| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Biology,University of Miami, Coral Gables, Florida 33124–0421,2
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Ardisiaescallonioides • hurricanes • life-historystrategy • Myrsinaceae • pioneer • shade-tolerant • regrowth • resiliency
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Within areas that experience frequent hurricane disturbance, dominantforest tree species often exhibit high survival, regrowth, andregeneration following hurricane disturbance (Zimmerman et al., 1994; Slater et al., 1995). Understory species maydiffer from canopy tree species in the direct effects of hurricanes onmortality, damage, and defoliation. Because of their small size andprotection from wind, understory species may suffer less directdefoliation and damage from wind than canopy tree species. Understoryspecies may suffer high damage levels, particularly in severehurricanes, due to debris rain from the canopy. Because many understoryspecies have multiple stems, however, they may be more resistant thantree species to mortality as a result of stem damage.
One of the most obvious effects of hurricanes is to increaseunderstory light levels. Understory shrubs are often consideredstrongly shade tolerant: able to germinate, grow, and reproduce underclosed-canopy conditions of usually less than 2% of full sunlight(Fetcher, Oberbauer, and Chazdon,1994). Some studies have shown, however, that low understorylight levels limit growth (Denslow et al.,1990). Thus, growth of understory species may respondstrongly to the increased understory light availability followinghurricane disturbance (Fernandez and Fetcher,1991). The ability to rapidly respond to increased lightfollowing hurricane disturbance may characterize dominant understoryshrubs in hurricane-prone forests.
Hurricane disturbance may also affect seed and seedling dynamics ofunderstory species. The quantity of leaf litter that is deposited on theforest floor limits immediate posthurricane germination and seedlingsurvival of pioneer species (Guzman-Grajales and Walker, 1991;Yih et al., 1991). As disturbanceintensity increases, tip-ups provide exposed bare soil, and germinationof pioneer species increases in response to high posthurricane heat andlight levels (Vazques-Yanes and Orozco-Segovia, 1994; Horvitz, McMann, and Freedman, 1995). Incontrast to pioneer species that have long-lived, dormant seeds,shade-tolerant understory shrubs with short-lived seeds must survivehurricane disturbance, recover from damage, and resume reproductionbefore postdisturbance seedlings can be produced. For understoryspecies that are limited in reproductive output by low understory lightlevels (Levey, 1988), these species mayproduce seeds and seedlings only after disturbance (see Pascarella, in press). Thus, hurricanedisturbance may not directly affect these life-history stages due to thetime delay between the disturbance and the production of seeds andseedlings.
On 24 August 1992, Hurricane Andrew passed over the southern Floridamainland (Mayfield, Avila, and Rappaport,1994). In south Florida subtropical forests, the shrubArdisia escallonioides Schlechtendahl and Chamisso(Myrsinaceae) is an abundant understory species. To determine theresiliency to hurricane disturbance and the response of differentlife-history stages to posthurricane environmental conditions, Istudied populations of A. escallonioides at four sites impactedto various degrees by Hurricane Andrew. I asked the followingquestions: (1) How much light reached the understory following thehurricane in the different populations? (2) Did A.escallonioides suffer less mortality, damage, and defoliation thancanopy species in these same forests? (3) Did plants in forestswith extensive canopy removal show increased growth, when compared toplants in forests with minimal canopy removal? and (4) Was germination,seedling survival, and seedling growth positively correlated withunderstory light availability?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The principal habitat of A. escallonioides in south Floridais subtropical hardwood forest on the Miami rock ridge and the FloridaKeys (Little, 1978). Stands ofsubtropical hardwood forests in south Florida occupy small (0.1–40ha) areas of moist to dry outcrops of limestone in a landscape of pinesavanna or wet prairie. They are similar floristically and structurallyto forests in the West Indies and the Yucatan Peninsula (Smith and Vankat, 1992). Canopy height rangesfrom 15 to 25 m with dominant canopy and subcanopy tree speciesincluding Coccoloba diversifolia, Sideroxylon(Mastichodendron) foetidissimum,Sideroxylon (Bumelia) salicifolium,Lysiloma latisiliquum, Quercus virginiana,Nectandra coriacea, and Bursera simaruba. Commonunderstory species include A. escallonioides,Psychotria sp., Eugenia sp., andCalyptranthes sp. (Snyder, Herndon, andRobertson, 1990).
The study region is in the subtropical moist forest life-zone(Dohrenwend and Harris, 1975), whichoccurs at the southern tip of Florida. Mean annual temperature in Miamiis 23°C, and the monthly mean temperature ranges from 20°C inDecember to 28°C in August. Mean annual rainfall is 1500 mm/yr,but is strongly seasonal, with 78% total rainfall occurringbetween May and October. The dry season is marked by rainless periodsof 15–25 d (Chen and Gerber, 1990). Soils in subtropical hardwood forests in the Miami rock ridge aretypically thin and sandy with extensive areas of surface limestone rockoutcroppings and sinkholes (Brown, Stone, andCarlisle, 1990).
