management Section. Tetra Tech, Inc. Owing Mills, Maryland, USA. Bedford, B.L., M.R. Wabridge, and A. Aldous. 1999. Patterns in nutrient availability and plant diversity of temperate North American wetlands. Ecology 80(7): 2151-2169. Blanch, S.J. and M.A. Brock. 1994. Effects of grazing and depth on two wetland plant species. Australian Journal of Marine and Freshwater Research 45: 1387-1394. Brown, M.T. and M.B. Vivas. 2005. Landscape Development Intensity Index. Environmental Monitoring and Assessment 101: 289-309. Brown, M.T. and S. Ulgiati. 2005. Emergy, transformity, and ecosystem health. Pages 333-352 in S.E. Jrgensen, R. Costanza, and F. Xu, editors. Handbook of ecological indicators for assessment of ecosystem health. Taylor and Francis, Boca Raton, Florida, USA. Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117-143. Cohen, M.J., S.M. Carstenn, and C.R. Lane. 2004. Floristic quality indices for biotic assessment of depressional marsh condition in Florida. Ecological Applications 14(3): 784-794. Cronk, J. K. and M.S. Fennessy. 2001. Wetland plants: biology and ecology. Lewis Publishers. Boca Raton, Florida, USA. Crowder, A. and D.S. Painter. 1991. Submerged macrophytes in Lake Ontario: current knowledge, importance, threats to stability, and needed studies. Canadian Journal of Fisheries and Aquatic Sciences 48:1539-1545. 89

PAGE 101

90 Dahl, T.E. 2000. Status and trends of wetlands in the conterminous United States 1986 to 1997. United States Department of the Interior, Fish and Wildlife Service, Washington, D.C., USA. Danielson, T.J. 1998. Indicators for monitoring and assessing biological integrity of inland freshwater wetlands. EPA 843-R-98-002. Wetlands Division Office of Water, United States Environmental Protection Agency, Washington, D.C., USA. David, P.G. 1999. Response of exotics to restored hydroperiod at Dupuis Reserve, Florida. Restoration Ecology 7(4): 407-410. Davis, J.H. 1967. General map of natural vegetation of Florida. Accessed: 12/2004. Found at: http://www.geoplan.ufl.edu/fdgl/fgdl.html Devine, R. 1998. Alien invasion. National Geographic Society, Washington, D.C., USA. Dufrne, M. and P. Legendre. 1997. Species assemblages and indicator species the need for a flexible asymmetrical approach. Ecological Monographs 67(3): 345-366. Ehrenfeld, J.G. and J.P. Schneider. 1991. Chamaecyparis thyoides wetlands and suburbanization: effects of non-point source water pollution on hydrology and plant community structure. Journal of Applied Ecology 28(2): 467-490. Ewel, K.C. 1990. Swamps. Pages 281-323 in R.L. Myers and J.J. Ewel. Ecosystems of Florida. University of Central Florida Press. Orlando, Florida, USA. Exotic Plant Pest Council (EPPC). 2003. Accessed 3/2005. Found at: http://www.fleppc.org/ FAC 62-340. Florida Administrative Code, Chapter 62-340. Accessed 3/2005. Found at http://www.dep.state.fl.us/legal/rules/surfacewater/62-340.pdf. Florida Department of Environmental Protection. Fennessy, S., R. Geho, B. Elifritz and R. Lopez. 1998. Testing the floristic quality assessment index as an indicator of riparian wetland quality. Final report to U.S. EPA. Ohio Environmental Protection Agency, Division of Surface Water, Columbus, Ohio, USA. Fennessy, S., M. Gernes, J. Mack, and D.H. Waldrop. 2001. Methods for evaluating wetland condition: using vegetation to assess environmental conditions in wetlands. EPA 822-R-01-007j. U.S. Environmental Protection Agency, Office of Water, Washington, D.C., USA. Fernald, E.A. and E.D. Purdum, editors. 1992. Atlas of Florida. University Press of Florida. Gainesville, Florida, USA. Fore, L.S. 2004. Development and testing of biomonitoring tools for macroinvertebrates in Florida streams. Statistical Design, Seattle, Washington. A report for the Florida Department of Environmental Protection, Tallahassee, Florida, USA. Fore, L.S. and C. Grafe. 2002. Using diatoms to assess the biological condition of large rivers in Idaho (U.S.A.). Freshwater Biology 47: 2015-2037. Francis, C.M., M.J.W. Austen, J.M. Bowles, W.B. Draper. 2000. Assessing floristic quality in southern Ontario woodlands. Natural Areas Journal 20:66-77. Galatowitsch, S.M., N.O. Anderson, and P.D. Ascher. 1999b. Invasiveness in wetland plants in temperate North America. Wetlands 19(4): 733-755. Galatowitsch, S.M., D.C. Whited, R. Lehtinen, J. Husveth, and K. Schik. 2000. The vegetation of wet meadows in relation to their land-use. Ecological Monitoring and Assessment 60: 121-144.

PAGE 102

91 Galatowitsch, S.M., D.C. Whited, and J.R. Tester. 1999a. Development of community metrics to evaluate recovery of Minnesota wetlands. Journal of Aquatic Ecosystem Stress and Recovery 6: 217-234. Gernes, M.C. and J.C. Helgen. 1999. Indexes of biotic integrity for wetlands, section B: wetland vegetation IBI for depressional wetlands. Final Report to the United States Environmental Protection Agency Assistance Number CD995525-01, April 1999. Minnesota Pollution Control Agency, St. Paul, Minnesota, USA. Gerristen, J., M.T. Barbour, and K. King. 2000. Apples, oranges, and ecoregions: on determining pattern in aquatic assemblages. Journal of the North American Benthological Association 19(3): 487-496. Gerristen, J. and J. White. 1997. Development of a biological index for Florida lakes. A Report to the Florida Department of Environmental Protection. Tetra Tech, Inc. Owing Mills, Maryland, USA. Godfrey, R.K. and J.W. Wooten. 1981. Aquatic and wetland plants of the southeastern United States. University of Georgia Press, Athens, Georgia, USA. Grace, J.B. and H. Jutila. 1999. The relationship between species density and community biomass in grazed and ungrazed coastal meadows. Oikos 85: 398-408. Griffith, G.E., J.M. Omernik, C.M. Rohm, and S.M. Pierson. 1994. Florida Regionalization Project. Environmental Research Laboratory, United States Environmental Protection Agency, Corvallis, OR. Griffith, G. E., D. E. Canfield, Jr., C. A. Horsburgh, J. M. Omernik, and S. H. Azevedo. 1997. Florida lake regions. Environmental Research Laboratory, United States Environmental Protection Agency, Corvallis, OR. Herman, K.D., A.A. Reznicek, L.A. Masters, G.S. Wilhelm, M.R. Penskar and W.W. Brodowicz. 1997. Floristic quality assessment: development and application in the state of Michigan (USA). Natural Areas Journal 17:265-279. Hobbs, R.J. and L.F. Hueneke. 1992. Disturbance, diversity, and invasion: implications for conservation. Conservation Biology 6(3): 324-337. James, M.O. and K.M. Kleinow. 1994. Trophic transfers of chemicals in the aquatic environment. Pages 1-35 in D.C. Malins and G.K. Ostrander, editors. Aquatic toxicology and cellular perspectives. Lewis Publishers. Boca Raton, Florida, USA. Karr, J.R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6: 21-27. Karr, J.R. 1993. Defining and assessing ecological integrity: beyond water quality. Environmental Toxicology and Chemistry 12: 1521-1531. Karr, J.R. and E.W. Chu. 1997. Biological monitoring and assessment: using multimetric indexes effectively. EPA 235-R-97-001. University of Washington, Seattle, Washington, USA. Karr, J.R. and E.W. Chu. 1999. Restoring life in running waters. Island Press. Washington, D.C., USA. Karr, J.R. and D.R. Dudley. 1981. Ecological perspectives on water quality goals. Environmental Management 5: 55-68. Kent, D.M. 2000. Evaluating wetland functions and values. Chapter 3 in D.M. Kent, editor. Applied wetlands science and technology. Lewis Publishers. Boca Raton, Florida, USA.

PAGE 103

92 Kerans, B.L. and J.R. Karr. 1994. A benthic index of biotic integrity (B-IBI) for rivers in the Tennessee valley. Ecological Applications 4(4): 768-785. Kruskal, J.B. 1964. Multidimensional scaling by optimizing goodness of fit to a non-metric hypothesis. Psychometrika 29: 115-129. Lane, C.R. 2000. Proposed ecological regions for freshwater wetlands of Florida. Masters Thesis, University of Florida, Gainesville, Florida, USA. Lane, C.R. 2003. Biological indicators of wetland condition for isolated depressional herbaceous wetlands in Florida. Ph.D. Dissertation, University of Florida, Gainesville, Florida, USA. Lane, C.R., M.T. Brown, M. Murray-Hudson, and M.B. Vivas. 2003. The Wetland Condition Index (WCI): biological indicators of wetland condition for isolated depressional herbaceous wetlands in Florida. A Report to the Florida Department of Environmental Protection. Howard T. Odum Center for Wetlands, University of Florida, Gainesville, Florida, USA. Micacchion, M. 2004. Integrated Wetland Assessment Program. Part 7: Amphibian Index of Biotic Integrity (AmphIBI) for Ohio Wetlands. Ohio EPA Technical Report WET/2004-7. Ohio Environmental Protection Agency, Wetland Ecology Group, Division of Surface Water, Columbus, Ohio. Mack, J. 2001. Vegetation Index of Biological Integrity (VIBI) for wetlands: ecoregional, hydrogeomorphologic, and plant community comparisons with preliminary wetland aquatic life use designations. Final Report to the United States Environmental Protection Agency Grant No. CD985875, Volume 1. Wetland Ecology Group, Division of Surface Water, Ohio Environmental Protection Agency, Columbus Ohio, USA. Found at: http://www.epa.state.oh.us/dsw.wetlands/wetlands_bioasses.html. McCune, B. and J.B. Grace. 2002. Analysis of ecological communities. MJM Software Design. Gleneden Beach, Oregon, USA. McCune, B., R. Rosentreter, J.M. Ponzetti, and D.C. Shaw. 2000. Epiphyte habitats in an old conifer forest in western Washington, USA. Bryologist 103: 417-427. Miller, R.E., Jr. and B.E. Boyd. 1999. Wetland rapid assessment procedure. South Florida Water Management District, Technical Publication REG-001. West Palm Beach, Florida, USA. Minitab Statistical Software, version 13.1. 2000. Found at: http://www.minitab.com. State College, Pennsylvania, USA. Mitsch, W.J. and J.G. Gosselink. 1993. Wetlands, 2nd edition. John Wiley and Sons, Inc. New York, New York, USA. Mushet, D.M., N.H. Euliss, and T.L. Shaffer. 2002. Floristic quality assessment of one natural and three restored wetland complexes in North Dakota, USA. Wetlands 22(1): 126-138. Myer, R.L. and J.J. Ewel, eds. 1990. Ecosystems of Florida. University of Central Florida Press. Orlando, Florida, USA. OConnell, T.J., L.E. Jackson, and R.P. Brooks. 1998. A bird community index of biotic integrity for the mid-Atlantic highlands. Environmental Monitoring and Assessment 51: 145-156. Odum, H.T. 1995. Environmental Accounting: Emergy and Environmental Decision Making. John Wiley and Sons, New York, New York, USA.

PAGE 104

93 Ott, R.L. and M. Longnecker. 2001. An introduction to statistical methods and data analysis, 5th edition. Duxbury, Wadsworth Group. Pacific Grove, California, USA. PCORD, version 4.1. MJM Software.. Found at: http://home.centurytel.net/~mjm/. Gleneden Beach, Oregon, USA Reiss, K.C. 2004. Developing biological indicators for isolated forested wetlands in Florida. Ph.D. Dissertation, University of Florida, Gainesville, Florida, USA. Reiss, K.C. and M.T. Brown. 2005. The Florida Wetland Condition Index (FWCI): developing biological indicators for isolated depressional forested wetlands. A Report to the Florida Department of Environmental Protection. Howard T. Odum Center for Wetlands, University of Florida, Gainesville, Florida, USA. Schindler, D.W. 1987. Detecting ecosystem responses to anthropogenic stress. Canadian Journal of Fisheries and Aquatic Sciences 44:6-25. Schulz, E.J., M.V. Hoyer, and D.E. Canfield, Jr. 1999. An index of biotic integrity: a test with limnological and fish data from sixty Florida lakes. Transactions of the American Fisheries Society 128: 564-577. ter Braak, C.J.F. 1987. The analysis of vegetation-environmental relationships by canonical correspondence analysis. Vegetatio 69: 69-77. Tobe, J.D., K. Craddock Burks, R.W. Cantrell, M.A. Garland, M.E. Sweeley, D.W. Hall, P. Wallace, G. Anglin, G. Nelson, J.R. Cooper, D. Bickner, K. Gilbert, N. Aymond, K. Greenwood, and N. Raymond. 1998. Florida wetland plants: an identification manual. Florida Department of Environmental Protection, Tallahassee, Florida, USA. United States Department of Agriculture, Natural Resource Conservation Service (USDA NRCS). 2002. The PLANTS Database, Version 3.5. National Plant Data Center, Baton Rouge, Louisiana, USA. Available on-line at: http://plants.usda.gov. Accessed 2001-2004. United States Environmental Protection Agency (USEPA). 1990. Feasibility report on environmental indicators for surface water programs. Office of Water Regulations and Standards and Office of Policy, Planning and Evaluation. Washington, D.C., USA. United States Environmental Protection Agency (USEPA). 1998a. Wetland Biological Assessments and HGM Functional Assessment. EPA 843-F-98-001f Wetland Bioassessment Fact Sheet 6. Washington, D.C., USA. Available on-line at: http://www.epa.gov/owow/wetlands/wqual/bio_fact/fact6.html. Accessed 2005. United States Environmental Protection Agency (USEPA). 1998b. Lake and reservoir bioassessment and biocriteria. EPA 841-B-98-007 Technical Guidance Document. Washington, D.C., USA. Available on-line at: http://www.epa.gov/owow/monitoring/tech/lakes.html. Accessed 2002-2004. United States Environmental Protection Agency (USEPA). 2001. Better Assessment Science Integrating Point and Nonpoint Sources (BASINS), Version 3.0. EPA 823-B-01-001. Users Manual. Office of Water. Technical Support and software download available at: http://www.epa.gov/ost/basins/. Accessed 2003-2005. United States Environmental Protection Agency (USEPA). 2002. Methods for evaluating wetland condition: introduction to biological assessment. EPA-822-R-02-014. Office of Water, Washington, D.C., USA.

PAGE 105

94 United States Environmental Protection Agency (USEPA). 2003. Biological Indicators of Watershed Health. Available on-line at: http://www.epa.gov/bioindicators. Accessed 2003-2004. Wharton, C.H., H.T. Odum, K. Ewel, M. Duever, A. Lugo, R. Boyt, J. Bartholomew, E. DeBellevue, S. Brown, M. Brown, and L. Deuver. 1976. Forested wetlands of Florida their management and use. Center for Wetlands, University of Florida, Gainesville, Florida, USA. Wienhold, C.E. and A.G. Van der Valk. 1989. The impact of duration of drainage on the seed banks of northern prairie wetlands. Canadian Journal of Botany 67: 1878-1884. Wilhelm, G. and D. Ladd. 1988. Natural Area Assessment in the Chicago Region. Pages 361-375 in Transactions of the 53rd North American Wildlife and Natural Resource Conference, Louisville, Kentucky. Wildlife Management Institute, Washington D.C., USA. Wunderlin, R.P. 1998. Guide to the Vascular Plants of Florida. University Press of Florida, Gainesville, Florida, USA. Wunderlin, R. P., and B. F. Hansen. 2003. Atlas of Florida Vascular Plants. S. M. Landry and K. N. Campbell, application development. Florida Center for Community Design and Research. Institute for Systematic Botany, University of South Florida. Found at: http://www.plantatlas.usf.edu/. Tampa, Florida, USA. Zimmerman, G.M., H. Goetz, and P.W. Mielke, Jr. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66(2): 606-61.



PAGE 1

Pilot Study The Florida Wetland Condition Index (F WCI): Preliminary Development of Biological Indicators for Fore sted Strand and Floodplain Wetlands Report Submitted to the Florida Department of Environmental Protection Under Contract #WM-683 Kelly Chinners Reiss and Mark T. Brown Howard T. Odum Center for Wetlands University of Florida Gainesville, Florida 32611-6350 June 2005

PAGE 2

ACKNOWLEDGMENTS Research on biological indicators was supported by a grant to Mark T. Brown, principle investigator, from the Florida Department of Environmental Protection (FDEP). FDEP staff provided support for this research, particularly Russ Frydenborg, Ellen McCarron, Ashley ONeal, Erica Hernandez, Julie Espy, Tom Frick, Joy Jackson, Liz Miller, Johnny Richardson, and Lori Wolfe. FDEP staff served as reviewers for the draft report, including Russ Frydenborg, Connie Bersock, Nia Wellendorf, Julie Espy, and Joy Jackson. Additionally, acknowledgement is due to the systems ecology research group at the Howard T. Odum Center for Wetlands. In particular, Chuck Lane (who developed biological indicators for Florida marshes) provided a framework for this analysis. Assistance in field-data collection, laboratory analysis, data entry, and/or feedback on statistical analyses from Eliana Bardi, Matt Cohen, Tony Davanzo, Melissa Friedman, Kristina Jackson, Joanna Reilly-Brown, Vanessa Rumancik, Kris Sullivan, Jim Surdick, and Casey Chinners Virata, was particularly valuable. Acknowledgement is due to the Florida botanists who participated in the Floristic Quality Assessment index surveys from 1999-2004, including: Guy Anglin, Anthony Arcuri, Dan Austin, Keith Bradley, Kathy Burks, David Hall, Ashley ONeal, Jim Poppleton, Nina Raymond, Bruce Tatje, John Tobe, and Wendy Zomlefer. This project and the preparation of this report were funded in part by a Section 319 Nonpoint Source Management grant from the U.S. Environmental Protection Agency through a contract with the Florida Department of Environmental Protection. ii

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS................................................................................................ii LIST OF TABLES...............................................................................................................v LIST OF FIGURES..........................................................................................................vii EXECUTIVE SUMMARY...............................................................................................ix CHAPTER 1 INTRODUCTION AND OVERVIEW...................................................................1 Forested Strand and Floodplain Wetlands...............................................................1 Historical Perspective.............................................................................................. 5 Biological Indicators of Ecosystem Integrity..........................................................6 Quantifying Anthropogenic Influence..................................................................... 8 Landscape Development Intensity Index..................................................... 8 Human Disturbance Gradient...................................................................... 9 Project Overview...................................................................................................11 2 METHODS............................................................................................................13 Study Area.............................................................................................................13 Site Selection......................................................................................................... 14 Gradients of Landscape Development Intensity....................................................17 Field-data Collection..............................................................................................21 Transect Sampling Design.........................................................................21 Floristic Quality Assessment.....................................................................24 Data Analysis ........................................................................................................27 Summary Statistics.....................................................................................27 Regional Compositional Analysis..............................................................28 Community Composition...........................................................................29 Metric Development..............................................................................................30 Florida Wetland Condition Index..........................................................................31 3 RESULTS..............................................................................................................33 Gradients of Anthropogenic Activity.....................................................................33 Landscape Development Intensity.............................................................33 Human Disturbance Gradient....................................................................37 iii

PAGE 4

Water Quality.............................................................................................40 Data Analysis.........................................................................................................40 Summary Statistics.....................................................................................42 Regional Compositional Analysis..............................................................42 Community Composition...........................................................................44 Metric Selection.........................................................................................46 Tolerance metrics...........................................................................46 Floristic Quality Assessment Index metric....................................49 Exotic species metric.....................................................................52 Native perennial species metric.....................................................52 Florida Wetland Condition Index..............................................................53 Cluster Analysis.........................................................................................56 Landscape Development Intensity Index and the Florida Wetland Condition Index..................................................................................................58 Human Disturbance Gradient and the Stream Condition Index............................61 4 DISCUSSION ........................................................................................................64 Describing Biological Integrity..............................................................................65 Richness, Evenness, and Diversity........................................................................65 Measuring Anthropogenic Activity.......................................................................67 Regionalization of the Florida Wetland Condition Index......................................69 Florida Wetland Condition Index Independent of Wetland Type.........................69 Limitations and Further Research..........................................................................71 Conclusions............................................................................................................71 APPENDIX A Standard Operating Procedures..............................................................................72 B Coefficient of Conservatism Scores.......................................................................80 C Summary Statistics.................................................................................................87 D Metric Scoring Criteria..........................................................................................88 REFERENCES..................................................................................................................89 iv

PAGE 5

LIST OF TABLES Table Page 1-1 Non-renewable energy use and LDI coefficients per land use used in the calculation of the LDI index..................................................................................10 1-2 Categorical scoring criteria used to calculate the Human Disturbance Gradient (HDG).....................................................................................................12 2-1 Site characterization for 24 freshwater forested wetlands.....................................16 3-1 Landscape Development Intensity (LDI) index scores for 10 forested strands using 1995 and 2000 land use coverages...................................................34 3-2 Landscape Development Intensity (LDI) index scores for 14 forested floodplain wetlands using 1995 and 2000 land use coverages..............................35 3-3 Human Disturbance Gradient (HDG) for 13 storet stations, which correspond with the forested floodplains sampled................................................39 3-4 Water quality (chemical and physical parameters) for 13 storet stations, which correspond with the forested floodplains sampled......................................41 3-5 Richness, evenness, and diversity of the macrophyte assemblage among a priori land use categories....................................................................................43 3-6 Richness, evenness, and diversity of the macrophyte assemblage between low (LDI < 2.0) and high (LDI 2.0) LDI groups................................................43 3-7 Macrophyte community composition similarity among Florida wetland regions (Lane 2000) and bioregions (Griffith et al. 1994) with MRPP.................44 3-8 Spearmans correlation coefficients for the macrophyte metrics and FWCI with LDI_F/wo_200m................................................................................46 3-9 Comparisons among five macrophyte metrics and the FWCI between low (LDI < 2.0) and high (LDI 2.0) LDI groups (LDI_F/wo_200m)................47 3-10 Tolerant indicator species for forested strand and floodplain wetlands................48 3-11 Sensitive indicator species for forested strand and floodplain wetlands...............50 3-12 Exotic species identified at 24 forested strand and floodplain wetlands...............54 3-13 FWCI scores and LDI values for wetland clusters based on macrophyte community composition.........................................................................................59 3-14 Correlations of metrics and FWCI scores with 20 variations of the LDI index....60 v

PAGE 6

3-15 Forested Wetland Condition Index (FWCI), Landscape Development Intensity Index (LDI_F/wo_200m), Human Disturbance Gradient (HDG), and Stream Condition Index (SCI) data available for 13 forested floodplain wetlands................................................................................................62 3-16 Correlations among four measures of ecosystem condition or anthropogenic activity, including the Human Disturbance Gradient (HDG), Stream Condition Index (SCI), Landscape Development Intensity Index (LDI_F/wo_200m), and the Florida Wetland Condition Index (FWCI).....63 4-1 The five metrics of the preliminary Florida Wetland Condition Index for freshwater forested strand and floodplain wetlands based on the macrophyte species assemblage.................................................................................................65 vi

PAGE 7

LIST OF FIGURES Figure Page 1-1 Photograph showing the interior of a freshwater forested strand wetland in Osceola County, Florida......................................................................................3 1-2 Photograph showing the interior of a freshwater forested floodplain wetland in Polk County, Florida..............................................................................4 2-1 Florida wetland regions defined by climatic and physical variables (solid line; Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994)..................14 2-2 Study site location of 24 forested wetlands in Florida...........................................15 2-3 Boundaries of the 100 and 200 m buffers drawn around the downstream transect for the LDI_T calculation at Forested Floodplain 3 (FF3).......................18 2-4 Boundaries of 100 and 200 m buffers around a wetland feature and designated land use................................................................................................19 2-5 Boundary of the watershed buffer around Forested Floodplain 3 (FF3) for the LDI_WS calculation...................................................................................20 2-6 Idealized transect layout for forested strand wetlands...........................................22 2-7 Idealized transect layout for forested floodplain wetlands....................................25 3-1 Comparison between the LDI calculations including (1995_LDI_T/w_100m) and excluding (1995 LDI_T/wo_100m) wetland area...........................................36 3-2 Comparison between LDI calculations at the transect (1995_LDI_T/wo_200m) and feature (1995 LDI_F/wo_200m) scale..................3 6 3-3 Comparison between LDI calculations at the watershed scale using equal weighting for all area within a watershed (1995 LDI_WS_ED/wo) and a distance weighted approach using linear weighting (1995 LDI_WS_DW_lin)....38 3-4 Comparison between 1995 and 2000 LDI calculations at the feature scale (without the wetland area) for 13 forested wetlands..............................................38 3-5 NMDS ordination bi-plot of 24 sample wetlands in macrophyte species space....45 3-6 The proportion of tolerant indicator species at wetlands increased with increasing development intensity...........................................................................49 3-7 The proportion sensitive indicator species at wetlands decreased with increasing development intensity...........................................................................5 1 3-8 FQAI scores decreased with increasing landscape development intensity............51 3-9 The proportion of exotic species at a wetland increased with increasing development intensity............................................................................................52 vii

PAGE 8

3-10 The proportion of native perennial species decreased with increasing development intensity (LDI)..................................................................................55 3-11 Forested Wetland Condition Index (FWCI) scores decreased with increasing development intensity (LDI)................................................................55 3-12 Change in average species p-value from the randomized Monte Carlo tests at each step in clustering........................................................................................57 3-13 Change in the number of significant indicator species from the indicator species analysis performed at each step in clustering............................................57 3-14 FWCI scores for three wetland clusters based on macrophyte community composition............................................................................................................58 viii

PAGE 9

EXECUTIVE SUMMARY PILOT STUDY THE FLORIDA WETLAND CONDITION INDEX (FWCI): PRELIMINARY DEVELOPMENT OF BIOLOGICAL INDICATORS FOR FORESTED STRAND AND FLOODPLAIN WETLANDS Over 30 years ago, the federal Water Pollution and Control Act obliged states to protect and restore the chemical, physical, and biological integrity of waters, and charged states with establishing water quality standards for all waters within state boundaries including wetlands. Criteria for defining water quality could be narrative or numeric, and it could be addressed through chemical, physical, or biological standards. Initially, states used chemical and physical criteria (testing waters for chemical concentrations or physical conditions that exceeded criteria), assuming losses in ecosystem integrity if the criteria were exceeded (Danielson 1998). The United States Environmental Protection Agency (USEPA) recognized the potential of biological criteria to assess water quality standards and in the late 1980s required states to use biological indicators to accomplish the goals of the Clean Water Act (USEPA 1990). In effect, biological assessment has evolved into one of the standard monitoring tools of water resource-protection agencies over the past 2 decades (Gerristen et al. 2000). Such biological assessment programs have been created for lakes and streams throughout the United States ( Barbour et al. 1996a ; Karr and Chu 1999 ; Gerristen et al. 2000), and more recently efforts to assess wetland condition have been initiated (Mack 2001; USEPA 2002). Within Florida, biological indices have been created based on macroinvertebrate community composition for streams (Barbour et al. 1996a; Fore 2004), lakes (Gerristen and White 1997), and isolated depressional freshwater herbaceous (Lane et al. 2003) and forested (Reiss and Brown 2005) wetlands. Biological indices have also been created based on the community composition of the diatom and macrophyte assemblages for Florida freshwater herbaceous and forested wetlands (Lane et al. 2003; Reiss and Brown 2005). The primary objective of this research was to develop a preliminary Florida Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland study sites were sought in various a priori designated land use categories that included natural, agricultural, and urban land uses. An independent measure of anthropogenic activity in the landscape was calculated for each wetland using the Landscape Development Intensity index (LDI) (Brown and Vivas 2005). The contribution of this research to our understanding of changes in the macrophyte community composition of forested strand and floodplain wetlands in relation to different anthropogenic activities in the surrounding landscape can be summarized in five main points: 1. Five macrophyte based metrics including proportion tolerant indicator species, proportion sensitive indicator species, Floristic Quality Assessment Index (FQAI) score, proportion exotic species, and proportion native perennial species, were useful biological indicators for defining biological integrity for forested strand and floodplain wetland vegetation; ix

