MANGROVES AS ESSENTIAL NURSERY HABITAT FOR GOLIATH

BULLETIN OF MARINE SCIENCE, 80(3): 567–586, 2007

567

Bulletin of Marine Science

of the University of Miami

MANGROVES AS ESSENTIAL NURSERY HABITAT FOR

GOLIATH GROUPER EPINEPHELUS ITAJARA

Christopher C. Koenig, Felicia C. Coleman,

Anne-Marie Eklund, Jennifer Schull, and Jeﬀrey Ueland

ABSTRACT

We evaluated goliath grouper’s [Epinephelus itajara (Lichtenstein, 1822)] use of

mangroves as essential nursery habitat by estimating absolute abundance, density,

survival, age structure, home range, mangrove habitat association, habitat quality,

and recruitment to the adult population. Densities (numbers km

–1

mangrove shore-

line) were calculated using Jolly-Seber mark-recapture methods for mangrove-lined

rivers and mangrove islands of the Ten ousand Islands (TTI) and Everglades Na-

tional Park, which includes Florida Bay, Florida, USA. Juveniles had smaller home

ranges around islands (170 m) than in rivers (586 m), as determined from observa-

tions on telemetered ﬁsh. Goliath grouper remained in mangrove habitats for 5–6

yrs (validated ages from dorsal spine sections), then emigrated from mangroves at

about 1.0 m total length. In the TTI, juvenile densities around mangrove islands

were higher (mean = 25 km

–1

, SE = 6.2, CV = 0.5) and less variable than those in riv-

ers (mean = 11 km

–1

, SE = 4.2, CV = 1.2). Density was negatively correlated with the

frequency of dissolved oxygen and salinity minima. Mean growth rate of recaptured

ﬁsh around mangrove islands (0.358 mm d

–1

, 95% CL = 0.317–0.398) was signiﬁ-

cantly higher than that in rivers (0.289 mm d

–1

, 95% CL = 0.269–0.308). e annual

survival rate, as estimated by the Kaplan-Meier method on telemetered ﬁsh, was

0.947 (95% CL = 0.834–1.0). Very low densities in Florida Bay were probably related

to other water-quality variables in this human-altered system. e oﬀshore abun-

dance of adults was largely explained by abundance of mangrove, but not seagrass

habitat. Mangrove habitat with suitable water conditions, which appears essential to

the recovery and sustainability of goliath-grouper populations, should be protected

and/or restored.

Nearshore estuarine habitats are thought to provide “essential”

nursery habitat

for many ﬁsh species (Kapetsky, 1985; Koenig and Coleman, 1998; orrold et al.,

1998; Fitzhugh et al., 2005), but deﬁnitive tests of the nursery role rarely occur (Beck

et al., 2001). Conﬁrmation of the nursery role requires an understanding of the focal

species’ life history and a thorough evaluation that culminates in demonstration of

the habitat’s greater contribution per unit area to the production of the adult popula-

tion (Beck, 1995; Beck et al., 2001; Laegdsgaard and Johnson, 2001; Nagelkerken et

al., 2002; Halpern, 2004a,b). is conﬁrmation is critical from a conservation per-

spective because it provides a distinction between source and sink habitats; source

habitats are currently the primary focus of eﬀorts to protect ﬁshery production

(Crowder et al., 2000).

In this study, we tested the hypothesis that mangroves are integral and essential to

the early life stages of goliath grouper [Epinephelus itajara (Lichtenstein, 1822)] by

pursuing four objectives: (1) estimating juvenile density, survival, growth, and move-

roughout the present paper, “essential” is used in the sense of the Sustainable Fisheries Act, wherein

“essential ﬁsh habitat” is deﬁned as “those waters and substrate necessary to ﬁsh for spawning, feeding,

or growth to maturity” and “necessary” means “required to support a sustainable ﬁshery and the man-

aged species’ contribution to a healthy ecosystem.”