All four study sites were located between 80° and 81° W and25° and 26° N. Two populations (Castellow Hammock[CAS] and Deering Estate [DEE]) were initiallycensused in June 1992 prior to Hurricane Andrew, one immediately afterthe hurricane in September 1992 (Matheson Hammock [MAT]) andone in December 1993 (John Pennekamp State Park [JPS]). Threepopulations (MAT, DEE, and CAS) were in subtropical forests managed asnatural areas by Metro-Dade County Parks, and one population (JPS) wasin south Key Largo in a forest managed by the Florida State Parks.
Three of the four study sites that were impacted by the hurricane(MAT, DEE, and CAS) were within a 24-km radius of each other, butdiffered in their location relative to the path of the hurricane. DEEwas located in the northern eyewall, CAS was south of the northerneyewall but in the eye center, MAT was north of the northern eyewall,and JPS was located south of the hurricane path. Location relative tothe northern and southern eyewalls is important because sustained andgust wind speeds are at a maximum in the northern eyewall, less in thesouthern eyewall, and drop off rapidly outside the eyewalls (Pielke, 1990). Damage to forests is stronglyassociated with both maximum sustained winds and maximum gusts (Boose, Foster, and Fluet, 1994). Sustained windspeeds in the northern eyewall exceeded 241 km/h with gusts up to285 km/h, making Andrew a category 4 hurricane on the Simpson-Saffirscale (Mayfield, Avila, and Rappaport,1994). Because maximum sustained winds occur in thenortheastern quadrat of west-moving hurricanes and wind speeds decreasesignificantly outside of the zone of circulation, it is believed thatDEE and CAS experienced the highest sustained winds and gusts, with MATexperiencing intermediate sustained winds and gusts (Mayfield, Avila, and Rappaport, 1994; Wakimoto and Black, 1994). Due to its locationsouth of the hurricane, JPS was minimally affected by the hurricane;wind speeds were estimated to be <100 km/h (Ross et al., 1995; R. Skinner, Florida State Parks,personal communication).
Generalplot and census
methods
In each population, five 5 x 10 m plots (total0.025
ha) were established in study areas that spanned the naturalvariation in
density of A. escallonioides. Adults wereseparated from
juveniles at the arbitrary height of 1 m based on initialobservation of
flowering adults in 1992. Within each plot, all adultswere mapped and
marked with uniquely numbered plastic bird bands. Threeindirect
measures of growth (plant height, summed diameter at 1.3 m inheight, and
number of stems) were used in this study. Before thehurricane, in June
1992, I measured the number of stems at 1 m heightarising from a common
base, height of the tallest stem, and diameter at1.3 m in height of all
stems 
1 m tall for DEE and CAS populations. Height of the tallest
stem was measured using a graduated tree-measuringpole. Adults were
divided into four size classes: 4.5 cm diameter,3 and <4.5 cm
diameter, 1.5 and <3 cm diameter, and<1.5 cm diameter. Most
stems marked before the hurricane retainedtheir coded tags during
Hurricane Andrew; others were relocated usingprehurricane plot
maps.
Within each plot, ten 1 x 1 m subplots (50 subplots perpopulation) were randomly located to sample juveniles (<1 m tall) andseedlings (identified using unique morphological characters). Heightand number of stems were measured during 1993 while height, diameter atthe ground, and number of stems were measured in 1994. All adults andjuveniles at MAT, DEE, and CAS were recensused in December 1992, 1993,and 1994; adults and juveniles at JPS were recensused in December 1994.
Estimating understory
lightlevels
Light availability in the understory was estimated
usinghemispherical canopy photographs (Rich,1990; Weiss et al., 1991).These photographs were taken
at the dry/rainy season transition periodin May 1993 (9 mo
posthurricane) for MAT, DEE, and CAS and for allpopulations in May 1994
(21 mo posthurricane). Photographs were takenusing a Nikon FM2 camera
with a Nikkor Hemispherical Lens (8 mm). Thecamera was mounted on
self-leveling gimbals so that a circular imageextending from horizon to
horizon was taken with the camera pointingstraight up at the canopy.
The lens was focused on infinity. Thegimbals were attached to a
monopod, and all photos were taken at theheight of 130 cm.
For each site, 40 photos were taken. Eight photo points were used ineach of the five plots. Within each plot, six photos (all photos >2m apart) were taken above a seed germination experiment and theremaining two photos were taken at the SW and NW corners of the plots. The program CANOPY (Rich, 1989) was usedto analyze the light availability and calculate the direct site factor(DSF) and indirect skylight factor (ISF) for each photo point. CANOPYdigitized the photographic image, corrected for magnetic declination andlens distortion, and calculated the effects of the sun tracks over thecanopy openings. DSF and ISF are unitless measures that calculate theamount of light reaching the forest understory relative to how muchlight would reach a point that was totally in the open (range 0–1with 0 completely shaded and 1 completely open). DSF and ISF valueshave been shown to be strongly correlated with more precise sensormeasurements of photosynthetically active radiation (Weiss et al., 1991; Clark,Clark, and Rich, 1993; Rich et al.,1993). ISF and DSF values were highly correlated in my studysites (Horvitz, McMann, and Freedman,1995). In this paper, ISF values were used to representcanopy openness of the different sites affected by Hurricane Andrew. Because of unequal variances among sites, nonparametric Kruskal-WallisANOVA was used to determine whether significant differences existedamong the sites. Photos were not taken for JPS in 1993.