PAGE 10

2. Vegetation richness, evenness, and diversity were not sensitive to a priori land use categories or development intensities in the surrounding landscape for forested strand and floodplain wetlands; 3. The Landscape Development Intensity (LDI) index was a useful tool correlating with the measured biological condition of vegetation for forested strand and floodplain wetlands; 4. Regional species lists for metrics would enhance the forested strand and floodplain Florida Wetland Condition Index (FWCI); 5. An FWCI with a set of core metrics could be developed for Florida freshwater wetlands, which includes separate species lists for indicator species by wetland type and ecoregions and separate Floristic Quality Assessment Index (FQAI) scores for species by wetland type. The FWCI provided a quantitative measure of the biological integrity of forested strand and floodplain wetlands in Florida. Comprised of five metrics, the FWCI was developed based on the community composition of the macrophyte species assemblages. Metrics were selected for inclusion in the FWCI based on the correlation (nonparametric Spearmans correlation coefficient) of each metric with a quantitative gradient of Landscape Development Intensity (LDI); based on a metrics visually distinguishable correlation with LDI in a scatter plot; and based on a statistical difference of metric values between low and high LDI groups (Mann-Whitney U-test). The FWCI was composed of individual metrics, which were scaled and added together, creating the preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest score of 50 reflecting the highest biological integrity and the lowest score of zero reflecting a lack of biological integrity or no similarity to the reference wetland condition. The five macrophyte metrics that met the three selection criteria (Spearmans correlation coefficient (|r| 0.50, p < 0.05), visually distinguishable scatter plots, and Mann-Whitney U-test between LDI groups (p<0.10)) were proportion tolerant indicator species; proportion sensitive indicator species; Floristic Quality Assessment Index (FQAI) score; proportion exotic species; and proportion native perennial species. Forested Strand and Floodplain FWCI Metrics for Wetland Vegetation 1. Proportion Tolerant Indicator Species 2. Proportion Sensitive Indicator Species 3. Floristic Quality Assessment Index (FQAI) Score 4. Proportion Exotic Species 5. Proportion Native Perennial Species The variable sensitivities of three different independently derived indices compared to the forested strand and floodplain FWCI, including the Landscape Development Intensity index (LDI; Lane et al. 2003; Brown and Vivas 2005), the Human Disturbance Gradient (HDG; Fore 2004), and the Stream Condition Index (SCI; Fore 2004), suggest that multiple measures of biological integrity may be more effective at describing ecosystem wide biological integrity than any single measure based on an individual species assemblage or surrounding land use activity. However, the strong x

PAGE 11

correlations among FWCI, LDI, and HDG (Spearmans correlation coefficient |r| 0.58, p < 0.05), and lack of correlations of SCI with both FWCI and LDI, suggest that in-stream macroinvertebrate based measures of biological condition and surrounding forested wetland macrophyte based measures of biological condition did not respond in a consistent manner to changes in anthropogenic activity. Using both the in-stream macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte FWCI biological assessment methods may provide a more complete picture of the overall condition of a wetland and associated stream at a particular spatial location. While agreement in the ranking of the biological condition of study wetlands using the FWCI and SCI was anticipated, discrepancies among the ranking from the different assemblages may provide great insight into biological condition as different species assemblages respond to changes in anthropogenic activities and the associated changes in inflows (e.g. nutrient enrichment) over different time scales. Additionally, use of the forested strand and floodplain FWCI may lead to specific conclusions as to the biological condition of local or nearby anthropogenic activity, while use of the SCI may enhance understanding of larger watershed scale influences from anthropogenic activity (i.e. due to the convergence of water within the watershed associated with stream flow). The quantitative score of biological integrity established through the FWCI can be used as an objective, quantitative means of comparing changes in macrophyte community composition for wetlands, including those impacted by varying degrees of anthropogenic influence. While the forested strand and floodplain FWCI for flowing water systems can not be used to predict changes in the physical and chemical parameters of a wetland, its strength lies in providing an overview of biological integrity through the integration of changes in macrophyte community composition from cumulative effects. xi

PAGE 12

CHAPTER 1 INTRODUCTION AND OVERVIEW Assessment techniques for categorizing ecosystem condition have been established for many Florida ecosystems, including freshwater lakes (Lake Condition Index, LCI: Gerristen and White 1997), streams (BioRecon and Stream Condition Index, SCI: Fore 2004), and depressional freshwater wetlands (Florida Wetland Condition Index, Lane et al. 2003; Reiss and Brown 2005). This research furthers the development of the Florida Wetland Condition Index (FWCI), establishing preliminary metrics with species lists specific to forested strand and floodplain wetlands, referred to as flowing water systems. The overall goal of the FWCI is to use changes in community composition to characterize the biological condition of wetland ecosystems. Wetland condition is scored on a numeric scale developed from the addition of scores from individual metrics. A metric is defined as a biological attribute having a consistent and predictable response to anthropogenic activities (Karr and Chu 1997). Each metric represents an indication of biological integrity, or a signal of ecosystem condition, based on a change in community composition from the reference standard condition. The reference standard condition is defined as the condition of wetlands surrounded by undeveloped landscapes and without apparent human induced alterations. By designating a measure of ecosystem condition we refer to what others have described as ecosystem integrity, defined by Karr and Dudley (1981) as the ability of an aquatic ecosystem to support and maintain a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to that of the natural habitats of the region. Forested Strand and Floodplain Wetlands Wetlands have been categorized in many different ways based on any number of community attributes including, but not limited to, dominant vegetation, hydrology, soil type, and location in the landscape (Mitsch and Gosselink 1993; Keddy 2000; Kent 2000). One of the most widely recognized classification systems in North America is that by Cowardin et al. (1979). Our study focused on what Cowardin et al. (1979) categorize palustrine wetlands, commonly described as freshwater marshes and swamps. More specifically, palustrine wetlands are defined as nontidal wetland ecosystems with trees, shrubs, persistent emergents, or emergent mosses or lichens as the dominant vegetation type, and tidal wetlands with these vegetation types with ocean-derived salinity levels below 0.5 (Cowardin et al. 1979). Palustrine wetlands occur throughout the landscape as small, shallow, permanent or intermittent water bodies; shoreward of lakes, river channels, or estuaries; on river floodplains; in isolated catchments; on slopes; or as islands in lakes or rivers (Cowardin et al. 1979). The category of palustrine wetlands includes eight classes: aquatic bed, emergent, forested, moss-lichen, rock bottom, scrub-shrub, unconsolidated bottom, and unconsolidated shore. Wetlands for this study were primarily in the forested wetland class, which includes wetlands in all water regimes, except subtidal wetlands, that are characterized by 1

PAGE 13

2 woody vegetation 6m tall or greater. The structure of palustrine forested wetlands typically includes an overstory of trees with an understory of young trees and shrubs and an understory of herbaceous species (Cowardin et al. 1979). The class of palustrine forested wetlands includes six subclasses and dominance types, including broad-leaved deciduous, needle-leaved deciduous, broad-leaved evergreen, needle-leaved evergreen, dead, and indeterminate deciduous. Four categories of water modifiers are used to describe the palustrine forested wetlands in this study (as categorized on the National Wetlands Inventory GIS coverage available from the Florida Geographic Data Library at http://www.fgdl.org ) including temporarily flooded, seasonally flooded, semipermanently flooded, and permanently flooded. During the growing season, temporarily flooded wetlands have surface water present for brief periods with a water table typically well below the soil surface. Vegetation in temporarily flooded wetlands consists of facultative species, including those that grow in both uplands and wetlands. Early in the growing season seasonally flooded wetlands have surface water standing for extended periods most often without surface water late in the season but with a water table near the soil surface. Similarly, semipermanently flooded forested wetlands generally have standing surface water throughout the growing season and a water table at or near the soil surface when not flooded. At the flood extreme, permanently flooded wetlands have standing surface water throughout the year with a vegetation community of obligate wetland species (Cowardin et al. 1979). Within the text Ecosystems of Florida (Myers and Ewel, eds. 1990), Ewel (1990) describes approximately 10 distinctive types of swamps. The freshwater forested strands in this study most closely resemble wetlands in the cypress pond and strand category. Wharton et al. (1976) described cypress strands as a diffuse freshwater stream flowing through a shallow depression on a greatly sloping plain. While Mitsch and Gosselink (1993) suggest that cypress strands are found primarily in south Florida, Ewel (1990) notes that cypress strands are common throughout Florida and are found where water flow is sufficient to create a depression channel in areas with little slope but where actual flow is seldom observed. The definition of forested strands for this study broadly encompasses all of these definitions with the primary distinction of forested strands including evidence of channelized flow (Figure 1-1), though actual flow was rarely observed at any of the strands during the 2003 growing season during the period of sampling. The forested floodplain wetlands in this study most closely resemble the river swamp category in Ecosystems of Florida (Myers and Ewel, eds. 1990); however the floodplain forests in this study were associated with smaller river systems than those typically characterized by Ewel (1990). Forested floodplain wetlands in this study were associated with low order streams and rivers and were not associated with the main channels of the largest river systems in Florida (ex. Apalachicola, Suwannee, etc.). The forested floodplain wetlands in this study can also be categorized as riparian wetlands such named for the influence on the wetland environment by the adjacent stream or river system. Mitsch and Gosselink (1993) note vast differences among riparian wetlands, with the common link being the interconnection between the riparian zone, the river or stream, and the adjacent upland environment. Riparian wetlands in the southeastern United States are characterized by low-lying, low slope, and broad floodplain areas with seasonally pulsing hydrologic influences on well developed soils (Mitsch and Gosselink

PAGE 14

3 Figure 1-1. Photograph showing the interior of a freshwater forested strand wetland in Osceola County, Florida. The dark organic layer in the center of the photo shows evidence of flowing water during times of high water. 1993). The floodplain wetlands in this study are similar to the strands, however standing water was always observed in the channelized stream (Figure 1-2). The average hydroperiod (the seasonal pattern and length of saturated soils or standing water level during a year (Ewel 1990; Mitsch and Gosselink 1993)) varies among strand and floodplain wetlands, with strands having a moderate length hydroperiod with saturated soils or standing water for six to nine months a year and floodplain forests having a short length hydroperiod with generally less than six months of saturated soils or standing water during a year (Ewel 1990). Fire frequency in forested strand and floodplain wetlands ranges from moderate frequency (approximately one per 20 years) for strands to low frequency (approximately one per 100 years) for floodplain wetlands (Ewel 1990). Strands and floodplains also differ in their organic matter accumulation depths, with strands having high organic matter accumulation with an organic soil layer greater than 1 m deep and floodplain wetlands having low organic matter accumulation with an organic soil layer less than 1 m deep (Ewel 1990). Additionally, strand and floodplain wetlands have different primary water sources with strands receiving most water from shallow groundwater sources and the main source of water for floodplain wetlands from surface water originating from the associated stream or river (Ewel 1990). Despite these differences in fire frequency, organic matter accumulation, and water source, the species composition is similar among forested strand and floodplain wetlands. Common shared tree species in strand and floodplain wetlands include Acer

PAGE 15

4 Figure 1-2. Photograph showing the interior of a freshwater forested floodplain wetland in Polk County, Florida. The channelized stream is visible in the center of the photo showing the presence of a permanently flooded and flowing stream adjacent to the floodplain wetland. rubrum (red maple), Fraxinus caroliniana (water ash), Gordonia lasianthus (loblolly bay), Liquidambar styraciflua (sweetgum), Magnolia virginiana (sweet bay), Nyssa sylvatica (black gum), Persea palustris (swamp bay), Quercus laurifolia (swamp laurel oak), Sabal palmetto (cabbage palm), Salix caroliniana (coastal plain willow), and Taxodium distichum (baldcypress). Additional tree species common to strands include Annona glabra (pond apple), Pinus elliottii (slash pine), and Pinus palustris (longleaf pine). The tree stratum of floodplain wetlands has greater species richness with additional common tree species including Alnus serrulata (hazel alder), Betula nigra (river birch), Carpinus caroliniana (American hornbeam), Carya aquatica (water hickory), Carya glabra (pignut hickory), Celtis laevigata (hackberry), Chamaecyparis thyoides (Atlantic white cedar), Diospyros virginiana (persimmon), Fraxinus pennsylvanica (green ash, red ash), Fraxinus profunda (pumpkin ash), Gleditsia aquatica (water locust), Magnolia grandiflora (southern magnolia), Nyssa aquatica (water tupelo), Pinus glabra (spruce pine), Pinus taeda (loblolly pine), Planera aquatica (planer tree), Platanus occidentalis (American sycamore), Quercus lyrata (overcup oak), Quercus michauxii (basket oak, swamp chestnut oak), Quercus nigra (water oak), Quercus virginiana (live oak), Rhapidophyllum hystrix (needle palm), Sabal minor (bluestem, dwarf palmetto), Salix nigra (black willow), and Ulmus americana (American elm).

PAGE 16

5 Forested strand and floodplain wetlands also share a number of species in the shrub stratum including Cephalantus occidentalis (buttonbush), Clethra alnifolia (sweet pepperbush), Cliftonia monophylla (black titi), Cyrilla racemiflora (titi), Ilex cassine (dahoon holly), Itea virginica (Virginia willow), Lyonia lucida (fetterbush), Myrica cerifera (wax myrtle), and Rubus argutus (blackberry). Additional common shrub species in forested strands include Chrysobalanus icaco (coco plum), Ilex glabra (gallberry), Leucothoe racemosa (fetterbush), Myrica heterophylla (northern bayberry), Myrsine guianensis (myrsine), Psychotria sulzneri (wild coffee), Psychotria undata (wild coffee), and Vaccinium arboretum (sparkleberry). Other common shrub species in river swamps include Aronia arbutifolia (red chokeberry), Crataegus marshallii (parsley haw), Ilex decidua (possum haw), Ilex vomitoria (yaupon), Leucothoe axillaries (dog-hobble), Rhododendron viscosum (swamp honeysuckle), Rubus betulifolius (blackberry), Sambucus canadensis (elderberry), Sebastiana fruticosa (Sebastian bush), Viburnum nudum (swamp haw), and Viburnum obovatum (small viburnum, black haw). The species composition of woody vines in strand and floodplain wetlands vary a great deal according to Ewel (1990), as there is only one shared species, Smilax laurifolia (bamboo-vine, catbrier). Common woody vine species in strands include Ampelopsis arborea (pepper vine), Ficus aurea (strangler fig), Ficus citrifolia (wild banyan tree), Vitis aestivalis (summer grape), and Vitis shuttleworthii (calusa grape); and common woody vine species in floodplain wetlands include Ampelopsis arborea (pepper vine), Aster carolinianus (climbing aster), Smilax walteri (coral greenbrier), Toxicodendron radicans (poison ivy), and Vitis rotundifolia (muscadine grape). Species lists for common tree, shrub, and woody vine species were adopted from Ewel (1990). Forested strand and floodplain wetlands provide important habitat for wildlife such as invertebrates, amphibians, reptiles, birds, and mammals. Benthic invertebrates form the base of the forested wetland food chain, and water quality is strongly related to the diversity of the benthic macroinvertebrate community (Ewel 1990). Strands and floodplain wetlands provide valuable habitat for bird and mammal species characterized by low vegetation density and high cavity density. Though these wetlands differ somewhat in their relative contributions to bird and mammal habitat as strands have low canopy insect production, low production of edible fruits and seeds, and high presence of water, whereas forested floodplain wetlands have high canopy insect production, high production of edible fruits and seeds, and low presence of water (Ewel 1990). Historical Perspective Over 30 years ago, the Water Pollution and Control Act (later referred to as the Clean Water Act, 1972) required states to restore and maintain the chemical, physical, and biological integrity of the Nations waters (USEPA 1990). This legislation included establishing water quality standards for all waters within state boundaries, including wetlands. Such water quality criteria could be qualitative or quantitative, and it could be addressed through chemical, physical, or biological standards. Initially, states used chemical and physical criteria (testing waters for chemical concentrations or physical conditions that exceeded known standards), assuming losses in ecosystem integrity if these standards were exceeded (Danielson 1998).

PAGE 17

6 Several shortcomings have been noted when deriving ecosystem integrity based on exceeding established limits for chemical and physical parameters. Such criteria have been considered incomplete in their ability to reflect more than the temporal concentration of substances within a water body (Karr 1993). For instance, the use of toxicity parameters for determining ecosystem integrity may falsely indicate high ecosystem integrity when a single toxicity parameter went overlooked. This same water body could have elevated levels of other toxins or metals that went untested, or be physically altered so that it has lost functions typically associated with a fully functioning water body (Karr and Chu 1997). Furthermore, chemical and physical sampling may not occur during specific loading events and may therefore incompletely describe the ecological condition of the system. Adams (2002) points out that other environmental factors such as sedimentation, alterations to habitat, varying temperature and oxygen levels, and changes in ecological aspects like food availability and predator-prey relationships are not reflected with chemical criteria alone. James and Kleinow (1994) note that different organisms respond in different ways to the amount, persistence, and exposure of chemical compounds otherwise foreign to an organism; and single-valued chemical and physical criteria of water quality may overlook important biological implications. Alternatively, biological indicators integrate the spatial and temporal effects of the environment on resident organisms, and are suitable for assessing the possible effects of multifaceted changes in ecosystems (Adams 2002). Karr and Chu (1997) and Adams (2002) note that biological indicators signal changes in the environment that might otherwise be overlooked or underestimated by methods that depend on chemical criteria alone. Organisms have an intricate relationship with their environment, reflecting current and cumulative ecosystem condition (Karr 1981). The presence of biological organisms reveals chemical exposure, expressing changes in the physical, chemical, and biological components of the ecosystem through changes in community composition (Adams 2002). The United States Environmental Protection Agency (USEPA) recognized the potential of biological criteria to assess water quality standards and in the late 1980s required states to use biological indicators to accomplish the goals of the Clean Water Act (USEPA 1990). In effect, biological assessment has evolved into one of the standard monitoring tools of water resource protection agencies over the last two decades (Gerristen et al. 2000). Biological criteria and monitoring programs through the USEPA have been created for lakes and streams throughout the United States (Barbour et al. 1996a; Karr and Chu 1999; Gerristen et al. 2000), and more recently efforts to assess wetland condition have been initiated (USEPA 2002). Indicators of Biological Integrity Biological monitoring to assess ecosystem condition has been applied widely in ecological research. The primary aim of biological monitoring is to detect changes in abundance, structure, and diversity of target species assemblages. One trend in biological monitoring has led to the development of indices of biotic integrity (referred to as IBIs), for different species assemblages including diatoms (Fore and Grafe 2002; Fore 2004); macrophytes (Galatowitsch et al. 1999a; Gernes and Helgen 1999; Mack 2001; Lane

PAGE 18

7 2003); macroinvertebrates (Kerans and Karr 1994; Barbour et al. 1996b); amphibians (Micacchion 2004); fish (Schulz et al. 1999); and birds (OConnell et al. 1998). Perhaps the most common species assemblage chosen for use in the development of IBIs is the macroinvertebrate assemblage, because many of the macroinvertebrate species rely entirely on the conditions of their aquatic environment for habitat, food, and reproductive activities. In Florida there are currently three biological indices that use the community composition of the macroinvertebrate assemblage to detect changes in biological integrity including the Lake Condition Index (LCI, Gerristen and White 1997), Stream Condition Index (SCI, Fore 2004), and the Florida Wetland Condition Index for depressional freshwater wetlands (FWCI, Lane et al. 2003; Reiss and Brown 2005). However, the forested strand and floodplain systems targeted in this study have varying hydrologic regimes from temporarily to permanently flooded, complicating macroinvertebrate collection due to variable hydrologic conditions. As such, we have chosen the macrophyte species assemblage for use in the preliminary forested strand and floodplain FWCI. Wetland macrophytes are defined as aquatic emergent, submergent, or floating plants growing in or near water (USEPA 1998); and are described as distinguishing landscape features. The spatial distribution of macrophytes in the landscape occurs according to a multitude of factors, including hydroperiod, water chemistry, and substrate type, as well as other factors such as available seed source and climate. Fennessy et al. ( 2001 ) state that the community composition of wetland macrophytes typifies the physical, chemical, and biological wetland dynamic in time and space. Macrophytes play a vital role in supporting the structure and function of wetlands by providing food and habitat for other assemblages including algae, macroinvertebrates, fish, amphibians, reptiles, birds, and mammals; and macrophyte populations can be used as a diagnostic tool to assess other aspects of the wetland environment. Crowder and Painter (1991) state that a lack of macrophytes where they are otherwise expected to grow suggests reduced wildlife populations from lack of food or cover and/or water quality concerns such as toxic chemical constituents, increased turbidity, or increased salinity. In contrast, an overgrowth of particular macrophytes may signify increased nutrient loading (USEPA 1998). Many advantages of studying macrophytes as indicators of wetland condition have been noted, including their large, obvious size; ease of identification, to at least some useful taxonomic level; known response to toxicity tests; and general lack of ability to move to avoid unfavorable conditions (Danielson 1998; Cronk and Fennessy 2001). Additionally, macrophytes readily respond to changes in nutrient, light, toxic contaminant, metal, herbicide, turbidity, water, and salt levels. They can also be sampled in the field with transects, or remotely from aerial photography; and well-established field methods of sampling macrophytes exist (USEPA 2003). Furthermore, the USEPA (2003) states additional advantages of using the macrophyte assemblage, including that they do not require laboratory analysis, can easily be used for calculating simple abundance metrics, and are superb integrators of environmental condition. In general, macrophytes represent a useful assemblage for describing wetland condition (Mack 2001). Schindler (1987) alleges that macrophytes can provide a more integrated picture of wetland function than measures such as nutrient cycling, productivity, decomposition, or chemical and physical composition.

PAGE 19

8 There are however some noted shortcomings of using macrophytes as biological indicators. These include the potential delay in response time for perennial shrub and tree species, difficulty identifying taxa to the species level in certain seasons and for some genera, uneven herbivory patterns, and varied pest-management practices (Cronk and Fennessy 2001). Despite these limitations, macrophytes have provided strong signals of anthropogenic influence (USEPA 2003). In fact, many states have begun using macrophytes in their wetland biological assessment programs, including, Minnesota (Galatowitsch et al. 1999a; Gernes and Helgen 1999), Montana (Apfelbeck 2000), North Dakota (Mushet et al. 2002), Ohio (Mack 2001), and Florida (Lane et al. 2003; Reiss and Brown 2005). Quantifying Anthropogenic Influence Wetlands occupy a large portion of the Florida landscape. An estimate from the 1780s reported 8,225,000 ha of wetlands in Florida (Dahl 2000). By the mid-1980s, the National Wetlands Inventory estimated Florida had 4,467,000 ha of wetlands remaining, translating into a loss in Florida of roughly 45% of the pre-1780s wetland area (Mitsch and Gosselink 1993; Dahl 2000). Throughout the continental United States, similar trends were apparent, with a drastic decline in the surface area of wetlands. More specifically, Dahl ( 2000 ) reported that 98% of all wetland losses throughout the continental United States from 1986 to 1997 were losses to freshwater wetlands. Of the remaining freshwater wetlands, 40% were adjacent to agricultural lands and therefore potentially affected by land use practices such as herbicide and pesticide application, irrigation, livestock watering and wastes, soil erosion, and deposition. An additional 17% were adjacent to urban or rural development. Freshwater non-tidal wetlands experienced the greatest development pressure just inland from coastlines as the demand for housing, transportation infrastructure, and commercial and recreational facilities increased ( Dahl 2000 ). These changes in land use were proportionally more widespread in Florida than much of the continental United States due to the remarkable length of coastline along both the Atlantic Ocean and Gulf of Mexico coasts of Florida. Agricultural and urban development activities influence an array of changes to the physical, chemical, and biological characteristics in nearby ecosystems. There have been numerous attempts at quantifying anthropogenic influence based on quantitative indices, for example, the Wetland Rapid Assessment Procedure (WRAP; Miller and Boyd 1999), the Minnesota disturbance index (Gernes and Helgen 1999), the Human Disturbance Gradient (HDG; Fore 2004), and the Landscape Development Intensity (LDI) index (Brown and Vivas 2005). Our study incorporates the LDI index as a measure of anthropogenic influence. Additionally, for thirteen floodplain wetlands, HDG scores have been obtained from previous studies (Fore 2004; Florida Department of Environmental Protection Geographic Information Systems map layers). Landscape Development Intensity Index The LDI index has been used as a gauge of human activity based on a development intensity measure derived from nonrenewable energy use in the surrounding landscape. The underlying concept behind calculating the LDI (quantifying the nonrenewable energy use per unit area in the surrounding landscape) stems from earlier

PAGE 20

9 works by Odum (1995), who pioneered emergy analysis for environmental accounting. [Emergy is an environmental accounting term referring to expressing energy use in solar equivalents (Odum 1995).] Brown and Ulgiati (2005) suggest that landscape condition, or ecosystem health, is strongly related to the surrounding intensity of human activity, and that ecological communities are affected by the direct, secondary, and cumulative impacts of activities in the surrounding landscape. Healthy ecosystems are defined as those with integrity and sustainability, which correlate to limited development in the surrounding landscape and the maintenance of ecosystem structure and function, even when stressors (e.g. flooding, drought, etc.) are present (Brown and Ulgiati 2005). The LDI scale encompasses a gradient from completely natural to highly developed land use intensity, and is calculated based on the percent of the area in a particular land use within the designated area surrounding the wetland (ex. 100 meters, 200 meters, etc.) multiplied times the LDI coefficient (Table 1-1), which is defined by the amount of nonrenewable energy use for a given land use (Brown and Vivas 2005). The LDI coefficient does not account for any individual causal agent directly, but instead, may represent the combined effects of air and water pollutants, physical damage, changes in the suite of environmental conditions (ex. groundwater levels, increased flooding), or a combination of such factors, all of which enter the natural ecological system from the surrounding developed landscape. Wetlands surrounded by more intense activities such as highways and multi-family residential land uses receive higher LDI index values, as the highest LDI coefficient of 10.0 is assigned to the urban land use category of Central Business District. Undeveloped land uses such as wetlands, lakes, and upland forests are assigned an LDI coefficient of 1.0, the lowest possible value, based on no use of nonrenewable energy in these ecosystems. Human Disturbance Gradient The Human Disturbance Gradient (HDG) is a quantitative measure used to assess the level of human disturbance to an ecosystem based on the cumulative score of four independent measures of the environment including ammonia concentration in the water (as mg N/L), hydrologic index, habitat assessment, and LDI for the buffer which included an area of 100 m on each side of the stream and 10 km upstream of the sampling point ( Fore 2004 ). The first measure included in the HDG, ammonia concentration, is included as a summary variable defining water quality because of its consistent correlation with other water quality parameters (i.e. total phosphate) and its availability in the dataset ( Fore 2004 ). Additionally, increases in ammonia concentration are thought to be evidenced from both agricultural (e.g., fertilizer and farming practices) and urban activities, whereas changes in total phosphorus concentrations are thought to be solely associated with agricultural operations ( Fore 2004 ), so using ammonia concentration accounts for changes in both agricultural and urban land use activities. The second measure included in the HDG is an estimate of hydrologic condition of the stream site, which is scored on-site at the discretion of the biologist conducting the sampling effort. Scores for the hydrologic condition range from 1-10, with a score of 1-2 Excellent representing a natural, undisturbed system with few impervious surfaces, high connectivity with ground water and surface features, and a natural flow regime ( Fore 2004 ). Intermediate categories include 3-4 Good, 5-6 Moderate, and 7-8 Poor. The final category, 9-10 Very Poor, is reserved for those systems with a flow regime entirely

