The Interplay between Mangroves and Saltmarshes at the Transition between Temperate and Subtropical Climate in Florida (Stevens et al, 2006)
Ó Springer 2006
Wetlands Ecology and Management (2006) 14:435–444
DOI 10.1007/s11273-006-0006-3
-1
The interplay between mangroves and saltmarshes at the transition between
temperate and subtropical climate in Florida
Philip W. Stevens1,3,*, Sandra L. Fox1,2 and Clay L. Montague1
1
Department of Environmental Engineering Sciences, University of Florida, P.O. Box 116450, Gainesville, FL
32611, USA; 2St. Johns River Water Management District, 4049 Reid Street, Palatka, Florida 32177, USA;
3
Present address: Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute,
Charlotte Harbor Field Laboratory, 1481 Market Circle, Unit 1, Port Charlotte, FL 33953, USA; *Author
for correspondence (e-mail: philip.stevens@myfwc.com; phone: +1-941-255-7403; fax: +1-941-255-7400)
Received 23 November 2005; accepted in revised form 17 January 2006
Key words: Avicennia, Cedar Keys, Florida, Freezes, Mangrove crabs, Spartina, Succession
Abstract
The interplay between mangroves and saltmarshes at the temperate to subtropical transition in Florida
results in dramatic changes to the appearance of the coastal landscape. In the 1980s, freezes killed entire
mangrove forests dominated by black mangroves, Avicennia germinans (L.). Following the freezes, salt-
marshes dominated by smooth cordgrass, Spartina alterniflora Loisel, revegetated the intertidal zone. After
a decade of mild winters, however, mangroves are beginning to reclaim the area. The rate of mangrove
expansion was determined by comparing aerial photography (change from 1995 to 1999), and from
monitoring transects (over a 3 year period) on three of the Cedar Keys, Florida (Lat. 29°08¢). The rate of
mangrove expansion varied among islands, and the mechanism of expansion ranged from propagule-
trapping by saltmarshes along the edges of mangrove clumps to widespread dispersal and growth of
existing or newly imported propagules. A freeze occurred during the study, which may have set back
mangrove expansion by defoliating mangrove trees and resetting mangrove reproduction. Mangrove
expansion was projected to take 20–30 years for complete seedling cover. Given the possibility of global
climate change and its potential influence on the distribution of coastal vegetation, the timeframes and
implications to coastal wetland ecosystems involved in this regular interplay will provide valuable baseline
information for future studies.
mangrove survival and growth. At the transition
Introduction
between temperate and subtropical climate,
The distribution of mangroves in the northern however, an interplay occurs between mangroves
Gulf of Mexico is controlled primarily by freezing dominated by black mangroves, Avicennia
temperatures (Davis 1940; Lugo and Patterson- germinans (L.), and saltmarshes dominated by
Zucca 1977; Kangas and Lugo 1990). Saltmarshes smooth cordgrass, Spartina alterniflora Loisel.
dominate intertidal zones in more temperate Freezes favor the marsh, but mangroves prevail
climates, but mangroves replace saltmarshes at during periods of mild winters. In Florida, this
lower latitudes where warm temperatures allow transition zone occurs along the east coast north of
436
Cape Canaveral to St. Augustine, and along the vegetation, quantifying the timeframes involved in
west coast north of Tampa Bay to the Cedar Keys this regular interplay would provide valuable
(Kangas and Lugo 1990). baseline information for future studies. The pur-
During periods of mangrove succession, a pose of this study is to document the ongoing
variety of factors influence the distribution of succession from saltmarshes to mangrove forests
mangrove and saltmarsh ecotones. Saltmarsh veg- within the temperate to subtropical transition in
etation cannot grow in the shade of mangrove trees Florida and to estimate the time frame for man-
(Kangas and Lugo 1990). Freezes open the forest grove replacement.
floor to light, which allows fast pioneering salt-
marsh species to vegetate the intertidal zone
(Patterson et al. 1993). In the process, saltmarshes
remove available nutrients, possibly causing root Materials and methods
competition for nutrients among the saltmarsh
species and new mangrove seedlings (Patterson Study location
et al. 1993). Saltmarshes, however, may also facil-
The Cedar Keys are located in the Big Bend region
itate mangrove colonization by trapping mangrove
of Florida (Lat. 29°08¢ N; Figure 1), which has a
propagules (Lewis and Dunston 1975). The com-
low-energy coastline with a mean tidal range of
bination of propagule trapping and nutrient com-
1 m (NOAA 1988). Salinity is influenced by the
petition makes the overall effect of the marsh on
Gulf of Mexico, local rainfall, surface runoff, and
mangrove recolonization unclear. Other factors
freshwater discharge from the Suwannee River
influencing the distribution of mangrove and salt-
(average 296 m3 sÀ1) 26 km to the north (NOAA
marsh ecotones at local scales are hydroperiod,
1985). Mangrove forests persist on the Cedar Keys
salinity, sediment characteristics, and propagule
because the intensity of freezes is reduced by the
predation (e.g., Patterson and Mendelssohn 1991;
surrounding water (Laessle and Wharton 1959;
Clarke and Allaway 1993; Clarke and Myercough
Lugo and Patterson-Zucca 1977; Montague and
1993; Patterson et al. 1993). At regional scales,
Odum 1997). Patches of closed-canopy mangrove
mangrove transgressions into saltmarshes occur-
forest on the Cedar Keys exist as nearly mono-
ring in South Florida (Wanless et al. 1994) and
specific stands of A. germinans. Red mangrove,
Australia (Saintilan and Williams 1999) have been
Rhizophora mangle L., and white mangrove,
explained by changes in precipitation patterns, in-
Laguncularia racemosa (L.), only occur as scat-
creases in nutrient levels and sedimentation,
tered individuals, probably because they do not
revegetation of disturbed areas, altered tidal re-
tolerate low-temperatures as well as A. germinans
gimes, and sea-level rise; however, the overall dis-
(Markley et al. 1982; McMillan and Sherrod
tribution of mangroves in the northern Gulf of
1986). The canopy height of A. germinans forests
Mexico appears to be dominated by freeze fre-
prior to the mangrove kill in the 1980s was
quency (Kangas and Lugo 1990), with other fac-
approximately 3–4 m (Lugo and Patterson-Zucca
tors explaining local variation (e.g., Patterson and
1977; Watson 1986).
Mendelssohn 1991).
Small intertidal wetlands on Snake Key, North
Recent climate history in Florida provides an
Key, and Seahorse Key (Figure 1) were selected
opportunity to investigate the interplay between
for study. Sites on Snake Key and North Key were
mangroves and saltmarshes. Hard freezes during
in pockets of marsh nearly surrounded by upland
the 1980s killed entire mangrove forests in the
forests and naturally occurring sand berms. These
northern Gulf of Mexico (McMillan and Sherrod
barriers effectively block wind in the marshes.
1986; Watson 1986; Montague and Weigert 1990;
Water enters these pocket marshes via creeks. The
Montague and Odum 1997). Saltmarshes replaced
site on Seahorse Key, on the other hand, was fully
the intertidal zone within 4–5 years. However,
exposed to open water and north wind; water filled
6 years of mild winters have allowed mangroves to
and drained the marsh from an adjacent lagoon.
begin reclaiming the area. Mangroves now grow in
The saltmarsh vegetation on the islands consisted
clumps among saltmarsh vegetation. Given the
almost entirely of S. alterniflora except on Sea-
possibility of global climate change and its
horse Key, which had a mixture of S. alterniflora,
potential influence on the distribution of coastal
437
Figure 1. Map of the Cedar Keys showing study sites.
saltwort, Batis maritima L., and glasswort, Sali- pink, and beach and spoil appeared white. Hard
cornia virginica L. copies of the same US Geological Survey aerial
photography were obtained from local resource
managers and placed under a stereoscope to fur-
ther assist in the delineation of the vegetation.
Aerial photography/change analysis
From the digitizing process, maps were gener-
Gross changes in vegetation were determined by ated depicting mangrove and saltmarsh distribu-
comparing 1995 and 1999 digital orthophotos. tion within the study area for 1995 and 1999.
