RESPONSES OF TWELVE TREE SPECIES COMMON IN EVERGLADES TREE
ISLANDS TO SIMULATED HYDROLOGIC REGIMES
David T. Jones
1
, Jay P. Sah
1
, Michael S. Ross
1
, Steven F. Oberbauer
2
, Bernice Hwang
1
,
and Krish Jayachandran
1,3
1
Southeast Environmental Research Center
2
Department of Biological Sciences
3
Department of Environmental Studies
Florida International University
Miami, Florida, USA 33199
Abstract:
Twelve tree species common in Everglades tree islands were subjected to three hydrologic
regimes under controlled conditions for 25 weeks and assessed for growth and physiological responses.
Treatments representing high, low, and no flood were maintained in pools of water to mimic seasonal
variation in water depths at different positions in tree islands. Soil inundation under the high flood
treatment resulted in reduced tree growth (height, basal diameter, crown volume) that was more
pronounced and occurred earlier in mesic forest species than in swamp forest species. Physiological
responses differed less among species, although stomatal conductance was a better predictor of the effects
of flood stress on growth than either relative water content or chlorophyll fluorescence (F
v
/F
m
). Some
swamp species appeared to be better adapted to rising water levels than others; Annona glabra, Morella
cerifera, and Salix caroliniana responded more positively to flooding, while Magnolia virginiana, Persea
borbonia, Chrysobalanus icaco, and Ilex cassine were less flood-tolerant. The highest mortalities and
lowest growth were observed in the five upland species: Bursera simaruba, Coccoloba diversifolia, Eugenia
axillaris, Sideroxylon foetidissimum, and Simarouba glauca. Of these, Sideroxylon and Simarouba did not
survive to the end of the study under the high flood treatment. The moist soil conditions simulated by the
low flood treatment resulted in greater growth in all species compared to soil inundation under high
flood, except for the most flood-tolerant (Annona, Morella, Salix). The arrangement of species according
to their responses to experimental flooding roughly paralleled their spatial distribution in the tree islands.
The gradient in species responses demonstrated in this experiment may help guide responsible water
management and tree island restoration in the Everglades.
Key Words: flood tolerance, tree island, Florida Everglades, hydrologic regime, restoration
INTRODUCTION
Tree islands are among the most distinctive
features of the Florida Everglades (USA) and have
been described by various workers over the past
90 years (Harshberger 1914, Harper 1927, Davis
1943, Loveless 1959, Craighead 1971, Sklar and van
der Valk 2002a). They generally occur on elevated
limestone outcrops or above bedrock depressions
embedded within a freshwater or brackish marsh or
swamp. Because surface elevation at the center of
the tree island is typically higher than the surround-
ing marsh, a vegetation gradient can usually be
identified, with tropical and subtropical hardwood
species inhabiting the better-drained interior posi-
tions and swamp species of mainly temperate origin
dominating the frequently flooded edge locations.
Tree islands cover less than five percent of the
Everglades, yet they perform many vital ecosystem
functions, including nutrient cycling and provision
of wildlife habitat, and have historical and cultural
significance as sites of human habitation (van der
Valk and Sklar 2002).
Tree island vegetation is one of the most sensitive
components of the Everglades landscape to changes
in regional hydrology, where extremes in marsh
water levels can have serious consequences (Loveless
1959, Craighead 1971, 1984, McPherson 1973,
Alexander and Crook 1984, Brandt et al. 2000,
Sklar and van der Valk 2002a). Prolonged periods of
high water may adversely affect the condition of tree
island vegetation via death or dieback in flood-
intolerant species. Similarly, persistent low water
may create conditions of extreme fire risk, during
which vegetation may be catastrophically damaged.
Management-oriented changes in water-flow pattern
WETLANDS, Vol. 26, No. 3, September 2006, pp. 830–844
’
2006, The Society of Wetland Scientists
830
in the Everglades have resulted in such hydrologic
extremes, particularly in some of the Water Conser-
vation Areas (WCA) north of Everglades National
Park (ENP), where the number of tree islands has
decreased (Schortemeyer 1980) and the vegetation
on the existing islands has been altered (Wetzel
2002). Sklar and van der Valk (2002b) reported that
tree island number declined 87% in WCA 2A
between 1953 and 1995, and 61% in WCA 3
between 1940 and 1995.
The loss of tree islands and their associated
historical, cultural, and biological values has raised
awareness of the fragility of these habitats and
stimulated a resurgence of interest in their study and
preservation.