Damage, defoliation, mortality,and
regrowth
To determine the direct effects of Hurricane Andrew on
mortality,damage, and defoliation, poststorm measurements were made
during theperiod from September to December 1992, 1–4 mo after
thehurricane, for adults in the three damaged populations (MAT, DEE,
andCAS). All prehurricane stems that had not reflushed by December
1992were followed through the following dry season until May 1993, when
theywere assigned a dead or resprouting status. Cause of mortality
wasrecorded as tip-up, snapped, or unknown. Structural damage
wasclassified as follows: (1) stems upright with no apparent
structuraldamage apart from the loss of side branches (normal
undamagedcondition); (2) stems bent over at an angle between 90°
and 15°from the horizontal; (3) stems knocked flat to the ground at
an angle<15° horizontal; (4) stems either upright or at an
angle>15° horizontal, but snapped. Defoliation was estimated for
alladult stems, regardless of damage class. Defoliation of stems
wasvisually divided into two categories: (1) 50% and
(2)<50% defoliation. Chi-square analysis was used to
examinewhether variation in mortality, damage, and defoliation was
independentof sites and to examine if these variables were associated
with sizeclass and prehurricane height (<2, 2–3, 3–4, and
>4m) at each site. A two-sample t test was used to test
forsignificant differences between DEE and CAS in height lost as a
resultof the hurricane.
Because light levels and changes in plant size did not varysignificantly among the five plots within sites (ANOVA), individualplants were used as the unit of analysis within sites (i.e.,experimental units). Plant size variables were based on a relativemeasure of growth to standardize potential differences in preexistingpopulation size-class structure among sites. Relative changes in plantsize variables were calculated as
Values of
relative change in plant size can range from -1 to
;
negative values indicate reduction in size, 0 means no change,and
positive values mean an increase in size. Because of unequal
samplesizes in the field and variance among sites, all relative changes
inplant size variables were analyzed using the
nonparametricKruskal-Wallis ANOVA followed by among-site comparisons at
the 0.05confidence level overall.
Seedfates and
seedling dynamics
To determine germination phenology, percentage
germination, andseedling survival under natural conditions, seeds were
planted in thefield. Within the study plots, seeds for the field
germinationexperiments were collected from the ripe fruits of 13 plants
from MATand 27 plants from DEE in February 1993 and of 20 plants each at
MAT,DEE, and CAS in February 1994. Due to a lack of seeds at CAS
inFebruary 1993, a mixture of DEE and MAT seeds were used for CAS.
Pulpwas removed from the fruit. Seeds from different plants within the
samestudy site were mixed together to minimize individual plant
effectsbefore planting them into experimental gardens in the field. In
lateFebruary 1993, ten seeds were placed underneath wire mesh cages
(meshsize ~1 cm, cages 10 x 10 cm) that were hammeredinto the
soil, and a group of ten seeds was placed ~1 m from thecage and
marked with a wire flag (a total of 30 replicates of eachtreatment per
site, with each of the five plots having six replicates ofeach
treatment). A canopy photograph was taken in 1993 over each of thecaged
treatment seeds. In 1994, this experiment was repeated using onlythe
caged treatment and no canopy photographs were taken over the
seeddepots. Due to concerns about introducing seeds from other
populations,no seeds were planted at JPS.
To reduce the effect of variation in leaf litter on seed germinationand seedling establishment (Guzman-Grajales and Walker, 1991), allsites had the top 2 cm of leaf litter removed, and seeds were notcovered by additional litter. Posthurricane leaf litter falling on theseed cages was minimal during the experiment because extensive canopycover was not present at the three sites.
Germination was monitored monthly at each site until December 1993and every 2 mo in 1994. Upon germination, seedlings were marked withindividually numbered bird bands, and the number of leaves per seedlingwas recorded monthly. Seedling height was recorded at the final census(December 1995). Monthly precipitation data were provided by FairchildTropical Garden, Coral Gables, Florida, which is located next to MAT. Monthly precipitation is highly correlated between Miami InternationalAirport and Homestead climatic data centers, which encompass MAT, DEE,and CAS, suggesting that there were no climatic differences between thethree sites (National Oceanic andAtmospheric Administration, 1992–1994).