Digital orthophotos (1-m resolution digital ortho- Mangrove area for each study site was calculated
quads) taken by US Geological Survey National for each year based on the total area digitized for
Aerial Photography Program at the Cedar Keys each vegetation class. Change analyses were per-
were obtained from the Florida Land Boundary formed in ArcMap using the Union feature, which
Information System (www.labins.org). For each computes a geographic intersection of the input
study site, the digital orthophotos were displayed features from both years producing output fea-
at 1:1000 scale in ArcMap 8.3. Mangrove and tures attributed with the vegetation class of both
saltmarsh vegetation were digitized based primar- time periods. The changes in mangrove distribu-
ily on height, color and texture variance (e.g., tion that occurred over the 5-year period were
Higinbotham et al. 2004). The vegetation was displayed in ArcMap using symbology based on
easily distinguished in the orthophotos; marsh the attribution for both time periods (e.g., persis-
vegetation appeared black, mangrove vegetation tent mangrove, mangrove expansion, mangrove
appeared dark red, upland vegetation appeared loss). For example, mangrove areas present in
438
both 1995 and 1999 aerial photographs were regional differences (Wilcoxon signed rank test,
termed persistent mangroves, and mangrove 0.05 level of significance).
expansion referred to areas where mangroves were An estimate of mangrove seedling recruitment
present in 1999 but not in the earlier 1995 photos. was used as an indicator of mangrove expansion
Polygons less than 10 m2 were dropped from the along the transects to supplement results from
change analyses because these small areas were aerial photography. The estimate of seedling
considered to be within the error of the digitizing recruitment was determined by the following
process. The results of the change analyses were equation: Seedling Recruitment = D seedling
used to provide a mangrove expansion rate (new density + D tree density + tree death + seedling
mangrove area yearÀ1) at each study site. death. Changes in seedling and tree densities were
calculated by subtracting densities in November of
one year from densities in November of the fol-
lowing year. Tree death was easily measured by
Ground level transects
counting the number of leafless trees within a
Thirty-meter transects oriented perpendicular to quadrat, but seedling death was not measured. As
mangrove/saltmarsh ecotones were established at a result, seedling recruitment was underestimated.
the Cedar Keys to provide a finer scale for moni-
toring changes in mangrove and saltmarsh densi-
ties, and to supplement results of GIS analyses. A Results
total of five transect locations were established,
which were placed in areas where definitive man- Freeze observations
grove and saltmarsh ecotones were apparent from
A severe winter occurred between the 1995 and
1995 aerial photography. The approximate eleva-
1996 growing seasons (minimum air temperature
tions along transects with respect to Mean Lower
of À8 °C in January 1996 at a station near the
Low Water (MLLW) ranged between 65 and
Cedar Keys; NOAA 1915–1997). Many trees had
88 cm, which resulted in daily inundation by the
lost some or all of their leaves when measured in
1-m tides characteristic of the area.
Increases in mangrove tree and seedling densi- May 1996, but most recovered by November 1996.
Tree mortality along transects following the freeze
ties were important indicators of mangrove
was 12% on Seahorse Key, 5% on North Key, and
expansion, and saltmarsh stem densities were
4% on Snake Key. Although mangroves in this
expected to decrease as mangroves expanded and
region typically flower during spring and drop
shaded the underlying saltmarsh. Mangrove trees
propagules during fall, flowering and propagule
and seedlings (identified by lack of branching)
were counted in 4-m2 quadrats at 5-m intervals drop was not observed on the islands during 1996.
Mangroves were flowering again by May 1997.
along the transects for 2 years (1995 and 1996) in
May and November, which approximate the usual
growing season for this region. Densities were also
measured in May 1997 along Seahorse Key and Aerial photography/change analysis
North Key transects and in November 1997 along
The study sites were at different stages of man-
transects on Snake Key. Densities of saltmarsh
stems were counted in 0.25-m2 quadrats at 5-m grove succession at the beginning of the study
period in 1995 (3% cover on Seahorse Key, 20%
intervals along the transects. Saltmarsh stem den-
sities within the seven 0.25-mÀ2 quadrats along cover on North Key, and 33% cover on Snake
Key, Table 1). Mangrove cover expanded to
each transect were summed to give number of
stems 1.75 mÀ2. Seasonal and annual changes in occupy ca. another 20% of the intertidal zones at
each study site during the 5-year period. From the
saltmarsh stem density was calculated for each
change analyses, the site on North Key expanded
transect by subtracting May values from November
at a rate of about 300 m2 yearÀ1, and the sites on
values (seasonal) or November values from sub-
Seahorse Key and Snake Key expanded at rates
sequent November values (annual). The five tran-
exceeding 840 m2 yearÀ1. Mangrove expansion
sects were used as replicates (for non-parametric
on North Key and Snake Key occurred along
analysis) to evaluate statistical significance of
439
the boundaries of persistent mangrove clumps season was defined as winters in which the mini-
mum temperature fell below À5 °C at several
(Figure 2). The mangrove expansion on Seahorse
Key also occurred along the boundaries of per- temperature stations within an area. During 1932–
sistent mangrove clumps, but much of the expan-
sion at this site was represented by many small
mangrove clumps distributed widely throughout
the marsh (Figure 2). Some mangrove loss was
evident at the North Key and Snake Key study
sites, but this loss was greatly outweighed by
mangrove expansion.
Ground level transects
Initial canopy height (used simply as an indicator of
where mangrove and saltmarsh boundaries occur)
and seedling recruitment for the five transects are
shown in Figure 3. Seedling recruitment along the
two Snake Key transects was highest around
mangrove clumps and declined with distance.
Mangrove seedling recruitment along Snake Key–
Transect 2 occurred up to 5 m from mangrove
clumps into saltmarsh areas. Seedling recruitment
along Snake Key–Transect 1 was as high as
76 seedlings mÀ2 yearÀ1 and seedlings recruited at
least 15 m into saltmarsh areas. Although seedling
recruitment on Seahorse Key occurred, it primarily
occurred within the mangrove clump (Transect 1),
and along the area fringing the lagoon (Transect 2).
On the North Key transect, seedling recruitment
did not occur. Seedling recruitment did not occur
along any transect in 1997.
Seasonal changes in saltmarsh stem density
along transects were not significant during 1995
(p = 0.35), but were significant during 1996
(p = 0.04). Annual changes in saltmarsh stem
density along transects were higher in 1996 than in
1995 (p = 0.04). These increases occurred along
transect locations that were farthest from man-
grove clumps.
Discussion
Frequency of freezes in Northern Florida
Figure 2. GIS change analysis between 1995 and 1999 on Snake
Freeze cycles in Florida have been explored by Key (a), North Key (b), and Seahorse Key (c). White refers to
upland or beach, blue refers to water, pink refers to persistent
scientists interested in the risks associated with
mangrove, red refers to mangrove expansion, opaque refers to
citrus production. In one such study, Miller and
mangrove loss, grey refers to remaining intertidal zone, dark
Downton (1992) defined a local killing freeze as grey refers to change in intertidal zone (low tide in 1995 aerial
winters in which the minimum temperature fell photography vs. high tide in 1999 aerial photography). Loca-
below À6.7 °C, and a regionwide severe freeze tions of 30-m transects are also shown.
440
Table 1. Total intertidal area, mangrove area, and percent mangrove cover at each study site for 1995 and 1999 determined from aerial
photography. Also shown are mangrove expansion rates at each study site determined from GIS change analysis.
Aerial photography
Location 1995 1999 Change analysis
mangrove
Intertidal Mangrove Mangrove Mangrove Mangrove expansion
area (m2) area (m2) area (m2)
cover (%) cover (%) rate (m2 yearÀ1)
Snake Key 15,333 5114 33 9044 59 862
North Key 6948 1356 20 2565 37 300
Seahorse 21,186 609 3 4822 23 842
Key
1980, a regionwide severe freeze season occurred catastrophic freeze in the 19th century, for
approximately every 8 years in Florida. The freeze example, involved a series of severe winters dur-
frequency changed substantially during the 1980s ing 1895–1905, and the series of severe winters
as tree killing freezes occurred in 4 out of 5 years during 1977–1989 comprised the catastrophic
between 1981 and 1985 (Miller and Downtown freeze of the 20th century.
1992).