Maintaining
and/or
restoring
the
health of tree islands (and other Everglades habitats)
are components of the Comprehensive Everglades
Restoration Plan (CERP), a multi-agency project
designed to restore and enhance the freshwater
resources and natural environments of southern
Florida (USACE 1999). Consequently, there is
a need within CERP for tools to assess the health
of tree islands and to relate these measures to the
hydrologic regime to which they are exposed. Flood
tolerance and other ecophysiological characteristics
of tree island tree species are a critical element in the
formulation of these performance measures.
Flood tolerances of tree island tree species are not
well-documented in the literature. Guerra (1997)
and Jones et al. (1997) evaluated tree island
vegetation in the southern Everglades after a period
of prolonged high water in 1994–95 and noted its
effects on individual tree and shrub species. Conner
et al. (2002) reviewed flood tolerance in ten common
tree island species based on flood impact studies
conducted largely in bottomland forests in other
parts of the southern United States. In the only
reported greenhouse study, Gunderson et al. (1988)
examined the effects of a range of hydrologic
conditions, including flooding, on seedling growth
and morphology in five Everglades tree island
species. None of these studies attempted to assess
physiological responses of plants, and between the
latter two, only three subtropical hardwood forest
species were examined compared to nine swamp
forest species. In order to gain insight into species
responses in the field, a controlled study combining
morphological and physiological measurements on
a broad range of hydric and mesic tree island species
growing under different hydrologic regimes is
needed.
Studies in other ecosystems have successfully used
morphological (e.g., root and stem biomass, height,
stem diameter, leaf area, comparative anatomy) and
physiological parameters (e.g., chlorophyll fluores-
cence, leaf water potential, relative water content,
gas exchange, stomatal conductance) to elucidate
tree responses to water stress induced by flooding
and/or drought (Regehr et al. 1975, Pereira and
Kozlowski 1977, O
¨ gren and O
¨ quist 1985, Ewing
1996, Schmull and Thomas 2000, Anderson and
Pezeshki 2001, Davanso et al. 2002). Previous
studies have shown that flooding to or above the
soil surface may result in a range of adverse
responses, from diminished growth and photosyn-
thesis to death, in seedlings and saplings (Keeley
1979, Pezeshki and Chambers 1986, Ewing 1996,
Lopez and Kursar 1999, Schmull and Thomas 2000,
Davanso et al. 2002) and trees (Broadfoot and
Williston 1973, Harms et al. 1980, Vu and Yele-
nosky 1991, Ewing 1996, McKevlin et al. 1998).
The objectives of our study were twofold. First,
we compared tree height, basal stem diameter,
crown volume, plant condition, mortality, stomatal
conductance, chlorophyll fluorescence, and leaf
relative water content to assess growth, survival,
and physiological performance in several common
tree island tree species exposed to varying hydrologic
conditions under shadehouse conditions. Second, we
compared the relative flood tolerances of these
species grown in the shadehouse to their observed
distribution along the hydrologic gradient in tree
islands under natural conditions.
We hypothesized that increased soil flooding
would adversely affect the growth and physiological
performance of plant species adapted to tree islands.
We predicted that the adverse responses would be
more pronounced and occur earlier in the tropical
hardwood tree species compared to the swamp
species, and the greatest effects would be seen under
conditions of prolonged soil flooding compared to
little or no soil flooding. We also expected that
relative flood tolerances of the species tested in the
shadehouse would reflect their observed distribution
along the hydrologic gradient in tree islands of the
southern Everglades.
METHODS
Species Studied
The names and distributions of the 12 tree species
used in this study, all native to Florida, are listed in
Table 1. Plant species are referred hereafter by their
genus name. Seven swamp forest species were
selected, all temperate in origin, except Annona and
Chrysobalanus, which are largely tropical. In south-
ern Florida, these species prefer wet habitats and are
common elements of the seasonally flooded portions
of tree islands. The remaining five upland forest
Jones et al., TREE RESPONSES TO HYDROLOGIC REGIMES
831
species are broadly distributed in the American
tropics, with southern Florida at the northern limit
of their ranges. Within the Everglades, they can be
found in the most elevated portions of the tree
islands, commonly referred to as hardwood ham-
mocks, as well as in other mesic forest sites. All 12
species are evergreen, with the exception of Salix
(deciduous) and Annona and Bursera (semi-decidu-
ous) (Tomlinson 1980).
Plant Acquisition and Experimental Design
During May and June of 2001, a minimum of 100
seedlings of each species was collected from tree
islands in Shark Slough, ENP. Seedlings were
transferred to small peat pots containing commercial
organic potting soil and placed in a glasshouse.
Eight weeks later, the plants were transferred to
26.5-L plastic pots (one plant per pot) containing
garden soil (pH 6.4) and placed in a shadehouse that
provided 50% full sun. After eight months of
growth, the plants were transferred to inflatable
swimming pools placed in the shadehouse in
preparation
for
the
flooding
experiment.