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

568

ment patterns in the Ten ousand Islands (TTI), southwest Florida; (2) comparing

our estimates of density in the TTI with abundance estimates from published and

unpublished sampling records from other estuarine and oﬀshore habitats; (3) com-

paring the eﬀects of water condition on densities in mangrove habitats of the TTI, in

rivers of the Everglades National Park (ENP), and in Florida Bay (FB), a system altered

by human activities; and (4) investigating the relative importance of juvenile habitat

to adult production by correlating regional near-shore habitat coverage with oﬀshore

adult abundance.

S A

e goliath grouper is the largest grouper in the western North Atlantic. It ma-

tures in 5–7 yrs, is long-lived (surviving up to 37 yrs), and reaches sizes exceeding

400 kg and 3.0 m total length (TL) (Robins et al., 1986; Bullock et al., 1992). Its life

cycle includes ontogenetic shifts in habitat use: larvae move from the pelagic envi-

ronment to shallow coastal areas, benthic juveniles live in estuaries, and adults occur

on oﬀshore reefs (< 50 m deep) (Eklund and Schull, 2001).

Goliath grouper is considered by the World Conservation Union (IUCN) to be

critically endangered throughout its range which extends from the west coast of Af-

rica to the west coast of Central America including the Caribbean, and from Brazil

to the southeastern U.S. (Heemstra and Randall, 1993; IUCN Red List, 2006, [http://

www.iucnredlist.org/search/search-basic]). A ﬁshing ban was put into eﬀect in the

southeastern U.S. in 1990 by the South Atlantic and the Gulf of Mexico ﬁshery man-

agement councils and in the Caribbean in 1993 by Caribbean Fishery Management

Council because it was considered to be severely overﬁshed (GMFMC, 1990; Sadovy

and Eklund, 1999). Although the National Marine Fisheries Service (NMFS) contin-

ues to list this species as overﬁshed (NMFS, 2004), the actual status of the popula-

tion remains unknown and diﬃcult to assess (Porch and Eklund, 2004), as is often

the case with fully protected species for which ﬁsheries data collection has not been

replaced with reliable population monitoring. is situation hinders development of

management measures aimed at rebuilding the population, ending overﬁshing, or

both, as required by the Sustainable Fisheries Act (Public Law 104-297).

S S

Study sites in the TTI and ENP (Fig. 1A) occurred in red mangrove (Rhizopho-

ra mangle Linnaeus) habitat along the shorelines of islands, tidal rivers and passes,

and shallow bays. e TTI system covers 1924 km

between latitudes 25°40´N and

26°00´N and longitudes 081°20´W and 081°40´W. e ENP covers the southern tip

of the Florida peninsula and extends through FB, the largest (2072 km

)

subtropical

estuary in North America (Loftus, 2000).

In both systems, we haphazardly chose widely dispersed study locations primar-

ily on the basis of accessibility (sites were often accessible only at high tides). In

the TTI, we sampled seven rivers (Littlewood, Palm, Blackwater, Pumpkin, Wood,

Whitney, and Ferguson), three man-made canals (92 Canal East, 92 Canal West, and

Faka Union Canal), and four mangrove-island systems (Remuda Ranch, Russell Key,

Fakahatchee, and Rabbit Key) (Fig. 1B). Water depths averaged < 0.1 to 2.0 m (MLW),

reaching 5 m in passes. Mixed diurnal tides range from 2.9 ft at Naples to 4.5 ft

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

569

Figure 1. Study sites in southwest Florida, USA. A. Areas covered by the present study and Flor-

ida coastal regions as delimited for Reef Environmental Education Foundation surveys. B. Ten

Thousand Islands system, showing sites used in the present study (those marked with asterisks are

man-made). River and canal sites: (A) 92 Canal West,* (B) 92 Canal East,* (C) Palm River, (D)

Blackwater River, (E) Whitney River, (F) Pumpkin River, (G) Little Wood River, (H) Faka Union

Canal,* (I) Ferguson River. Passes: (1) Remuda Ranch Channel, (2) Fakahatchee Pass, (3) Russell

Pass, (4) Rabbit Key Pass. C. Everglades National Park system. Rivers: (A) Lostmans River, (B)