The freezes that affect the citrus industry appear
to coincide with freezes that affect mangroves and Impact of freezes on mangroves
probably subtropical species in general. Major
impacts to mangroves resulting from freezes According to Lugo and Patterson-Zucca (1977),
occurred in 1962 (Dr Frank Maturo, University of freeze damage to mangroves is greatest in areas
Florida, personal communication), 1977 (Lugo more exposed to cold wind. The Seahorse Key
and Patterson-Zucca 1977), 1981, 1983, 1985, 1989 marsh is exposed to cold north winds that follow
(personal observation) and 1996 (this study). cold fronts in this region. This may explain the
These accounts are consistent with an annual high mangrove mortality (12%) at Seahorse Key
minimum temperate of less than À6.7 °C (defini- following the January 1996 freeze (compared to
tion of a local killing freeze for citrus; Miller and 5% on North Key and 4% on Snake Key). The
Downtown 1992) recorded at the Cedar Key marshes on Snake Key and North Key, where
temperature station (NOAA 1915–1997). How- mangroves are more extensive, are isolated from
ever, temperatures on the outermost islands north wind by natural barriers such as berms and
(Snake Key, North Key, and Seahorse Key) are upland forests. Even on these islands, however,
slightly warmer than Cedar Key (Laessle and some mortality was evident from direct observa-
Wharton 1959). Therefore, the lowest temperature tions along transects and from results of the
change analyses that show some small net losses
that can be tolerated by mangroves at this latitude
may be close to Davis’ (1940) original suggestion of mangrove, which likely occurred from areas
of À4 °C. affected by freeze that had not yet fully recovered.
A single severe winter, however, may only set Although Seahorse Key appears to be the
back mangrove development without killing a most vulnerable to freezes, the impacts of the
substantial number of trees. At the Cedar Keys, January 1996 freeze reset mangrove development
mangroves have recovered within one year fol- on all three islands. The failure of mangroves to
lowing a severe freeze (Lugo and Patterson-Zucca flower and produce propagules during the 1996
growing season presumably resulted from deple-
1977; personal observation 1981). However, a
series of severe freezes is catastrophic, as evi- tion of the plants’ resources during recovery from
denced by the mangrove kill in the 1980s. Such a the previous winter. The lack of seedling recruit-
catastrophic series of severe winters may recur ment in 1997 suggests that mangrove expansion at
only once every century in Florida (Winsberg the Cedar Keys depends largely on a local source
1990; Miller and Downton 1992). The of propagules rather than imports from more
441
Snake Key - Transect 1
2.5 100
(Seedlings m-2 yr-1)
Canopy Height (m) 80
2
Recruitment
60
1.5 canopy height
40
20 recruitment 96
1
0 recruitment 97
0.5
-20
0 -40
0m 5m 10m 15m 20m 25m 30m
2.5 80
Snake Key - Transect 2
Canopy Height (m)
(Seedlings m-2 yr-1)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
2.5 80
North Key
(Seedlings m-2 yr-1)
Canopy Height (m)
2 60
Recruitment canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
Seahorse Key - Transect 1
2.5 80
(Seedlings m-2 yr-1)
Canopy Height (m)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
2.5 80
Seahorse Key - Transect 2
Canopy Height (m)
(Seedlings m-2 yr-1)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
Distance Along Transect
Figure 3. Canopy height and seedling recruitment along each transect.
442
southerly locations, such as Tampa Bay (a net scattered throughout the marsh that grew to be-
northward transport of Gulf of Mexico waters come trees during the study period. Some of the
occurs in this region; Clarke 1997). Hence, freezes smallest ‘clumps’ apparent in the digital ortho-
that impact mangrove reproduction at the Cedar photos may result from one or two trees that
Keys cause a major setback to mangrove succession. developed a broad canopy (possibly an advanta-
Freezes that impact mangroves, however, may geous strategy when competing alone with salt-
indirectly affect saltmarshes. Saltmarsh stem densi- marsh). Some of these small clumps among
ties at the Cedar Keys were higher in 1996 than in saltmarsh vegetation were also evident on Snake
1995, even though mangrove expansion was Key.
expected to result in lower saltmarsh stem densities.
Although the increases followed a freeze that defo-
Time-frame for mangrove succession
liated mangroves, increased light availability cannot
explain the changes in saltmarsh stem density
The freezes in 1983 and 1985 killed 98% of man-
because they occurred primarily in areas away from
groves at the Cedar Keys (Montague and Odum
mangrove clumps. However, recycled nutrients
1997), which may have eliminated the local man-
from fallen mangrove leaves might explain the
grove propagule source. It is not known whether
increase. Freezes may have resulted in greater
local seedlings survived, which would have pro-
amounts of detritus as mangrove biomass was
vided a seed bank for subsequent mangrove suc-
defoliated or killed, and subsequent detrital break-
cession, or if new recruitment was initially
down increased nutrients available for plant growth.
dependent on propagules from areas further south.
Regardless, as more mangroves continue to reach
reproductive maturity at the Cedar Keys, the rate
Mangrove expansion
of mangrove colonization should increase. In the
absence of major freezes, the time it will take for
The mechanism of mangrove seedling recruitment
mangroves to completely cover the marsh on Snake
and expansion appeared to differ among the study
Key is another 8 years (15,333 m2 total area –
sites. On Snake Key, trapping of mangrove propa-
9044 m2 1999 mangrove area/mangrove expansion
gules by saltmarsh (Lewis and Dunston 1975) could
rate of 862 m2 yearÀ1). Thus, the time frame for
account for the high seedling recruitment near
complete mangrove recolonization on Snake Key
mangrove clumps. Propagules produced by the
since the freeze kill during the 1980s is nearly
mangrove clumps on Snake Key were immediately
20 years. Using the same calculation for North
trapped in saltmarsh within 5–15 m of their source.
Key and Seahorse Key, these study sites will take
Without the adjacent saltmarsh, more mangrove
25–30 years for mangroves to again reclaim the
propagules may have been exported rather than
marsh. Periodic freezes, however, ensure that some
being retained in the vicinity of the parent trees.
saltmarsh cover will persist. This mixture of man-
High rates of seedling recruitment along the edges
grove forests and saltmarshes may continue for
of the mangrove clumps and subsequent growth of
several decades until another catastrophic series of
these mangroves into trees probably explains why
freezes kill mangrove forests, and allow saltmarshes
Snake Key has the greatest mangrove cover and
to once again occupy the intertidal zone.
highest expansion rates among the study sites.
On Seahorse Key, mangrove seedling recruit-
ment occurred within the existing mangrove clump Implications for coastal wetland communities
and along the transect adjacent to the lagoon also
As the intertidal zone changes between mangrove
suggesting that mangrove seedlings may remain
forests and saltmarshes, faunal changes should
near parent trees (Blanchard and Prado 1995;
occur. Both mangroves and saltmarshes provide
McKee 1995). However, numerous mangrove
food and refuge for fiddler crabs, periwinkles,
clumps on Seahorse Key appeared to ‘sprinkle in’
clams, and oysters (Odum et al. 1982; Stout 1984).
throughout the marsh (apparent from change
However, the greater volume and structural com-
analysis), particularly at the southern end of the
plexity of mangroves supports additional fauna
study site adjacent to the lagoon. This mechanism
such as the mangrove tree crab, Aratus pisonii
of expansion may result from single seedlings
443
(Milne Edwards), the coffee-bean snail, Melampus of severe winters, however, the combined energy
coffeus (Linnaeus), the mangrove crab, Goniopsis drains of salinity and frost limit the amount of
cruentata (Latreille), the ladder-horn snail, Ceri- structure that can be maintained. As a result,
thidea scalariformis (Say), the mangrove rivulus mangroves are killed and the coastal wetlands
fish, Rivulus marmoratus Poey, and the mangrove rapidly retrogress to saltmarshes. Following a
cuckoo, Coccyzus minor (Gmelin) (Odum et al. mangrove kill, saltmarshes occupy the intertidal
1982; Davis et al. 1995; Taylor et al. 1995). In zone within a period of only 5–10 years. The
mangroves, the periwinkle Littoraria angulifera saltmarsh vegetation, however, may stabilize
(Lamarck) replaces the Littoraria irrorata (Say) of marsh soils and maintain certain wetland func-
saltmarshes (Odum et al. 1982). Also, mangrove tions (e.g., juvenile fish habitat, detritus produc-
forests may attract nesting colonial birds (Watson tion). Whether the overall effect of saltmarshes
1986). Conversely, saltmarsh-dependent bird spe- facilitates mangrove colonization by trapping
cies such as seaside sparrows, Ammospiza maritima mangrove propagules or retards mangrove rees-
(Wilson), and long-billed marsh wrens, Cistotho- tablishment by outcompeting mangrove seed-
rus palustris (Wilson), may leave the area when lings remains equivocal. Nevertheless, saltmarshes
saltmarsh is overtaken by mangroves (Post and maintain emergent habitat in the intertidal zone in
Greenlaw 1994). the interim between occupations by mangroves.