The influence of site and treatment on percentage seed germinationwas tested using a nested ANCOVA, with ISF as a covariate. In 1994, Iused a one-way ANOVA to investigate the effects of site on percentageseed germination. I used linear regression analysis to examine theinfluence of site, treatment, date of germination, and ISF on finalseedling height and leaf number, and I used logistic regression toexamine the influence of these variables on seedling survival. I alsoused linear regression analysis to examine the relationship betweenmonthly variation in seed germination and seedling survival and monthlyprecipitation totals. Subplots were monitored monthly for naturalseedling recruitment.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
=72.57, df = 6, P < 0.0001). CAS had
the highestpercentage of Ardisia escallonioides adults that
were bentover, and DEE had the highest percentage of plants knocked
flat. MAThad the highest percentage of adults remaining upright, and no
plantswere flattened at MAT. The percentage of plants that suffered
snappedtops was identical in all three populations. The type of damage
wassignificantly associated with size at CAS ( =29.38, df
= 9, P < 0.001) and DEE (= 25.55, df
= 9, P < 0.01), but not at MAT( = 7.51,
df = 4, P =0.11). At CAS and DEE, plants <2 m
in height suffered greater thanexpected snapped tops, while plants >4
m suffered greater thanexpected knockdowns. Average adult height lost
to Hurricane Andrew wassimilar at DEE (105 cm) and CAS (106 cm)
(t test =1.51, df = 153.8, P =
0.13). In these sites,there was a significant positive relation between
height lost during thehurricane and prehurricane height (Pearson
correlation, r= 0.72, N = 282,
P <0.0001).
|
= 33.51, df = 2, P < 0.0001).
Ardisiaescallonioides plants at MAT suffered the least
defoliation, CASwas intermediate, and DEE suffered the highest
defoliation levels. Defoliation was significantly associated with size
at MAT (= 21.43, df = 2, P < 0.0001), but
wasnot associated with size at CAS ( = 5.79, df= 3,
P = 0.12) or at DEE (= 4.95, df = 9,
P = 0.84). At MAT, plants>3 m suffered greater
defoliation, while plants <2 m retainedleaves more than expected by
chance. At both DEE and CAS, almost allseverely defoliated plants had
begun resprouting by 6–8 wk afterHurricane
Andrew.
Structural damage was not significantly related to
defoliation level( = 3.34, df = 2, P <
0.19)at MAT, but was significantly related to defoliation level at
DEE( = 26.10, df = 3, P <0.0001) and CAS
( = 21.45, df = 3,P < 0.0001). At these
sites, both knocked over and bentover plants had more individuals
retaining leaves than expected, whileupright and top-snapped individuals
suffered greater defoliation thanexpected.
Ardisia escallonioides began releafing from dormantepicormic buds axillary to old leaf scars at any point along the stemwithin 2 mo after the hurricane. Leaf morphology changed after thehurricane. In severely defoliated individuals at CAS and DEE, newleaves were typical sun leaves, differing from the prehurricane shadeleaves. Sprouting of new stems primarily occurred on plants that werebent over, flattened, or snapped, while few undamaged plants sproutednew stems. In A. escallonioides individuals that wereflattened, most of the old stems had broken off and decayed by December1994, 27 mo posthurricane.
Change in plant size
Change in adult plant
size, measured as height, diameter at 1.3 m inheight, and stem number,
varied significantly among sites in both 1993and 1994 (Table 3). In1993, plants at the two severely
damaged sites (CAS and DEE) had greaterincreases in these variables than
the lightly damaged (MAT) or undamagedsite (JPS). There was a
significantly greater increase in plant size atall sites in 1993 vs.
1994 (P < 0.05), except for height atMAT and diameter at
DEE. Following the hurricane, plants produced manynew stems, and total
shrub diameter increased as a result of both newstem production and
radial increases in older stems. At DEE and CAS,average stem height
rapidly returned to almost prehurricane levels byDecember 1994, only 27
mo after Hurricane Andrew.
|
|
|
After 14 mo, the average seedling had 6.2 ± 0.22 (±1SE) leaves and was 5.27 ± 0.12 (±1 SE) cm tall (N= 205 seedlings). Linear regression analysis found significanteffects of date of germination on height (N = 158,P < 0.05), with earlier germinating seedlings taller, andsignificant effects of site on number of leaves (N =159, P < 0.05), with MAT having more leaves per seedlingthan the other two sites. However, total variation explained was only13% for height and 4% for number of leaves. Lightavailability (ISF) was not significantly related to either seedlingheight or number of leaves.
There was no natural seedling recruitment in 1992 or 1993 at anysite. In 1994, seedling recruitment occurred at all threehurricane-damaged sites, but no seedling recruitment was observed at theundamaged site. There was no flowering of plants at the undamaged site. CAS had the greatest total seedling recruitment (366 seedlings),compared to only 41 at DEE and 42 at MAT. However, mean seedlingrecruitment per subplot (1 x 1 m) did not differ significantlyamong sites, as seedlings at CAS were spatially clumped(F2,147 = 2.33, P =0.10).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although mortality was low, populations of A. escallonioidessuffered relatively high levels of damage and defoliation, particularlyin the two sites closest to the northern eyewall. Horvitz, McMann, and Freedman (1995) found thatcanopy trees experienced greatest structural damage at CAS, intermediatedamage at DEE, and the least damage at MAT, a pattern similar to thatnoted for A. escallonioides. Hurricane-related mortality anddamage to A. escallonioides appeared to be primarily fromdebris rain from canopy trees, not to direct wind damage. This findingwas consistent with observations of other understory species at otherhammocks within south Florida (Ross et al.,1995; Slater et al., 1995). However, individuals at the forest edge at DEE were exposed and may havebeen damaged directly from the strong winds, accounting for the highlevel of defoliation in this population.