Broad River, (C) Harney River, (D) North Prong River, (E) Shark River, (F) Roberts River. Bay

sites: (1) Bob Allen Key, (2) Black Betsey Key, (3) Little Buttonwood, (4) Blackwater Sound. Gray

= land; white = water.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

570

at the Shark River (NOAA tide tables). Undercuts (= scoured shorelines) formed a

structurally complex habitat in sampled areas, especially around mangrove islands

where mangrove roots formed “curtains” and submerged dead trees provided ad-

ditional habitat structure. e ENP study sites included six rivers (Harney, Broad,

Shark, Roberts, Lostman’s, and North Prong) entering the Gulf of Mexico and four

mangrove-island sites (Blackwater Sound, Little Buttonwood Sound, E. Bob Allen

Key, and Black Betsy Key) in FB (Fig. 1C). e river sites in the TTI and ENP diﬀered

little in appearance, whereas the TTI mangrove island sites diﬀered from those in

the ENP in being shallower and more widely spaced and having fewer undercuts.

M  M

J

Capture.— In the rivers, we captured ﬁsh using blue-crab traps and commercial grouper

ﬁsh traps baited with dead striped mullet (Mugil cephalus Linnaeus, 1758). Blue-crab traps

were 61 cm wide by 61 cm long by 46 cm high and made of 3.8-cm, 17.5-gauge plastic-coated

wire mesh; two funnels, each 13 cm long (proximal opening 19 × 12.7 cm; distal opening, 18 ×

7.6 cm), led from the outside into a lower chamber, and two funnels, each 10 cm long (proxi-

mal and distal openings, 18 × 7.6-cm) led from the lower chamber into an upper chamber.

e ﬁsh traps were 61 cm wide by 107 cm long by 46 cm high. e top and bottom panels

were 3.8-cm, 12.5-gauge plastic-coated wire mesh, and the side panels 2.5 × 5.2-cm, 14-gauge

wire mesh. e single entry funnel was 53 cm long (proximal opening, 45.7 × 30.5 cm; distal

opening, 30.5 × 10.2 cm).

We deployed 50 traps in each river at 93-m (0.05 NM) intervals in a linear sequence 4.6 km

(2.5 NM) long (Figs. 2A,B). e line of traps extended up each river from the mouth, except in

rivers shorter than 4.6 km, such as the Pumpkin, Little Wood (Fig. 2B), and Whitney, where

it extended past the river mouth into the bay. Trap spacing was arbitrary, but small enough

to maximize catch and large enough to sample most of the river or canal. Each deployment

lasted 21 d (sampled at 7 d intervals), with inter-deployment intervals about 42 d. Capture-ef-

ﬁciency studies during the ﬁrst year identiﬁed the peak capture period for subsequent sam-

pling: captures per set from June through November (105/1365 = 0.077) were signiﬁcantly

greater (binomial test, P < 0.0001) than those from January through May (6/460 = 0.013).

We sampled mangrove-island shorelines monthly during the summer and fall, mostly at

neap tides, using setlines (Fig. 2C). Each setline was composed of a baited 14-0 circle hook

attached by a 10-cm piece of 600-lb. test monoﬁlament (which minimized hook electrolysis)

to a 500-lb. test stainless steel cable attached to a 4-cm long section of 20 cm diam PVC pipe

(as a reel and an aid in relocating the setlines), which was attached to a mangrove limb. At

least three setlines were used per site (24–27 lines per sampling area). Each sampling event

included at least two bait-and-sample rounds, each 4–12 hrs in duration.

We tagged each newly captured ﬁsh ventrally with an individually numbered stainless-

steel-core internal-anchor tag (Floy Tag Company) and removed the third dorsal-ﬁn spine

and second and third dorsal-ﬁn rays for age determination (Fig. 3). Records for each ﬁsh cap-

tured or recaptured included the time, date, tag number, weight (g), size (mm total length,

TL), capture location (Trimble GeoExplorer GPS receiver in Arcview 8 GIS), and evidence

of previous spine and ray removal. We recaptured ﬁsh ourselves, but we included a toll-free

telephone number on each tag so that others could also report captures.