During the present study, saltmarsh-dependent Although coastal wetlands alternate between
A. maritima and C. palustris were only encoun- mangroves and saltmarshes at the temperate to
tered at the Seahorse Key study site where salt- subtropical transition in Florida, overall produc-
marsh cover remains extensive (personal tion and survival of wetland-dependent biota are
observation). Although pelicans were observed maintained despite periodic disturbances.
nesting in mangroves on Seahorse Key prior to the
catastrophic freezes in the 1980s (Watson 1986), Acknowledgements
colonial birds were not observed in the developing
mangroves of the present study. The only fauna We are grateful to F. Maturo and the University
unique to mangroves on the islands was A. pisonii. of Florida Marine Lab at Seahorse Key for pro-
Aratus pisonii reached densities of 1 crab mÀ3 viding logistical support during this study. We also
prior to the 1996 freeze (unpublished data), which thank K. Litzenberger, refuge manager of the
are within the range reported for mature mangrove Lower Suwannee National Wildlife Refuge, for
forests of South Florida (1–4 crabs mÀ3; Beever permitting us to conduct research within the Cedar
et al. 1979). Mangrove fauna may not only be Keys National Wildlife Refuge. Florida Sea Grant
dependent on the habitat present, but also on and the Aylesworth Foundation provided financial
other environmental factors, especially tempera- support. T. Crisman and S. Vince gave thoughtful
ture (e.g., R. marmoratus, Taylor 1993). A long advice and guidance throughout this study, and
period of time without freezes may allow a richer commented on early drafts. Finally, we acknowl-
mangrove fauna to develop. Alternatively, the lag edge all of the volunteers who assisted with
in development of mangrove fauna may exceed the the fieldwork especially G. Stevens, J. Stevens,
frequency of freezes at this latitude. B. Clarke, and A. Wilson.
References
Implications for coastal wetland ecosystems
Beever J.W., Simberloff D. and King L.L. 1979. Herbivory and
Odum (1983) describes succession and retrogres- predation by the mangrove tree crab Aratus pisonii. Oecolo-
sion as ‘the self-organizational process by which gia 43: 317–328.
ecosystems develop structure and processes from Blanchard J. and Prado G. 1995. Natural regeneration of
Rhizophora mangle in strip clearcuts in Northwest Ecuador.
available energies.’ In the absence of freezes,
Biotropica 27: 160–167.
mangroves succeed saltmarshes perhaps because
Clarke P.J. and Allaway W.G. 1993. The regeneration niche of
they are able to channel more energy into greater the grey mangrove (Avicennia marina): effects of salinity,
stature and biomass, thereby shading out salt- light, and sediment factors on establishment, growth, and
marshes (Kangas and Lugo 1990). During periods survival in the field. Oecologia 93: 548–556.
444
Clarke P.J. and Myerscough P.J. 1993. The intertidal distribu- from the Gulf of Mexico. St. Lucie Press, Delray Beach, FL,
tion of the grey mangrove (Avicennia marina) in southeastern pp. 1–33.
Australia: the effects of physical conditions, interspecific National Oceanic and Atmospheric Administration (NOAA)
competition, and predation on propagule establishment and 1915–1997. Climatological Data. Florida Section, National
survival. Aust. J. Ecol. 18: 307–315. Climatic Center, Asheville, North Carolina.
Clarke J. 1997. Atlantic Pilot Atlas. International Marine, National Oceanic and Atmospheric Administration (NOAA)
Camden, ME. 1985. Gulf of Mexico: Coast and Ocean Zones: Strategic
Davis J.H. 1940. The ecology and geologic role of mangroves in Assessment: Data Atlas.
Florida. Papers from Tortugas Lab, 32. National Oceanic and Atmospheric Administration (NOAA).
Davis W.P., Taylor D.S. and Turner B.J. 1995. Does the aut- 1988. Sea-level variations for the U.S. 1855–1986.
ecology of the mangrove rivulus fish (Rivulus marmoratus) Odum W.E. 1983. Systems Ecology. John Wiley and Sons, New
reflect a paradigm for mangrove ecosystem sensitivity? Bull. York, NY, 644 pp.
Mar. Sci. 57: 208–214. Odum W.E., McIvor C.C. and Smith T.J. III 1982. The ecology
Higinbotham C.B., Alber M. and Chalmers A.G. 2004. Analysis of the mangroves of south Florida: a community profile. U.S.
of tidal marsh vegetation patterns in two Georgia estuaries Fish and Wildlife Service, Office of Biological Services,
using aerial photography and GIS. Estuaries 27: 670–683. Washington, DC, FWS/OBS-81/24, 144 pp.
Kangas P.C. and Lugo A.E. 1990. The distribution of man- Patterson C.S. and Mendelssohn I.A. 1991. A comparison of
groves and saltmarsh in Florida. J. Trop. Ecol. 31: 32–39. physiochemical variables across plant zones in a mangal/salt
Laessle A.M. and Wharton C.H. 1959. Northern extensions in marsh community in Louisiana. Wetlands 11: 139–161.
the recorded ranges of plants on Sea Horse Key and associ- Patterson C.S., Mendelssohn I.A. and Swenson E.M. 1993.
ated Key, Levy County, Florida. Q. J. Fla. Acad. Sci. 22: Growth and survival of Avicennia germinans seedlings in a
105–113. mangal/salt marsh community in Louisiana, USA. J. Coast.
Lewis R.R. and Dunston F.M. 1975. The possible role of Res. 9: 801–810.
Spartina alterniflora Loisel in establishment of mangroves in Post W. and Greenlaw J.S. 1994. Seaside sparrow, No. 127. In:
Florida. In: Lewis R.R. (ed.), Proceedings 2nd Annual Poole A. and Gill F. (eds), The Birds of North America. The
Conference on Restoration of Coastal Vegetation in Florida. American Ornithologists Union and the Academy of Natural
Hillsborough Community College, Tampa, FL, pp. 82–100. Science of Philadelphia, Philadelphia, PA.
Lugo A.E. and Patterson-Zucca C.P. 1977. The impact of low Saintilan N. and Williams R.J. 1999. Mangrove transgression
temperature stress on mangrove structure and growth. Trop. into saltmarsh environments in south-east Australia. Glob.
Ecol. 18: 149–160. Ecol. Biogeogr. 8: 117–124.
Markley J.L., McMillian C. and Thompson G.A. Jr. 1982. Stout J.P. 1984. The ecology of irregularly flooded saltmarshes
Latitudinal differentation in response to chilling temperatures of the northeastern Gulf of Mexico: a community profile.
among populations of three mangroves, Avicennia germinans, U.S. Fish and Wildlife Service, Office of Biological Services,
Laguncularia racemosa, and Rhizophora mangle, from the Washington DC, 85 (7.1), 98 pp.
western tropical Atlantic and Pacific Panama. Can. J. Bot. Taylor D.S. 1993. Notes on the impact of the December 1989
60: 2704–2715. freeze on local populations of Rivulus marmoratus in Florida,
McKee K.L. 1995. Seedling recruitment patterns in a Belizean with additional distribution records in the state. Fla. Sci. 56:
mangrove forest: effects of establishment ability and physico- 129–134.
chemical factors. Oecologia 101: 448–460. Taylor D.S., Davis W.P. and Turner B.J. 1995. Rivulus mar-
McMillan C. and Sherrod C.L. 1986. The chilling tolerance of moratus: ecology of distributional patterns in Florida and the
black mangrove, Avicennia germinans, from the Gulf of central Indian River Lagoon. Bull. Mar. Sci. 57: 202–207.
Mexico coast of Texas, Louisiana, and Florida. Contrib. Wanless H.R., Parkinson R.W. and Tedesco L.P. 1994. Sea level
Mar. Sci. 29: 9–16. control on stability of Everglades wetlands. In: Davis S.M.
Miller K.A. and Downton M.W. 1992. The freeze risk to Florida and Ogden J.C. (eds), Everglades: The Ecosystem and its
citrus. Part I: investment decisions. J. Clim. 6: 354–363. Restoration. St. Lucie Press, Delray Beach, FL, pp. 199–223.
Montague C.L. and Weigert R.G. 1990. Saltmarshes. In: Myers R. Watson A.M. 1986. Nutrient-production relations at Seahorse
and Ewel J. (eds), Ecosystems of Florida. University of Key Lagoon, Florida: some consequences of shorebird
Central Florida Press, Orlando, FL, pp. 481–516. accumulations. M.S. Thesis, University of Florida, Gaines-
Montague C.L. and Odum H.T. 1997. The intertidal marshes ville, FL, 87 pp.
of Florida’s Gulf Coast. In: Coultas C.L. and Hsieh Y.P. Winsberg M.D. 1990. Florida Weather. University of Central
(eds), Ecology and Management of Tidal Marshes: A Model Florida Press, Orlando, FL.