Damage and defoliation were related at the two severely damagedsites. This relationship is likely due to the timing of damage inrelation to the passage of the hurricane. Plants that remained uprightthroughout the hurricane experienced the highest probabilities of severedefoliation, while plants that were knocked down or bent over early inthe storm experienced lower wind speeds and retained leaves. Size hadsignificant effects on damage levels at the two severely damaged sites,with larger plants being knocked over and small plants experiencingsnapped tops from the intense debris rain from the canopy. Defoliationwas so strong at the two most severely damaged sites that size was notrelated to defoliation. Only at the lightly disturbed site did tallerplants experience significantly more defoliation. This is alsoconsistent with defoliation patterns found in the canopy layer at thesame sites by Horvitz, McMann, and Freedman(1995).
The ability to resprout following hurricane disturbance is animportant characteristic of tropical trees in hurricane-prone areas(Boucher et al., 1994; Zimmerman et al., 1994; Howard and Schokman, 1995; Slater et al., 1995). Ardisiaescallonioides rapidly resprouted, particularly on severely damagedindividuals. Observations of tropical trees following hurricanes havealso noted increased sprouting in the most severely damaged trees(Bellingham, Tanner, and Healey,1994). Other Ardisia species are also goodresprouters following hurricane disturbance, suggesting this may be awidespread characteristic of the genus. Following a typhoon in theBonin Islands of Japan, Ardisia sieboldii resproutedextensively from bent over stems. Compared to other species, A.sieboldii had a very high capacity for resprouting, which allowedit to dominate the subcanopy following severe hurricane disturbance(Shimizu, 1994). Unlike someunderstory shrubs in tropical forests that reproduce vegetativelythrough fragmentation (Gartner, 1989;Greig, 1993; Sagers, 1993), A. escallonioides was notobserved to resprout in the field from vegetative fragments.
Understory light levels
andposthurricane growth
One of the most direct effects of
Hurricane Andrew was a dramaticincrease in light availability in the
understory of closed-canopyforests (Loope et al., 1994). Lightlevels in the three
damaged forests (MAT, DEE, and CAS) increasedgreatly immediately
following the hurricane, as a result of extensivedefoliation and damage
to many canopy trees (Horvitz,
McMann, and Freedman, 1995; Pascarella,personal
observation). By 9 mo after the hurricane, posthurricane lightlevels
differed among sites, with the two sites closest to the northerneyewall
having the greatest light availability. Significantly lowerlevels of
mortality and less extensive damage to canopy trees (Horvitz, McMann, and Freedman,
1995) at MAT thanat DEE or CAS resulted in more canopy cover
at the former site. In allhurricane-disturbed sites, posthurricane
light levels were similar to ormuch higher than levels in large canopy
gaps in tropical rain forests(Dirzo
et al., 1992).
Between 9 and 21 mo posthurricane, resprouting of standing trees,seedling recruitment of fast-growing pioneer species, vine tangles, andheight growth of advance regeneration saplings quickly reducedunderstory light levels (Horvitz, McMann, andFreedman, 1995; Pascarella,1997b). However, light levels at all three hurricane-damagedstudy sites were still in the range of levels in medium to large gaps(Dirzo et al., 1992). Depending on therate of canopy recovery, it is likely that all three damaged populationswill experience several additional years of significantly greater lightlevels than will populations in undamaged forest.
In spite of the highest levels of damage and defoliation,posthurricane growth was greatest at the two sites (CAS and DEE) withthe highest light availability. This result is similar to Denslow et al. (1990), who found that relativegrowth rates of tropical understory shrubs were significantly greater ingap centers (highest light levels) than at gap edges or within closedcanopy forests (both with lower light levels). In their experimentalstudy, nutrient addition through fertilization had no effect on growthrate in any light environment.
Seed and seedling
dynamics
Lacking long-term dormancy, the dormancy pattern for
A.escallonioides can be classified as innate seasonal
dormancy(Angevine and Chabot,
1979; Garwood,
1989), with dormancy enforced duringthe dry season and the
beginning of the rainy season. Dormancy isadvantageous because seeds
are dispersed at the peak of the dry season(Pascarella, unpublished
data). Germination by October may be criticalfor seedlings of A.
escallonioides to establish sufficientroots to survive the first
dry season.