PAGE 21

10 Table 1-1. Non-renewable energy use and LDI coefficients per land use used in the calculation of the LDI index. Land Use Nonrenewable Energy Use (E14 solar equivalent joules/ha/yr) LDI Coefficient Natural Land / Open Water 0.0 1.00 Pine Plantation 5.1 1.58 Low Intensity Open Space / Recreational 6.7 1.85 Unimproved Pastureland (with livestock) 8.3 2.06 Improved pasture (no livestock) 19.5 2.89 Low Intensity Pasture (with livestock) 36.9 3.51 High Intensity Pasture (with livestock) 51.5 3.83 Citrus 65.4 4.06 Medium Intensity Open Space / Recreational 67.3 4.09 Row crops 117.1 4.63 High Intensity Agriculture (dairy farm) 201.0 5.15 Single Family Residential (Low-density) 1077.0 6.79 Recreational / Open Space (High-intensity) 1230.0 6.92 Single Family Residential (Med-density) 2461.5 7.59 Low Intensity Transportation 3080.0 7.81 Single Family Residential (High-density) 3729.5 7.99 Low Intensity commercial (Comm Strip) 3758.0 8.00 Institutional 4042.2 8.07 Highway (4 lane) 5020.0 8.28 Industrial 5210.6 8.32 Multi-family residential (Low rise) 7391.5 8.66 High intensity commercial (Mall) 12661.0 9.18 Multi-family residential (High rise) 1285.0 9.19 Central Business District (Avg 2 stories) 16150.3 9.42 Electric Power Facility 29401.3 10.00 controlled by human modification, with a flashy hydrograph and extreme alteration of the natural ecosystem ( Fore 2004 ). Habitat assessment (or habitat condition index as termed in Fore 2004 ) is the third measure included in the HDG. The Florida Department of Environmental Protection (FDEP) has developed Standard Operating Procedures (SOPs) for river and stream habitat assessments (FDEP-SOP-001/01: Form FD 9000-5 June 1, 2001 available from the Bureau of Laboratories at http://www.dep.state.fl.us/labs/library/forms.htm ). The habitat assessment is separated into four primary habitat components: substrate diversity (number of diverse, productive habitats), substrate availability (percent of stream reach composed of productive habitat), water velocity (based on the maximum observed

PAGE 22

11 velocity, where higher velocities receive higher scores), and habitat smothering (as the percent of the stream reach covered by sand or silt accumulation). The four secondary habitat parameters include artificial channelization (as an assessment of anthropogenic modification of the stream reach), bank stability (evidence of bank stability/instability from erosion or bank failure), riparian buffer zone width (estimate of the width of the vegetation on the least buffered side), and riparian zone vegetation quality (estimate of community composition and structure). These eight habitat parameters are scored by the field biologist on a scale of 1-20, with 20 representing the highest habitat quality. Three of the secondary habitat components (bank stability, riparian buffer zone width, and riparian zone vegetation quality) are scored separately on a scale of 1-10 (with 10 representing the highest habitat quality) for both the right and left banks (when added together these three categories are still given equal weighting, with half of the score reflecting the condition of each bank). Scores for each of the eight parameters are summed, and a stream is then assigned a categorical score of optimal (134-160), suboptimal (91-123), marginal (54-80), or poor (11-43) based on the total score ( Fore 2004 ). The fourth measure of the HDG is an LDI score calculated within a 100 m buffer on each side of the stream for 10 km upstream from the sampling point (LDI_BF) ( Fore 2004 ). Each measure included in the HDG is categorically assigned a score based on the score for each measure ( Table 1-2 ). Ammonia concentration in the water (as mg N/L), habitat assessment, and LDI for the buffer are assigned scores of 0, 1, or 2; whereas the hydrologic index is assigned scores of 0, 1, 2, or 3. Overall HDG values potentially range from 0 (no detectable human induced disturbance) to 9 (extreme anthropogenic disturbance) ( Fore 2004 ). Original categorical scoring criteria for the HDG were established from correlations of the HDG categories (ammonia, hydrologic condition, habitat index, and LDI_BF) with the macroinvertebrate metric EPT (Ephemeroptera, Plecoptera, and Trichoptera) taxa richness. All HDG scores used in this analysis were taken directly from HDG scores provided by FDEP. Project Overview The community composition of the macrophyte assemblage was sampled in flowing wetland systems including forested strand (n = 10) and forested floodplain (n = 14) wetlands throughout Florida. Our primary goal was to develop a preliminary Florida Wetland Condition Index (FWCI) for freshwater flowing water wetlands, building upon the FWCI for isolated depressional herbaceous ( Lane et al. 2003 ) and forested ( Reiss and Brown 2005 ) freshwater wetlands. Our secondary goal was to determine the appropriate scale for LDI buffers for flowing water wetlands that would best capture the current wetland condition, based on correlations with the FWCI. Our third goal was to correlate measures of the HDG ( Fore 2002 ) and macroinvertebrate community data from the Florida SCI with our macrophyte FWCI and landscape LDI measures of ecosystem condition. Wetland study sites were sought in various landscape settings that included natural, agricultural, and urban land uses. The elongated forested wetlands associated with flowing waters were in many areas continuous for extended geographic distances, and the forested edges bordered a great variety of land use activities. Therefore, it was

PAGE 23

12 Table 1-2. Categorical scoring criteria used to calculate the Human Disturbance Gradient (HDG). The HDG is the sum of the categorical scores for the four individual measures ammonia concentration in the water (as mg N/L), hydrologic index, habitat assessment, and LDI for the buffer. The HDG range is from 0 (no detectable human induced disturbance) to 9 (extreme anthropogenic disturbance). HDG Categorical Score HDG Measure 0 1 2 3 Ammonia concentration (mg N/L) < 0.1 0.1 2.0 > 2.0 Hydrologic Condition < 6 6 7 8 9 10 Habitat Assessment > 65 50 65 < 50 LDI for the buffer < 2.0 2.0 3.5 > 3.5 difficult to ascertain the specific land use activity contributing most heavily or frequently to perturbations in community composition. While a priori land use categories were assigned (reference, agricultural, urban), wetlands were also divided into two groups representing low and high landscape development intensities.

PAGE 24

CHAPTER 2 METHODS Twenty-four freshwater forested wetlands, including 10 strand and 14 floodplain wetlands, were sampled during 2003. This chapter describes site selection, Landscape Development Intensity (LDI) index calculations, field-data collection, and methods of statistical analyses. Study Area Florida has been dissected into smaller geographic zones using a number of different methods, including bioregions ( Griffith et al. 1994 ) with 13 subecoregions (grouped into four ecoregions for small wadeable streams of panhandle, northeast, peninsula, and Everglades); lake ecoregions of Florida ( Griffith et al. 1997 ) with 47 subecoregions; and Florida wetland regions determined with a hydrologic model by Lane ( 2000 ), which included physical (surficial geology, soils, digital elevation model, slope) and climatic (precipitation, potential evapotranspiration, runoff, annual days of freezing) variables with four ecoregions (panhandle, north, central, and south). When boundaries for the bioregions of Griffith et al ( 1994 ) and wetland regions of Lane ( 2000 ) were compared, there were similarities among the panhandle and south/Everglades boundaries; however, the north/northeast and central/peninsula ecoregions differed in extent ( Figure 2-1 ). We adopted the four wetland regions developed by Lane ( 2000 ) to partition the state during site selection. However, Human Disturbance Gradient (HDG) and Stream Condition Index (SCI) scores available for the floodplain forests were designated according to the Florida bioregions ( Griffith et al. 1994 ); as such we use both wetland region and bioregion categories in our analyses. The four wetland regions defined by Lane ( 2000 ) are categorized as panhandle, north, central, and south. The panhandle wetland region is characterized by less human development than the other ecoregions. Streams in the panhandle wetland region typically run from north to south, discharging into the Gulf of Mexico. The primary ecosystems are scrub and high pine, temperate hardwood forests, pine flatwoods and dry prairies, and swamps ( Fernald and Purdum 1992 ). The major population centers of Tallahassee, Panama City, and Pensacola are included within the panhandle wetland region ( Davis 1967 ). The north wetland region has similar ecosystem types, primarily with pine flatwoods and dry prairies and less scrub and high pine, temperate hardwood forests, and swamps ( Fernald and Purdum 1992 ). Two major drainage features in the north wetland region include the Santa Fe and Suwannee River system and the lower St. Johns River, with Jacksonville being the major city situated along the northeastern Atlantic coast. The central wetland region has a distinguishing central ridge feature with a higher maximum elevation than the north and south wetland regions ( Fernald and Purdum 1992 ). The central wetland region is characterized by pine flatwoods and dry prairies with scrub and high pine along the ridge. There is less area in temperate hardwood forests, less area in swamps, and an increase in the area of inland lakes and freshwater marshes ( Fernald and Purdum 1992 ). The largest population centers include 13

PAGE 25

14 Figure 2-1. Florida wetland regions defined by climatic and physical variables (solid line; Lane 2000) and Florida bioregions (dashed line; Griffith et al. 1994). Orlando and Tampa. The south wetland region is unique with the nearly flat terrain of the Everglades (less than 2 m relief; Fernald and Purdum 1992). Much of the south is swamps and freshwater marsh, and few streams remain unaltered. Lake Okeechobee is in this wetland region as are the urban centers of Miami and Ft. Myers located on the densely populated east and west coasts, respectively. Site Selection During the 2003 growing season (May-July), 24 forested wetlands were sampled throughout the state of Florida (Figure 2-2). Random site selection was not feasible given the necessity of obtaining permission to access private lands and the non-random pattern of land development in Florida. The forested strand wetlands (n=10) were arranged spatially throughout Florida, so that the distribution was spread among three of the four Florida wetland regions (Lane 2000), including north (n=2), central (n=5), and south (n=3). No strands were sampled in the panhandle wetland region. The forested floodplain wetlands (n=14) were selected as a subset of systems that have been sampled using the SCI (Fore 2004). Many of the streams that had received a score of poor according to the SCI no longer had floodplain forest vegetation along the banks, but rather had mowed grass or paved banks. As such, these streams did not fit our criteria for

PAGE 26

15 Figure 2-2. Study site location of 24 forested wetlands in Florida (strands n=10; floodplains n=14). Wetland region boundaries follow Lane ( 2000 ). site selection, and where therefore excluded from sampling. Forested floodplain wetlands sampled were located in the panhandle (n=2), north (n=8), central (n=3), and south (n=1) wetland regions. A priori categories were used to define sample wetlands as reference, agricultural, and urban. The suite of reference wetlands included wetlands surrounded by intact native ecosystems. These wetlands were thought to represent the best possible wetland condition currently in Florida. These wetlands could be categorized as reference standard wetlands by the USEPA ( 1998a ), and have been described as wetlands considered to be the least altered that are reflective of characteristic levels of wetland function. Agricultural wetlands were defined as those currently surrounded by cattle pasture, rangeland, row crops, citrus, and silvicultural land uses. Urban wetlands included those embedded within commercial, industrial, and residential land uses. Many of the urban wetlands sampled were suspected to previously have been embedded in agricultural land uses due to the development patterns throughout Florida. Hereafter wetlands embedded in primarily undeveloped landscapes are called reference; wetlands embedded in primarily agricultural land uses are called agricultural; and wetlands embedded in primarily urban land uses are called urban. Table 2-1 provides general information about each sample wetland, including site code, sample date, wetland type, wetland region ( Lane 2000 ), county, bioregion ( Griffith

PAGE 27

Table 2-1. Site characterization for 24 freshwater forested wetlands. Site Code* Sample Date Wetland Type Wetland Region ^ County Bioregion ^ A Priori Category Associated Stream FF1 May 19 2003 Floodplain North Putnam Peninsula Urban Orange Creek FF2 May 20 2003 Floodplain North Marion Peninsula Reference Juniper Creek FF3 May 22 2003 Floodplain North Volusia Peninsula Urban Groover Branch FF4 May 23 2003 Floodplain Central Lake Peninsula Agricultural Blackwater Creek FF5 May 25 2003 Floodplain North Nassau Northeast Urban Pigeon Creek FF6 May 25 2003 Floodplain Central Polk Peninsula Reference Livingston Creek FF7 May 25 2003 Floodplain Central Polk Peninsula Reference Tiger Creek FF8 May 26 2003 Floodplain North Nassau Northeast Urban Alligator Creek FF9 May 29 2003 Floodplain North Clay Northeast Urban Green Creek FF10 June 8 2003 Floodplain South Lee Everglades Urban Leitner Creek FF11 June 13 2003 Floodplain Panhandle Hamilton Northeast Agricultural Rock Creek FF12 June 14 2003 Floodplain North Baker Northeast Agricultural/Urban Turkey Creek FF13 June 24 2003 Floodplain North Martin Peninsula Reference Kitchen Creek FF14 July 1 2003 Floodplain Panhandle Walton Panhandle Agricultural Limestone Creek FS1 May 24 2003 Strand Central Osceola Peninsula Reference NA FS2 May 26 2003 Strand Central Osceola Peninsula Agricultural NA FS3 May 27 2003 Strand Central Osceola Peninsula Agricultural NA FS4 May 31 2003 Strand Central Osceola Peninsula Urban NA FS5 June 1 2003 Strand Central Osceola Peninsula Urban NA FS6 June 16 2003 Strand South Palm Beach Peninsula Urban NA FS7 June 24 2003 Strand South Palm Beach Peninsula Reference NA FS8 July 7 2003 Strand North Alachua Peninsula Agricultural NA FS9 July 8 2003 Strand North Alachua Peninsula Reference NA FS10 July 14 2003 Strand South Martin Peninsula Reference NA *Site Codes correspond to the wetland type (FF = floodplain forest; FS = forested strand) and were numbered in the order sampling occurred. ^ Wetland region from Lane ( 2000 ); Bioregion from Griffith et al. ( 1994 ). Associated Stream = Stream Condition Index (SCI) stream data; NA = Not Applicable, forested strands were not associated with SCI streams.

PAGE 28

17 et al. 1994 ), a priori category of surrounding land use, and associated stream. Site codes were assigned to preserve the anonymity of individual land owners. Gradients of Landscape Development Intensity Digital orthophoto imagery, available from Labins The Land Boundary Information System from the Florida Department of Environmental Protection (FDEP) (available at http://www.labins.org/2003/index.cfm ), were used to georectify transect locations from GPS coordinates using ArcGIS 8.3 from Environmental Systems Research Institute, Inc. Polygons were established for wetlands delineated from digital orthophoto imagery. LDI scores were calculated using 1995 and 2000 land use coverages available for download through the Florida Geographic Data Library (available at http://www.fgdl.org/ ). Coverages for 1995 land use were available separately for each Florida Water Management District (Northwest Florida, NWFWMD; Suwannee River, SRWMD; St. Johns River, SJRWMD; Southwest Florida, SWFWMD; and South Florida, SFWMD). More recent coverages (2000 land use) were available for a limited number of sample wetlands including only those within the boundaries of the SJRWMD. LDI index scores were calculated at the transect (LDI_T), wetland feature (LDI_F), and watershed (LDI_WS) scale. At the transect scale, a 100 m buffer was created around the downstream sample transect. Two LDI_T scores were then calculated based on the area of the land uses within the 100 m buffer ( Figure 2-3 ), one including the wetland area and the other excluding the wetland area. The LDI_T score including the wetland area will generally be lower than the LDI_T score excluding the wetland area, based on the increase in the percent of land assigned a 1.0 Natural Land/Open Space LDI coefficient (because of the inclusion of the wetland area, which is assigned an LDI coefficient of 1.0). Including the wetland area in the calculation was thought to weight the LDI_T based on the size of the wetland sampled, which could be an important factor influencing wetland condition. LDI_T calculations were repeated using a 200 m buffer ( Figure 2-3 ). LDI_T 1995 identification of land uses within the 100 m buffer were taken from land use/land cover coverages and checked with 1995 digital orthophoto imagery. LDI_T 2000 identification of land uses were taken from 2000 land use/land cover coverages and checked with 1999 digital orthophoto imagery. Land uses that had changed since the photos were taken were updated based on land use maps drawn during field site visits. This step was not possible for larger buffer areas where visibility of surrounding land uses on site was hindered due to distance and barriers, and so discrepancies may be apparent in 100 m buffer calculations (hand corrected for each buffer) and larger 200 m buffers (not corrected, land use/land cover straight from available GIS coverages). LDI_F and LDI_WS were also calculated once including and once excluding the wetland area. The boundary for the LDI_F was established as the area making up the 200 m buffer upstream of the downstream transect ( Figure 2-4 ). The boundary of the feature was established when a distance of at least 30 m showed a break in wetland vegetation established from photo-interpretation of the digital orthophoto imagery. The upstream boundary for the LDI_WS was established using the Better Assessment Science Integrating Point and Nonpoint Sources 3.0 (BASINS 3.0) environmental analysis system ( USEPA 2001 ) ( Figure 2-5 ). BASINS is a multi-purpose environmental assessment tool

PAGE 29

18 Figure 2-3. Boundaries of the 100 and 200 m buffers drawn around the downstream transect for the LDI_T calculation at Forested Floodplain 3 (FF3). Two LDI_T values were calculated for each sample wetland, one including the wetland area and the other excluding the wetland area.

PAGE 30

19 Figure 2-4. Boundaries of 100 and 200 m buffers around a wetland feature and designated land use. The 200 m buffer was used for the LDI_F calculation at Forested Floodplain 3 (FF3).

PAGE 31

20 Figure 2-5. Boundary of the watershed buffer around Forested Floodplain 3 (FF3) for the LDI_WS calculation. Four LDI_WS values were calculated for each sample wetland: two LDI_WSs were calculated based on an equal weighting by area of all land uses rshed including a calculation inclusive of the wetland area e of the wetland area (LDI_WS_ED/wo); and two dista within the upstream wate(LDI_WS_ED/w) and exclusiv nce weighted LDIs were calculated in which land uses nearest to the sample wetland were more highly weighted, including linear weighting (LDI_WS_DW_lin) and weighting based on an exponential decay function (LDI_WS_DW_exp).

PAGE 32

21 designed for watershed and water quality studies that integrates a GIS (ArcView 3.2), national watershed database, and environmental assessment and modeling tools (USEPA 2001). The digital terrain model from the National Elevation Dataset, which is a 30 m raster-bere weighted more than those occurring farther away (WS_LDI_DW). LDI_WS_ED scores were calculated both including (LDI_WS_ED/w) and excluding (LDI_WS_ED/wo) the wetland area. The distance weighted calculations were done in two ways. First, a calculation was made using a linear weighting of land uses (LDI_WS_DW_lin); second, a calculation was made using an exponential decay function, where land uses nearest the sampling point were weighted most significantly (LDI_WS_DW_exp). GIS analyses were performed in ArcGIS 8.3 (Environmental Systems Research Institute, Inc. 2002). The following equation was used to calculate LDI: LDITotal = % LUi LDIi (2-1) where %LUi is the percent of a land use within the buffer of interest and LDIi is the LDI coefficient for a particular land use based on the amount of nonrenewable energy use per unit area in the surrounding landscape (Table 1-1). The LDI coefficient values and LDI equation were based on work by Brown and Vivas (2005). Potential LDI coefficients ranged from a minimum of 1.0 (Natural Land/Open Space) to a maximum of 10.0 (Central Business District). Field-data Collection A concise summary of field-data collection procedures follows. Appendix A provides more detailed descriptions of field-data collection techniques in the format of Standard Operating Procedures (SOPs) for field use. Transect Sampling Design Transects were established perpendicular to the hydrologic flow of the system. he wetland/upland boundary was determined based on the Florida Unified Wetland plafacMowe strands consisted of establishing four transects that extended from the wetland/upland boundary to the middle of the channelized flow in the center of the strand (Figure 2-6a). Transects were established at 25 m intervals. While it is recognized that strands were not perfectly symmetrical in nature, an effort was made to establish transects that represented a cross-section of the strand along the gradient of ased dataset produced by the United States Geological Survey (USGS), was used for watershed delineation. Two separate types of LDI_WSs were calculated based on an equal weighting by area of all land uses within the upstream watershed (LDI_WS_ED) and a distance weighted calculation, where land uses occurring closer to the sampling point w T Delineation Methodology (Chapter 62-340, F.A.C.), using a combination of wetland nt presence according to wetland plant status (e.g. obligate, facultative wetland, ultative, or upland) and wetland hydrologic indicators (e.g. lichen lines, moss collars). difications of transect establishment were implemented for strand and floodplain tlands types, described below. Sampling in forested

PAGE 33

22 (A) Figure 2-6. Idealized transect layout for forested strand wetlands. (A) Standard transect layout included four transects that represented a cross-section of the strand along the gradient of wetland/upland boundary to the water channel, with transect establishment along both sides of the water course. (B) Alternative transect placement included four transects established on the same side of the water channel due to limited access or permission restrictions. 0-5m 0-5m 5-10 5-10 10-15 10-15 15-2015-2020-2520-2525-3025-3030-35T1 T2Upland/Wetland Boundary Main Water Channel 0-5m 5-10 10-15 15-20 20-25T3 0-5m 5-10 10-15 15-2020-2525-30T425-3030-3535-40 25m Flow Direction Forested Strand Wetland Forested Strand Wetland Upland Upland

PAGE 34

23 (B) Upland/Wetland Boundary Main Water Channel Flow Direction Forested Strand Wetland Upland 0-5m5-1010-15T1 15-2020-25 T3 0-5m5-1010-1515-2020-2525-30 T4 25m 0-5m5-1010-1515-2020-25 25-30 30-35 35-40 25-30 25m 30-35T2 30-35 25m 0-5m5-1010-1515-2020-25 25-30 30-35 35-40

PAGE 35

24 upland/wetland boundary to the water channel. In some sampling efforts wetland forests were sampled along both sides of the water course. However, on some occasions, limited access or lack of permission restricted sampling to the same side of the water course. In these instances, four transects were established on the same side of the water channel (Figure 2-6b). Forested floodplain wetlands varied in size, with many wetlands spanning an area broader than 100 m across. To encourage comparable sampling efforts among the different width strand and floodplain systems, a maximum 50 m transect length was established for floodplain systems. For forested floodplains narrower than 50 m wide, four transects were established that spanned the area from the upland/wetland boundary, as determined by wetland plant presence and hydrologic indicators, to the edge of the stream channel. The initial transect was established downstream at a point closest to the SCI sampling location. Consecutive transects were established upstream at 25 m intervals. In forested floodplain wetlands wider than 50 m, the initial transect was established at the upland/wetland boundary and extending 50 m inward perpendicular to the stream. The second transect was established 25 m upstream, starting at the waters edge and extending a full 50 m towards the upland/wetland boundary, perpendicular to the water course. The pattern was repeated for the third and fourth transects (Figure 2-7). While four transects of 50 m length were considered the optimal sampling effort, this was not always achieved due to limited access and development pressures resulting in limited remaining areas of forested floodplain. Along each transect, a series of 1 m wide by 5 m long quadrats was established back to back. Living macrophytes rooted within each quadrat were identified to the lowest taxonomic level possible. When field identification was limited, a sample specimen was collected, pressed, and identified in the laboratory. An expert Florida botanist (Dr. David Hall) was consulted for identification of unknown specimens. Taxonomic information including species, genus, and family were compiled for all of the macrophytes identified. Additional characteristics were collected for use in metric development, including category (annual or perennial, evergreen or deciduous, indigenous or exotic) and growth form (aquatic, fern, grass, herb, sedge, shrub, tree, or vine). References specific to Florida were consulted first (Tobe et al. 1998; Wunderlin and Hansen 2003), and additional information was supplemented from other sources (in the following order: Godfrey and Wooten 1981, Wunderlin 1998, and USDA NRCS 2002). Wetland indicator status (i.e. obligate, facultative wetland, etc.) for each species was obtained from FAC Ch. 62-340. Wetland indicator statuses for species not listed in FAC Ch. 62-340 were obtained from Tobe et al. (1998), Wunderlin and Hansen (2003), and USDA NRCS (2002), in that order. When species were not assigned a Florida specific indicator status, the National Wetlands Inventory indicator status was used (Wunderlin and Hansen 2003; USDA NRCS 2002, in that order). When information was still unavailable for plant characteristics in published literature, Florida botanists (who also participated in the Floristic Quality Assessment Index) were consulted. muwaois. Floristic Quality Assessment Index A Floristic Quality Assessment Index (FQAI) has been included in many of the ltimetric biotic indices created for the macrophyte assemblage. The concept of FQAI s developed by Wilhelm and Ladd (1988) for vegetation around Chicago, Illin

PAGE 36

25 Figure 2-7. Idealized transect layout for forested floodplain wetlands. Transect 1 began at the upland/wetland boundary and extended 50 m inward, perpendicular to the channelized flow. Transect 2 began at the edge of the channel and extended 50 m towards the wetland/upland boundary. This pattern repeated for transects 3 and 4. Upland/Wetland Boundary Main Water Channel Forested Strand Wetland Upland 0-5m-10 510-15 15-2020-2525-3030-35 35-40 40-4545-50 T4 25m 0-5m 5-10 10-15 15-2020-25T1 0-5m 5-10 10-15 15-20 20-2525-30 25m Flow Direction 25-30 30-35T3 35-4040-4545-50 25m 30-35 35-4040-4545-50T2 0-5m5-1010-15 15-20 20-25 25-3030-35 35-40 40-4545-50

PAGE 37

26 This method of scoring plant species based on expert botanist opinion has been used in Michigan (Herman et al. 1997), Ohio (Andreas and Lichvar 1995; Fennessy et al. 1998; Mack 2001). 2002), and Floridaanehe FQAI provides a quaparticular environbotanists assigned coefficd coefficients of conservatism) to individual species. This techniqsmts is more accurate thnefor vegin to each species. The FQidual wetland was calculated as: (2-2) species richness as the denominator. Theories on the importance f species richness suggest that higher species richness signifies a more valuable ecosystem, which can be quantifieduare root function (Fennessy et al 998). However, a recent study by Cohen et al. (2004) found a stronger relationship radient for the mean CC (Eq. 2-2) than with the traditional FQAI equatiot al. 1994), categories, or other unspecified differences. onservatism scores were obtained from Florida botanist survefrom the depressional forested and herbaceous wetlands were strong ncluded in previous Florida FQAI surveys. As such, each botanis Ontario (Francis et al. 2000), North Dakota (Mushet et al (Le t al. 2003; Cohen et al. 2004; Reiss and Brown 2005). Tntitative means of assessing the fidelity of a plant to a ment through the Delphi technique (Kent 2000), where individual ients (terme ue asues that the collective decision by a group of expert botanisa th professional judgment of one individual (Kent 2000). The Florida FQAI enlisted regional expert botanists to provide quantitative scores etation the form of a coefficient of conservatism assigned individuallyAI score for an indiv FQAI = [ (C + C + C) ] / N 1 2 n where C = species specific coefficient of conservatism score (CC), and N = species richness. This equation was considered a modified FQAI because previous studies used the square root of native o by using the sq 1 along a disturbance g n (using the square root of native species richness in the denominator). They also reported that using total species richness (i.e. including exotic species in the calculation of species richness) improved the relationship of mean CC with the human disturbance gradient (measured with LDI). The sum of the species CC scores was divided by total species richness ( Eq. 2-2 ) in this study to account for potential differences in species richness due to differences in wetland regions ( Lane 2000 ), bioregions ( Griffith e a priori land useCoefficient of cys to compliment previous FQAI efforts for Florida. FQAI scores were adopted first from depressional forested wetlands ( Reiss and Brown 2005 ) and second from depressional herbaceous wetlands ( Lane et al. 2003 ; Cohen et al. 2004 ). Correlations between CC scores for species (unpublished data). In an effort to streamline the acquisition of species specific CC scores for the FQAI for strand and floodplain wetlands, CC scores were only obtained for those species not it was sent a list of species identified in the forested strand and floodplain wetlands that did not already have a CC score for the previous Florida FQAI efforts (n = 86). Botanists participating in the 2003 FQAI effort for forested strand and floodplain wetlands included Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and Bruce Tatje. Botanists