Age, Growth, and Tag Loss.—Collected spines were cleaned and dried, and several trans-

verse 0.5 mm sections were cut from each (Beuhler low speed isomet saw) 5–10 mm above the

spine base. e sections were mounted on slides, and two readers independently determined

age by counting translucent bands (= annuli) under compound dissecting microscopes. e

spine-aging method was validated by comparison with ages determined from otoliths and

by determination of incremental change over time from additional spines collected from re-

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

571

Figure 2. Examples of capture sites for juvenile goliath grouper, Epinephelus itajara in the Ten

Thousand Islands. (A) trap sampling in the Palm River. (B) trap sampling in the Little Wood

River. (C) setline sampling in Russell Pass. Traps were set at 0.093-km intervals. Sites of capture

(small closed circles) and no capture (small open circles) are indicated along with home range

(larger open circles) around capture sites. Gray = land; white = water.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

572

captured ﬁsh that were either marked with oxytetracycline (50 IU kg

–1

) or not (Brusher and

Schull, unpubl. data). Growth rates (average daily increase in length, mm day

–1

) were esti-

mated from recaptured ﬁsh. To minimize the eﬀects of measurement error on growth rate we

used only recaptured ﬁsh at liberty for > 30 d (n = 259). Growth was assumed to be linear over

the size range of juveniles captured in the study (following Bullock et al., 1992). Growth rate

diﬀerences among sites were tested for signiﬁcance with the Kruskal-Wallis non-parametric

test (no normalizing transformation was found). Where data were normal (Shapiro-Wilk

test), ANOVA was used to test for diﬀerences. Tag loss was indicated by evidence of previous

spine and ray removal in untagged ﬁsh.

Absolute Abundance and Survival.—To estimate juvenile absolute abundance and survival,

we used Jolly-Seber (J-S) mark-recapture methods (Krebs, 1999), the assumptions of which are

(1) that every individual in the sampled population has the same capture probability (equal

catchability); (2) that every marked individual has the same probability of surviving between

sampling times; (3) that no tag loss occurs during the sampling period. It also requires that

sampling time be short relative to intervals between samples (a feature built into our sam-

pling design), that at least three sampling times occur for each site, and that a relatively high

proportion (~ 0.50) of ﬁsh be marked.

We knew a priori that satisfying the equal catchability assumption would be diﬃcult be-

cause no single gear type uniformly sampled the entire juvenile size range (20–1000 mm

TL). (We collected newly settled juvenile goliath grouper at about 20 mm TL in multiple

locations in mangrove leaf litter, but could not include them in the J-S abundance estimate

due to collection gear limitations. However, they were included in the estimates of overall

absolute abundance and density.) Traps worked in rivers but not around mangrove island

undercuts, where they were dislodged by swift currents and rapidly fouled. Despite this bias

and the potential diﬃculty of distinguishing habitat-speciﬁc size (or age) distributions from

gear selectivity, we pursued multigear sampling to ensure capture of a greater size range (and

therefore age spectrum) of ﬁsh.

We tested for equal catchability within the selection limits of the gear with the zero-trun-

cated Poisson test and then compared the observed and expected frequency distributions

using a chi-squared goodness-of-ﬁt test (Krebs, 1999).

Figure 3. Juvenile goliath grouper Epinephelus itajara (38 cm TL) caught by sh trap in the Ten

Thousand Islands. Note tag on abdomen. Photograph by M. Finn.

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

573

We estimated absolute abundance for all sizes, 20–1000 mm TL and for all ages, 0–6, pres-

ent in the mangroves (N

) (1) for trap sampling, by determining site-speciﬁc abundances of

2-yr-olds, the age most often selected by traps (= mean proportion of 2-yr-olds caught × N

J-S

each river), and (2) for setline sampling, by determining the site-speciﬁc abundances of 4-yr-

olds, the age most often selected by setlines (= mean proportion of 4-yr-olds caught × N

J-S

each mangrove island site). We calculated the number in each age group by reconstructing all

age groups (ages 0–5) based on M = 0.05, our Kaplan-Meier estimate of survival (see below),

then summing to obtain the absolute abundance of all age groups at each sampling site during

the sampling period. We did not include age 6 in the calculation because most ﬁsh appeared

to leave the mangroves in their ﬁfth year (Table 1).