Life-history strategy
Throughout
its pan-Caribbean range, A. escallonioides occursin forests
that are subject to hurricane disturbance. SoutheasternFlorida
experiences a hurricane every 6–8 yr on average and asevere
hurricane every 14 yr (Simpson
andLawrence, 1971; Simpson and Riehl,1981). This is the highest
frequency of hurricane disturbancein the continental United States and
higher than many areas of theCaribbean. Under these frequencies,
hurricane disturbance may supersedelocal treefall gaps as the major
natural disturbance affecting foreststructure and dynamics (Tanner, Kapos, and Healey,1991;
Boose, Foster, and
Fluet,1994; Lugo,
1995). This highfrequency of disturbance is likely to
influence local speciescomposition in relation to a species'
ability to resist and respondto hurricane disturbance.
Ardisia escallonioides was resilient to severe hurricanedisturbance, with low direct mortality despite relatively high levels ofdefoliation and damage. Rapid, positive growth response to thedisturbance was found, particularly in forests with severe canopydamage. In contrast, adults and juveniles under closed-canopy forestshad minimal or negative growth. This suggests that hurricanedisturbance results in a release of suppressed adults. Further evidencefor suppression was observed in the undisturbed, closed-canopy site JPS. Most adults at this site had very few leaves; many stems were decayingand leaning on canopy tree stems for physical support (Pascarella,personal observation). Although low-light conditions limit growth,adults can persist under low-light conditions in mature forests, due tolow annual mortality and clonal growth through rhizomes. Rhizomes mayproduce clonal shoots in small light gaps caused by localized canopytree mortality (Pascarella, personal observation), allowing plants toseek out increased light levels at the forest floor.
The response of A. escallonioides to hurricane disturbanceat the seed and seedling stage is distinct from both pioneer andadvance-regeneration shade-tolerant species. Unlike native pioneerspecies that recruited seedlings from long-term dormant seed banks(Pascarella, 1997b), A.escallonioides lacked long-term seed banks. Unlike shade-tolerantunderstory shrubs such as Picramnia pentandra (Picramniaceae)and Eugenia axillaris (Myrtaceae), which had extensive seedlingbanks in mature closed-canopy forest prior to Hurricane Andrew (Pascarella, 1996, personal observation),A. escallonioides lacked advance-regeneration seedling banks. Instead, it reproduced after disturbance opened the canopy, andseedlings were recruited several years following the hurricane (Pascarella, in press). Direct effects ofhurricane disturbance in the form of an increase in understory lightlevels had no positive effects on seed germination and seedling growth.Because seeds are not usually present in the soil prior to a hurricaneand A. escallonioides lacks soil seed banks, there may belittle selection on these life-history traits. Although seedlingsurvival was positively associated with increased light levels, seedlinggrowth was not. Seedling dynamics may be determined by an interactionbetween light levels, seasonal water availability, and otherenvironmental variables not experimentally manipulated in thisstudy.
This study emphasizes how tropical understory shrub life-historystrategies may not easily be characterized as strictly pioneer orshade-tolerant species (Ellison et al.,1993). Understory shrubs experience a wide range of lightenvironments throughout their lives. In tropical forests prone torecurrent hurricane disturbance, strong plasticity in growth responsesto variable light environments and rapid resprouting ability may betypical traits of the dominant understory species. A demographic modelof the effects of hurricane disturbance on population dynamics of A.escallonioides (Pascarella and Horvitz,1998) suggests that hurricanes are important in maintaininglarge populations of this species in south Florida forests. For manyunderstory species in hurricane prone forests, posthurricaneenvironmental conditions may be critical periods for growth in betweenperiods of suppression from taller canopyspecies.
|
FOOTNOTES |
|---|
2 Current address: Department of Biology, Valdosta State
University, Valdosta, Georgia
31698.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND
METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Basnet, K., F. Scatena, G. E. Likens, and A.E. Lugo.1993.Ecological consequences of root graftingin Tabunuco (Dacryodes excelsa) trees in the LuquilloExperimental Forest, Puerto Rico. Biotropica25 :28–35.
Bellingham, P. J. , V. Kapos, N. Varty,J. R. Healey, E. V. J. Tanner, D. L. Kelly, J. W. Dalling, L. S. Burns,D. Lee, and G. Sidrak.1992.Hurricanes need not causehigh mortality: the effects of Hurricane Gilbert on forests in Jamaica.Journal of Tropical Ecology8 :217 –223.
———, E. V. J. Tanner,and J. R. Healey.1994.Sprouting of trees in Jamaicanmontane forest after a hurricane. Journal of Ecology82 :747–758.
Boose, E. R., D. R. Foster, and M.Fluet.1994.Hurricane impacts to tropical andtemperate landscapes. Ecological Monographs64 :369–400.[CrossRef][ISI]
Boucher, D. H., J. H. Vandermeer, M. A.Mallona, N. Zamora, and I. Perfecto.1994.Resistanceand resilience in a directly regenerating rain forest—Nicaraguantrees of the Vochysiaceae after Hurricane Joan. Forest Ecology andManagement 68:127 –136.
Brown, R. B., E. L. Stone, and V. W.Carlisle.1990.Soils. In R. L. Myers and J.J. Ewel [eds.], Ecosystems of Florida, 35–69.University of Central Florida Press, Orlando, FL.