The
experimental
design
was
randomized
complete
block, with 12 species and 3 treatment combinations
represented twice in each of four blocks. Plants were
stratified among the three treatments according to
height and randomly placed within blocks.
The three treatments – high flood (HF), low flood
(LF), no flood (NF) – simulated the range of
hydrologic conditions found on tree islands in
Shark
Slough
as
determined
by
topographic
surveys. The bottoms of pots under HF and LF
were positioned at 0, 27.1, and 57 cm, respectively,
above the pool bottom. HF and LF represented
realistic hydrologic regimes found at the lower
and higher ends, respectively, of the tree island
swamp forest environmental gradient, while NF
represented hydrologic conditions found in the
relatively higher tropical hardwood forest portion
of a tree island.
Water levels, equal in all pools, were managed to
simulate variation in mean weekly water depths in
Shark Slough derived from hydrologic data re-
corded at U.S. Geological Survey ground-water
Table 1.
List of species and their distributions. Species are grouped by their habitat preference in southern Florida.
Species (Common Name)
1
Family
Distribution
2,3
Swamp Forest Species
Annona glabra L. (Pond Apple)
Annonaceae
southern Florida, West Indies, Mexico to
South America, West Africa
Chrysobalanus icaco L. (Coco Plum)
Chrysobalanaceae
southern Florida, West Indies, Mexico to
South America, West Africa
Ilex cassine L. (Dahoon)
Aquifoliaceae
Virginia to southern Florida, Cuba,
Bahama Islands
Magnolia virginiana L. (Sweetbay)
Magnoliaceae
eastern U.S. from Massachusetts to
southern Florida
Morella cerifera (L.) Small (Wax Myrtle)
Myricaceae
Bermuda, Greater Antilles, Central
America, eastern U.S. from New
Jersey to southern Florida
Persea borbonia L. (Red Bay)
Lauraceae
Gulf and Atlantic States of U.S.
Salix caroliniana Michx.
(Carolina Willow)
Salicaceae
southeastern U.S. from Virginia to southern
Florida, Cuba
Upland Forest Species
Bursera simaruba (L.) Sarg.
(Gumbo-Limbo)
Burseraceae
southern Florida, West Indies, Mexico to
northern South America
Coccoloba diversifolia Jacq. (Pigeon Plum)
Polygonaceae
southern Florida, West Indies
Eugenia axillaris (Sw.) Willd.
(White Stopper)
Myrtaceae
southern Florida, Bermuda, West Indies,
Central America
Sideroxylon foetidissimum Jacq. (Mastic)
Sapotaceae
southern Florida, West Indies, Mexico,
Belize
Simarouba glauca DC. (Paradise Tree)
Simaroubaceae
southern Florida, West Indies, Central
America
1
Wunderlin (1998),
2
Little (1978),
3
Tomlinson (1980).
832
WETLANDS, Volume 26, No. 3, 2006
hydrostation G620 (ENP) for the period 1990–1999.
At the start of the experiment (April 2002), co-
inciding with the beginning of the growing season in
southern Florida, pools were filled with piped water
(pH 7.7) to a depth of 9.4 cm. Every seven days
thereafter, water levels were adjusted by adding or
removing water according to the appropriate weekly
water depth. Following this schedule, under HF,
water levels exceeded the bottoms of pots on day 1
and reached the soil surface at week 10. Under LF,
the bottoms of pots first encountered the rising
water at week 10 and the soil surface would have
been inundated at week 28 had the experiment not
been terminated prematurely (see below). Because
pots would not be subjected to any flooding under
NF, they were randomly placed in a single ring
along the outer circumference of their assigned
pools. Under NF, plants required regular watering,
while all others were watered by hand whenever the
soil surface appeared dry, especially in the earlier
weeks of the experiment when water levels were still
low. For logistical reasons associated with rapid
growth in several species 3 treatment combinations
(e.g., competition among individuals for light, out-
growing
the
shadehouse), the
experiment
was
terminated after 25 weeks of treatment.
Mortality and Overall Plant Condition
Plant condition was determined by subjectively
assessing the overall health and vigor of each
individual tree and assigning a numerical score from
0 to 5. Visual observations on the conditions of
stems and foliage (e.g., coloration, growth) and the
occurrence of pests and disease (e.g., gall formation
and other insect damage) were used to make each
assessment. Individuals that displayed mostly green
foliage, leafy stems, active stem growth, and a lack
of pests and diseases were perceived as being healthy
and vigorous and assigned a value of 5. Individuals
that displayed yellowing leaves, premature leaf fall,
little or no stem growth, and/or insect or disease
damage were assigned a value of 1 to 4, depending
on the extent of the observed condition or condi-
tions. Dead individuals were assigned a value of 0.