Wetlands Ecology and Management (2006) 14:435–444
DOI 10.1007/s11273-006-0006-3
-1
The interplay between mangroves and saltmarshes at the transition between
temperate and subtropical climate in Florida
Philip W. Stevens1,3,*, Sandra L. Fox1,2 and Clay L. Montague1
1
Department of Environmental Engineering Sciences, University of Florida, P.O. Box 116450, Gainesville, FL
32611, USA; 2St. Johns River Water Management District, 4049 Reid Street, Palatka, Florida 32177, USA;
3
Present address: Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute,
Charlotte Harbor Field Laboratory, 1481 Market Circle, Unit 1, Port Charlotte, FL 33953, USA; *Author
for correspondence (e-mail: philip.stevens@myfwc.com; phone: +1-941-255-7403; fax: +1-941-255-7400)
Received 23 November 2005; accepted in revised form 17 January 2006
Key words: Avicennia, Cedar Keys, Florida, Freezes, Mangrove crabs, Spartina, Succession
Abstract
The interplay between mangroves and saltmarshes at the temperate to subtropical transition in Florida
results in dramatic changes to the appearance of the coastal landscape. In the 1980s, freezes killed entire
mangrove forests dominated by black mangroves, Avicennia germinans (L.). Following the freezes, salt-
marshes dominated by smooth cordgrass, Spartina alterniflora Loisel, revegetated the intertidal zone. After
a decade of mild winters, however, mangroves are beginning to reclaim the area. The rate of mangrove
expansion was determined by comparing aerial photography (change from 1995 to 1999), and from
monitoring transects (over a 3 year period) on three of the Cedar Keys, Florida (Lat. 29°08¢). The rate of
mangrove expansion varied among islands, and the mechanism of expansion ranged from propagule-
trapping by saltmarshes along the edges of mangrove clumps to widespread dispersal and growth of
existing or newly imported propagules. A freeze occurred during the study, which may have set back
mangrove expansion by defoliating mangrove trees and resetting mangrove reproduction. Mangrove
expansion was projected to take 20–30 years for complete seedling cover. Given the possibility of global
climate change and its potential influence on the distribution of coastal vegetation, the timeframes and
implications to coastal wetland ecosystems involved in this regular interplay will provide valuable baseline
information for future studies.
mangrove survival and growth. At the transition
Introduction
between temperate and subtropical climate,
The distribution of mangroves in the northern however, an interplay occurs between mangroves
Gulf of Mexico is controlled primarily by freezing dominated by black mangroves, Avicennia
temperatures (Davis 1940; Lugo and Patterson- germinans (L.), and saltmarshes dominated by
Zucca 1977; Kangas and Lugo 1990). Saltmarshes smooth cordgrass, Spartina alterniflora Loisel.
dominate intertidal zones in more temperate Freezes favor the marsh, but mangroves prevail
climates, but mangroves replace saltmarshes at during periods of mild winters. In Florida, this
lower latitudes where warm temperatures allow transition zone occurs along the east coast north of
436
Cape Canaveral to St. Augustine, and along the vegetation, quantifying the timeframes involved in
west coast north of Tampa Bay to the Cedar Keys this regular interplay would provide valuable
(Kangas and Lugo 1990). baseline information for future studies. The pur-
During periods of mangrove succession, a pose of this study is to document the ongoing
variety of factors influence the distribution of succession from saltmarshes to mangrove forests
mangrove and saltmarsh ecotones. Saltmarsh veg- within the temperate to subtropical transition in
etation cannot grow in the shade of mangrove trees Florida and to estimate the time frame for man-
(Kangas and Lugo 1990). Freezes open the forest grove replacement.
floor to light, which allows fast pioneering salt-
marsh species to vegetate the intertidal zone
(Patterson et al. 1993). In the process, saltmarshes
remove available nutrients, possibly causing root Materials and methods
competition for nutrients among the saltmarsh
species and new mangrove seedlings (Patterson Study location
et al. 1993). Saltmarshes, however, may also facil-
The Cedar Keys are located in the Big Bend region
itate mangrove colonization by trapping mangrove
of Florida (Lat. 29°08¢ N; Figure 1), which has a
propagules (Lewis and Dunston 1975). The com-
low-energy coastline with a mean tidal range of
bination of propagule trapping and nutrient com-
1 m (NOAA 1988). Salinity is influenced by the
petition makes the overall effect of the marsh on
Gulf of Mexico, local rainfall, surface runoff, and
mangrove recolonization unclear. Other factors
freshwater discharge from the Suwannee River
influencing the distribution of mangrove and salt-
(average 296 m3 sÀ1) 26 km to the north (NOAA
marsh ecotones at local scales are hydroperiod,
1985). Mangrove forests persist on the Cedar Keys
salinity, sediment characteristics, and propagule
because the intensity of freezes is reduced by the
predation (e.g., Patterson and Mendelssohn 1991;
surrounding water (Laessle and Wharton 1959;
Clarke and Allaway 1993; Clarke and Myercough
Lugo and Patterson-Zucca 1977; Montague and
1993; Patterson et al. 1993). At regional scales,
Odum 1997). Patches of closed-canopy mangrove
mangrove transgressions into saltmarshes occur-
forest on the Cedar Keys exist as nearly mono-
ring in South Florida (Wanless et al. 1994) and
specific stands of A. germinans. Red mangrove,
Australia (Saintilan and Williams 1999) have been
Rhizophora mangle L., and white mangrove,
explained by changes in precipitation patterns, in-
Laguncularia racemosa (L.), only occur as scat-
creases in nutrient levels and sedimentation,
tered individuals, probably because they do not
revegetation of disturbed areas, altered tidal re-
tolerate low-temperatures as well as A. germinans
gimes, and sea-level rise; however, the overall dis-
(Markley et al. 1982; McMillan and Sherrod
tribution of mangroves in the northern Gulf of
1986). The canopy height of A. germinans forests
Mexico appears to be dominated by freeze fre-
prior to the mangrove kill in the 1980s was
quency (Kangas and Lugo 1990), with other fac-
approximately 3–4 m (Lugo and Patterson-Zucca
tors explaining local variation (e.g., Patterson and
1977; Watson 1986).
Mendelssohn 1991).
Small intertidal wetlands on Snake Key, North
Recent climate history in Florida provides an
Key, and Seahorse Key (Figure 1) were selected
opportunity to investigate the interplay between
for study. Sites on Snake Key and North Key were
mangroves and saltmarshes. Hard freezes during
in pockets of marsh nearly surrounded by upland
the 1980s killed entire mangrove forests in the
forests and naturally occurring sand berms. These
northern Gulf of Mexico (McMillan and Sherrod
barriers effectively block wind in the marshes.
1986; Watson 1986; Montague and Weigert 1990;
Water enters these pocket marshes via creeks. The
Montague and Odum 1997). Saltmarshes replaced
site on Seahorse Key, on the other hand, was fully
the intertidal zone within 4–5 years. However,
exposed to open water and north wind; water filled
6 years of mild winters have allowed mangroves to
and drained the marsh from an adjacent lagoon.
begin reclaiming the area. Mangroves now grow in
The saltmarsh vegetation on the islands consisted
clumps among saltmarsh vegetation. Given the
almost entirely of S. alterniflora except on Sea-
possibility of global climate change and its
horse Key, which had a mixture of S. alterniflora,
potential influence on the distribution of coastal
437
Figure 1. Map of the Cedar Keys showing study sites.
saltwort, Batis maritima L., and glasswort, Sali- pink, and beach and spoil appeared white. Hard
cornia virginica L. copies of the same US Geological Survey aerial
photography were obtained from local resource
managers and placed under a stereoscope to fur-
ther assist in the delineation of the vegetation.
Aerial photography/change analysis
From the digitizing process, maps were gener-
Gross changes in vegetation were determined by ated depicting mangrove and saltmarsh distribu-
comparing 1995 and 1999 digital orthophotos. tion within the study area for 1995 and 1999.