Chen, E., and J. F.Gerber.1990.Climate. In R. L. Myers and J.J. Ewel [eds.], Ecosystems of Florida, 11–34.University of Central Florida Press, Orlando, FL.
Clark, D. B., D. A. Clark, and P. M.Rich.1993.Comparative analysis of microhabitatutilization by saplings of nine tree species in neotropical rain forest.Biotropica 25:397 –407.
Denslow, J. S., J. C. Schultz, P. M.Vitousek, and B. R. Strain.1990.Growth responses oftropical shrubs to treefall gap environments. Ecology71 :165–179.
Dirzo, R., C. C. Horvitz, H. Quevedo, and M.A. Lopez.1992.The effects of gap age and size on theunderstorey herb community of a tropical Mexican rain forest. Journal of Ecology80 :809 –822.
Dohrenwend, R. E., and L. D.Harris.1975.A climatic change impact analysis ofpeninsular Florida life zones. U.S. Department of Transportation,Washington, DC.[CrossRef][ISI]
Ellison, A. M., J. S. Denslow, B. A.Loiselle, and D. M. Brenés.1993.Seed andseedling ecology of Neotropical Melastomataceae. Ecology74 :1733–1749.
Elmqvist, T., W. E. Rainey, E. D. Pierson,and P. A. Cox.1994.Effects of tropical cyclones Ofaand Val on the structure of a Samoan lowland rain forest.Biotropica 26:384 –391.
Fernandez, D. S., and N.Fetcher.1991.Changes in light availability followingHurricane Hugo in a subtropical montane forest in Puerto Rico.Biotropica 23:393 –399.
Fetcher, N., S. F. Oberbauer, and R. L.Chazdon.1994.In L. A. McDade, K. S. Bawa, H.A. Hespenheide, and G. S. Hartshorn [eds.], La Selva: ecologyand natural history of a neotropical rain forest, 128–141.University of Chicago Press, Chicago, IL.[CrossRef][ISI]
Gartner, B.L.1989.Breakage and regrowth of Piperspecies in rain forest understory. Biotropica21 :303–307.
Garwood, N.C.1989.Tropical soil seed banks: a review.In M. A. Leck, V. T. Parker, and R. L. Simpson[eds.], Ecology of soil seed banks, 149–210. AcademicPress, San Diego, CA.[CrossRef][ISI]
Greig, N.1993.Regenerationmode in neotropical Piper: habitat and species comparisons.Ecology74 :2125 –2135.[CrossRef][ISI]
Guzman-Grajales, S. M., and L. R.Walker.1991.Differential seedling responses to litterafter Hurricane Hugo in the Luquillo Experimental Forest, Puerto Rico.Biotropica 23:407 –413.
Horvitz, C. C., S. McMann, and A.Freedman.1995.Exotics and hurricane damage in threehardwood hammocks in Dade County, Florida. Journal of CoastalResearch 21:145 –158.
Howard, R. A., and L.Schokman.1995.Recovery responses of tropical treesafter Hurricane Andrew. Harvard Papers in Botany6 :37–74.
Levey, D. J.1988.Tropicalwet forest treefall gaps and distributions of understory birds andplants. Ecology 69:1076 –1089.
Little, E. L.,Jr.1978.Atlas of United States trees, vol. 5,Florida. U.S. Forest Service Miscellaneous Publication Number 1361. U.S.Government Printing Office, Washington, DC.
Loope, L., M. Duever, A. Herndon, J. Snyder,and D. Jansen.1994.Hurricane impact on uplands andfreshwater swamp forest. BioScience44 :238–246.
Lugo, A.E.1995.Reconstructing hurricane passages overforests—a tool for understanding multiple-scale responses todisturbance. Trends in Ecology and Evolution10 :98–99.
Lundell, C.L.1966.Myrsinaceae. In P. C. Standley and L.O. Williams [eds.], Flora of Guatemala, Fieldiana: Botany 24,Nos. 1 and 2, 137–199. Chicago Natural History Museum, Chicago,IL.
Mayfield, M., L. Avila, and N.Rappaport.1994.Atlantic Hurricane Season of 1992.Preliminary report (updated 2 March 1993) Hurricane Andrew 16–28August 1992. Monthly Weather Review122 :517–538.
National Oceanic and AtmosphericAdministration.1992–1994.Local climatologicaldata. Annual summary with comparative data. Miami, Florida. NationalClimatic Data Center, Asheville, NC.
Pascarella, J.B.1995.The effects of Hurricane Andrew on thepopulation dynamics and mating system of the tropical understory shrubArdisia escallonioides (Myrsinaceae). Ph.D. dissertation,University of Miami, Coral Gables, FL.
———.1996.Reproductiveecology of Picramnia pentandra (Picramniaceae) in SouthFlorida. Caribbean Journal of Science32 :99–104.
———.1997a.Pollinationecology of Ardisia escallonioides (Myrsinaceae).Castanea62 :1 –7.
———.1997b.Hurricanedisturbance and the regeneration of Lysiloma latisiliquum(Fabaceae): a tropical tree in south Florida. Forest Ecology andManagement 92:97 –106.