PAGE 38

27 scored cosystems hat were widely distributed and occurred in disturbed cal of well-established ecosystems that sustain only le diversity in a wetland (McCune and Grace 2002), was alculated as the Shannon diversity (H') value divided by the natural log of richness: (2-3) s of taxon i, and N was the total number of occurre each species based on its faithfulness to Florida forested strand and floodplain wetlands. Potential CC scores ranged from 0 to 10: 0 Exotic and native species that act as opportunistic invaders, included species that commonly occur in disturbed e 1-3 Species t ecosystems 4-6 Species with a faithfulness to a particular ecosystem, but also tolerant of moderate levels of disturbance 7-8 Species typi minor disturbances 9-10 Species occurring within a narrow set of ecological conditions Species with low CC scores were considered tolerant of many disturbances, whereas species with high CC scores were considered to occur within a narrow set of stable ecological condition. Appendix B lists the CC scores for the macrophyte species identified in this study. Data Analysis Summary Statistics Summary statistics for the macrophyte assemblage were calculated at the species level, including richness (R), evenness (E), Shannon diversity (H'), Simpson diversity (D), and Whittakers beta diversity ( W ). Richness (R) was defined as the total number of distinct taxa identified within the sample wetland. Evenness (E), described as the fraction of maximum possib c E = H' / ln (R) The Shannon diversity index has been described as measuring the information content of a sample unit where maximum diversity yields maximum uncertainty ( McCune and Grace 2002 ). For Shannon diversity calculations (H'), the sample unit was an individual forested wetland: H' = p i log(p i ) (2-4) p i = n i / N (2-5) where n i was the number of occurrence nces of all taxa at a wetland. The number of occurrences represented the number of quadrats a species occurred in, and the total number of occurrences of all taxa at a wetland represented the sum of the total number of quadrats of all of the species identified. Simpson diversity (the compliment of Simpsons original index, which was a measure of dominance) was a measure of the likelihood species that are randomly chosen in sampling will be different. The equation used to calculate Simpson diversity was: D = 1 p i 2 (2-6)

PAGE 39

28 Whittakers beta diversity ( W ) was computed as a calculation of overall beta diversity, or the compositional change represented in a sample. Whittakers beta diversity was calculated as the number of species at a particular forested wetland (S c ) divided by the average species richness per quadrat (S), minus one: w = [S c / S] 1 (2-7) The resulting value for Whitakers beta diversity was described as the number of distinct communities ( McCune and Grace 2002 ). When w equaled zero, all of the sample units contained all of the species. Some multivariate methods strongly depend on beta diversity, and as a general rule beta diversity greater than 5 was considered high ich described within-group similarity. When A equaled one, all items were identical within groups, and when A equaled zero, differences within-groups equaled that expected by chance. 02) found that most values of A ere less than 0.1 in community ecology. MRPP was calculated across all wetland region versus south, and central versus south) when sample size was appropriate. An appropriate (McCune and Grace 2002). Summary statistic means within a priori land use categories were compared with Fishers Least Significant Difference (LSD) pair wise comparison test using Minitab ( Version 13.1, Minitab Statistical Software ). The strength of using Fishers LSD was in the comparison of unequal group sizes ( Ott and Longnecker 2001 ; Minitab 2000 ). Sample wetlands were divided into two groups based on 1995 LDI_F/wo scores including low (LDI < 2.0) and high (LDI 2.0) LDI groups. Comparisons were made for summary statistics between low and high LDI groups using the non-parametric Mann-Whitney U-Test in Minitab ( Ott and Longnecker 2001 ). Overall calculations of beta and gamma diversity were calculated for sample wetlands in the three a priori land use categories. Gamma diversity was calculated as the overall number of taxa encountered at all sample wetlands per a priori land use category. Higher gamma diversity for an a priori land use category suggested a greater difference among the species composition of wetlands within that a priori land use category, assuming a similar number of wetlands were sampled within each a priori land use category. Beta diversity was calculated as a priori category gamma diversity divided by the mean site taxa richness. Regional Compositional Analysis The Multi-Response Permutation Procedure (MRPP) was used to test the similarity of macrophyte community composition among Lanes (2000) four Florida wetland regions (see further application of MRPP in Zimmerman et al. 1985; McCune et al. 2000; McCune and Grace 2002). MRPP is a nonparametric technique which tests for no difference between groups (the null hypothesis) and is available in PCORD (Version 4.1 from MJM Software, Gleneden Beach, Oregon). It is an appropriate procedure for ecological community data as it does not require distributional assumptions of normality and homogeneity of variances. The Srensen distance measure was used to calculate the average weighted within-group distance. MRPP analysis provided a test statistic (T), p-value, and chance-corrected within-group agreement (A), wh McCune and Grace (20 w groups (panhandle versus north versus central versus south; wetland regions according to Lane 2000 ) as well as for multiple pair wise comparisons (panhandle versus north, panhandle versus central, panhandle versus south, north versus central, north

PAGE 40

29 sample size was defined by a minimum of two wetlands within each category for comparison. For example, if there were seven wetlands being compared with six in one region and only one in a different region, the MRPP comparison analysis could not be executenorthease comparisons (panhandle versus northeast, panhanot accurately characterize ecological trends. This has been described as a compen85% of the cells within the matrix were assigne final run was comple d. Additional MRPP tests were calculated across all bioregions (panhandle versus st versus peninsula versus Everglades; bioregions according to Griffith et al. 1994) as well as for multiple pair wi dle versus peninsula, panhandle versus Everglades, northeast versus peninsula, northeast versus Everglades, and peninsula versus Everglades) when sample size was appropriate. Community Composition Community composition of the macrophyte assemblage was summarized in a non-metric multidimensional scaling (NMDS) ordination. NMDS, an ordination technique designed to compress multi-dimensional space, is particularly agreeable with ecological data because it does not rely on linear relationships among variables, which generally do n sation for the zero-truncation problem through the use of ranked distances (a common characteristic of non-parametric analyses) and the use of an appropriate distance measure ( McCune and Grace 2002 ). The zero-truncation problem refers to the extraordinary number of zeros (denoting a species is absent from a sampling location) typical in community ecology data sets. For example, in the species by site matrix used in our data analysis (283 species by 24 sites), d a zero (i.e. the species does not occur at the site). Other ordination techniques depend upon a value for each measured variable for each sample unit. However NMDS is useful for the presence/absence data sets, where many species are not present, or receive a zero in the species by site matrix. By ranking the variables, NMDS caters to the non-parametric community ecology data set. NMDS was used to explore the dissimilarities of the community composition of the macrophyte assemblage among sample wetlands. The Srensen (Bray-Curtis) distance measure was used for ordination. The dimensionality was chosen based on an initial 6-dimensional run in autopilot mode, which suggested an optimal 2-dimensional solution. To find the optimal 2 dimensional solution, 50 runs with real data and 50 randomized runs were performed with the instability criterion set at 0.00001 and the maximum number of iterations to reach a stable solution set at 500. This procedure was repeated 20 times to insure stability and reproducibility in results. The ted with the starting point set as the results from the best experimental 2-dimensional run, with the lowest stress and best overall fit. LDI_F/wo, species richness, Shannon diversity, Simpson diversity, Whittakers beta diversity, and sampling date (as Julian date) were tested for correlation with the NMDS ordination axes with Pearson correlation coefficients. Metric Development In the context of this study, metrics are defined as biological attributes which have a consistent and predictable response to anthropogenic activities (Karr and Chu 1997). Metrics were summarized in three main categories, including tolerance indicators,

PAGE 41

30 community structure, and community balance. Metrics were as the proportion (P) by dividing the number of species (n) fitting a particular metric category (ex. exotic species) dividedistent and predictable trends along the gradiene adopted from previous FWCI studies (Lane et al. 2003; Reiss aus, and basGrace 2002). Calculated indicator values are based on two standards, faithfulness and exclusion. Faithfulness is defined mathematically by a particular species always being present in a erfect indicator species is xclusive to that group, meaning it never occurs in other groups. Indicator values range from 0nt indicator species; species signific by the total species richness (N) for each sample wetland: P i = n i / N i (2-8) where i represents an individual wetland. Scatter plots were constructed for each candidate metric versus 1995 LDI_F/wo_200m to ensure correlations were visually distinguishable. Candidate metrics were accepted if they successfully distinguished between low (LDI < 2.0) and high LDI (LDI 2.0) groups using the Mann-Whitney U-test (p < 0.10) calculated with Version 13.1, Minitab Statistical Software, and showed cons t of anthropogenic land use activity according to the strength and significance of the Spearmans correlation coefficient (p < 0.05) calculated with Analyze-It software v. 1.67 for Microsoft Excel. The Spearman rank correlation tested for an association between two related variables, and was a non-parametric alternative to the Pearson correlation. Macrophyte metrics wer nd Brown 2005 ), included tolerance (tolerant species; sensitive species), exotic, FQAI, and longevity and plant growth form (native perennial). Additional metrics were explored for summary statistics (richness, evenness, and diversity), as well as dominance, annual to perennial ratio, wetland indicator status, native evergreen, native deciduo al area (total, deciduous, or evergreen). Tolerance metrics were calculated with Indicator Species Analysis (ISA) in PCORD ( Version 4.1 from MJM Software, Gleneden Beach, Oregon ). Sample wetlands were categorized into low and high LDI groups and analyzed with ISA, which evaluates the abundance and faithfulness of species in a defined group ( McCune and Grace 2002 ). ISA is used to detect and describe the value of species indicative of environmental conditions. It requires a priori groups and data on the presence of taxa in each group. Groups are commonly defined by categorical environmental variables, levels of disturbance, experimental treatments, presence and absence of a target species, or habitat types ( McCune and particular group. Additionally, the p e (no indication) to 100 (a perfect indication of a particular group). Multiple ISA iterations were conducted to determine sensitive and tolerant indicator species. Sample wetlands were categorized based on 1995 LDI_F/wo_200m scores, and ISA calculations were completed for consecutive LDI breaks from 1.0 through 3.5 at each 0.5 increment using the presence/absence of taxa at each wetland. In total ISA was run six times. Species significantly associated (p < 0.10) with the higher LDI group (LDI > 1.0; LDI > 1.5; LDI > 2.0; LDI > 2.5; LDI > 3.0; LDI > 3.5) for each iteration were included in the overall list of tolera antly associated (p < 0.10) with the lower LDI group (LDI = 1.0; LDI 1.5; LDI 2.0; LDI 2.5; LDI 3.0; LDI 3.5) for each iteration were included in the overall list of sensitive indicator species. Exotic species were excluded from the sensitive indicator species list, as this list was constructed to represent reference biological condition, which

PAGE 42

31 should exclude the presence of nonnative species. A random number seed was used for each ISA iteration. Calculated indicator values were tested for statistical significance using a Monte Carlo randomization technique with 1000 randomized runs. Indicator species categorized as tolerant species were associated with a higher LDI group; indicator species categorized as sensitive species were associated with a lower LDI group. Tolerant and sensitive indicator species lists were compiled based on a collection of significant indicator species at each incremental LDI break. This method was employed due to the small sample size of wetlands available to construct indicator species lists. The proportion indicator species metrics were calculated for each wetland as the number of indicator species (either tolerant or sensitive) divided by species richness. The proportion exotic species metric was calculated as the number of species that were exotic to Florida divided by species richness. The timeline for determining the exotic status of a species was set near the beginning of European settlement in North America. Sources consulted to determine the exotic status of a species included Godfrey and Wooten ( 1981 ), Tobe et al. ( 1998 ), Wunderlin ( 1998 ), USDA NRCS ( 2002 ), and Wunderlin and Hanse n (2003). determined that disturbance often favors annual species over perennial species and promotes the invasion of nonnative perennials in wetlands. Galatowitsch et al. (2000) found that while native perennial cover was reduced in wetlands impacted by cultivation, the occurrence of introduced perennials rather than annuals increased in stormwater impacted wetlands. Florida Wetland Condition Index The Florida Wetland Condition Index (FWCI) consists of individual metrics, which were scaled and added together. Metric scoring was based on an approach from the Stream Condition Index (SCI), a Florida based biological index of the macroinvertebrate assemblage used to discern stream condition (Fore 2004) and also used in the development of the FWCI for depressional forested wetlands (Reiss 2004). Metrics were natural log transformed to improve the normality of distribution. The 5th to 95th percentile values of each metric were normalized from 0 to 10, with 10 always representing the reference biological wetland condition. An agglomerative cluster analysis in PCORD was used to determine wetland groups (McCune and Grace 2002), for wetlands with similar macrophyte community composition. The dissimilarity matrix was constructed using the Srensen distance measure and the flexible beta linkage method ( = -0.25), which is a flexible clustering FQAI scores for each wetland were calculated based on the sum of CC scores for each species identified divided by total species richness at each wetland (Eq. 2-2). FQAI scores had the potential to range from 0.0-10.0. Wetlands with higher FQAI scores represented wetlands with species indicative of more stable environmental conditions. The proportion native perennial species was calculated as the number of native perennial species encountered divided by wetland species richness. Each species was categorized as native or exotic and annual or perennial according to Godfrey and Wooten (1981), Tobe et al. (1998), Wunderlin (1998), USDA NRCS (2002), and Wunderlin and Hansen (2003). Wienhold and Van der Valk (1989) and Ehrenfeld and Schneider (1991)

PAGE 43

32 setting designed to reduce chaining in the dendrogram. The resulting dendrogram was pruned using ISA, which provided an objective, quantitative means of selecting the optimum number of clusters represest ecologically meaningful point of pruning (McCune and Grace 2002). Gership determined in each step of the agglomerative cluster analysis was used ashe grouping variable for ISA. Using the ecies by site presence/absence matrix, indicator values were calculated for each species occurring in three or mom. This is an entirely separate application of ISA as that used to construct the metrics for tolerant and sensitive indicatoom the randomized Monte Carlo tests were averaged among all species at each step of the dendrogram, and the cluster step with the lowest average p-value was used to ram yielding the greatest ecologically meaningful informaor forested strands). The LDI_F/wo_200m and HDG represented measures of anthrop nting the mo roup memb t sp re sites for each step of the dendrogra r species. This method of using ISA to determine the most appropriate and meaningful point on a dendrogram for understanding a significant number of groups for a data set is described more completely by McCune and Grace (2002). The resulting p-values fr determine the level in the dendrog tion ( McCune and Grace 2002 ). FWCI scores and LDI index values among clusters were then compared using Fishers LSD pair wise comparison (p < 0.05) available in Minitab To determine how the biological data compared to the intensity of landscape development, Spearmans rank correlation was used to compare individual metrics and the FWCI with the multiple LDI calculations. For forested floodplain wetlands, comparisons of the FWCI and LDI_F/wo_200m were made with the HDG and SCI scores using the non-parametric Spearmans rank correlation coefficient (HDG and SCI scores were not available f ogenic influence; whereas the FWCI and SCI represented biological indices based on the macrophyte community composition in the forested wetland and macroinvertebrate community composition in the channelized stream flow, respectively.

PAGE 44

33 CHAPTER 3 RESULTS nic Activity Two measures of anthropoge nic activity in the lands cape were correlated to d the Hu man Disturbance Gradient cape Development Intensity LD e forested strand ( Table 3-1 ) and tions were calculated for all 24 we tlands using 1995 land use coverages and 2000 se co eDistrict (SJRWMD). Land uses in the 100 m buffers for LDI_T tioner a l orthophoto imagery and assigned land use ries hed (WS) scale LDI calculations were com a ulations including (1995_LDI_T/w_100m) and 5 LDI_T/wo_100m) wetland area ) show that calculations ltural and urban wetlands sistently higher than those incl uding the wetland area (1995 LDI_T/w_100m) re curring above the 1:1 r ethods. In calculation of LD I_T/w compensates foferences in wetland area between la ay not corr elate to differences ice unity composition to enic activ i lsc at the transect and feature scales buffers ( ause the forested strand and ticipated that the feature gnal of organisms that are a priori land ct (1995 LDI_T/wo_200m) and and three urba n wetlands; whereas the rema ining 11 wetlands received scores, including five agricultural and six urban wetlands. equal weighting for all area within a nd area, (1995 LDI_WS_ED/wo) and a distance weighted physical, chem Landscape Developm (HDG). Lands floodplain ( calcula land u Managem calcula catego Waters wetlands. A com excluding excluding the wetland area ( were con for the fo line. Fou effect the study wet in biolog anthropog The correlation between 1995 LDI calculations with 200 m floodplain wetlands involve channe scale LDI would better reflec affected by upstream use category, received identi feature scales (1995 LDI_F/wo_200m LDI_F/wo_200m agricultural, higher 1995 LDI_T/wo_200m Watershed scale LDI calculations using watershed, excluding the wetla G r adi en ts o f A nt hr opo ge ic al, a e nd nt b In iol ten ogi s ity cal ( m LD ea I) su ind r es of wetland condition, including the ex an I in Ta de b x c le alc 3-2 ula ) wetlands. Transect (LDI_T) and feature (LDI_F) scale tio ns we re co mp let ed for th ver nt ag es f or we tla nd s lo ca ted wi thi n the boundary of the St Johns River Water s w bas e d on eli g nea rou te nd d f tr rom uth d ed igit la ed nd us e m a ps drawn during the field site visit. pleted for the 14 forested floodplain ( Figure 3-1 ) for agricu r dif ay or m cores, including six reference, one p riso n bet we en LD I calc (199 19 95 LD I_ T/w o_ 10 0m sted refe w re et nce lan w ds, etla as nd ca s r n b ece e s ive een d b a s y the values generally oc c ore of 1.0 using both calculation m nd al r s. Ho spo we nse ver s th uc is h dif as fer ch enc ang e i e s in macrophyte comm n a r ea m ities Figure 3-2 anthropogenic activities. Three wetlands, one in each an n the 99 and ) wa t the biological response si LD ape. /w s low (R lized flowing water, it is an I_T 2 = 0m 0.32). Bec s cal scores for the transe 5 ) Ten wetlands received higher 1995 20 th 1 o_

PAGE 45

Table 3-1. Landscape Development Intensity (LDI) index scores for 10 forested strands using 1995 and 2000 land use coverages. LDI calculations were completed at the transect (T) and wetland feature (F) scale, and including (/w) and excluding (/wo) the wetland area. See text for further details of individual calculations. FS1 FS2 FS3 FS4 FS5 FS6 FS7 FS8 FS9 FS10 A Priori Group R A A U U U U A R R 1995 LDI_T/w_100m1.22.42.31.21.93.71.02.61.3 DI_T/w0m1.33.53.31.42.34.61.03.51.4 LDI_T/w_200m1.32.62.41.22.13.51.02.91.4 DI_T/w0m1.33.23.51.33.53.91.03.21.5 LDI_F/w_200m2.62.53.11.93.03.01.03.01.7 DI_F/w0m3.02.93.11.93.53.21.03.12.1 LDI_T/w_100m1.22.62.34.35.45.81.06.11.0 DI_T/w0m1.23.82.58.07.07.51.08.01.0DI_T/wm 1.52.62.3na na na na 6.91.4na I_T/w0m1.53.22.3na na na na 7.41.6na I_F/wm 2.52.52.6na na na na 6.91.7na DI_F/w0m2.92.93.0na na na na 7.22.1na 1.2 1995L o_10 1.2 1995 1.0 1995L o_20 1.0 1995 1.0 1995L o_20 1.0 2000 1.0 2000L o_10 1.2 2000 L _200 2000 LDLD o_20 2000 _200 2000 L o_20 na 20 00 land userage naila thtla cove es wer ot av ble for ese we n ds.

PAGE 46

ent Intensity (LDI) index scores for 14 forested floodplain wetlands using 1995 and 2000 land usecoverages. LDI calculations were completed at the transect (T), w and watershed (WS) scale, and including (/w) and excluding (/wo) the wetland area. See text for further details ations. Site CoFF1 FF2 FF3 FF4 FF5 8 FF9 FF10 FF11FF12FF13FF14 etland feature (F),of individual calculFF6 FF7 FF de A Prioe U R U A U U U A A R A ri Land Us R R U 1995 1.9 1.0 5.2 1.5 4.8 1.2 4.2 1.0 1.2 1995 3.2 1.0 7.6 2.2 6.9 1.6 6.3 1.0 1.5 1995 2.2 1.0 5.9 1.2 4.0 1.3 4.3 1.0 1.7 1995 3.1 1.0 7.1 1.3 4.7 1.6 6.5 1.0 2.1 1995 1.5 1.1 3.2 1.6 5.7 1.6 2.4 1.9 1.7 1995 2.0 1.2 3.7 2.3 7.1 1.6 2.8 2.1 1.8 1995 2.4 1.5 3.8 2.2 2.1 1.4 1.9 1.8 2.0 2.5 1.5 2.9 2.1 1.4 1.9 1.8 2.1 1995 2.2 1.4 2.2 2.2 2.1 2.7 2.8 2.3 1.5 2.2 1.4 2.0 1.7 2.0 2.3 1.4 2.7 2.8 1.4 2.8 1.6 1.9 2000 1.9 1.0 5.2 1.5 5.6 1.2 2.9 1.0 1.2 2000 3.2 1.0 7.6 2.2 8.1 1.6 5.2 1.0 1.5 2000 2.2 1.0 5.9 1.0 2.2 na 1.3 na na 4.1 na na 2000 3.1 1.0 7.1 1.1 3.1 na 1.4 na na 5.6 na na 2000 1.5 1.1 3.0 1.6 1.8 na 1.5 na na 2.3 na na 2000 2.0 1.3 3.5 2.2 2.0 na 1.7 na na 2.7 na na 2000 2.3 1.5 2.3 2.2 1.7 na 1.6 na na 1.9 na na 2000 2.4 1.5 2.3 2.3 1.8 na 1.8 na na 1.9 na na 2000 2.2 1.5 2.4 2.2 1.9 na 1.7 na na 2.0 na na 2000 LDI_WS_DW_exp 2.2 1.4 2.8 2.2 1.9 na1.6 na na 2.8 na na 1.9 1.1 1.6 3.2 1.0 3.3 1.1 2.1 4.4 1.0 2.3 1.0 1.0 3.3 1.1 3.1 1.0 1.0 3.9 1.1 1.7 1.4 1.9 2.0 1.5 1.9 1.4 2.5 2.5 1.8 1.8 2.7 3.1 2.2 1.7 2.3 1.9 2.7 3.4 2.5 1.7 2.2 2.1 2.6 2.6 2.6 1.5 1.9 1.1 1.6 3.3 1.0 3.3 1.1 2.1 4.5 1.0 na 3.4 na 4.8 na 2.0 na 2.5 na 2.3 na 2.6 na 2.3 na 2.6 Table 3-2. Landscape Developm na 2000 laoverages were not available for these wetlands. LDI_T/w_100m LDI_T/wo_100m LDI_T/w_200mLDI_T/wo_200mLDI_F/w_200mLDI_F/wo_200mLDI_WS_ED/w 1995 LDI_WS_ED/wo LDI_WS_DW_lin 1995 LDI_WS_DW_exp LDI_T/w_100mLDI_T/wo_100mLDI_T/w_200mLDI_T/wo_200mLDI_F/w_200mLDI_F/wo_200mLDI_WS_ED/w LDI_WS_ED/wo LDI_WS_DW_lin nd use c

PAGE 47

36 R2 = 0.9810 0246802468101995 LDI_T/w_100m1995 LDI_T/wo_100m Strands Floodplains 1:1 Line 1:1 Line Figure 3-1. Comparison between the LDI calculations including (1995_LDI_T/w_100m) and excluding (1995 LDI_T/wo_100m) wetland area (R 2 = 0.98). Calculations excluding the wetland area (1995 LDI_T/wo_100m) for agricultural and urban wetlands were consistently higher than those including the wetland area (1995 LDI_T/w_100m) for the forested wetlands. Four reference wetlands received a score of 1.0 using both calculation methods. R2 = 0.3246810LDI_T/wo_200m 1995 0202468101995 LDI_F/wo_200m Strands Floodplains 1:1 Line Figure 3-2. Comparison between LDI calculations at the transect (1995_LDI_T/wo_200m) and feature (1995 LDI_F/wo_200m) scale for 24 forested wetlands (R 2 = 0.32).