To ensure equal probability of survival from one sampling time to the next we made sure

that all released ﬁsh were in good condition. We estimated survival in two ways: (1) with

mark-recapture and the J-S method, and (2) with telemetry and the Kaplan-Meier method

(Krebs, 1999). J-S survival estimates, which depend only on the marked population, can be

confounded by many factors related to the behavior of the ﬁsh and/or the sampling design.

For example, if marked individuals emigrate, or for some reason cannot be recaptured (e.g.,

they outgrow the sampling gear or actively avoid recapture), they are assumed by the model to

have died. e Kaplan-Meier survival estimate does not have such limitations and makes use

of all the relocation data until the ﬁsh either dies or is lost from the sample.

Density, Habitat Association, and Home Range.—We estimated home range by tagging 22

juveniles (size range 430–940 mm TL) intraperitoneally with individually coded transmitters

(Sonotronics; battery life 48 mo) and relocating them during our normal sampling activities

over the period of the study. We assumed that tracking ﬁsh telemetrically left them relatively

undisturbed and therefore provided more reliable information on movement patterns and

home ranges than recaptures with traps or setlines, which could aﬀect their behavior.

So that telemetered ﬁsh could be recognized externally, a numbered internal-anchor tag

with external streamer was placed on either side of the incision, which was closed with stain-

less steel medical staples and covered with antibiotic salve before the ﬁsh was released.

We relocated telemetered ﬁsh with a directional hydrophone (Sonotronics), recorded posi-

tions, and conﬁrmed that detected ﬁsh were still alive by direct observation using SCUBA or

by prodding. We then estimated home range (= distance traversed by an individual during

the sampling period) by averaging the greatest straight-line distances between relocation sites

(White and Garrott, 1990). Half of the home range was equivalent to the radius of a circle

centered on the sampling site (Seber, 1982). Home ranges were estimated separately for rivers

and islands. Densities (D) were determined from

DnH

(1)

Table 1. Size and age distribution of juvenile goliath grouper, Epinephelus itajara, in mangrove-

lined rivers and mangrove islands of the Ten Thousand Islands, southwest Florida, USA, 1998–

2000. All sh caught in rivers were taken in traps; those caught near mangrove islands were taken

with setlines. Data are numbers of individuals; TL = total length.

TL (mm) Age 0 Age 1 Age 2 Age 3 Age 4 Age 5 Age 6 Tot a l

101–200 1 8 3 25 37

201–300 3 90 107 108 3 311

301– 400 52 239 100 9 400

401–500 1 81 48 17 2 149

501– 600 8 21 36 3 68

601–700 4 8 34 11 1 58

701–800 3 22 17 2 44

801–900 15 17 2 34

901–1,000 6 7 2 15

Total n 4 151 442 313 142 57 7 1,116

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

574

where n

= estimated absolute abundance for all size classes (20–1000 mm TL) and H =

kilometers of mangrove shoreline within the home range about each sampling area.

Patterns of dispersion were estimated by use of the coeﬃcient of dispersion (CD, variance

to mean ratio), where means and variances were calculated from the number of ﬁsh captured

at each sampling site over the 2 yr sampling period.

Water Condition.—We evaluated the relationship between dissolved oxygen (DO, mg L

–1

)

and salinity and ﬁsh density using 5 yrs of water quality data (1997–2001) collected monthly

near or at our sampling sites from the surface and bottom (total samples = 116–120 per site)

by the South Florida Water Management District and archived on websites [http://www.sf-

wmd.gov/org/ema/envmon/wqm/coastal/tti.html and http://www.sfwmd.gov/org/ema/en-

vmon/wqm/ﬂab/ﬂabindex.html]. We assumed that salinity and DO were among the most

important variables aﬀecting ﬁsh distribution in these systems. We did not include tempera-

ture minima in this analysis—even though goliath grouper are sensitive to low temperatures

(Sadovy and Eklund, 1999)—because any cold fronts moving through the area would aﬀect

the entire SW Florida system, not just individual sites within the system. e State of Florida

minimum DO water quality standard for marine waters is 4 mg L

–1

, but we used 3 mg L

–1

reﬂect the dynamic diurnal and seasonal conditions in estuarine habitats. e State of Florida

has no salinity standards for estuaries, so we used lower (salinity of 5) and upper (salinity of

40) limits known for other serranid ﬁshes found in estuaries (e.g., gag, black sea bass, and red

grouper) (Koenig and Coleman, 1998; Atwood et al., 2001).