Plant condition was the only measurement used in
this study in which all individuals were assessed
weekly. Mean condition of surviving individuals was
calculated weekly for each species and treatment,
and mortality was tracked on the same interval.
Plants that appeared dead were kept in their pools
and observed for several more weeks; dead trees
were removed from blocks.
Growth Measurements
Height, basal diameter, and crown volume of all
individuals were measured during the week prior to
treatment (week 0), then at 6, 12, 18, and 24 weeks
after initiation of the experiment. Height from the
top of the soil in each pot to the highest point (leaf,
stem, or meristem) of the tree was recorded to the
nearest cm. Basal diameter was measured using
a plastic dial caliper to the nearest tenth of
a millimeter. The diameter of single-stemmed trees
was
measured
at
a
position
along
the
stem
approximately 2 cm above the soil surface; red paint
was applied to the stem at this position to ensure
consistency when measuring. For individual trees
that produced multiple stems arising at or near soil
level, as in Morella and Salix, the most vigorous
stem
was
identified;
this
stem
was
measured
throughout the study. Crown volume was estimated
by modeling the tree crown as a series of conic
frustums of 10 cm height and terminal cones of
smaller height (Figure 1). Crown volumes of indi-
viduals with crown depths of at least 30 cm were
measured by taking crown-width measurements at
Figure 1.
Tree crown volume diagrams and equations for calculating volumes of (a) conic frustums (used for crown
depths greater to or equal to 30 cm) and (b) conic frustums and cones (used for crown depths less than 30 com).
V5volume, R5radius at the widest base, S5corresponding perpendicular radius, r5radius at the other base,
s5corresponding perpendicular radius, H5height of frustum and cone, h5height of cone.
Jones et al., TREE RESPONSES TO HYDROLOGIC REGIMES
833
intervals of 10 cm along the stem and totaling the
volume of each conic frustum (Figure 1a); volumes
of conic frustums were then summed to estimate
total crown volume. For individuals with crown
depths of less than 30 cm, lengths of the basal and
widest portions of the crowns, their perpendicular
widths, and their distance to crown tops were
measured (Figure 1b). Volumes of the conic frustum
and cone were then summed to estimate total crown
volume.
Physiological Measurements
Stomatal conductance, chlorophyll fluorescence,
and leaf relative water content were measured, at
weeks 3, 6, 12, 18, and 24. Four individuals from
each species-treatment combination (one from each
block) were selected for the measurements. Mea-
surements were conducted in the shadehouse on
young, fully expanded, intact, sun leaves. A different
leaf on the same tree was selected each sampling
week. We attempted to standardize measurements
by taking data under sunny, dry conditions between
the hours of 0900 and 1500, whenever possible.
A LI-1600 steady state porometer (LI-COR
Biosciences, Lincoln, NE, USA) was used to
measure conductance to water vapor. Attached
leaves were inserted in the porometer and conduc-
tance was measured after 60 s.
Fluorescence yield, expressed in terms of the ratio
of variable fluorescence to maximal fluorescence
(F
v
/F
m
), was measured using an OS1-FL pulse
modulated chlorophyll fluorometer (Opti-Sciences,
Inc., Tyngsboro, MA, USA). Leaves were dark-
adapted for 10 min before measurement.
To determine relative water content, two 6-mm-
diameter discs were punched from a leaf and
immediately weighed to obtain fresh weight (fw).
Discs
were then
wrapped
in saturated
paper
toweling and kept in small, sealed petri dishes in
the lab at room temperature for 16–20 h and
reweighed after blotting dry to obtain saturated
weight (sw). Discs were then placed in an oven at 60
uC for 24 h and reweighed to obtain the dry weight
(dw). Relative water content was calculated using
the formula of J. C
ˇ atsky´ (in Slavı´k 1974):
Relative water content %
ð Þ ~
fw
À dw
ð
Þ= sw À dw
ð
Þ | 100:
.Statistical Analyses
A split-plot design approach was used to analyze
the main effects of species, hydrologic treatment, and
time. In standard repeated measures ANOVA, which
resembles multivariate analysis of variance (MAN-
OVA), a single missing value causes the entire subject
to be omitted from analysis in a listwise deletion
procedure. When missing values are common, split-
plot ANOVA is an effective alternative approach to
analyze repeated measures data (Maceina et al.
1994). In our study, missing values resulted primarily
from the mortality of individual plants. In addition,
on rare occasions, physiological data could not be
collected because of equipment malfunction, or in
the aftermath of leaf shedding events or herbivore
outbreaks that affected few species.
In preliminary analyses, the effect of Block (the
Dostları ilə paylaş: |