PAGE 48

37 approach using linear weighting (1995 LDI_WS_DW_lin) showed a stronger correlation (R2 = 0.82) for the 14 forested floodplain wetlands (Figure 3-3). LDI calculations were also completed for multiple years, using land use/land cover data from 1995 and 2000. Comparison between 1995 and 2000 LDI calculations at the feature scale (without the wetland area) for 13 forested wetlands showed a moderate correlation (R2 = 0.41) (Figure 3-4). However, 12 of the wetlands had nearly equal LDI scores for the 1995 and 2000 LDI_F/wo_200m calculations with one clear disagreement, FS8, which had a significant area of the 200m buffer categorized as Low Intensity Pasture (with livestock) in 1995 and Institutional in 2000. The land use was not actually changed, only the classification on the GIS coverage was changed, as the property is part of the University of Floridas (thus the Institutional classification) Beef Research Unit (thus the Low Intensity Pasture (with livestock) classification). Removing FS8 because of erroneous differences in the land use/land cover coverages, the remaining 12 wetlands had strongly correlated 1995 and 2000 LDI_F/wo_200m calculations (R2 = 0.99 without FS8). allhas as described above (Figure 3-4). Using the older land use/land cover data from 1995 may have limited implications in changes to LDI scores as changes in land use are localized to t development. In these areas it may be argued that more recent overages are needed, however at the statewide scale the 1995 land use coverage rovides a baseline for LDI calculations. We are not suggesting that the most recent and complete coverage should not be used, but simply that the LDI calculation should not be disregarded because the coverage is ten years old. Third, because the 24 wetlands surround channelized flowing water, the feature scale LDI calculations were used to capture the human activity influencing the upstream wetlands. Fourth, the LDI_T_100m calculations included hand delineation of land use according to interpretation of digital orthophoto imagery and ground truthing; however, the application of LDI as a remote (GIS computer based) tool for assessing human activity in the landscape, using widely available coverages, was an important consideration for future application. Human Disturbance Gradient Thirteen of the forested floodplain wetlands sampled corresponded to storet stations established by FDEP for water quality monitoring. Data were available for the four components of the Human Disturbance Gradient (HDG), including ammonia concentration, hydrologic condition, habitat assessment (habitat condition index), and LDI for the buffer (Table 3-3). Using the scoring criteria established by FDEP (Table 1-), each of the HDG components were scored individually and summrange fromimpaired condition. The maximum HDG score for the 13 forested floodplain wetlands was a 4 at the urban wetland FF10 (storet 28020234) for the most recent sampling period LDI_F/wo_200m LDI calculations for 1995 land use/land cover were selected for further analysis for four reasons. First, 1995 land use/land cover data were available for of the 24 forested wetlands sampled. Second, calculations for 13 forested wetlands ving both 1995 and 2000 LDI_F/wo_200m calculations showed similar value areas with recenc p 2 ed to get the HDG for each storet station. HDG scores were adopted directly from previous FDEP assessments, and no scoring was done directly for this research. HDG scores potentially 0-9, with 0 representing the reference condition and 9 representing a highly

PAGE 49

38 R2 = 0.8201231995 LDI_WS_ED/wo19LDI_WSin 4 0 1 2 3 4 95 _DW_l Floodplains 1:1 Line ll area within a watershed (1995 LDI_WS_ED/wo) and a distance weighting (1995 LDI_WS_DW_lin) (R2 = 0.82). Figure 3-3. Comweightingweighted approach using linear Figure 3-4. Com(without the wetland area) fo p for a arison between LDI calculati ons at the watershed scale using equal parison between 1995 and 20 00 LDI calculations at the feature scale r 13 forested wetlands (R 2 = 0.41). R 2 = 0.41 0 262000 LDI_F/200m 4 8 02 460m 1 995 LD I_F /wo _20 wo_ Strands 1:1 Line F lood pla ins

PAGE 50

man Disturbance Gradient (HDG) for 13 storet stations, which correspond with the forested floodplains sampled. Some foulns fo9thncentratiobinsc Code 3 RE rested floodplain wetlands have multiple HDG calce HDG were also provided, including ammonia cooring criteria were presented in Table 1-2. STORET HDG Date HDG Ammonia (NH atio r different sampling dates (19n, Hydrologic Condition, HaHydrologic Condition Hydro SCORE Habitat Assessmen 2-2tat It 00 0). The four components of dex, and LDI buffer. HDG Habitat SCORE LDI Buffer LDI_BF SCORE 3 ) (mg N/L) NHSCO FF1 20020404 8/15/2000 0 0 4 0 84 0 1.38 0 FF1 FF2 FF3 FF4 FF5 FF5 FF5 FF5 FF5 FF6 FF7 FF8 FF9 FF10 FF10 FF11 FF12 FF14 0 1.38 00 1.06 01 1.79 00 1.02 0 1 1.27 00 1.27 0 0 1.27 00 1.27 00 1.27 0 0 1.51 0 0 1.04 00 1.40 00 1.21 00 4.73 20 4.73 20 1.16 0 0 1.56 00 1.33 0 4 0 841 0 884 0 582 0 84 3 0 653 0 88 3 0 913 0 903 0 91 3 0 86 2 0 897 1 882 0 997 1 817 1 863 0 97 2 0 865 0 75 20020404 8/15/2000 0 0.04 0 20010454 7/19/1999 0 0.01 0 27010580 8/22/2000 2 0.10 120010455 2/23/1999 0 19010099 8/17/1993 1 019010099 3/5/1998 0 19010099 2/1/2000 0 0.04 0 19010099 7/9/1996 0 0.02 019010099 8/18/1992 0 -26011019 3/8/1999 0 26011020 7/20/1999 0 0 19020027 7/15/1996 1 0.08 0 20030246 4/17/1997 0 0 28020234 10/9/2000 4 0.12 1 28020234 9/11/1996 3 0.06 021010032 1/17/1995 0 19010072 3/22/1999 0 0 32010024 7/14/1998 0 0.02 0 no*bel data available ow detection limits Table 3-3. Hu

PAGE 51

40 ints were available for many of the SCI points used in this ject, and the most complete da ta from the most recent sampling period was oderate to low range of scores suggests that the ds sled represented wetlands on the lower scale of impairment. This may be u pling locations with SCI e si ss ncceptable for this study been comple ted altered, often because the area had or paved as part of adjacent human activity in the surrounding landscape. s nsidered for macrophyte Wquality data (chemical and physical parameters) were obtained for 13 of the d ater table included temperature ( C), pH, specific conductivity (umhos/cm), eygen (mg O2/L), turbidity (NTU), Total Kjeldahl Nitrogen (mg/L), ammonia N/L), and total phosphorus (mg P/L) ( Table 3-4 ). FF5 an urban wetland had the alur ratur7.oC), snductivi(49 umhos/cm), and the t va rast FF8, another urban (5.7 NTU), Total Kjeldahl Nitrogen (1.50 nd total phosphorus (0.28 mg P/ L). A third urban wetland FF10 had the v st values for temperature C Data Analysis S milies identified. The most abundant species was the vine Vitis li uscadine), which was found rooted within vegetation quadrats at 23 of 24 Acer rubrum (red maple) at 7o s Woodwardia virginica ia chain fern) found at 54% of the wetlands; the most common graminoid was (variable witchgrass) found at 46% of the wetlands. Boehmeria ica (f shru dentalis (buttonbush) and (wax myrtle), each found at 71% of the wetlands; the second most of the wetlands; and the o 39%) occurred at a um of three sam of the species identified ies or 1 nds; approximately twoonly one wetland. In f forested floodplain ds hosted e forested on. (10/9/2000). [Multiple data po research pro used for correlation purposes.] The m wetlan an acc data w because the floodplain wetland had been sodded These sampling and FW Water Quality foreste parame dissolv (as mg lowest v highes wetland, had the highest values for turbidity mg N/L), a lowest (25.8 amp rate s re vi tate site m d en du t w rin he g n con te si rec der on ing nai t hat an m ce, an b y s ut tre fou am s d u am na ites, c on sid C ere I d d ev hig elo hl pm y im en p t, air as ed the o y n th no e lon SC ge I, r h wer ost e ed no w t co etland vegetation. ater floo ers av d ox dp ail lain we tla nds s am pled that corresponded to storet stations. Wo es lue fo fo tem dis pe so e ( ox 1 pe m cif g ic O co2 ty ont r lve d yg en (1 0.2 /L ). In c alue ) and for am di m sso on lv ia ed nitr ox og yg en en (0 (2. .12 6 0 mg mg O N 2 /L /L) ), an d t he hig he o 190 genera and 97 fa rotundifo study wetlands. The second m found (Virgin Panicum commutatum cylindr wetlands; the two m Myrica cerifera common tree was second m at71% of the wetlands. Of the species en minim (48 spec fifths of the species total the wetlan in comm tatew a (m id e, 24 we tla nd s w e re sam p led w ith 28 3 ma crophyte species, representing os tla t nd abu s nd Th an e t s mo pec st ies co w mm as o th n f e tr ern ee wa 5% f th e s tu dy we a lse ne ttl o e) st c wa om s t m he on m os t c bs om we mo re n Ce he ph rba ala ce nt ou hu s species, found at 63% of the s o cci S ab al p pa le lm we et tla to nd (ca Toxicodendron radicans s bb (12 ag .5 e p countered, 110 species ( %) alm ) N e fou arl nd y o a ne t 7 (Eastern poison ivy) found -fi 1% fth st common vine was 7% stran 21 ) identified (126 species 4 s we pe re d wetlands cie id s. en Th tifi ed at ho on ste ly or 44%) were found at d strand and floodplain wetlands had 91 species tw 161 species, and the o sa mp le we tla orested

PAGE 52

T al s pl a lae n 0 C ur able 3-4. Water quality (chemic sampled. Three forested floodpode Date STORET Temp an in werat d p hys ical parameter tlands (FF1, FF5, ae pH Specific Conductivit ) for 13 storet stations which d FF10) have multiple data fry Dissolved Oxygen Turbidity To correspond with the forested flood om different sampling dates (1993-20tal Kjeldahl Nitrogen Ammonia Nitrogen Total Phosphorus ins 0). oF1 8/15/2000 20020404 2 C 3.7 umhos/cm 6.60 121 mg /L NTU 6.2 mg/L mg/L mg/L 0.75 0.004 0.08 F F F F F F F F 1 F .7 F .5 F .6 F .7 F .7 F .6 F .8 F .6 F .4 F .3 F1 8/15/2000 20020404 2 F2 7/19/1999 20010454 2 F3 8/22/2000 27010580 2 F4 2/23/1999 20010455 1 F5 8/17/1993 19010099 2 F5 3/5/1998 19010099 F5 2/1/2000 19010099 7 F5 7/9/1996 19010099 24 F6 3/8/1999 26011019 18 F7 7/20/1999 26011020 25 F8 7/15/1996 19020027 25 F9 4/17/1997 20030246 16 F10 10/9/2000 28020234 24 F10 9/11/1996 28020234 25 F11 1/17/1995 21010032 11 F12 3/22/1999 19010072 16 F14 7/14/1998 32010024 23 3.7 3.2 5.0 2.8 5.0 6.61 121 7.10 1580 7.40 746 6.90 230 6.36 70 5.10 75 5.25 49 6.90 219 6.72 172 5.90 101 5.60 68 7.20 399 7.40 451 3.97 79 6.10 77 7.70 189 6.2 1.2 6.8 1.8 2.7 7.2 3.2 10.2 1.9 5.3 7.0 6.6 1.5 4.2 5.7 7.7 3.5 2.6 2.0 3.1 8.3 6.8 7.1 3.0 0.75 0.038 0.08 0.27 0.012 0.04 0.62 0.100 0.13 0.34 0.000 0.04 0.47 0.02 0.25 0.039 0.02 0.66 0.024 0.02 0.26 0.000 0.00 1.50 0.080 0.28 0.46 0.000 0.00 0.88 0.120 0.18 0.75 0.059 0.11 0.46 0.000 0.05 0.12 0.017 0.02 no data available

PAGE 53

42 Summary Statistics Species richness (R), species evenness (E), Shannon diversity (H'), Simpson iversity (D), and Whittakers beta diversity (W) were calculated based on the mblage for each samd. ness rarom 21 sence floodplain fotland),cies at FF5 (a wetland eized landscape). evenneonsisng all a p, and speciesly showed differences at the hundred t, explainingd vaendecimal places. Shannon diversity ranged fromy; these w with the) a(7ichness, r thpsod5), and FF5 and FF8 (another urban wetland) shared the highest Simpson diversity index value (ers beta diversity ranged fromow of 1.8 at (a reference wetland), to a high of 7.4 at FF4 (an agricultural wetland). A completee of summtatistics fable 3-5 summarizes comparisons of mean richness, evenness, and diversity alculations by a priori land use category. Urban wetlands had the mean greatest species ifferent among a priori land use categorta diversity (4.56 1.44), and overall beta (4.9) a (233) diversity. grouped according to wetland regions ), as well as for multiple pair wise comparisons. Sample size limitations prevented all possible wetland region pair par a umf two wetlands per wetland region and land use caere nblsor the MRPP statistic (T),ecroret (A), and significance value (p) (). R crof the panhandle and north wetland regions for et (ing a priori and urbanigthe level, suggesignresp csition across all wetland regwia priori land use categories, the MRPP analysis did not suggest a difference among species composition in north and central reference wetlands or agricultural wetlands for all d acrophyte assem ple wetlan Species rich nged f pecies at FF2 (a refermbedded in an urban rested we Species to 77 spess was c tent amo riori land use groups evenness on housandth decimal place why reporte lues were id tical for two 3.05 to 4.34 at FF2 and FF5, respectiv lowest (21 el7) species r ere the same wetlands nd highest espectively. Similarly, FF2 had e lowest Sim n diversity in ex value (0.9 0.99). Whittak a l FF6 tabl ary s or each wetland is presented in Appendix C. T c richness (50 17) followed by agricultural wetlands (41 10). This trend was evident for both Shannon diversity, with urban wetlands having greater mean values (3.86 0.35). Whittakers beta diversity was greatest in agricultural wetlands (4.94 1.07). Beta and gamma diversity were calculated for overall a priori land use categories, with urban wetlands having the highest beta and gamma diversity (4.0 and 201, respectively). None of the summary statistics were significantly d ies (Fishers LSD pair wise comparison, = 0.05) as the standard deviations overlap for most of the summary statistics. Similarly, summary statistics (including species richness, species evenness, Shannon diversity, and Whittakers beta diversity) were not significantly different between low (1995 LDI_F/wo_200m < 2.0) and high (1995 LDI_F/wo_200m 2.0) LDI groups (Mann-Whitney U-Test; Table 3-6 ). However, wetlands in the high LDI group had higher (though not statistically significant) mean species richness (47 17), Shannon diversity (3.80 0.34), Whittakers be and gamm Regional Compositional Analysis MRPP was calculated across all wetlands (Lane 2000 wise com isons when minim o tegory w ot availa e. Re ults f tests included the test chance-corrOnly the M ted within-gPP pair wise up agompa emenison Table 3-7 the compiled pool of all w lands includ reference, agricultural, ) was not s nificant at = 0.05 ting that there were regionally sions. Of the possible pair wise com ificant diffe nces among paris eciesthin ompo ons

PAGE 54

43 Table 3-5. Richness, evenness, and diversity of the macrophyte assemblage among a priori land use categories. Referenceculturrban Agri al U Spec36 11a41 10a50 17a ies richness (R) Species evenness (E) 1.00 0.00a1.00 0.00a1.00 0.00aShanon div.32a3 0 0Simpons d0.97 0.01a0 0 0Whittaker's 3.53 1.03a4.9 1 1Betaiversi3.7Gamma dive134 n ersity (H') 3.55 0 .69 .2a 5 3.86 .35 a s iversity (D) .97 .01 a 0.98 .01 a Beta diversity (W) 4 .07 a 4.57 .38 a d ty 3.7 4.0 rsity 153 201 Categnot significantly different (Fisher's LSD, 0.05) wetlanmong urban wetlands suggested a difference in thethe wetland regions (multiple comarison panhandle versus north versus central versus south; pair wise north versus central; and pair wse norations to interpre pue to small sampleh wetland region, the direnc s sition among urban wetland showed the greatest variability. using wetlands grouped according to bioregions (Griffe forested wetland sampled, only one wapanhandle bioregion and one in the Everglades bioregion. All of the reference wetlands (n=7) were located in the peninsula bioregion. Thus the only comarisons possible were between a purban wetlands thehaninsula ecoregiofference in species coms not found among agricultural or urban wetlands in the northeast and peni Table 3-6. Richness, evenness, and diversity of the macrophyte assemblage between low I 2.0) LDI groups. ories with similar letters were = d regio ns. However, all MRPP tests aion of wetlands among species composit p i th versus south)c While limit tation are ap arent d sizes within ea ffe es in specie compo The MRPP analysis was repeated ith et al. 1994). Of th s located in the p riori agricultural and in nort east d pen ns, limiting the MRPP analysis to two comparisons. A significant diposition wa nsula bioregions (Table 37). (LDI < 2.0) and high (LD Low LDIHigh LDI W ^ p` Species richness (R) 39 1147 17 120 0.32 Species evenness (E) 1.00 0.001.00 0.00 153 0.38Shannon diversity (H') 3.62 0.313.80 0.34 120 0.32Simpsons diversity (D) 0.97 0.010.98 0.01 120 0.32Whittaker's Beta diversity (W) 4.16 1.444.56 1.44 123 0.42B eta diversity 4.34.9 Gamma diversity 169233 W = Mann-Whitney U-test statistic p` = significance value ^

PAGE 55

44 Table 3-7. Macrophyte community composition similarity among wetland regions (Lane 2000) and bioregions (Griffith et al. 1994) with MRPP. Sites (n)T^A` p# Florida Ecoregions (Lane 2000) All wetlands All regions (P vs N vs C vs S) 24-4.2 0.07 0.00* Panhandle vs north 120.2 -0.56* Panhandle vs central 10-1.9 0.04 0.04* Panhandle vs south 6-2.4 0.12 0.02* North vs central 18-3.2 0.04 0.01* North vs south 14-3.7 0.07 0.01* Central vs south 12-4.0 0.07 0.00*Reference wetlands North vs central 6-1.2 0.04 0.10*Agricultural wetlands All regions (P vs N vs C vs S) 70.4 -0.65*Urban wetlands All regions (P vs N vs C vs S) 10-4.4 0.18 0.00* North vs central 7-2.8 0.13 0.01* North vs south 8-3.9 0.15 0.01*Bioregions (Griffith et al. 1994) Agricultural wetlands Northeast vs peninsula 6-1.5 0.12 0.07*Urban Wetlands Northeast vs peninsula 9-1.6 0.05 0.07 A high |T| value and significant p-value (p<0.05) suggests a difference in species composition. ^ T = the MRPP test statistic ` A = the chance corrected within-group agreement # p = the signi ficance value. Community Composition Macrophyte community composition was summarized in a 2-dimensional nonmetric multidimensional scaling (NMDS) ordination to explore gradients in macrophyte community composition (Figure 3-5). The final solution had an overall stress of 17.8 with a final stability of 0.0002, which is considered a fair stress value useful for ordinations with community data sets (Kruskal 1964; Clarke 1993; McCune and Grace 2002). Axis 1 explained 53.3% variance, and axis 2 explained 29.2% additional variance. No measured variables were correlated with the ordination axes, including 1995 LDI_F/wo_200m, species richness, Shannon diversity, Simpson diversity, Whittakers beta diversity, and Julian date. Wetlands appeared to be broadly grouped on axis 1 according to wetland region, and perhaps the gradients of latitude and longitude would have been appropriate

PAGE 56

45 Figure 3-5. NMDS ordination bi-plot of 24 sample wetlands in macrophyte species space; wetlands are labeled according to site code; symbols correlate to wetland regions ( Lane 2000 ). Axis 1 explained 53.3% variance; axis 2 explained an additional 29.2% variance. xplanatory variables for the ordination axes. For purposes of speculation, the second fell above the horizontal line of axis 2 at the center of the plot; whereas the remaining ight strands fell on or below the horizontal line. In contrast, only three of the forested floodplain wetlands (FF2, FF11, and FF13) fell below the horizontal line; while, the remaining 11 floodplain wetlands fell on or above the horizontal line. Perhaps there were distinct dnity(strand vMDS, which could be supported by the observation that the forested strand and floodplain wetlands inthis stred just 91 species. of the species ied aany o4 forested floodplain wetlands also occurring at any of the 10 forested strand wetlands, which had fewer speg that 57% of the species identified within the forested strands also occurred in at least one floodplain wetland. However, e axis may be correlated to wetland type, as only two of the forested strands (FS5 and FS7) e ifferences in the macrophyte commu composition between wetland types ersus floodplain) detected with the N udy sha This correlates to 43% identif t f the 1 cies identified (161 species) meanin FF1 FF10 FF11 FF12 FF13 FF14 FF2 FF3 FF4 FF5 FF6 FF7 FS5 FF8 FF9 FS1 FS10 FS2 FS3 FS4 FS6 FS7 FS8 FS9 Axis 1: 53.3% A xis 2: 29.2 % Ecoregions (Lane 2000)SouthCentralNorthPanhandle

PAGE 57

46 c onsidering that 126 species (45% of the entire 283 species identified for this study) were speis not as significant s some other unmeasured environmental variable. A meaningful grouping of wetlands based ori land use classificationparen Metricifilatedeation coeffic p<0.05) were senclusrelimCI for foresteodplain wetlan8). Medn were the proportion tolerant indicator spe propivepecies (SEN); Floristic Quality Assessment ); pric species (EX); and proporive perennial species roantipecies and procreasing development intensity; whereas, the pro FQAI, and proportion native perennial species decreascape development intensity. All metrics significantly (p < .05) differentiated between low (LDI<2.0) and high (LDI2.0) 1995 LDI_F/wo_200m roups (Table 3-9). ad the highest proport only sampled at one of the 24 wetlands included in this study, perhaps the dissimilarity of cies composition among wetland type (strand versus floodplain) a n a prio was not ap t. Selection Five metrics that were sign cantly corre with LDI (Sp rmans correla ient |r|0.50, lected for i ion in the p ina ry FW d strand and flo ds (Table 3-cies (TOL); etrics selectortion sensit for in indi clusiocator s Index (FQAI oportion exot tion nat (NP). The p portion of toler ind cator s portion exotic species increased with in portion sensitive indicator species, ed with increased lands 0 g Tolerance metrics Tolerant and sensitive indicator species were determined statewide using Indicator Species Analysis ( PCORD ). Table 3-10 provides a list of 19 tolerant indicator species. Figure 3-6 shows a scatter plot of proportion tolerant indicator species versus 1995 LDI_F/wo_200m. The proportion tolerant indicator species increased with increasing development intensity. FF10 (an urban floodplain forest in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ) h ion tolerant indicator species (0.29), followed by FS5 (0.28) and FF7 (0.25), which were both in the central wetland region ( Lane 2000 ) or peninsula bioregion ( Griffith et al. 1994 ). Two wetlands in the peninsula bioregion ( Griffith et al. 1994 ) had less than 0.05 proportion tolerant indicator species, including FS10 a reference strand in the south wetland region ( Lane 2000 ) and FF2 a reference floodplain in the north wetland region ( Lane 2000 ). Table 3-8. Spearmans correlation coefficients for macrophyte metrics and FWCI with 1995 LDI_F/wo_200m. Metric Spearmans r p-value Proportion tolerant indicator species 0.65 0.001 Proportion sensitive indicator species -0.83 <0.001 FQAI score -0.50 0.012 Proportion exotic species 0.54 0.006 Proportion native perennial species -0.56 0.004 FWCI -0.75 <0.001

PAGE 58

47 Table 3-9. Comparisons among five macrophyte metrics and the FWCI between low (LDI < 2.0) and high (LDI 2.0) LDI groups (LDI_F/wo_200m). Metric Low LDIHigh LDI W^p` Tolerant indicator species 0.11 0.060.19 0.06 115.5 0.007 Sensitive indicator species 0.15 0.080.05 0.05 14 0.001FQAI score 4.72 0.684.10 0.64 34 0.036Exotic species 0.03 0.040.10 0.08 109 0.022Native perennial species 0.95 0.06 0.86 0.11 34 0.036FWCI 38.33 7.9121.54 12.00 16 0.001 Values represent the mean standard deviation W^ = the Mann-Whitney U-Test statistic p` = the significance value Table 3-11 provides a list of the 16 sensitive indicator species. Figure 3-7 shows that the proportion sensitive indicator species decreased with increasing development intensity. The wetland with the highest proportion sensitive indicator was FF2 (0.38; a reference wetland), followed by and FF9 (0.18; an urban wetland) and FS7 (0.17; a reference wetland). Three wetlands had no sensitive indicator species including two in the central wetland region (Lane 2000) or peninsula bioregion (Griffith et al. 1994) (the urban strand FS5 and the agricultural strand FS3) and one wetland in the south wetland region (Lane 2000) or Everglades bioregion (Griffith et al. 1994) (FF10). Shrub and tree species were included in the ISA for both tolerant and sensitive metrics. Metrics developed based on the macrophyte community composition included woody species rooted within the sampling quadrats, as structure was thought to play an important role in the biological condition of flowing water forested wetlands. Excluding the tree and shrub layers would seemingly underscore the importance of these woody species. In fact, tree and shrub species comprised 52.6% of the tolerant and 62.5% of the sensitive indicator species lists (Tables 3-10 and 3-11). The five tolerant indicator tree species were Acer rubrum (red maple), Ilex cassine (dahoon holly), Liquidambar styraciflua (sweetgum), Nyssa sylvatica var. biflora (swamp tupelo), and Prunus caroliniana (Carolina laurelcherry) (Table 3-10); the five tolerant indicator shrub species were Callicarpa americana (American beautyberry), Cephalanthus occidentalis (buttonbush), Ludwigia peruviana (water-primrose), Sambucus canadensis (elderberry), and Viburnum obovatum (Walter viburnum). Vines also comprised a high percentage of the tolerant indicator species list, including Berchemia scandens (rattan vine), Smilax auriculata (earleaf greenbrier), Smilax tamnoides (bristly greenbrier), and Toxicodendron radicans (Eastern poison ivy). Of the sensitive indicator species, 43.8% were shrubs, 18.8% trees, 18.8% herbs, 12.5% graminoids, 6.3% ferns, and 0% vines (Table 3-11). The three sensitive indicator tree species included Fraxinus caroliniana (Carolina ash), Persea palustris (swamp bay) and Pinus elliottii (slash pine). Seven shrub species were categorized as sensitive indicator species, including Agarista populifolia (Florida hobble-bush), Hypericum hypericoides (St. Andrews cross), Ilex coriacea (bay-gall holly), Lyonia lucida (fetter

PAGE 59

Table 3-10. Tolerant indicator species for forested strand and floodplain wetlands. Botanical Name Common Name LDI Break (Indicator Value, p-value) Growth Form Acer rubrum Red Maple 1.0 (81.8, 0.057) Tree 1.5 (90,0.005) Berchemia scandens Rattan Vine 3.5 (58.3, 0.062) Vine Bidens alba Hairy Beggar-Ticks 3.5 (54.9, 0.097) Herb Callicarpa americana American Beautyberry 3.5 (75, 0.053) Shrub Cephalanthus occidentalis Buttonbush 1.0 (77.3,0.071) Shrub 1.5 (61,0.072) Commelina diffusa Dayflower 2.0 (50.6,0.034) Herb 3.0 (67.3,0.015) Hydrocotyle verticillata Pennywort 3.0 (43.7,0.068) Herb Ilex cassine Dahoon Holly 1.5 (55, 0.095) Tree Liquidambar styraciflua Sweetgum 1.5 (60,0.092) Tree Ludwigia peruviana Water-Primrose 3.5 (54.9, 0.087) Shrub Nyssa sylvatica var. biflora Swamp Tupelo 1.5 (60, 0.092) Tree Panicum rigidulum Red-Top Panicum 3.0 (47.4, 0.031) Graminoid Prunus caroliniana Carolina Laurelcherry 2.5 (27.3, 0.092) Tree Rhynchospora miliacea Millet Beakrush 3.0 (43.7, 0.072) Graminoid Sambucus canadensis Elderberry 2.5 (27.3,0.09) Shrub Smilax auriculata Earleaf Greenbrier 3.5 (72.4, 0.086) Vine Smilax tamnoides Bristly Greenbrier 3.0 (37.7, 0.051) Vine Toxicodendron radicans Eastern Poison Ivy 1.0 (77.3, 0.07) Vine 1.5 (61,0.068) 2.0 (57.3,0.069) Viburnum obovatum Walter viburnum 3.5 (58.3, 0.052) Shrub

PAGE 60

49 0.00.10.20.302468101995 LDI_F/wo_200mProportion Tolerant Indicator Species Floodplains Strands Figure 3-6. The proportion of tolerant indicator species at wetlands increased with increasing development intensity (LDI). bush), Rhododendron viscosum (swamp azalea), Vaccinium arboreum (sparkleberry), and Vaccinium corymbosum (highbush blueberry). The two sensitive indicator graminoid species were Cladium jamaicense (saw-grass) and Panicum hemitomon (maidencane). Floristic Quality Assessment Index metric Wetland FQAI scores decreased with increasing 1995 LDI_F/wo ( Figure 3-8 ). Of the 283 species identified in the flowing water wetlands, 17 were assigned coefficient of conservatism (CC) scores of zero, and 11 additional species received scores less than 1.0. Of the species receiving a zero CC score, nine (53%) were listed as Category I invasive exotics, and two (12%) were listed as Category II invasive exotics ( EPPC 2003 ). [The rankings of Category I or II invasive exotics are from the Florida Exotic Pest Plant Council (EPPC), which focuses on identifying exotic pest species. Category I species include those invasive exotic species considered responsible for changes to native plant communities through the displacement of natives, changes in community structure or ecological functions, and hybridizing with natives, based on documented ecological damage ( EPPC 2003 ). Category II species have not yet altered native plant communities, but have increased in abundance or frequency and may become ranked as Category I with confirmed ecological damage ( EPPC 2003 ).] The species with the highest CC score was Pieris phyllyreifolia (climbing fetter-bush) (9.5), followed by Panicum abscissum (cut-throat grass) (9.22), Taxodium ascendens (pond cypress) (8.8), Asplenium heterochroum (bicolored spleenwort) (CC = 8.5), and Pinckneya bracteata (fever-tree) (8.3). A complete list of CC scores is available in Appendix B Wetland FQAI scores were significantly correlated with LDI

PAGE 61

Table 3-11. Sensitive indicator species for forested strand and floodplain wetlands. Botanical Name Common Name LDI Break (Indicator Value, p-value) Growth Form Agarista populifolia Florida Hobble-Bush 2.5 (30.8, 0.092) Shrub Centella asiatica Coinwort 2.5 (38.6,0.082) Herb Cladium jamaicense Saw-Grass 1.5 (45.5,0.059) Graminoid Eupatorium capillifolium Dog Fennel 1.0 (88, 0.034) Herb Fraxinus caroliniana Carolina Ash 2.5 (38.6, 0.064) Tree Hypericum hypericoides St. Andrew's Cross 2.5 (30.8, 0.095) Shrub Ilex coriacea Bay-Gall Holly 2.5 (38.5, 0.037) Shrub Lyonia lucida Fetter-Bush 2.5 (53.6,0.017) Shrub 3.0 (52.9,0.054) Panicum hemitomon Maidencane 2.0 (38.6,0.042) Graminoid Persea palustris Swamp Bay 2.5 (61.2, 0.006) Tree Phlebodium aureum Golden Polypody 1.0 (95.7, 0.008) Fern 1.5 (45.5,0.065) Pinus elliottii Slash Pine 1.0 (95.7, 0.008) Tree 1.5 (45.5,0.065) Rhododendron viscosum Swamp Azalea 2.5 (38.6, 0.081) Shrub Saururus cernuus Lizard's Tail 3.0 (64.7, 0.01) Herb Vaccinium arboreum Sparkleberry 2.5 (30.8,0.098) Shrub Vaccinium corymbosum Highbush Blueberry 3.0 (52.9, 0.038) Shrub

PAGE 62

51 0.00.10.20.30.402468101995 LDI_F/wo_200mProportion Sensitive Indicator Species Floodplains Strands Figure 3-7. The proportion sensitive indicator species at wetlands decreased with increasing development intensity (LDI). 012345602468101995 LDI_F/wo_200mFQAI Score Floodplains Strands Figure 3-8. FQAI scores decreased with increasing landscape development intensity.