Digital orthophotos (1-m resolution digital ortho- Mangrove area for each study site was calculated
quads) taken by US Geological Survey National for each year based on the total area digitized for
Aerial Photography Program at the Cedar Keys each vegetation class. Change analyses were per-
were obtained from the Florida Land Boundary formed in ArcMap using the Union feature, which
Information System (www.labins.org). For each computes a geographic intersection of the input
study site, the digital orthophotos were displayed features from both years producing output fea-
at 1:1000 scale in ArcMap 8.3. Mangrove and tures attributed with the vegetation class of both
saltmarsh vegetation were digitized based primar- time periods. The changes in mangrove distribu-
ily on height, color and texture variance (e.g., tion that occurred over the 5-year period were
Higinbotham et al. 2004). The vegetation was displayed in ArcMap using symbology based on
easily distinguished in the orthophotos; marsh the attribution for both time periods (e.g., persis-
vegetation appeared black, mangrove vegetation tent mangrove, mangrove expansion, mangrove
appeared dark red, upland vegetation appeared loss). For example, mangrove areas present in
438
both 1995 and 1999 aerial photographs were regional differences (Wilcoxon signed rank test,
termed persistent mangroves, and mangrove 0.05 level of significance).
expansion referred to areas where mangroves were An estimate of mangrove seedling recruitment
present in 1999 but not in the earlier 1995 photos. was used as an indicator of mangrove expansion
Polygons less than 10 m2 were dropped from the along the transects to supplement results from
change analyses because these small areas were aerial photography. The estimate of seedling
considered to be within the error of the digitizing recruitment was determined by the following
process. The results of the change analyses were equation: Seedling Recruitment = D seedling
used to provide a mangrove expansion rate (new density + D tree density + tree death + seedling
mangrove area yearÀ1) at each study site. death. Changes in seedling and tree densities were
calculated by subtracting densities in November of
one year from densities in November of the fol-
lowing year. Tree death was easily measured by
Ground level transects
counting the number of leafless trees within a
Thirty-meter transects oriented perpendicular to quadrat, but seedling death was not measured. As
mangrove/saltmarsh ecotones were established at a result, seedling recruitment was underestimated.
the Cedar Keys to provide a finer scale for moni-
toring changes in mangrove and saltmarsh densi-
ties, and to supplement results of GIS analyses. A Results
total of five transect locations were established,
which were placed in areas where definitive man- Freeze observations
grove and saltmarsh ecotones were apparent from
A severe winter occurred between the 1995 and
1995 aerial photography. The approximate eleva-
1996 growing seasons (minimum air temperature
tions along transects with respect to Mean Lower
of À8 °C in January 1996 at a station near the
Low Water (MLLW) ranged between 65 and
Cedar Keys; NOAA 1915–1997). Many trees had
88 cm, which resulted in daily inundation by the
lost some or all of their leaves when measured in
1-m tides characteristic of the area.
Increases in mangrove tree and seedling densi- May 1996, but most recovered by November 1996.
Tree mortality along transects following the freeze
ties were important indicators of mangrove
was 12% on Seahorse Key, 5% on North Key, and
expansion, and saltmarsh stem densities were
4% on Snake Key. Although mangroves in this
expected to decrease as mangroves expanded and
region typically flower during spring and drop
shaded the underlying saltmarsh. Mangrove trees
propagules during fall, flowering and propagule
and seedlings (identified by lack of branching)
were counted in 4-m2 quadrats at 5-m intervals drop was not observed on the islands during 1996.
Mangroves were flowering again by May 1997.
along the transects for 2 years (1995 and 1996) in
May and November, which approximate the usual
growing season for this region. Densities were also
measured in May 1997 along Seahorse Key and Aerial photography/change analysis
North Key transects and in November 1997 along
The study sites were at different stages of man-
transects on Snake Key. Densities of saltmarsh
stems were counted in 0.25-m2 quadrats at 5-m grove succession at the beginning of the study
period in 1995 (3% cover on Seahorse Key, 20%
intervals along the transects. Saltmarsh stem den-
sities within the seven 0.25-mÀ2 quadrats along cover on North Key, and 33% cover on Snake
Key, Table 1). Mangrove cover expanded to
each transect were summed to give number of
stems 1.75 mÀ2. Seasonal and annual changes in occupy ca. another 20% of the intertidal zones at
each study site during the 5-year period. From the
saltmarsh stem density was calculated for each
change analyses, the site on North Key expanded
transect by subtracting May values from November
at a rate of about 300 m2 yearÀ1, and the sites on
values (seasonal) or November values from sub-
Seahorse Key and Snake Key expanded at rates
sequent November values (annual). The five tran-
exceeding 840 m2 yearÀ1. Mangrove expansion
sects were used as replicates (for non-parametric
on North Key and Snake Key occurred along
analysis) to evaluate statistical significance of
439
the boundaries of persistent mangrove clumps season was defined as winters in which the mini-
mum temperature fell below À5 °C at several
(Figure 2). The mangrove expansion on Seahorse
Key also occurred along the boundaries of per- temperature stations within an area. During 1932–
sistent mangrove clumps, but much of the expan-
sion at this site was represented by many small
mangrove clumps distributed widely throughout
the marsh (Figure 2). Some mangrove loss was
evident at the North Key and Snake Key study
sites, but this loss was greatly outweighed by
mangrove expansion.
Ground level transects
Initial canopy height (used simply as an indicator of
where mangrove and saltmarsh boundaries occur)
and seedling recruitment for the five transects are
shown in Figure 3. Seedling recruitment along the
two Snake Key transects was highest around
mangrove clumps and declined with distance.
Mangrove seedling recruitment along Snake Key–
Transect 2 occurred up to 5 m from mangrove
clumps into saltmarsh areas. Seedling recruitment
along Snake Key–Transect 1 was as high as
76 seedlings mÀ2 yearÀ1 and seedlings recruited at
least 15 m into saltmarsh areas. Although seedling
recruitment on Seahorse Key occurred, it primarily
occurred within the mangrove clump (Transect 1),
and along the area fringing the lagoon (Transect 2).
On the North Key transect, seedling recruitment
did not occur. Seedling recruitment did not occur
along any transect in 1997.
Seasonal changes in saltmarsh stem density
along transects were not significant during 1995
(p = 0.35), but were significant during 1996
(p = 0.04). Annual changes in saltmarsh stem
density along transects were higher in 1996 than in
1995 (p = 0.04). These increases occurred along
transect locations that were farthest from man-
grove clumps.
Discussion
Frequency of freezes in Northern Florida
Figure 2. GIS change analysis between 1995 and 1999 on Snake
Freeze cycles in Florida have been explored by Key (a), North Key (b), and Seahorse Key (c). White refers to
upland or beach, blue refers to water, pink refers to persistent
scientists interested in the risks associated with
mangrove, red refers to mangrove expansion, opaque refers to
citrus production. In one such study, Miller and
mangrove loss, grey refers to remaining intertidal zone, dark
Downton (1992) defined a local killing freeze as grey refers to change in intertidal zone (low tide in 1995 aerial
winters in which the minimum temperature fell photography vs. high tide in 1999 aerial photography). Loca-
below À6.7 °C, and a regionwide severe freeze tions of 30-m transects are also shown.
440
Table 1. Total intertidal area, mangrove area, and percent mangrove cover at each study site for 1995 and 1999 determined from aerial
photography. Also shown are mangrove expansion rates at each study site determined from GIS change analysis.
Aerial photography
Location 1995 1999 Change analysis
mangrove
Intertidal Mangrove Mangrove Mangrove Mangrove expansion
area (m2) area (m2) area (m2)
cover (%) cover (%) rate (m2 yearÀ1)
Snake Key 15,333 5114 33 9044 59 862
North Key 6948 1356 20 2565 37 300
Seahorse 21,186 609 3 4822 23 842
Key
1980, a regionwide severe freeze season occurred catastrophic freeze in the 19th century, for
approximately every 8 years in Florida. The freeze example, involved a series of severe winters dur-
frequency changed substantially during the 1980s ing 1895–1905, and the series of severe winters
as tree killing freezes occurred in 4 out of 5 years during 1977–1989 comprised the catastrophic
between 1981 and 1985 (Miller and Downtown freeze of the 20th century.
1992).
The freezes that affect the citrus industry appear
to coincide with freezes that affect mangroves and Impact of freezes on mangroves
probably subtropical species in general. Major
impacts to mangroves resulting from freezes According to Lugo and Patterson-Zucca (1977),
occurred in 1962 (Dr Frank Maturo, University of freeze damage to mangroves is greatest in areas
Florida, personal communication), 1977 (Lugo more exposed to cold wind. The Seahorse Key
and Patterson-Zucca 1977), 1981, 1983, 1985, 1989 marsh is exposed to cold north winds that follow
(personal observation) and 1996 (this study). cold fronts in this region. This may explain the
These accounts are consistent with an annual high mangrove mortality (12%) at Seahorse Key
minimum temperate of less than À6.7 °C (defini- following the January 1996 freeze (compared to
tion of a local killing freeze for citrus; Miller and 5% on North Key and 4% on Snake Key). The
Downtown 1992) recorded at the Cedar Key marshes on Snake Key and North Key, where
temperature station (NOAA 1915–1997). How- mangroves are more extensive, are isolated from
ever, temperatures on the outermost islands north wind by natural barriers such as berms and
(Snake Key, North Key, and Seahorse Key) are upland forests. Even on these islands, however,
slightly warmer than Cedar Key (Laessle and some mortality was evident from direct observa-
Wharton 1959). Therefore, the lowest temperature tions along transects and from results of the
change analyses that show some small net losses
that can be tolerated by mangroves at this latitude
may be close to Davis’ (1940) original suggestion of mangrove, which likely occurred from areas
of À4 °C. affected by freeze that had not yet fully recovered.