———.Inpress.Hurricane disturbance, plant-animal interactions, and thereproductive success of a tropical shrub.Biotropica.
———, and C.C.Horvitz.1998.Hurricane disturbance and the populationdynamics of a tropical understory shrub: megamatrix elasticity analysis.Ecology 79:547 –563.
Pielke, R. A.1990.Thehurricane. Routledge, New York, NY.
Rich, P. M.1989.A manualfor analysis of hemispherical canopy photography. Los Alamos NationalLaboratory Report LA-11733-M.
———.1990.Characterizing plantcanopies with hemispherical photographs. Remote Sensing Reviews5 :13 –29.
———, D. B. Clark, D. A.Clark, and S. F. Oberbauer.1993.Long-term study ofsolar radiation regimes in a tropical wet forest using quantum sensorsand hemispherical photography. AgriculturalandForestry Meteorology65 :107 –127.
Ross, M. S., J. F. Meeder, G. Telesnicki, andC. Weekley.1995.Terrestrial ecosystems of theCrocodile Lakes National Wildlife Refuge: the effects of HurricaneAndrew. Southeast Environmental Research Program, Florida InternationalUniversity, Final report to the Department of Interior, U.S. Fish andWildlife Service.
Sagers, C.L.1993.Reproduction in neotropical shrubs: theoccurrence and some mechanisms of asexuality. Ecology74 :615–618.
Shimizu,Y.1994.Regeneration of mesic forest at Sekimon,Hahajima in the Bonin (Ogasawara) Islands with reference to disturbanceof the forest by the Typhoon No. 17 (1983). Regional Views(Komazawa University) 7: 3–32.
Simpson, R. H., and M. B.Lawrence.1971.Atlantic hurricane frequencies alongthe U.S. coastline. National Oceanic and Atmospheric AdministrationTechnical Memo Number NWS ST-58.
———, and H.Riehl.1981.The hurricane and its impact. LouisianaState University Press, Baton Rouge, LA.
Slater, H. H., W. J. Platt, B. B. Baker, andH. A. Johnson.1995.Effects of Hurricane Andrew ondamage and mortality of trees in subtropical hardwood hammocks of LongPine Key, Everglades National Park, Florida, USA. Journal of CoastalResearch 21:197 –207.
Smith, I. K., and J. L.Vankat.1992.Dry evergreen forest (coppice)communities of North Andros Island, Bahamas. Bulletin of the TorreyBotanical Club 119:181 –191.
Snyder, J. R., A. Herndon, and W. B.Robertson, Jr.1990.South Florida rockland.In R. L. Myers and J. J. Ewel [eds.], Ecosystems ofFlorida, 230–274. University of Central Florida Press, Orlando,FL.
Tanner, E. V. J., V. Kapos, and J. R.Healey.1991.Hurricane effects on forest ecosystems inthe Caribbean. Biotropica 23:513 –521.
Tomlinson, P.B.1980.The biology of trees native to tropicalFlorida. Published by the author. Petersham, MA.
Vasquez-Yanes, C., and A.Orozco-Segovia.1994.Signals for seeds to sense andrespond to gaps. In M. M. Caldwell and R. M. Pearcy[eds.], Exploitation of environmental heterogeneity by plants,209–236. Academic Press, Orlando, FL.
Wakimoto, R. M., and P. G.Black.1994.Damage survey of Hurricane Andrew and itsrelationship to the eyewall. Bulletin of the American MeteorologicalSociety 75:189 –200.
Walker, L. R.1991.Treedamage and recovery from Hurricane Hugo in Luquillo Experimental Forest,Puerto Rico. Biotropica 23:379 –386.
Weiss, S. B., P. M. Rich, D. D. Murphy, W. H.Calvert, and P. R. Ehrlich.1991.Forest canopystructure at overwintering monarch butterfly sites: measurements withhemispherical photography. Conservation Biology5 :165–175.
Whigham, D. F., I. Olmsted, E. C. Cano, andM. E. Harmon.1991.The impact of Hurricane Gilbert ontrees, litter fall, and woody debris in a dry tropical forest in thenortheastern Yucatan peninsula. Biotropica23 :434–442.
Yih, K. D., H. Boucher, J. H. Vandermeer, and N.Zamora.1991.Recovery of the rain forest ofsoutheastern Nicaragua after destruction by Hurricane Joan.Biotropica 23:106 –111.
Zimmerman, J. K., E. M. Everham III, R. B.Waide, D. Jean Lodge, C. M. Taylor, and N. V. L.Brokaw.1994.Responses of tree species to hurricanewinds in subtropical wet forest in Puerto Rico: implications fortropical tree life histories. Journal of Ecology82 :1–12.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ISF) in an undamaged
forest (JPS) and three forests (MAT, DEE, CAS) damaged by Hurricane
Andrew in August 1992. Means within a column followed by the same letter
are not significantly different at the 0.05 level (Kruskal-Wallis
nonparametric ANOVA). F
values are presented for each
comparison. * P
< 0.0001.