PAGE 63

52 gradient (|r| = 0.50; p = 0.012; Table 3-8 ). When wetlands were divided into two groups based on 1995 LDI_F/wo_200m (LDI < 2.0 and LDI 2.0), there was a significant difference between FQAI scores (U = 34; p < 0.05) ( Table 3-9 ). The range of wetland FQAI scores was 3.0, though the scale of species CC scores ranged from 0.0-9.5. The wetland with the highest FQAI was FF2 (5.6), a reference floodplain forest in the north ( Lane 2000 ) or peninsula stream ( Griffith et al. 1994 ) ecoregions. The wetland receiving the lowest FQAI score was FF10 (2.7), an urban floodplain forest in southwest Florida in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ). Three wetlands in the low LDI group with low FQAI scores included FS7 (4.3), FS10 (3.9), and FF6 (3.4). Sixty-seven percent of the wetlands in the low LDI group had an FQAI score greater than 4.5; whereas 73% of wetlands in the high LDI group had an FQAI score less than 4.5. Exotic species metric The proportion of exotic species at a wetland was significantly correlated with the gradient of development intensity in the surrounding landscape (|r| = 0.54, p = 0.006; Table 3-8 ). Figure 3-9 shows that the proportion of exotic species increased with increasing LDI. The south wetland region ( Lane 2000 ; or Everglades bioregion from Griffith et al. 1994 ) hosted the wetland with the greatest proportion exotic species, FF10 (0.29), which also received the lowest FQAI score (2.7). The wetland with the second highest proportion exotic species was FS3 (0.18), an agricultural strand surrounded by cattle pasture. Six floodplain wetlands had no exotic species present, including two reference wetlands (FF2 and FF7), two agricultural wetlands (FF4 and FF11), and two urban wetlands (FF5 and FF9). The proportion of exotic species at wetlands was significantly different between low and high LDI groups (W = 109, p = 0.022; Table 3-9 ). Table 3-12 lists the 35 exotic species encountered throughout the forested strand and floodplain wetlands. The most common exotic species was the herbaceous species Commelina diffusa (dayflower) found at 10 (42%) of the 24 forested wetlands, which was also categorized as a tolerant indicator species ( Table 3-10 ). Sixty-nine percent of the exotic species (n=24) were found at only one forested wetland. Thirty-seven percent of the species (n=13) were listed as Category I invasive exotic species, and 14% (n=5) were listed as Category II invasive exotic species ( EPPC 2003 ). Forty-three percent of the exotic species were herbs, 20% vines, 14% shrubs, 11% trees, 9% graminoids, and 3% ferns. Native perennial species metric Of the 283 macrophyte species identified, 234 (83%) were classified as native perennials. Figure 3-10 shows that the proportion of native perennial species at a wetland decreased with increasing development intensity. The native perennial species metric was significantly correlated with LDI (Spearman |r| = 0.56, p = 0.004; Table 3-8 ); and there was a significant difference between the proportion native perennial species at low and high LDI group wetlands (W = 34, p = 0.036; Table 3-9 ). FF10 had the lowest proportion native perennial species (0.60). Four wetlands hosted entirely native perennial species, including one reference wetland (FF2), two agricultural wetlands (FF4 and FF11), and one urban wetland (FF9).

PAGE 64

53 0.00.10.20.302468101995 LDI_F/wo_200mProportion Exotic Species Floodplains Strands Figure 3-9. The proportion of exotic species at a wetland increased with increasing development intensity. Florida Wetland Condition Index Five metrics were included in the preliminary FWCI for forested strand and floodplain wetlands, including proportion tolerant indicator species, proportion sensitive indicator species, FQAI score, proportion exotic species, and proportion native perennial species. Metrics were natural log transformed to improve normality. Metric scoring was done on a continuous scale spreading between the 5 th to 95 th percentiles of the values for the sampled wetlands. Scores for each metric were then added together to create the preliminary forested strand and floodplain FWCI with a scale of 0-50, with 50 representing the reference condition of high biological integrity. Appendix D provides metric scoring criteria. Figure 3-11 shows that FWCI scores decreased with increasing development intensity. Correlations between macrophyte metrics and FWCI with LDI were significant (p < 0.05) for all of the metrics and the FWCI (|r| = 0.75, p < 0.001) ( Table 3-8 ). One wetland, FF2 a reference forested floodplain wetland, received a perfect 50 on the FWCI scale, also receiving a perfect 10 score for all five metrics. One wetland scored the lowest potential score of zero, FF10 which was an urban forested floodplain wetland in the south wetland region ( Lane 2000 ) or Everglades bioregion ( Griffith et al. 1994 ). All of the wetlands in the low LDI group scored above the midpoint of 25 on the FWCI scale; while 64% of wetlands in the high LDI group scored below 25.

PAGE 65

54 Table 3-12. Exotic species identified at 24 forested strand and floodplain wetlands. Botanical Name Common Name EPPC Growth Form Abrus precatorius Rosary Pea I Vine Alternanthera philoxeroides Alligator Weed II Herb Alternanthera sessilis Sessile Alligator Weed Herb Begonia cucullata Wax Begonia II Herb Cinnamomum camphora Camphor Tree I Tree Colocasia esculenta Elephant's Ear I Herb Commelina communis Asiatic Dayflower Herb Commelina diffusa Dayflower Herb Cuphea carthagenensis Columbia waxweed Herb Cynodon dactylon Bermudagrass Graminoid Cyperus difformis Variable Flatsedge Graminoid Emilia fosbergii Florida Tasselflower Herb Eugenia uniflora Surinam Cherry I Shrub Koelreuteria elegans Flamegold II Tree Ligustrum sinense Chinese Privet I Shrub Lonicera japonica Japanese Honeysuckle I Vine Ludwigia peruviana Water-Primrose Shrub Lygodium japonicum Japanese Climbing Fern I Vine Lygodium microphyllum Small-leaf Climbing Fern I Vine Melaleuca quinquenervia Punk Tree; Melaleuca I Tree Merremia dissecta Alamo Vine Vine Mimosa pigra Black Mimosa I Shrub Momordica charantia Balsampear Vine Oeceoclades maculata Monk Orchid Herb Paspalum notatum Bahiagrass Graminoid Phyllanthus urinaria Water Leafflower Herb Sapium sebiferum Chinese Tallowtree I Tree Schefflera actinophylla Australian Umbrella Tree I Tree Schinus terebinthifolius Brazilian Pepper I Shrub Thelypteris dentata Downy Maiden Fern Fern Trifolium repens White Clover Herb Urena lobata Caesarweed II Herb Wisteria sinensis Chinese Wisteria II Vine Xyris jupicai Richard's Yellow-Eyed-Grass Herb Youngia japonica Oriental False Hawksbeard Herb Exotic Pest Plant Council (EEPC) categories from EEPC ( 2003 ).

PAGE 66

55 0.50.60.70.80.91.002468101995 LDI_F/wo_200mProportion Native Perennial Species Floodplains Strands Figure 3-10. The proportion of native perennial species decreased with increasing development intensity (LDI). 0102030405002468101995 LDI_F/wo_200mFWCI Floodplains Strands Figure 3-11. Forested Wetland Condition Index (FWCI) scores decreased with increasing development intensity (LDI).

PAGE 67

56 Cluster Analysis Cluster analysis determined wetland groupings based on macrophyte community composition using a species by site presence/absence matrix. Based on the lowest average p-value for all species from the randomized Monte Carlo tests used in the ISA analysis, the most ecologically meaningful cluster for dendrogram pruning was at cluster step 5 ( Figure 3-12 ) with the lowest average species p-value of 0.27. The highest number of significant (p < 0.05) indicator species was found at cluster step 3, which had 37 significant indicator species ( Figure 3-13 ). Cluster steps 4 and 5 had the second highest number of significant indicator species at 31. Exploration of the sample sites assigned to the groups at cluster step 3 suggested an association between group membership and spatial distribution of the wetlands throughout Florida. [Note: ISA used to determine the most ecologically meaningful clusters for dendrogram pruning is separate from ISA used in determining tolerant and sensitive indicator species. For ISA used to determine meaningful clusters for dendrogram pruning, group membership was assigned based on the wetland groupings established through cluster analysis. For ISA used to determine sensitive and tolerant indicator species, group membership was assigned based on the calculated LDI score for each wetland.] The groups defined through the agglomerative cluster analysis at cluster step 5 were roughly defined by wetland regions ( Lane 2000 ), bioregions ( Griffith et al. 1994 ), and a priori land use categories, including: Cluster 1: FF2, a north wetland region, peninsula bioregion, reference floodplain; Cluster 2: Nine floodplains, including two in the panhandle, six in the north, and one in the central wetland regions; one in the panhandle, five in the northeast, and three in the peninsula bioregions; and four agricultural and five urban wetlands; Cluster 3: Five strand and two floodplain wetlands; one in the north and six in the central wetland regions; all seven in the peninsula bioregion; four reference, one agricultural, and two urban wetlands; Cluster 4: Three strand and two floodplains, including one in the north and four in the south wetland regions; four in the peninsula and one in the Everglades bioregions; three reference and two urban wetlands; Cluster 5: Two floodplain wetlands with one in the north and one in the central wetland regions; both in the peninsula bioregion; and both agricultural wetlands. Wetland groups based on the third cluster step combined the aforementioned groups from the cluster step 5, so that: Cluster 1: Combines Clusters 1 & 2 Cluster 2: Composed of Cluster 3 Cluster 3: Combines Clusters 4 & 5 It appear that the initial grouping for cluster step three was based on spatial location and that further cluster steps separate out wetlands based on impairment. Figure 3-14 shows that based on wetland groups from cluster step 3, FWCI scores for wetlands in Cluster 1 (37.5 10.5) were significantly different from wetlands in Cluster 2 (24.4 10.0) and Cluster 3 (19.8 13.1) (p < 0.05). Wetland FWCI scores for Cluster 2 and Cluster 3 were not significantly different from one another. Regionalization was apparent in the wetland groupings of Cluster 3 (at cluster step 3); as Cluster 3 was comprised of the all of

PAGE 68

57 0.00.10.20.30.40.50.60.7234567891011121314151617181920Number of ClustersAverage p-value Figure 3-12. Change in average species p-value from the randomized Monte Carlo tests at each step in clustering. The minimum average p-value (0.27) was found at cluster step 5. 010203040234567891011121314151617181920Number of ClustersNumber Significant Indicators Figure 3-13. Change in the number of significant indicator species from the indicator species analysis performed at each step in clustering. The maximum number of significant indicator species (37) was found at cluster step 3 (p < 0.05), followed by 31 significant indicator species at cluster steps 4 and 5.

PAGE 69

58 Cluster 3Cluster 2Cluster 1 50403020100 FWCI a b b Figure 3-14. FWCI scores for three wetland clusters based on macrophyte community composition. Boxes represent 25 th and 75 th quartiles, lines represent median FWCI scores per cluster, dots represent the mean, and vertical lines represent the range. Clusters with similar letters were not significantly different (p<0.05). the south wetland region wetlands ( Lane 2000 ), including the only wetland sampled in the Everglades bioregion ( Griffith et al. 1994 ). Table 3-13 provides means and standard deviations for cluster FWCI and LDI_F/wo_200m scores. LDI_F/wo_200m scores were not significantly different among clusters. Landscape Development Intensity Index and the Florida Wetland Condition Index Nearly all of the correlation (110 of 120, or 92%) between the five macrophyte metrics and the FWCI with the 20 variations of LDI were significantly correlated at the flexible p < 0.10 level ( Table 3-14 ). Additionally, nearly half of the comparisons (58 of 120, or 48.3%) were significantly correlated at the strictest significance level (p < 0.01). Three of the LDI calculations (1995 LDI_F/w_200m, 2000 LDI_F/w_200m, and 2000 LDI_F/wo_200m) were significant at the strictest significance level (p < 0.01) for all five metrics and the FWCI, which should not be surprising given that metrics were selected for inclusion in the FWCI based on correlations with the 1995 LDI_F/wo_200m (using Spearman correlation with LDI, visually distinguishable and ecologically meaningful pattern when graphed with LDI, and differentiation among LDI groups with the Mann Whitney U-test). The proportion tolerant indicator species metric was significantly correlated with 1995 and 2000 LDI_F (wetland feature scale) calculations (p < 0.01). However, the proportion tolerant indicator species metric was not significantly correlated at the strictest significance level (p < 0.01) for any of the 1995 or 2000 LDI_T (transect scale)

PAGE 70

59 Table 3-13. FWCI scores and LDI values for wetland clusters (at the third cluster step) based on macrophyte community composition. Wetland Clusters FWCI LDI_F/wo_200m 1 37.5 10.5 a 2.2 0.7 a 2 24.4 10.0 b 2.5 0.7 a 3 19.8 13.1 b 2.9 2.1 a Clusters with similar letters within columns were not significantly different (p < 0.05). LDI_F/wo_200m represents the LDI calculated at the feature scale excluding the wetland area within a 200 m buffer. calculations. The proportion sensitive indicator species metric was more strongly correlated with the LDI calculations at the transect and feature scales (11 of 12 correlations with |r| 0.63, p < 0.01, remaining one at |r| = 0.57, p < 0.05) than the watershed scales (1995 and 2000 LDI_WS_DW_exp at p < 0.01; 2000 LDI_WS_DW_lin at p < 0.05; 2000 LDI_WS_ED/wo not significant; remaining four LDI_WS correlations at p < 0.10). The FQAI score metric was not significantly correlated with three of the 1995 transect level LDI calculations, but was strongly (|r| > 0.66) and significantly (p < 0.01) correlated with all of the 1995 and 2000 watershed scale LDI calculations. Watershed scale LDI calculations were completed for floodplain forested wetlands only, leaving speculation as to whether these results suggest that FQAI scoring reflect landscape level anthropogenic activity (e.g. exotic species with low CC scores entering a system due to anthropogenic activities) or whether FQAI scoring was biased towards larger flowing water systems such as floodplain forests. The proportion exotic species and proportion native perennial species metrics were more strongly correlated with the feature scale LDI calculations than transect or watershed scales. In fact, three watershed 1995 LDI calculations were not significant with the proportion exotic species or the proportion native perennial species metrics, including 1995 LDI_WS_ED/w, 1995 LDI_WS_ED/wo, and 1995 LDI_WS_DW_lin. The strongest correlation between the FWCI and LDI was with the 2000 LDI_WS_DW_exp (|r| = 0.95; p < 0.01), though the correlation with 2000 LDI_F/wo_200m (|r| = 0.94; p < 0.01) was also remarkably strong. The two correlations with the highest Spearmans correlation coefficients were between the proportion tolerant indicator species metric and the 2000 LDI_WS_DW_exp and between the FQAI score metric and the 2000 LDI_WS_DW_lin (|r| = 0.96; p < 0.01).

PAGE 71

Table 3-14. Correlations of metrics and FWCI scores with 20 variations of the LDI index. Differences in LDI calculations include 1995 or 2000 land use; transect (T), feature (F), or watershed (WS) scale buffers; including (/w) or excluding (/wo) the wetland area; 100 (100m) or 200 meter (200m) buffers; equal distance (ED) or distance weighted (DW); and linear (lin) or exponential (exp) weighting. Column headings refer to the five metrics and the FWCI: proportion tolerant indicator species (TOL), proportion sensitive indicator species (SEN), Floristic Quality Assessment Index (FQAI), proportion exotic species (EX), and proportion native perennial species (NP). LDI n= TOL SEN FQAI EX NP FWCI 1995 LDI_T/w_100m 24 0.44 ** -0.68 -0.39 *** 0.51 ** -0.48 ** -0.58 1995 LDI_T/wo_100m 24 0.41 ** -0.65 ns 0.46 ** -0.43 ** -0.52 1995 LDI_T/w_200m 24 0.38 *** -0.63 ns 0.49 ** -0.44 ** -0.49 ** 1995 LDI_T/wo_200m 24 0.45 ** -0.69 ns 0.51 ** -0.44 ** -0.52 ** 1995 LDI_F/w_200m 24 0.72 -0.90 -0.57 0.61 -0.63 -0.79 1995 LDI_F/wo_200m 24 0.66 -0.83 -0.50 ** 0.54 -0.55 -0.75 1995 LDI_WS_ED/w 14 0.60 ** -0.49 *** -0.71 ns -0.50 *** -0.57 ** 1995 LDI_WS_ED/wo 14 0.59 ** -0.49 *** -0.67 ns ns -0.53 ** 1995 LDI_WS_DW_lin 14 0.58 ** -0.49 *** -0.69 ns ns -0.53 *** 1995 LDI_WS_DW_exp 14 0.80 -0.72 -0.75 0.60 ** -0.65 ** -0.70 2000 LDI_T/w_100m 24 0.51 ** -0.72 -0.49 ** 0.48 ** -0.54 -0.63 2000 LDI_T/wo_100m 24 0.51 ** -0.72 -0.42 ** 0.44 ** -0.49 ** -0.59 2000 LDI_T/w_200m 13 0.54 *** -0.69 -0.74 0.83 -0.85 -0.81 2000 LDI_T/wo_200m 13 ns -0.57 ** -0.66 ** 0.73 -0.77 -0.72 2000 LDI_F/w_200m 13 0.78 -0.93 -0.88 0.79 -0.85 -0.91 2000 LDI_F/wo_200m 13 0.82 -0.92 -0.91 0.82 -0.84 -0.94 2000 LDI_WS_ED/w 8 0.73 ** -0.68 *** -0.95 0.77 ** -0.83 ** -0.88 2000 LDI_WS_ED/wo 8 0.65 *** ns -0.86 0.73 ** -0.69 *** -0.78 ** 2000 LDI_WS_DW_lin 8 0.83 ** -0.80 ** -0.96 0.77 ** -0.81 ** -0.92 2000 LDI_WS_DW_exp 8 0.96 -0.88 -0.88 0.83 ** -0.78 ** -0.95 *p<0.01 **p<0.05 ***p<0.10 ns not significant

PAGE 72

61 Human Disturbance Gradient and Stream Condition Index Thirteen of the forested floodplain wetlands had Human Disturbance Gradient (HDG) and Stream Condition Index (SCI) scores for sampling events taken within the channelized flow of the forested floodplain wetlands ( Table 3-15 ). Sample dates for the FWCI and LDI were during 2003 ( Table 2-1 ); sample dates for the HDG and SCI ranged from 1995-2000 ( Table 3-15 ). Table 3-15 provides storet identification numbers, HDG and SCI sample dates, a priori land use categories, bioregions ( Griffith et al. 1994 ), wetland regions ( Lane 2000 ), FWCI, LDI_F/wo_200m scores, HDG scores, and SCI scores for the 13 forested floodplain wetlands. Data for the most recent and complete sampling event for each wetland, as presented in Table 3-15 was selected for correlation purposes. The HDG was significantly correlated with the SCI (|r| = 0.58, p<0.05), LDI_F/wo_200m (|r| = 0.66, p<0.05), and FWCI (|r| = 0.74, p<0.01) ( Table 3-16 ). However, the SCI was not significantly correlated with either the LDI_F/wo_200m or the FWCI. The correlation between the LDI_F/wo_200m and FWCI (|r| = 0.68, p<0.05) ( Table 3-16 ) for the 13 forested floodplain wetlands (only those with HDG and SCI scores), was weaker than the correlation between them for the entire data set of 24 forested flowing water wetlands (|r| = 0.75, p<0.001) ( Table 3-8 ).

PAGE 73

Table 3-15. Forested Wetland Condition Index (FWCI), Landscape Development Intensity Index (LDI_F/wo_200m), Human Disturbance Gradient (HDG), and Stream Condition Index (SCI) data for 13 forested floodplain wetlands. Site STORET ID HDG/SCI Date Land Use Bioregion Wetland Region FWCI LDI_F/wo_200m HDG SCI FF1 20020404 8/15/2000 U Peninsula North 39.1 2.0 0 80 FF2 20010454 7/19/1999 R Peninsula North 50.0 1.2 0 70 FF3 27010580 8/22/2000 U Peninsula North 16.4 3.7 2 20 FF4 20010455 2/23/1999 A Peninsula Central 43.6 2.3 0 75 FF5 19010099 2/1/2000 U Northeast North 44.0 1.9 0 55 FF6 26011019 3/8/1999 R Peninsula Central 26.2 1.4 0 80 FF7 26011020 7/20/1999 R Peninsula Central 28.2 2.5 0 90 FF8 19020027 7/15/1996 U Northeast North 23.6 2.5 1 70 FF9 20030246 4/17/1997 U Northeast North 46.6 1.8 0 95 FF10 28020234 10/9/2000 U Everglades South 0.0 7.1 4 50 FF11 21010032 1/17/1995 A Northeast Panhandle 42.0 1.6 0 25 FF12 19010072 3/22/1999 A Northeast North 33.5 2.8 0 85 FF14 32010024 7/14/1998 A Panhandle Panhandle 36.5 1.8 0 80 STORET ID refers to Florida Department of Environmental Protection (FDEP) database ID. HDG/SCI Date refers to original sample date corresponding to HDG and SCI data. FWCI and LDI were calculated for the 2003 site visit. Land Use refers to a priori land use category (R-reference, A-agricultural, U-urban) Bioregion from Griffith et al ( 1994 ) Wetland Region from Lane ( 2000 ) LDI_F/wo_200m refers to the LDI calculated at the feature scale, excluding the wetland area, within a 200 m buffer.

PAGE 74

63 Table 3-16. Correlations among four measures of ecosystem condition or anthropogenic activity, including the Human Disturbance Gradient (HDG), Stream Condition Index (SCI), Landscape Development Intensity Index (1995 LDI_F/wo_200m), and the Florida Wetland Condition Index (FWCI) for freshwater forested floodplain wetlands. Values are Spearmans rank correlation coefficients. HDG SCI LDI_F SCI -0.58 ** LDI_F 0.66 ** ns FWCI -0.74 ns -0.68 ** *p<0.01, **p<0.05, ns=not significant

PAGE 75

CHAPTER 4 DISCUSSION The primary objective of this research was to develop a preliminary Florida Wetland Condition Index (FWCI) for forested strand and floodplain wetlands. Wetland study sites were sought in various a priori designated land use categories that included undeveloped, agricultural, and urban land uses. The preliminary FWCI provides a quantitative measure of the biological integrity of forested strand and floodplain wetlands in Florida. Comprised of five metrics, the preliminary FWCI was developed based on the community composition of the macrophyte species assemblages ( Table 4-1 ). Metrics were selected for inclusion in the FWCI based on the correlation (nonparametric Spearman correlation coefficient) of each metric with the Landscape Development Intensity (LDI), an independent measure of anthropogenic activity in the landscape calculated for each wetland ( Brown and Vivas 2005 ); based on a metrics visually distinguishable correlation with LDI in a scatter plot; and based on a statistical difference of metric values between low and high LDI groups (Mann-Whitney U-test). The FWCI was composed of individual metrics, which were scaled and added together, creating the preliminary forested strand and floodplain wetland FWCI (0-50 scale), with the highest score of 50 reflecting the highest biological integrity and the lowest score of zero reflecting a lack of biological integrity or no similarity to the reference wetland condition. The contribution of this research to our understanding of changes in the macrophyte community composition of forested strand and floodplain wetlands in relation to differing anthropogenic activities in the surrounding landscape can be summarized in five main points: 1. Five macrophyte based metrics including proportion tolerant indicator species, proportion sensitive indicator species, Floristic Quality Assessment Index (FQAI) score, proportion exotic species, and proportion native perennial species, were useful biological indicators for defining biological integrity for forested strand and floodplain wetland vegetation; 2. Vegetation richness, evenness, and diversity were not sensitive to a priori land use categories or development intensities in the surrounding landscape for forested strand and floodplain wetlands; 3. The Landscape Development Intensity (LDI) index was a useful tool correlating with the measured biological condition of vegetation for forested strand and floodplain wetlands; 4. Regional species lists for metrics would enhance the forested strand and floodplain Florida Wetland Condition Index (FWCI); 5. An FWCI with a set of core metrics could be developed for Florida freshwater wetlands, which includes separate species lists for indicator species by wetland type and ecoregions and separate Floristic Quality Assessment Index (FQAI) scores for species by wetland type. 64

PAGE 76

65 Table 4-1. The five metrics of the preliminary Florida Wetland Condition Index for freshwater forested strand and floodplain wetlands based on the macrophyte species assemblage. FWCI Metrics 1. Proportion Tolerant Indicator Species 2. Proportion Sensitive Indicator Species 3. Floristic Quality Assessment Index Score 4. Proportion Exotic Species 5. Proportion Native Perennial Species Describing Biological Integrity Biological indicators were useful in determining the biological integrity of freshwater forested strand and floodplain wetlands. For the purposes of this study, biological integrity has been defined quantitatively with the FWCI. The FWCI incorporated five metrics from the macrophyte species assemblages ( Table 4-1 ). Strong correlations between the FWCI and the intensity of development in the surrounding landscape (based on the use of nonrenewable energy and calculated with the LDI) suggest that changes in macrophyte community composition quantified as metrics were captured by the LDI. It has been proposed that organisms respond to environmental gradients by colonizing a range of feasible conditions beyond which the organisms fail to persist ( ter Braak 1987 ). By selecting species that occur throughout the range of measurable environmental parameters, the FWCI defined and detected deviations from the condition of reference wetlands based on macrophyte community composition. Each of the FWCI metrics addressed some disparity from the assumed range of feasible conditions. The proportion sensitive indicator species metric showed the strongest correlation with LDI, suggesting that the presence of a suite of taxa characteristic of wetlands with high biological integrity may be the most effective means of identifying changes in macrophyte community composition in freshwater forested strand and floodplain wetlands associated with changes in anthropogenic activities. Richness, Evenness, and Diversity Measures of richness, evenness, and diversity of the macrophyte assemblage were not sensitive to differences in land use or development intensity in the surrounding landscape within the forested strand and floodplain wetlands. Perhaps due to a limited sample size and high variability inherent in the landscape (e.g. regional differences, land use differences, etc.) no statistical test on summary statistics produced statistically significant results. Nevertheless, wetlands in the high LDI group (LDI 2.0) had higher (though not statistically significant) mean species richness and diversity (Shannon diversity, Whittakers beta diversity, and overall beta and gamma diversity). Species evenness and Simpsons diversity were remarkably similar among the sample wetlands; in effect no differences were detectable.