A single severe winter, however, may only set Although Seahorse Key appears to be the
back mangrove development without killing a most vulnerable to freezes, the impacts of the
substantial number of trees. At the Cedar Keys, January 1996 freeze reset mangrove development
mangroves have recovered within one year fol- on all three islands. The failure of mangroves to
lowing a severe freeze (Lugo and Patterson-Zucca flower and produce propagules during the 1996
growing season presumably resulted from deple-
1977; personal observation 1981). However, a
series of severe freezes is catastrophic, as evi- tion of the plants’ resources during recovery from
denced by the mangrove kill in the 1980s. Such a the previous winter. The lack of seedling recruit-
catastrophic series of severe winters may recur ment in 1997 suggests that mangrove expansion at
only once every century in Florida (Winsberg the Cedar Keys depends largely on a local source
1990; Miller and Downton 1992). The of propagules rather than imports from more
441
Snake Key - Transect 1
2.5 100
(Seedlings m-2 yr-1)
Canopy Height (m) 80
2
Recruitment
60
1.5 canopy height
40
20 recruitment 96
1
0 recruitment 97
0.5
-20
0 -40
0m 5m 10m 15m 20m 25m 30m
2.5 80
Snake Key - Transect 2
Canopy Height (m)
(Seedlings m-2 yr-1)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
2.5 80
North Key
(Seedlings m-2 yr-1)
Canopy Height (m)
2 60
Recruitment canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
Seahorse Key - Transect 1
2.5 80
(Seedlings m-2 yr-1)
Canopy Height (m)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
2.5 80
Seahorse Key - Transect 2
Canopy Height (m)
(Seedlings m-2 yr-1)
2 60
Recruitment
canopy height
1.5 40
recruitment 96
1 20
recruitment 97
0.5 0
0 -20
0m 5m 10m 15m 20m 25m 30m
Distance Along Transect
Figure 3. Canopy height and seedling recruitment along each transect.
442
southerly locations, such as Tampa Bay (a net scattered throughout the marsh that grew to be-
northward transport of Gulf of Mexico waters come trees during the study period. Some of the
occurs in this region; Clarke 1997). Hence, freezes smallest ‘clumps’ apparent in the digital ortho-
that impact mangrove reproduction at the Cedar photos may result from one or two trees that
Keys cause a major setback to mangrove succession. developed a broad canopy (possibly an advanta-
Freezes that impact mangroves, however, may geous strategy when competing alone with salt-
indirectly affect saltmarshes. Saltmarsh stem densi- marsh). Some of these small clumps among
ties at the Cedar Keys were higher in 1996 than in saltmarsh vegetation were also evident on Snake
1995, even though mangrove expansion was Key.
expected to result in lower saltmarsh stem densities.
Although the increases followed a freeze that defo-
Time-frame for mangrove succession
liated mangroves, increased light availability cannot
explain the changes in saltmarsh stem density
The freezes in 1983 and 1985 killed 98% of man-
because they occurred primarily in areas away from
groves at the Cedar Keys (Montague and Odum
mangrove clumps. However, recycled nutrients
1997), which may have eliminated the local man-
from fallen mangrove leaves might explain the
grove propagule source. It is not known whether
increase. Freezes may have resulted in greater
local seedlings survived, which would have pro-
amounts of detritus as mangrove biomass was
vided a seed bank for subsequent mangrove suc-
defoliated or killed, and subsequent detrital break-
cession, or if new recruitment was initially
down increased nutrients available for plant growth.
dependent on propagules from areas further south.
Regardless, as more mangroves continue to reach
reproductive maturity at the Cedar Keys, the rate
Mangrove expansion
of mangrove colonization should increase. In the
absence of major freezes, the time it will take for
The mechanism of mangrove seedling recruitment
mangroves to completely cover the marsh on Snake
and expansion appeared to differ among the study
Key is another 8 years (15,333 m2 total area –
sites. On Snake Key, trapping of mangrove propa-
9044 m2 1999 mangrove area/mangrove expansion
gules by saltmarsh (Lewis and Dunston 1975) could
rate of 862 m2 yearÀ1). Thus, the time frame for
account for the high seedling recruitment near
complete mangrove recolonization on Snake Key
mangrove clumps. Propagules produced by the
since the freeze kill during the 1980s is nearly
mangrove clumps on Snake Key were immediately
20 years. Using the same calculation for North
trapped in saltmarsh within 5–15 m of their source.
Key and Seahorse Key, these study sites will take
Without the adjacent saltmarsh, more mangrove
25–30 years for mangroves to again reclaim the
propagules may have been exported rather than
marsh. Periodic freezes, however, ensure that some
being retained in the vicinity of the parent trees.
saltmarsh cover will persist. This mixture of man-
High rates of seedling recruitment along the edges
grove forests and saltmarshes may continue for
of the mangrove clumps and subsequent growth of
several decades until another catastrophic series of
these mangroves into trees probably explains why
freezes kill mangrove forests, and allow saltmarshes
Snake Key has the greatest mangrove cover and
to once again occupy the intertidal zone.
highest expansion rates among the study sites.
On Seahorse Key, mangrove seedling recruit-
ment occurred within the existing mangrove clump Implications for coastal wetland communities
and along the transect adjacent to the lagoon also
As the intertidal zone changes between mangrove
suggesting that mangrove seedlings may remain
forests and saltmarshes, faunal changes should
near parent trees (Blanchard and Prado 1995;
occur. Both mangroves and saltmarshes provide
McKee 1995). However, numerous mangrove
food and refuge for fiddler crabs, periwinkles,
clumps on Seahorse Key appeared to ‘sprinkle in’
clams, and oysters (Odum et al. 1982; Stout 1984).
throughout the marsh (apparent from change
However, the greater volume and structural com-
analysis), particularly at the southern end of the
plexity of mangroves supports additional fauna
study site adjacent to the lagoon. This mechanism
such as the mangrove tree crab, Aratus pisonii
of expansion may result from single seedlings
443
(Milne Edwards), the coffee-bean snail, Melampus of severe winters, however, the combined energy
coffeus (Linnaeus), the mangrove crab, Goniopsis drains of salinity and frost limit the amount of
cruentata (Latreille), the ladder-horn snail, Ceri- structure that can be maintained. As a result,
thidea scalariformis (Say), the mangrove rivulus mangroves are killed and the coastal wetlands
fish, Rivulus marmoratus Poey, and the mangrove rapidly retrogress to saltmarshes. Following a
cuckoo, Coccyzus minor (Gmelin) (Odum et al. mangrove kill, saltmarshes occupy the intertidal
1982; Davis et al. 1995; Taylor et al. 1995). In zone within a period of only 5–10 years. The
mangroves, the periwinkle Littoraria angulifera saltmarsh vegetation, however, may stabilize
(Lamarck) replaces the Littoraria irrorata (Say) of marsh soils and maintain certain wetland func-
saltmarshes (Odum et al. 1982). Also, mangrove tions (e.g., juvenile fish habitat, detritus produc-
forests may attract nesting colonial birds (Watson tion). Whether the overall effect of saltmarshes
1986). Conversely, saltmarsh-dependent bird spe- facilitates mangrove colonization by trapping
cies such as seaside sparrows, Ammospiza maritima mangrove propagules or retards mangrove rees-
(Wilson), and long-billed marsh wrens, Cistotho- tablishment by outcompeting mangrove seed-
rus palustris (Wilson), may leave the area when lings remains equivocal. Nevertheless, saltmarshes
saltmarsh is overtaken by mangroves (Post and maintain emergent habitat in the intertidal zone in
Greenlaw 1994). the interim between occupations by mangroves.