PAGE 77

66 General ecological theory predicts a decrease in plant diversity resulting from an increase in anthropogenic assaults such as grazing ( Blanch and Brock 1994 ; Grace and Jutila 1999 ) and nutrient enrichment ( Bedford et al. 1999 ), though the forested strand and floodplain wetlands displayed the opposite trend. Mitsch and Gosselink ( 1993 ) report that freshwater forested wetlands have low species diversity as compared to other ecosystems, so perchance macrophyte species that entered wetlands in developed landscapes were merely taking advantage of available habitat and in fact increased the overall species diversity. However, Ewel ( 1990 ) notes that due to high topographic and soil variability, river swamps may be the most diverse type swamp in Florida. Clearly there is some uncertainty within the published literature as to anticipated and abnormal diversity for Florida forested strand and floodplain wetlands. Many of the species entering wetlands in developed landscapes were categorized as exotic species, and the increased incidence of exotic species has long been associated with disturbed ecosystems ( Galatowitsch 1999b ; Cronk and Fennessy 2001 ). An increase in the frequency of exotic species has been attributed to drainage and hydrologic alterations ( Hobbs and Heunneke 1992 ; David 1999 ; Galatowitsch et al. 1999b ), increased human development ( Cronk and Fennessy 2001 ), and ecosystem scale alterations such as clear-cut harvests ( Devine 1998 ). Within the study wetlands, the proportion of exotic species increased with increasing development intensity in the surrounding landscape. It appears that the influx of exotic species added to rather than diminished the species richness and diversity within the freshwater forested strand and floodplain wetlands. As such, the presence of exotic species alone may not be an ideal indicator of biological integrity. While many ecological theories have been established suggesting that the presence or occurrence of exotic species increases with anthropogenic disturbance ( Cronk and Fennessy 2001 ; Galatowitsch 1999b ), there was a inconsistent pattern of occurrence of exotic species in the forested strand and floodplain wetlands sampled. For example, six of the freshwater forested floodplain wetlands, including two reference, two agricultural, and two urban wetlands, hosted zero exotic species. That four floodplain wetlands in developed landscapes (including two agricultural and two urban wetlands) hosted no exotic species was somewhat contradictory to the theories of increased exotic species occurrence in disturbed ecosystems. However, for the complete dataset, the trend of an increase in the proportion of exotic species with increasing landscape development intensity held. Some concerns arise considering the discrepancies on the exotic status of some species. For example, not all of the species listed by the Exotic Pest Plant Council (EPPC) as invasive exotics altering native plant communities (Category I) or invasive exotics increasing in frequency or abundance with the potential to alter native plant communities (Category II) ( EPPC 2003 ) received an FQAI score of 0.0 corresponding to species that act as opportunistic invaders, including species that commonly occur in disturbed ecosystems. The disagreement on the status of a species as an invasive exotic translates into disagreement as to the meaning of a particular exotic species occurring within an ecosystem. As a case in point, two exotic species were included as tolerant indicator species including Commelina diffusa (dayflower) and Ludwigia peruviana (water-primrose). Neither of the two tolerant indicator species that are also exotic species was listed as an EPPC Category I or II species. In fact, none of the 13 Category I

PAGE 78

67 invasive exotic species identified in the forested strand and floodplain wetlands was categorized as a tolerant indicator species. The intent of this research was to assess the current biological integrity of Florida freshwater forested flowing water wetlands in order to develop a quantitative wetland condition index. As such the comparison was made against the current day reference standard of biological integrity apparent in the reference strand and floodplain wetlands sampled. While we may innately believe that reference wetlands host no exotic species, we found that five of nine wetlands surrounded by low development intensity (56%) hosted at least one exotic species. Increasingly, it may become apparent that even reference wetlands with the highest current standard of biological integrity may host some proportion of exotic species. This may be even more apparent in the southern half of the state, where drainage and development have altered nearly all of the Florida landscape. In fact, the south wetland region reference wetland (FS10) had the highest proportion exotic species of all wetlands in the low LDI group by nearly 10%. Consequently, establishing a baseline for the reference condition of biological integrity is crucial for the application of indices of biotic integrity. Consensus should be reached as to whether scientists proceed by establishing a present day baseline for future assessments or by maintaining and updating a moving baseline for future application of the FWCI. Implications for both methods are complex. Measuring Anthropogenic Activity The variable sensitivities of three different independently derived indices compared to the forested strand and floodplain FWCI, including the Landscape Development Intensity index (LDI; Lane et al. 2003 ; Brown and Vivas 2005 ), the Human Disturbance Gradient (HDG; Fore 2004 ), and the Stream Condition Index (SCI; Fore 2004 ), suggest that multiple measures of biological integrity may be more effective at describing ecosystem wide biological integrity than any single measure based on an individual species assemblage or surrounding land use activity. Measurements of anthropogenic activity such as the LDI and HDG seek to describe ecosystem biological integrity from the perspective of outside anthropogenic influences which act to alter an ecosystem. While the LDI was used as a remote based measure of human development intensity, the HDG used both remote and local conditions integrating four components ranging from the in-stream chemical water quality, habitat structure, and hydrologic alteration, to a remote landscape assessment (using a form of LDI). For the HDG each of the four components was given equal weighting in determining the influence of anthropogenic activities on the biological integrity of an ecosystem. The LDI and HDG were strongly correlated, suggesting that the LDI alone may capture the influence from human development intensity on ecosystem biological integrity without the need for additional sampling and sample processing associated with in-stream chemical water quality, habitat structure, and hydrologic alteration. The FWCI was strongly correlated with both the HDG and the LDI, though the significance of correlation was slightly stronger with the HDG than the LDI. The caveat here was that only three of the thirteen floodplain wetlands had an HDG greater than zero, and because these scores were tested using the non-parametric Spearmans rank correlation coefficient test the strong correlations may simply be a factor of zero HDG

PAGE 79

68 values. In fact, our findings of strong correlations of LDI and HDG must be interpreted with caution because so few of the wetlands received HDG values greater than zero (an HDG score of zero represents no detectable human induced disturbance). This was likely due to the fact that the HDG and SCI were calculated in the channelized water course, whereas the LDI and FWCI were calculated for the surrounding floodplain forest. Many of the streams with low SCI and high HDG scores (considered those with low biological integrity) were not sampled for the FWCI and LDI as no floodplain forest remained, with the banks of the channelized water course being either sodded and mowed or paved. Therefore we should cautiously interpret correlations and discrepancies between the SCI (no significant correlations with FWCI or LDI) and HDG with the FWCI and LDI because the range of impaired conditions used to develop the FWCI was much narrower than that used to develop the SCI. The lack of correlations of SCI with both FWCI and LDI, suggest that in-stream macroinvertebrate based measures of biological condition and surrounding forested wetland macrophyte based measures of biological condition did not respond in a consistent manner to changes in anthropogenic activity. Using both the in-stream macroinvertebrate SCI biological assessment and the surrounding wetland macrophyte FWCI biological assessment methods may provide a more complete picture of the overall condition of a wetland and associated stream at a particular spatial location. While agreement in the ranking of the biological condition of study wetlands using the FWCI and SCI was anticipated, discrepancies among the ranking from the different assemblages may provide great insight into biological condition as different species assemblages respond to changes in anthropogenic activities and the associated changes in inflows (e.g. nutrient enrichment) over different time scales. Additionally, use of the forested strand and floodplain FWCI may lead to specific conclusions as to the biological condition of local or nearby anthropogenic activity, while use of the SCI may enhance understanding of larger watershed scale influences from anthropogenic activity (i.e. due to the convergence of surface water within the watershed associated with stream flow). By correlating the community composition based FWCI and the landscape based LDI, we were better able to understand anthropogenic influence on biological integrity on forested wetland ecosystems. We found that the LDI index was a useful tool in approximating biological integrity. Its primary power was in the reproducible, objective, and quantitative methods employed to obtain a score based on the use on non-renewable energy in the surrounding landscape. A second strength of the LDI index was apparent in the practical application of a remote GIS based method of describing ecosystem condition as a starting point to identify potential areas for further biological sampling. Regionalization of the Florida Wetland Condition Index Pronounced differences in the local climate across Florida, such as differences in the amount of annual rainfall, seasonal maximum temperatures, and number of freeze days ( Fernald and Purdum 1992 ; Lane 2000 ) and the broad latitudinal and longitudinal ranges of sample wetlands (26.3N -30.8N latitude, 80.1W-82.1W longitude), suggest differences would be found among macrophyte community composition in Florida wetlands. In fact, compositional differences of the macrophyte species assemblages were

PAGE 80

69 found among Lane s ( 2000 ) Florida wetland regions. However, the study sample size limited the development of regional indicator species and metric scoring criteria. Wetland clusters based on macrophyte community composition (species presence by site) roughly correlated with wetland spatial distribution throughout the state. The distribution of wetlands in the cluster groups suggests that wetlands located in the northern area of Florida may have higher biological integrity than wetlands in the central or southern peninsula given a statewide scoring approach of the preliminary forested strand and floodplain FWCI. Accordingly, most of the human development in Florida has occurred along the east and west coastal areas of peninsular Florida ( Fernald and Purdum 1992 ), suggesting that while the reference wetlands selected in the south and central wetland regions were selected as the best possible examples of reference type conditions, they may be more affected by development in the surrounding landscape (such as compounded secondary effects) than their panhandle and north wetland region reference counterparts. While the ease and utility of a single statewide FWCI would seemingly prevail over regional indicator species and metric scoring criteria, the necessity of scoring each wetland region based on the best possible reference conditions cannot be overlooked (as suggested by Karr and Chu 1999 ). Regionalization of biological indices has been suggested throughout the literature. One of the main premises behind indices of biological integrity is a comparison of like to like ( Gerristen et al. 2000 ), that is, to reduce the noise in background variability in biological data, which could be accomplished through regionally based indices. The lower FWCI scores for wetlands in the central and southern peninsula may also be a factor of the smaller sample size for wetlands located in those wetland regions. Since the macrophyte community composition was found to be different between wetland regions, wetlands in the south and central wetland regions were underrepresented and therefore had less influence in the indicator species analysis and the metric scoring criteria. A larger sample size would improve uncertainty related to questions surrounding the biological integrity of wetlands in the south and central wetland regions. Florida Wetland Condition Index Independent of Wetland Type Recent works by Lane et al. ( 2003 ) and Reiss and Brown ( 2005 ) presented a five metric FWCI for isolated depressional herbaceous wetlands and a six metric FWCI for isolated depressional forested wetlands in Florida based on the community composition of the macrophyte assemblage. The depressional herbaceous FWCI was created based on a sample size of 75 freshwater marshes surrounded by reference (n=34) and agricultural (n=40) land uses throughout peninsular Florida. The depressional forested FWCI was created based on a sample size of 118 freshwater wetlands surrounded by reference (n=37), agricultural (n=40) and urban (n=41) land uses throughout the extent of Florida. The five metrics of the preliminary FWCI for forested strand and floodplain wetlands were nearly identical to the five metrics of the depressional herbaceous FWCI (with an adaptation from annual to perennial ratio from the depressional herbaceous FWCI changed to the proportion native perennial species for the depressional forested and forested strand and floodplain FWCIs) and similar to five of the depressional forested FWCI metrics.

PAGE 81

70 The five macrophyte metrics included in the three wetland type FWCIs (depressional herbaceous, depressional forested, forested strand and floodplain) were tolerant and sensitive indicator species, FQAI score, exotic species, and native perennial species (annual to perennial ratio in the depressional herbaceous FWCI). Tolerant and sensitive indicator species lists were constructed independently for each wetland type. Of the 16 tolerant indicator species for flowing water wetlands, five also occurred as tolerant indicator species for depressional forested wetlands ( Reiss and Brown 2005 ). Two of those species, Commelina diffusa (dayflower) and Ludwigia peruviana (water-primrose) also occurred as tolerant indicator species for depressional herbaceous wetland ( Lane et al. 2003 ). The additional three overlapping tolerant indicator species between the flowing water and depressional forested wetlands were Acer rubrum (red maple), Sambucus canadensis (elderberry), and Toxicodendron radicans (Eastern poison ivy). One tolerant indicator species for flowing water wetlands, Nyssa sylvatica var. biflora (swamp tupelo) was found to be a sensitive indicator species for depressional herbaceous wetlands ( Lane et al. 2003 ); and the tolerant indicator species for forested strand and floodplain wetlands Panicum rigidulum (red-top panicum) was found to be a sensitive indicator species for both depressional herbaceous and forested wetlands. Clearly a larger sample size and refinement of indicator species analysis for the forested strand and floodplain wetlands FWCI could reduce the inconsistencies among indicator species. Similarly, four sensitive indicator species were common among depressional herbaceous, depressional forested, and forested strand and floodplain wetlands, including Cladium jamaicense (saw-grass), Lyonia lucida (fetter-bush), Panicum hemitomon (maidencane), and Pinus elliottii (slash pine). Vaccinium corymbosum (highbush blueberry) was also considered a sensitive indicator species for both depressional herbaceous and forested strand and floodplain wetlands. However, three sensitive indicator species for forested strand and floodplain wetlands were listed as tolerant indicator species for depressional forested (Saururus cernuus, lizards tail) or depressional herbaceous and forested (Centella asiatica, coinwort; Eupatorium capillifolium, dog fennel) wetlands. The three additional metrics included in the FWCIs were Floristic Quality Assessment Index score, exotic species, and native perennial species (modified from the depressional herbaceous wetland FWCI that used annual to perennial ratio). The variant of the annual to perennial ratio (as native perennial species) was used to account for variable conditions at urban wetlands, which were not included in the study of depressional herbaceous wetlands but were studied in both the depressional forested and forested strand and floodplain wetland research. Wienhold and Van der Valk ( 1989 ) and Ehrenfeld and Schneider ( 1991 ) determined that disturbance often favors annual species over perennial species and promotes the invasion of nonnative perennials in wetlands. Galatowitsch et al. ( 2000 ) found that while native perennial cover was reduced in wetlands impacted by cultivation, the occurrence of introduced perennials rather than annuals increased in stormwater impacted wetlands. The sixth depressional forested FWCI metric was the wetland status species (including both obligate and facultative wetland species), which was not included for the preliminary forested strand and floodplain wetland FWCI as it did not meet selection criteria for inclusion. Overall, the depressional herbaceous, depressional forested, and flowing water forested wetland FWCIs included five similar metrics. Perhaps the strong similarity of

PAGE 82

71 metrics selected for inclusion in the FWCIs suggests that a universal assessment index with core metrics could be constructed regardless of wetland type. However, it would likely be necessary to maintain independent indicator species lists and metric scoring criteria for different wetland types within each wetland region. Limitations and Further Research Generally wetlands were visited only once, with a complete sample effort lasting just one day, which provided a mere snapshot of wetland condition. Visiting these wetlands only once did not allow insight into seasonal or yearly variations in the macrophyte assemblage; and the preliminary forested strand and floodplain FWCI would benefit from inter-seasonal validation. The FWCI would also benefit from validation based on a new set of wetlands to test the repeatability of this index. A larger sample size of wetlands will improve the scoring criteria of the FWCI based on wetland regions for metrics such as indicator species analysis. Funding for additional wetland sampling and FWCI refinement is in the application phase. Regionalization appears to be an important next step in refining the FWCI for all wetland types, as this study was limited to a statewide approach due to small sample sizes within each wetland region. Conclusions The use of the macrophyte assemblage for a biological assessment of Florida freshwater flowing water forested wetlands provided a useful tool for detecting changes in biological integrity associated with changes in anthropogenic activity. While richness and diversity measures of macrophyte community composition were not particularly sensitive to changes in landscape development intensity, metrics used as biological indicators of changes in macrophyte community composition were. In fact, the strong correlation between the landscape scale human disturbance gradient (LDI) and the local scale wetland condition index of biological integrity (FWCI) demonstrated the potential value of using the LDI index as an initial indication of biological integrity, which can be further tested with chemical and physical parameters and compared against assemblage specific biological indices. Due to similarities with metrics from the FWCIs for depressional herbaceous ( Lane et al. 2003 ) and depressional forested ( Reiss and Brown 2005 ) wetlands, a multi-metric multi-assemblage FWCI could be constructed for all freshwater wetlands throughout the state of Florida, with a set of core metrics and specific indicator species and metric scores based on wetland type within the wetland regions. While the forested strand and floodplain FWCI for flowing water systems can not be used to predict changes in the physical and chemical parameters of a wetland, its strength lies in providing an overview of biological integrity through the integration of changes in macrophyte community composition from cumulative effects. The quantitative score of biological integrity established through the FWCI should not be used as a surrogate for wetland value, but as an objective, quantitative means of comparing changes in community composition along gradients of human development intensity, which can be used objectively to assess the biological integrity of Floridas wetlands.

PAGE 83

APPENDIX A STANDARD OPERATING PROCEDURES CHECKLIST OF MATERIALS/FIELD EQUIPMENT Miscellaneous SOPs Large cooler with frozen ice bottles for unknown vegetation Waders Garmin III GPS unit Florida Gazetteer Machete Aerial photo & FLUCCS codes of site Vegetation Transects 2-3 100m transect tapes 1 m PVC with distance marks (cm) 2-3 compasses Clipboards Field data sheets a minimum of 10 per site Site Characterization & WRAP sheets 1 per person per site Pencils Sharpie Bag for unknown plants Masking tape Field ID manuals Prism for basal area Hand lens Index cards 72

PAGE 84

73 SOPS FOR FORESTED STRAND WETLANDS: VEGETATION 1. Note the direction of flow through the landscape. 2. Locate a line running through the center of the strand along the flow gradient. This is the center-line. 3. Randomly select a starting point for the initial transect. Each consecutive transect will begin approximately 25 m upstream of the initial transect, so that a stretch of approximately 100 m will be sampled along the length of the strand. Run transects perpendicular to the main channelized flow. 4. At the beginning of each transect, delineate the edge of the wetland using a combination of wetland plants and hydrologic indicators. Be conservative on the side of the wetland. 5. Establish the transect using a meter tape and a compass. Each transect will start with 0 meters at the wetland edge and run into the center-line (established in step 2). 6. Use a separate field data sheet for each transect. If the number of species located on a transect exceeds the number of columns on the data sheet, start a new data sheet. Be thorough in completing field data sheets including information on site, transect direction, date, and data recorder. Specify if there are multiple field data sheets for a single transect. 7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long, record all species rooted within these elongated quadrats. 8. Plant species names are recorded on the data sheets using the full genus and species names. Each unknown species is given a unique ID code using the transect number (ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.). 9. Collect voucher specimens for all unknown species being sure to get plant inflorescence and roots, tag samples with properly labeled masking tape, and put into a labeled collection bag. Note the color of the inflorescence on the label, as the flowers often do not preserve well. Index cards can be used to protect especially sensitive parts. When vegetation sampling is complete, store the collection bag in a cooler on ice until identification can be completed. 10. Voucher specimens are identified in the field on the day of sampling. Unidentified plants will be placed in a plant press for further clarification and identification. Plant nomenclature follows FDEPs Florida Wetland Plant Identification Manual (Tobe et al. 1998). If time prohibits immediate pressing, unknown plants should be stored in the cooler. 11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be recorded. Use the data sheet for basal area, and record basal area per species using variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent elbow. Looking through the prism count the number of trees per species that fall within the variable area plot. The prism shall be centered over the sampling point at all times, with the field person rotating around the prism so that the entire circular area (360 o ) around the point of sampling is included.

PAGE 85

Forested Strand Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands Site: Transect Number: Date: Data Recorder: Species 0-5 m 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 90-95 95-100

PAGE 86

Forested Strand Wetlands Field Data Sheet Basal Area UF Center for Wetlands Site: Transect Direction: Date: Data Recorder: Species 10 m 20 m 30 m 40 m 50 m 60 m 70 m 80 m 90 m 100 m

PAGE 87

76 SOPS FOR FORESTED FLOODPLAIN WETLANDS: VEGETATION 1. Note the direction of flow through the landscape. 2. Locate a line running through the center of the strand along the flow gradient. This is the center-line. 3. Randomly select a starting point for the initi al transect, preferab ly this point with coincide with a Stream Condition Index (SCI) sample point. Each consecutive transect will begin approximately 25 m upstr eam of the initial transect, so that a stretch of approximately 100 m will be samp led along the length of the strand. Run transects perpendicular to the main channelized flow. 4. Delineate the wetland line using a combin ation of wetland plants and hydrologic indicators. Be conservativ e on the side of the wetland. 5. Establish the transect using a meter tape and a compass. Transects will be limited to a maximum length of 50 m. The first transect will begin with 0 meters at the wetland edge and run towards the centerline (established in step 2) The second transect will begin 25 m upstream of the firs t transect. Transect 2 begins at the edge of the channelized flow and runs through the we tland perpendicular to the flow for a maximum of 50 m. Repeat placement for transects 3 and 4. 6. Use a separate field data sheet for each tr ansect. If the number of species located on a transect exceeds the number of columns on the data sheet, start a new data sheet. Be thorough in completing field data sheets including information on site, transect direction, date, and data recorder. Specify if there are multiple field data sheets for a single transect. 7. Create quadrats that are 0.5 m on either side of the transect (1-m wide) and 5-m long, record all species rooted w ithin these elongated quadrats. 8. Plant species names are recorded on the data sheets using the full genus and species names. Each unknown species is given a uni que ID code using the transect number (ex. 1-1, 1-2, 1-3, 2-1, 2-2, etc.). 9. Collect voucher specimens for all unknown species being sure to get plant inflorescence and roots, tag samples with properly labeled masking tape, and put into a labeled collection bag. Note the color of the inflorescence on the label, as the flowers often do not preserve well. Index cards can be us ed to protect especially sensitive parts. When vegetation sampling is complete, store the collection bag in a cooler on ice until identi fication can be completed. 10. Voucher specimens are identified in the fi eld on the day of sampling. Unidentified plants will be placed in a plant press for fu rther clarification and identification. Plant nomenclature follows FDEPs Florida Wetland Plant Identification Manual (Tobe et al. 1998). If time prohibits immediate pres sing, unknown plants should be stored in the cooler. 11. At each 10 meters along each transect, (i.e. 10 m, 20 m, etc.), tree basal area will be recorded. Use the data sheet for basal ar ea, and record basal area per species using variable area plots and a 10 factor prism. Hold the prism at eye level, with a bent

PAGE 88

77 elbow. Looking through the prism count the number of trees per species that fall within the variable area plot. The prism shall be centered over the sampling point at all times, with the field person rotating around the prism so that the entire circular area (360 o ) around the point of sampling is included.

PAGE 89

Forested Floodplain Wetlands Field Data Sheet Transects, 1 x 5 m quadrat presence UF Center for Wetlands Site: Transect Number: Date: Data Recorder: Species 0-5 m 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50

PAGE 90

Forested Floodplain Wetlands Field Data Sheet Basal Area UF Center for Wetlands Site: Transect Direction: Date: Data Recorder: Species 10 m 20 m 30 m 40 m 50 m

PAGE 91

APPENDIX B COEFFICIENT OF CONSERVATISM SCORES Table B-1. Coefficient of Conservatism (CC) scores for macrophyte species identified in forested strand and floodplain wetlands in Florida. CC scores were assigned from FQAI surveys from isolated depressional forested wetlands (Reiss and Brown 2005) first, followed by isolated depressional herbaceous wetlands (Lane et al. 2003). Species without scores from previous FQAI studies were scores according to their faithfulness and fidelity to freshwater flowing water wetlands (strands and floodplains) in 2003 by five expert Florida botanists (Tony Arcuri, Dan Austin, David Hall, Nina Raymond, and Bruce Tatje). Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Abrus precatorius 0.0 1 Acer rubrum 5.2 1 Acer saccharum 6.0 1 Agarista populifolia 6.5 1 Alnus serrulata 6.0 1 Alternanthera philoxeroides 0.0 1 Alternanthera sessilis 0.7 1 Amaranthus australis 2.6 1 Ambrosia artemisiifolia 0.7 1 Ampelopsis arborea 3.3 1 Amphicarpum muhlenbergianum 5.0 1 Andropogon virginicus 2.6 1 Annona glabra 6.8 1 Apios americana 3.1 1 Ardisia escallonioides 7.0 1 Arisaema triphyllum 6.5 1 Aronia arbutifolia 5.7 1 Arundinaria gigantea 5.3 1 Asimina parviflora 5.0 1 Asimina reticulata 4.4 1 Asplenium heterochroum 8.5 1 Aster elliottii 4.2 1 Bacopa caroliniana 6.0 1 Bacopa monnieri 4.3 1 Begonia cucullata 1.5 1 Berchemia scandens 5.1 1 Bidens alba 1.0 1 Bidens mitis 3.8 1 Bignonia capreolata 4.4 1 Blechnum serrulatum 5.5 1 Boehmeria cylindrica 4.5 1 Bumelia lycioides 6.0 1 Callicarpa americana 2.4 1 Campsis radicans 3.3 1 Carex albolutescens 3.6 1 80

PAGE 92

81 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Carex crus-corvi 6.0 1 Carex leptalea 6.8 1 Carex lupulina 6.3 1 Carex stipata 4.5 1 Carpinus caroliniana 6.8 1 Carya glabra 5.8 1 Celtis laevigata 5.0 1 Centella asiatica 1.9 1 Cephalanthus occidentalis 6.0 1 Chasmanthium laxum 6.0 1 Chasmanthium nitidum 6.3 1 Chrysobalanus icaco 6.3 1 Cinnamomum camphora 0.2 1 Cladium jamaicense 5.5 1 Clematis crispa 5.5 1 Colocasia esculenta 0.0 1 Commelina communis 2.5 1 Commelina diffusa 1.7 1 Conyza canadensis 0.3 1 Cornus foemina 6.6 1 Crataegus marshallii 6.3 1 Crinum americanum 7.6 1 Cuphea carthagenensis 1.4 1 Cynodon dactylon 0.0 1 Cyperus difformis 2.0 1 Cyperus globulosus 1.8 1 Cyperus haspan 2.6 1 Cyperus ligularis 3.2 1 Cyperus odoratus 3.6 1 Cyrilla racemiflora 4.5 1 Decumaria barbara 6.3 1 Dichondra carolinensis 1.9 1 Dichromena colorata 5.5 1 Digitaria serotina 1.8 1 Diodia virginiana 2.4 1 Dioscorea floridana 5.5 1 Diospyros virginiana 4.0 1 Eclipta alba 1.7 1 Emilia fosbergii 0.5 1 Erechtites hieraciifolius 2.1 1 Erianthus giganteus 6.0 1 Eryngium baldwini 4.4 1 Eryngium prostratum 4.0 1 Erythrina herbacea 4.0 1 Eugenia uniflora 1.3 1 Euonymus americanus 5.5 1 Eupatorium capillifolium 0.5 1 Eupatorium mikanioides 5.5 1

PAGE 93

82 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Eupatorium perfoliatum 5.9 1 Eustachys petraea 2.0 1 Fagus grandifolia 7.5 1 Fraxinus caroliniana 7.1 1 Galium tinctorium 3.1 1 Gaylussacia dumosa 5.4 1 Gaylussacia frondosa 6.7 1 Gelsemium sempervirens 4.0 1 Gordonia lasianthus 6.7 1 Gratiola hispida 6.0 1 Hamamelis virginiana 6.5 1 Hibiscus coccineus 6.6 1 Hydrocotyle bonariensis 3.3 1 Hydrocotyle ranunculoides 3.1 1 Hydrocotyle umbellata 2.9 1 Hydrocotyle verticillata 3.1 1 Hypericum brachyphyllum 6.8 1 Hypericum hypericoides 4.0 1 Hypericum myrtifolium 5.5 1 Hypericum tetrapetalum 5.0 1 Hypoxis curtissii 5.7 1 Hyptis alata 4.3 1 Ilex cassine 8.1 1 Ilex coriacea 6.0 1 Ilex glabra 4.3 1 Ilex opaca var. opaca 6.0 1 Ilex vomitoria 4.8 1 Ipomoea hederifolia 2.0 1 Ipomoea pandurata 4.8 1 Ipomoea sagittata 5.4 1 Itea virginica 7.9 1 Juncus effusus 1.9 1 Juncus megacephalus 3.3 1 Juncus polycephalos 3.3 1 Juniperus virginiana 5.2 1 Justicia ovata 5.5 1 Koelreuteria elegans 2.0 1 Lachnanthes caroliana 3.1 1 Lemna minor 1.0 1 Leucothoe axillaris 7.0 1 Leucothoe racemosa 6.2 1 Ligustrum sinense 0.0 1 Lindernia grandiflora 3.6 1 Liquidambar styraciflua 3.3 1 Liriodendron tulipifera 6.8 1 Lobelia cardinalis 6.8 1 Lonicera japonica 0.0 1 Ludwigia maritima 3.3 1

PAGE 94

83 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Ludwigia palustris 4.0 1 Ludwigia peruviana 1.2 1 Ludwigia pilosa 5.8 1 Ludwigia repens 2.9 1 Lycopus rubellus 5.2 1 Lygodium japonicum 0.0 1 Lygodium microphyllum 0.0 1 Lyonia fruticosa 6.0 1 Lyonia lucida 6.0 1 Lyonia mariana 6.8 1 Magnolia grandiflora 6.2 1 Magnolia virginiana var. australis 8.1 1 Matelea floridana 6.7 1 Melaleuca quinquenervia 0.0 1 Merremia dissecta 0.3 1 Micranthemum glomeratum 4.0 1 Micranthemum umbrosum 4.3 1 Micromeria brownei 4.8 1 Mikania scandens 2.4 1 Mimosa pigra 0.7 1 Mitchella repens 6.7 1 Momordica charantia 0.0 1 Morus rubra 4.4 1 Myrica cerifera 3.1 1 Myrsine guianensis 5.2 1 Nephrolepis exaltata 3.8 1 Nyssa ogeche 7.0 1 Nyssa sylvatica var. biflora 7.4 1 Oeceoclades maculata 0.4 1 Oplismenus setarius 3.3 1 Orontium aquaticum 7.6 1 Osmunda cinnamomea 5.5 1 Osmunda regalis 6.9 1 Oxypolis filiformis 6.7 1 Panicum abscissum 9.2 1 Panicum anceps 4.3 1 Panicum commutatum 4.5 1 Panicum dichotomum 4.0 1 Panicum ensifolium 5.0 1 Panicum erectifolium 5.7 1 Panicum hemitomon 5.0 1 Panicum rigidulum 4.5 1 Panicum spretum 5.4 1 Panicum tenue 4.2 1 Parietaria praetermissa 3.0 1 Parthenocissus quinquefolia 3.0 1 Paspalum conjugatum 3.1 1 Paspalum notatum 0.0 1

PAGE 95

84 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Passiflora incarnata 3.0 1 Peltandra virginica 5.8 1 Persea borbonia 6.3 1 Persea palustris 7.4 1 Phlebodium aureum 6.8 1 Phyla nodiflora 1.4 1 Phyllanthus urinaria 0.0 1 Phytolacca americana 1.2 1 Pieris phyllyreifolia 9.5 1 Pinckneya bracteata 8.3 1 Pinus clausa 5.6 1 Pinus elliottii 4.0 1 Pinus taeda 3.3 1 Pluchea foetida 3.8 1 Pluchea rosea 3.6 1 Polygala rugelii 8.2 1 Polygonum densiflorum 5.3 1 Polygonum hirsutum 8.2 1 Polygonum hydropiperoides 2.6 1 Pontederia cordata 5.0 1 Proserpinaca palustris 3.8 1 Prunus caroliniana 3.0 1 Prunus serotina 3.6 1 Psychotria nervosa 5.2 1 Psychotria sulzneri 5.5 1 Pteridium aquilinum 3.6 1 Ptilimnium capillaceum 3.1 1 Quercus laurifolia 3.6 1 Quercus michauxii 5.7 1 Quercus nigra 2.1 1 Quercus virginiana 4.2 1 Rhododendron canescens 6.8 1 Rhododendron viscosum 7.6 1 Rhus copallinum 2.4 1 Rhynchospora baldwinii 5.7 1 Rhynchospora inundata 6.0 1 Rhynchospora microcephala 4.8 1 Rhynchospora miliacea 7.1 1 Rubus argutus 2.1 1 Rubus trivialis 1.9 1 Ruellia caroliniensis 4.3 1 Rumex verticillatus 4.8 1 Sabal minor 6.2 1 Sabal palmetto 4.5 1 Sabatia calycina 6.2 1 Sagittaria filiformis 6.0 1 Sagittaria lancifolia 4.5 1 Sagittaria latifolia 5.0 1

PAGE 96

85 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Salvia lyrata 3.2 1 Sambucus canadensis 1.7 1 Samolus valerandi 5.6 1 Sanicula canadensis 5.7 1 Sapium sebiferum 0.0 1 Saururus cernuus 5.5 1 Schefflera actinophylla 0.5 1 Schinus terebinthifolius 0.0 1 Scleria triglomerata 4.8 1 Serenoa repens 4.5 1 Setaria geniculata 3.1 1 Sida rhombifolia 1.0 1 Smilax auriculata 3.8 1 Smilax bona-nox 2.6 1 Smilax glauca 3.3 1 Smilax laurifolia 5.2 1 Smilax pumila 6.0 1 Smilax smallii 4.5 1 Smilax tamnoides 3.6 1 Smilax walteri 6.0 1 Solidago fistulosa 3.6 1 Sparganium americanum 6.7 1 Sporobolus floridanus 7.1 1 Stenotaphrum secundatum 0.8 1 Stillingia aquatica 7.4 1 Symplocos tinctoria 6.0 1 Taxodium ascendens 8.8 1 Taxodium distichum 7.2 1 Thelypteris dentata 3.4 1 Thelypteris hispidula 4.5 1 Thelypteris palustris 5.8 1 Tilia americana 5.5 1 Toxicodendron radicans 1.9 1 Triadenum virginicum 5.0 1 Trichostema dichotomum 4.5 1 Trifolium repens 0.0 1 Tripsacum dactyloides 4.0 1 Ulmus americana 7.4 1 Urena lobata 0.0 1 Vaccinium arboreum 6.4 1 Vaccinium corymbosum 5.7 1 Vaccinium stamineum 5.8 1 Vaccinium tenellum 5.5 1 Viburnum dentatum 6.0 1 Viburnum nudum 5.0 1 Viburnum obovatum 4.7 1 Viola affinis 5.5 1 Vitis aestivalis 2.9 1

PAGE 97

86 Species CC Score 1999/2000 Herbaceous 2001/2002 Forested 2003 Flowing Vitis cinerea 2.0 1 Vitis rotundifolia 2.1 1 Vitis shuttleworthii 3.5 1 Wisteria sinensis 1.0 1 Woodwardia areolata 5.7 1 Woodwardia virginica 4.8 1 Xyris jupicai 1.7 1 Youngia japonica 0.0 1

PAGE 98

APPENDIX C SUMMARY STATISTICS Table C-1. Summary statistics of richness (R), evenness (E), Shannon diversity (H'), Simpson diversity (D), and Whittakers beta diversity (W) for the macrophyte assemblage (species level). Site R E H' D w FF1 59 1.0001 4.08 0.98 5.88 FF2 21 1.0002 3.05 0.95 3.73 FF3 56 0.9999 4.03 0.98 4.86 FF4 46 1.0001 3.83 0.98 7.36 FF5 77 1.0000 4.34 0.99 6.82 FF6 25 1.0000 3.22 0.96 1.78 FF7 32 1.0001 3.47 0.97 3.49 FF8 75 0.9999 4.32 0.99 6.11 FF9 45 1.0001 3.81 0.98 3.90 FF10 35 0.9999 3.56 0.97 3.18 FF11 29 0.9999 3.37 0.97 4.73 FF12 46 1.0001 3.83 0.98 5.15 FF13 48 0.9999 3.87 0.98 4.93 FF14 60 0.9999 4.09 0.98 5.94 FS1 39 1.0001 3.66 0.97 2.68 FS2 39 1.0001 3.66 0.97 4.90 FS3 33 1.0001 3.50 0.97 4.45 FS4 55 0.9999 4.01 0.98 5.03 FS5 39 1.0001 3.66 0.97 3.91 FS6 29 0.9999 3.37 0.97 3.02 FS7 30 0.9999 3.40 0.97 2.96 FS8 35 0.9999 3.56 0.97 2.12 FS9 47 1.0000 3.85 0.98 3.73 FS10 42 1.0001 3.74 0.98 4.43 87

PAGE 99

APPENDIX D METRIC SCORING FOR THE MACROPHYTE FLORIDA WETLAND CONDITION INDEX FOR FLOWING WATER SYSTEMS 1. Calculate values for the 5 metrics: 1 Proportion tolerant indicator species 2 Proportion sensitive indicator species 3 FQAI score 4 Proportion exotic species 5 Proportion native perennial species 2. Take the natural log of metrics to improve distribution. = ln (metric value + 1) 1 is added to avoid errors related to taking the natural log of a zero value 3. Use the scoring equations to normalize scores between 0 and 10. Metrics that increase with increasing LDI tolerant, exotic metrics: = 10 (( metric 5th percentile) ( 10 / ( 95th percentile 5th percentile))) Metrics that decrease with increasing LDI sensitive, FQAI, native perennial metrics: = (( metric 5th percentile) ( 10 / ( 95th percentile 5th percentile))) Below are the 5th and 95th percentiles for each metric (transformed values are presented, see step 2 above): 5th Percentile 95th Percentile Proportion tolerant indicator species 0.05 0.24 Proportion sensitive indicator species 0.00 0.16 FQAI score 1.44 1.84 Proportion exotic species 0.00 0.17 Proportion native perennial species 0.56 0.69 4. Rescore, so that the metrics in the outer 5th percentiles receive scores of 0 or 10. = IF ( score < 0, 0, ( IF ( score >= 10, score, 10))) 88

PAGE 100

LIST OF REFERENCES Adams, S.M. 2002. Biological indicators of aquatic ecosystem stress: introduction and overview. Pages 1-11 in S.M. Adams, editor. Biological indicators of aquatic ecosystem stress. American Fisheries Society. Bethesda, Maryland, USA. Analyse-it Software, Ltd. 1997-2003. version 1.67. Leeds, England, United Kingdom. Andreas, B.K. and R.W. Lichvar. 1995. A floristic assessment system for northern Ohio. Wetlands Research Program Technical Report WRP-DE-8. U.S. Army Corps of Engineers Waterways Experiment Station, Vicksburg, Mississippi, USA. Apfelbeck, R. 2000. Developing preliminary bioassessment protocols for Montana wetlands, State of Montana Department of Environmental Quality. Helena, Montana, USA. Arcview GIS 3.2 Environmental Systems Research Institute, Inc. 1999. Neuron Data, Inc. 1991-1996. Portions copyright 1991-1995 Arthur D. Applegate. Found at: http://www.esri.com/. Redlands, California, USA. ArcGIS 8.3 Environmental Systems Research Institute, Inc. 1999-2002. Found at: http://www.esri.com/arcgis. Redlands, California, USA. Barbour, M.T., J. Gerristen, G.E. Griffith, R. Frydenborg, E. McCarron, J.S. White, and M.L. Bastian. 1996a. A framework for biological criteria for Florida streams using benthic macroinvertebrates. Journal of the North American Benthological Society 15(2): 185-211. Barbour, M.T., J. Gerristen, and J.S. White. 1996b. Development of the Stream Condition Index (SCI) for Florida. A Report to the Florida Department of Environmental Protection, Stormwater and Nonpoint Source Management Section. Tetra Tech, Inc. Owing Mills, Maryland, USA. Bedford, B.L., M.R. Wabridge, and A. Aldous. 1999. Patterns in nutrient availability and plant diversity of temperate North American wetlands. Ecology 80(7): 2151-2169. Blanch, S.J. and M.A. Brock. 1994. Effects of grazing and depth on two wetland plant species. Australian Journal of Marine and Freshwater Research 45: 1387-1394. Brown, M.T. and M.B. Vivas. 2005. Landscape Development Intensity Index. Environmental Monitoring and Assessment 101: 289-309. Brown, M.T. and S. Ulgiati. 2005. Emergy, transformity, and ecosystem health. Pages 333-352 in S.E. Jrgensen, R. Costanza, and F. Xu, editors. Handbook of ecological indicators for assessment of ecosystem health. Taylor and Francis, Boca Raton, Florida, USA. Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117-143. Cohen, M.J., S.M. Carstenn, and C.R. Lane. 2004. Floristic quality indices for biotic assessment of depressional marsh condition in Florida. Ecological Applications 14(3): 784-794. Cronk, J. K. and M.S. Fennessy. 2001. Wetland plants: biology and ecology. Lewis Publishers. Boca Raton, Florida, USA. Crowder, A. and D.S. Painter. 1991. Submerged macrophytes in Lake Ontario: current knowledge, importance, threats to stability, and needed studies. Canadian Journal of Fisheries and Aquatic Sciences 48:1539-1545. 89

PAGE 101

90 Dahl, T.E. 2000. Status and trends of wetlands in the conterminous United States 1986 to 1997. United States Department of the Interior, Fish and Wildlife Service, Washington, D.C., USA. Danielson, T.J. 1998. Indicators for monitoring and assessing biological integrity of inland freshwater wetlands. EPA 843-R-98-002. Wetlands Division Office of Water, United States Environmental Protection Agency, Washington, D.C., USA. David, P.G. 1999. Response of exotics to restored hydroperiod at Dupuis Reserve, Florida. Restoration Ecology 7(4): 407-410. Davis, J.H. 1967. General map of natural vegetation of Florida. Accessed: 12/2004. Found at: http://www.geoplan.ufl.edu/fdgl/fgdl.html Devine, R. 1998. Alien invasion. National Geographic Society, Washington, D.C., USA. Dufrne, M. and P. Legendre. 1997. Species assemblages and indicator species the need for a flexible asymmetrical approach. Ecological Monographs 67(3): 345-366. Ehrenfeld, J.G. and J.P. Schneider. 1991. Chamaecyparis thyoides wetlands and suburbanization: effects of non-point source water pollution on hydrology and plant community structure. Journal of Applied Ecology 28(2): 467-490. Ewel, K.C. 1990. Swamps. Pages 281-323 in R.L. Myers and J.J. Ewel. Ecosystems of Florida. University of Central Florida Press. Orlando, Florida, USA. Exotic Plant Pest Council (EPPC). 2003. Accessed 3/2005. Found at: http://www.fleppc.org/ FAC 62-340. Florida Administrative Code, Chapter 62-340. Accessed 3/2005. Found at http://www.dep.state.fl.us/legal/rules/surfacewater/62-340.pdf. Florida Department of Environmental Protection. Fennessy, S., R. Geho, B. Elifritz and R. Lopez. 1998. Testing the floristic quality assessment index as an indicator of riparian wetland quality. Final report to U.S. EPA. Ohio Environmental Protection Agency, Division of Surface Water, Columbus, Ohio, USA. Fennessy, S., M. Gernes, J. Mack, and D.H. Waldrop. 2001. Methods for evaluating wetland condition: using vegetation to assess environmental conditions in wetlands. EPA 822-R-01-007j. U.S. Environmental Protection Agency, Office of Water, Washington, D.C., USA. Fernald, E.A. and E.D. Purdum, editors. 1992. Atlas of Florida. University Press of Florida. Gainesville, Florida, USA. Fore, L.S. 2004. Development and testing of biomonitoring tools for macroinvertebrates in Florida streams. Statistical Design, Seattle, Washington. A report for the Florida Department of Environmental Protection, Tallahassee, Florida, USA. Fore, L.S. and C. Grafe. 2002. Using diatoms to assess the biological condition of large rivers in Idaho (U.S.A.). Freshwater Biology 47: 2015-2037. Francis, C.M., M.J.W. Austen, J.M. Bowles, W.B. Draper. 2000. Assessing floristic quality in southern Ontario woodlands. Natural Areas Journal 20:66-77. Galatowitsch, S.M., N.O. Anderson, and P.D. Ascher. 1999b. Invasiveness in wetland plants in temperate North America. Wetlands 19(4): 733-755. Galatowitsch, S.M., D.C. Whited, R. Lehtinen, J. Husveth, and K. Schik. 2000. The vegetation of wet meadows in relation to their land-use. Ecological Monitoring and Assessment 60: 121-144.

PAGE 102

91 Galatowitsch, S.M., D.C. Whited, and J.R. Tester. 1999a. Development of community metrics to evaluate recovery of Minnesota wetlands. Journal of Aquatic Ecosystem Stress and Recovery 6: 217-234. Gernes, M.C. and J.C. Helgen. 1999. Indexes of biotic integrity for wetlands, section B: wetland vegetation IBI for depressional wetlands. Final Report to the United States Environmental Protection Agency Assistance Number CD995525-01, April 1999. Minnesota Pollution Control Agency, St. Paul, Minnesota, USA. Gerristen, J., M.T. Barbour, and K. King. 2000. Apples, oranges, and ecoregions: on determining pattern in aquatic assemblages. Journal of the North American Benthological Association 19(3): 487-496. Gerristen, J. and J. White. 1997. Development of a biological index for Florida lakes. A Report to the Florida Department of Environmental Protection. Tetra Tech, Inc. Owing Mills, Maryland, USA. Godfrey, R.K. and J.W. Wooten. 1981. Aquatic and wetland plants of the southeastern United States. University of Georgia Press, Athens, Georgia, USA. Grace, J.B. and H. Jutila. 1999. The relationship between species density and community biomass in grazed and ungrazed coastal meadows. Oikos 85: 398-408. Griffith, G.E., J.M. Omernik, C.M. Rohm, and S.M. Pierson. 1994. Florida Regionalization Project. Environmental Research Laboratory, United States Environmental Protection Agency, Corvallis, OR. Griffith, G. E., D. E. Canfield, Jr., C. A. Horsburgh, J. M. Omernik, and S. H. Azevedo. 1997. Florida lake regions. Environmental Research Laboratory, United States Environmental Protection Agency, Corvallis, OR. Herman, K.D., A.A. Reznicek, L.A. Masters, G.S. Wilhelm, M.R. Penskar and W.W. Brodowicz. 1997. Floristic quality assessment: development and application in the state of Michigan (USA). Natural Areas Journal 17:265-279. Hobbs, R.J. and L.F. Hueneke. 1992. Disturbance, diversity, and invasion: implications for conservation. Conservation Biology 6(3): 324-337. James, M.O. and K.M. Kleinow. 1994. Trophic transfers of chemicals in the aquatic environment. Pages 1-35 in D.C. Malins and G.K. Ostrander, editors. Aquatic toxicology and cellular perspectives. Lewis Publishers. Boca Raton, Florida, USA. Karr, J.R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6: 21-27. Karr, J.R. 1993. Defining and assessing ecological integrity: beyond water quality. Environmental Toxicology and Chemistry 12: 1521-1531. Karr, J.R. and E.W. Chu. 1997. Biological monitoring and assessment: using multimetric indexes effectively. EPA 235-R-97-001. University of Washington, Seattle, Washington, USA. Karr, J.R. and E.W. Chu. 1999. Restoring life in running waters. Island Press. Washington, D.C., USA. Karr, J.R. and D.R. Dudley. 1981. Ecological perspectives on water quality goals. Environmental Management 5: 55-68. Kent, D.M. 2000. Evaluating wetland functions and values. Chapter 3 in D.M. Kent, editor. Applied wetlands science and technology. Lewis Publishers. Boca Raton, Florida, USA.

PAGE 103

92 Kerans, B.L. and J.R. Karr. 1994. A benthic index of biotic integrity (B-IBI) for rivers in the Tennessee valley. Ecological Applications 4(4): 768-785. Kruskal, J.B. 1964. Multidimensional scaling by optimizing goodness of fit to a non-metric hypothesis. Psychometrika 29: 115-129. Lane, C.R. 2000. Proposed ecological regions for freshwater wetlands of Florida. Masters Thesis, University of Florida, Gainesville, Florida, USA. Lane, C.R. 2003. Biological indicators of wetland condition for isolated depressional herbaceous wetlands in Florida. Ph.D. Dissertation, University of Florida, Gainesville, Florida, USA. Lane, C.R., M.T. Brown, M. Murray-Hudson, and M.B. Vivas. 2003. The Wetland Condition Index (WCI): biological indicators of wetland condition for isolated depressional herbaceous wetlands in Florida. A Report to the Florida Department of Environmental Protection. Howard T. Odum Center for Wetlands, University of Florida, Gainesville, Florida, USA. Micacchion, M. 2004. Integrated Wetland Assessment Program. Part 7: Amphibian Index of Biotic Integrity (AmphIBI) for Ohio Wetlands. Ohio EPA Technical Report WET/2004-7. Ohio Environmental Protection Agency, Wetland Ecology Group, Division of Surface Water, Columbus, Ohio. Mack, J. 2001. Vegetation Index of Biological Integrity (VIBI) for wetlands: ecoregional, hydrogeomorphologic, and plant community comparisons with preliminary wetland aquatic life use designations. Final Report to the United States Environmental Protection Agency Grant No. CD985875, Volume 1. Wetland Ecology Group, Division of Surface Water, Ohio Environmental Protection Agency, Columbus Ohio, USA. Found at: http://www.epa.state.oh.us/dsw.wetlands/wetlands_bioasses.html. McCune, B. and J.B. Grace. 2002. Analysis of ecological communities. MJM Software Design. Gleneden Beach, Oregon, USA. McCune, B., R. Rosentreter, J.M. Ponzetti, and D.C. Shaw. 2000. Epiphyte habitats in an old conifer forest in western Washington, USA. Bryologist 103: 417-427. Miller, R.E., Jr. and B.E. Boyd. 1999. Wetland rapid assessment procedure. South Florida Water Management District, Technical Publication REG-001. West Palm Beach, Florida, USA. Minitab Statistical Software, version 13.1. 2000. Found at: http://www.minitab.com. State College, Pennsylvania, USA. Mitsch, W.J. and J.G. Gosselink. 1993. Wetlands, 2nd edition. John Wiley and Sons, Inc. New York, New York, USA. Mushet, D.M., N.H. Euliss, and T.L. Shaffer. 2002. Floristic quality assessment of one natural and three restored wetland complexes in North Dakota, USA. Wetlands 22(1): 126-138. Myer, R.L. and J.J. Ewel, eds. 1990. Ecosystems of Florida. University of Central Florida Press. Orlando, Florida, USA. OConnell, T.J., L.E. Jackson, and R.P. Brooks. 1998. A bird community index of biotic integrity for the mid-Atlantic highlands. Environmental Monitoring and Assessment 51: 145-156. Odum, H.T. 1995. Environmental Accounting: Emergy and Environmental Decision Making. John Wiley and Sons, New York, New York, USA.

PAGE 104

93 Ott, R.L. and M. Longnecker. 2001. An introduction to statistical methods and data analysis, 5th edition. Duxbury, Wadsworth Group. Pacific Grove, California, USA. PCORD, version 4.1. MJM Software.. Found at: http://home.centurytel.net/~mjm/. Gleneden Beach, Oregon, USA Reiss, K.C. 2004. Developing biological indicators for isolated forested wetlands in Florida. Ph.D. Dissertation, University of Florida, Gainesville, Florida, USA. Reiss, K.C. and M.T. Brown. 2005. The Florida Wetland Condition Index (FWCI): developing biological indicators for isolated depressional forested wetlands. A Report to the Florida Department of Environmental Protection. Howard T. Odum Center for Wetlands, University of Florida, Gainesville, Florida, USA. Schindler, D.W. 1987. Detecting ecosystem responses to anthropogenic stress. Canadian Journal of Fisheries and Aquatic Sciences 44:6-25. Schulz, E.J., M.V. Hoyer, and D.E. Canfield, Jr. 1999. An index of biotic integrity: a test with limnological and fish data from sixty Florida lakes. Transactions of the American Fisheries Society 128: 564-577. ter Braak, C.J.F. 1987. The analysis of vegetation-environmental relationships by canonical correspondence analysis. Vegetatio 69: 69-77. Tobe, J.D., K. Craddock Burks, R.W. Cantrell, M.A. Garland, M.E. Sweeley, D.W. Hall, P. Wallace, G. Anglin, G. Nelson, J.R. Cooper, D. Bickner, K. Gilbert, N. Aymond, K. Greenwood, and N. Raymond. 1998. Florida wetland plants: an identification manual. Florida Department of Environmental Protection, Tallahassee, Florida, USA. United States Department of Agriculture, Natural Resource Conservation Service (USDA NRCS). 2002. The PLANTS Database, Version 3.5. National Plant Data Center, Baton Rouge, Louisiana, USA. Available on-line at: http://plants.usda.gov. Accessed 2001-2004. United States Environmental Protection Agency (USEPA). 1990. Feasibility report on environmental indicators for surface water programs. Office of Water Regulations and Standards and Office of Policy, Planning and Evaluation. Washington, D.C., USA. United States Environmental Protection Agency (USEPA). 1998a. Wetland Biological Assessments and HGM Functional Assessment. EPA 843-F-98-001f Wetland Bioassessment Fact Sheet 6. Washington, D.C., USA. Available on-line at: http://www.epa.gov/owow/wetlands/wqual/bio_fact/fact6.html. Accessed 2005. United States Environmental Protection Agency (USEPA). 1998b. Lake and reservoir bioassessment and biocriteria. EPA 841-B-98-007 Technical Guidance Document. Washington, D.C., USA. Available on-line at: http://www.epa.gov/owow/monitoring/tech/lakes.html. Accessed 2002-2004. United States Environmental Protection Agency (USEPA). 2001. Better Assessment Science Integrating Point and Nonpoint Sources (BASINS), Version 3.0. EPA 823-B-01-001. Users Manual. Office of Water. Technical Support and software download available at: http://www.epa.gov/ost/basins/. Accessed 2003-2005. United States Environmental Protection Agency (USEPA). 2002. Methods for evaluating wetland condition: introduction to biological assessment. EPA-822-R-02-014. Office of Water, Washington, D.C., USA.

PAGE 105

94 United States Environmental Protection Agency (USEPA). 2003. Biological Indicators of Watershed Health. Available on-line at: http://www.epa.gov/bioindicators. Accessed 2003-2004. Wharton, C.H., H.T. Odum, K. Ewel, M. Duever, A. Lugo, R. Boyt, J. Bartholomew, E. DeBellevue, S. Brown, M. Brown, and L. Deuver. 1976. Forested wetlands of Florida their management and use. Center for Wetlands, University of Florida, Gainesville, Florida, USA. Wienhold, C.E. and A.G. Van der Valk. 1989. The impact of duration of drainage on the seed banks of northern prairie wetlands. Canadian Journal of Botany 67: 1878-1884. Wilhelm, G. and D. Ladd. 1988. Natural Area Assessment in the Chicago Region. Pages 361-375 in Transactions of the 53rd North American Wildlife and Natural Resource Conference, Louisville, Kentucky. Wildlife Management Institute, Washington D.C., USA. Wunderlin, R.P. 1998. Guide to the Vascular Plants of Florida. University Press of Florida, Gainesville, Florida, USA. Wunderlin, R. P., and B. F. Hansen. 2003. Atlas of Florida Vascular Plants. S. M. Landry and K. N. Campbell, application development. Florida Center for Community Design and Research. Institute for Systematic Botany, University of South Florida. Found at: http://www.plantatlas.usf.edu/. Tampa, Florida, USA. Zimmerman, G.M., H. Goetz, and P.W. Mielke, Jr. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66(2): 606-61.