During the present study, saltmarsh-dependent Although coastal wetlands alternate between
A. maritima and C. palustris were only encoun- mangroves and saltmarshes at the temperate to
tered at the Seahorse Key study site where salt- subtropical transition in Florida, overall produc-
marsh cover remains extensive (personal tion and survival of wetland-dependent biota are
observation). Although pelicans were observed maintained despite periodic disturbances.
nesting in mangroves on Seahorse Key prior to the
catastrophic freezes in the 1980s (Watson 1986), Acknowledgements
colonial birds were not observed in the developing
mangroves of the present study. The only fauna We are grateful to F. Maturo and the University
unique to mangroves on the islands was A. pisonii. of Florida Marine Lab at Seahorse Key for pro-
Aratus pisonii reached densities of 1 crab mÀ3 viding logistical support during this study. We also
prior to the 1996 freeze (unpublished data), which thank K. Litzenberger, refuge manager of the
are within the range reported for mature mangrove Lower Suwannee National Wildlife Refuge, for
forests of South Florida (1–4 crabs mÀ3; Beever permitting us to conduct research within the Cedar
et al. 1979). Mangrove fauna may not only be Keys National Wildlife Refuge. Florida Sea Grant
dependent on the habitat present, but also on and the Aylesworth Foundation provided financial
other environmental factors, especially tempera- support. T. Crisman and S. Vince gave thoughtful
ture (e.g., R. marmoratus, Taylor 1993). A long advice and guidance throughout this study, and
period of time without freezes may allow a richer commented on early drafts. Finally, we acknowl-
mangrove fauna to develop. Alternatively, the lag edge all of the volunteers who assisted with
in development of mangrove fauna may exceed the the fieldwork especially G. Stevens, J. Stevens,
frequency of freezes at this latitude. B. Clarke, and A. Wilson.
References
Implications for coastal wetland ecosystems
Beever J.W., Simberloff D. and King L.L. 1979. Herbivory and
Odum (1983) describes succession and retrogres- predation by the mangrove tree crab Aratus pisonii. Oecolo-
sion as ‘the self-organizational process by which gia 43: 317–328.
ecosystems develop structure and processes from Blanchard J. and Prado G. 1995. Natural regeneration of
Rhizophora mangle in strip clearcuts in Northwest Ecuador.
available energies.’ In the absence of freezes,
Biotropica 27: 160–167.
mangroves succeed saltmarshes perhaps because
Clarke P.J. and Allaway W.G. 1993. The regeneration niche of
they are able to channel more energy into greater the grey mangrove (Avicennia marina): effects of salinity,
stature and biomass, thereby shading out salt- light, and sediment factors on establishment, growth, and
marshes (Kangas and Lugo 1990). During periods survival in the field. Oecologia 93: 548–556.
444
Clarke P.J. and Myerscough P.J. 1993. The intertidal distribu- from the Gulf of Mexico. St. Lucie Press, Delray Beach, FL,
tion of the grey mangrove (Avicennia marina) in southeastern pp. 1–33.
Australia: the effects of physical conditions, interspecific National Oceanic and Atmospheric Administration (NOAA)
competition, and predation on propagule establishment and 1915–1997. Climatological Data. Florida Section, National
survival. Aust. J. Ecol. 18: 307–315. Climatic Center, Asheville, North Carolina.
Clarke J. 1997. Atlantic Pilot Atlas. International Marine, National Oceanic and Atmospheric Administration (NOAA)
Camden, ME. 1985. Gulf of Mexico: Coast and Ocean Zones: Strategic
Davis J.H. 1940. The ecology and geologic role of mangroves in Assessment: Data Atlas.
Florida. Papers from Tortugas Lab, 32. National Oceanic and Atmospheric Administration (NOAA).
Davis W.P., Taylor D.S. and Turner B.J. 1995. Does the aut- 1988. Sea-level variations for the U.S. 1855–1986.
ecology of the mangrove rivulus fish (Rivulus marmoratus) Odum W.E. 1983. Systems Ecology. John Wiley and Sons, New
reflect a paradigm for mangrove ecosystem sensitivity? Bull. York, NY, 644 pp.
Mar. Sci. 57: 208–214. Odum W.E., McIvor C.C. and Smith T.J. III 1982. The ecology
Higinbotham C.B., Alber M. and Chalmers A.G. 2004. Analysis of the mangroves of south Florida: a community profile. U.S.
of tidal marsh vegetation patterns in two Georgia estuaries Fish and Wildlife Service, Office of Biological Services,
using aerial photography and GIS. Estuaries 27: 670–683. Washington, DC, FWS/OBS-81/24, 144 pp.
Kangas P.C. and Lugo A.E. 1990. The distribution of man- Patterson C.S. and Mendelssohn I.A. 1991. A comparison of
groves and saltmarsh in Florida. J. Trop. Ecol. 31: 32–39. physiochemical variables across plant zones in a mangal/salt
Laessle A.M. and Wharton C.H. 1959. Northern extensions in marsh community in Louisiana. Wetlands 11: 139–161.
the recorded ranges of plants on Sea Horse Key and associ- Patterson C.S., Mendelssohn I.A. and Swenson E.M. 1993.
ated Key, Levy County, Florida. Q. J. Fla. Acad. Sci. 22: Growth and survival of Avicennia germinans seedlings in a
105–113. mangal/salt marsh community in Louisiana, USA. J. Coast.
Lewis R.R. and Dunston F.M. 1975. The possible role of Res. 9: 801–810.
Spartina alterniflora Loisel in establishment of mangroves in Post W. and Greenlaw J.S. 1994. Seaside sparrow, No. 127. In:
Florida. In: Lewis R.R. (ed.), Proceedings 2nd Annual Poole A. and Gill F. (eds), The Birds of North America. The
Conference on Restoration of Coastal Vegetation in Florida. American Ornithologists Union and the Academy of Natural
Hillsborough Community College, Tampa, FL, pp. 82–100. Science of Philadelphia, Philadelphia, PA.
Lugo A.E. and Patterson-Zucca C.P. 1977. The impact of low Saintilan N. and Williams R.J. 1999. Mangrove transgression
temperature stress on mangrove structure and growth. Trop. into saltmarsh environments in south-east Australia. Glob.
Ecol. 18: 149–160. Ecol. Biogeogr. 8: 117–124.
Markley J.L., McMillian C. and Thompson G.A. Jr. 1982. Stout J.P. 1984. The ecology of irregularly flooded saltmarshes
Latitudinal differentation in response to chilling temperatures of the northeastern Gulf of Mexico: a community profile.
among populations of three mangroves, Avicennia germinans, U.S. Fish and Wildlife Service, Office of Biological Services,
Laguncularia racemosa, and Rhizophora mangle, from the Washington DC, 85 (7.1), 98 pp.
western tropical Atlantic and Pacific Panama. Can. J. Bot. Taylor D.S. 1993. Notes on the impact of the December 1989
60: 2704–2715. freeze on local populations of Rivulus marmoratus in Florida,
McKee K.L. 1995. Seedling recruitment patterns in a Belizean with additional distribution records in the state. Fla. Sci. 56:
mangrove forest: effects of establishment ability and physico- 129–134.
chemical factors. Oecologia 101: 448–460. Taylor D.S., Davis W.P. and Turner B.J. 1995. Rivulus mar-
McMillan C. and Sherrod C.L. 1986. The chilling tolerance of moratus: ecology of distributional patterns in Florida and the
black mangrove, Avicennia germinans, from the Gulf of central Indian River Lagoon. Bull. Mar. Sci. 57: 202–207.
Mexico coast of Texas, Louisiana, and Florida. Contrib. Wanless H.R., Parkinson R.W. and Tedesco L.P. 1994. Sea level
Mar. Sci. 29: 9–16. control on stability of Everglades wetlands. In: Davis S.M.
Miller K.A. and Downton M.W. 1992. The freeze risk to Florida and Ogden J.C. (eds), Everglades: The Ecosystem and its
citrus. Part I: investment decisions. J. Clim. 6: 354–363. Restoration. St. Lucie Press, Delray Beach, FL, pp. 199–223.
Montague C.L. and Weigert R.G. 1990. Saltmarshes. In: Myers R. Watson A.M. 1986. Nutrient-production relations at Seahorse
and Ewel J. (eds), Ecosystems of Florida. University of Key Lagoon, Florida: some consequences of shorebird
Central Florida Press, Orlando, FL, pp. 481–516. accumulations. M.S. Thesis, University of Florida, Gaines-
Montague C.L. and Odum H.T. 1997. The intertidal marshes ville, FL, 87 pp.
of Florida’s Gulf Coast. In: Coultas C.L. and Hsieh Y.P. Winsberg M.D. 1990. Florida Weather. University of Central
(eds), Ecology and Management of Tidal Marshes: A Model Florida Press, Orlando, FL.