A

Abundance.—We used goliath grouper sighting frequency data as a proxy for adult popu-

lation abundance rather than population-density data to minimize the eﬀects of feeding or

reproductive aggregations on regional estimates. We determined sighting frequency based on

dive surveys conducted by e Reef Environmental Education Foundation (REEF) on Florida

reefs from 1993 to 2004 (REEF database = http://www.reef.org/data/twa/index.htm). REEF

surveyors use a standardized “roving diver method” in which divers swim freely through sites,

recording every observed ﬁsh species that can be positively identiﬁed. Because adult goliath

grouper are distinct in terms of size, shape, and color pattern, we are conﬁdent that both

novice and expert surveyors made correct identiﬁcations.

We compared goliath grouper sighting frequency oﬀ Florida with the abundance of two

major estuarine habitat types in coastal Florida, mangrove habitat (data from the Florida Nat-

ural Areas Inventory, available at http://www.fnai.org/about.cfm) and seagrass habitat (data

from Florida Fish and Wildlife Conservation Commission’s Fish and Wildlife Research Insti-

tute), using a logistic regression model to determine the signiﬁcance (P < 0.05), direction, and

strength of the relationship (R

) between the variables.

R

A, G,  T L.—Juvenile goliath grouper (n = 1116) were found in

mangrove habitat up to age 6 and 1000 mm TL (Table 1). e overall mean growth

rate for all recaptured ﬁsh was 0.300 mm d

–1

(95% CL = 0.282–0.318, n = 259). In the

TTI, growth rates were not signiﬁcantly diﬀerent among rivers (Kruskal-Wallis, P >

0.05) nor among mangrove-island sites (Kruskal-Wallis, P > 0.05), but growth rates

around islands (mean = 0.358 mm d

–1

, 95% CL = 0.317–0.398, n = 41) were signiﬁ-

cantly greater (Kruskal-Wallis, P < 0.01) than growth rates in rivers (mean = 0.289

mm d

-1

, 95% CL = 0.269–0.308, n = 218). We found no signiﬁcant (Kruskal-Wallis, P

> 0.05) relationship between growth rate and population density among rivers, nor

among island sites.

Tag loss was negligible (n = 1), all ﬁsh were released in good condition, and recap-

tured ﬁsh did not have infected tagging wounds.

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

575

A A  S

Equal Catchability.—Although TTI goliath grouper met the general assumption

of equal catchability (within the selective limits of the gear; zero-truncated Poisson

test; χ

, P > 0.05), all three gears showed some size (= age) selectivity (ANOVA, P <

0.05; Fig. 4). Crab traps selected primarily ﬁsh 300–400 mm TL because the fun-

nel opening excluded larger individuals and the mesh size allowed smaller ones to

escape during trap retrieval. Fish traps were less selective at the upper size range but

also allowed ﬁsh < 200 mm TL to escape. Setlines captured primarily ﬁsh 500–1000

mm TL and appeared to exclude ﬁsh < 400 mm TL.

Absolute Abundance.—Because of gear selectivity, the J-S method only estimated

the absolute abundance of the selected sizes for each gear type. ese were calcu-

lated directly by the J-S method and indirectly with a predictive model (Table 2; Fig.

5) based on catch per unit eﬀort (CPUE; one unit of eﬀort = one set of traps 4.6 km

long), such that:

..YX79 51=- +

(2)

where Y is the absolute abundance and X is the mean CPUE for the sampling site

(Linear Regression, P < 0.05, R

= 0.89). e CPUE model was developed from annual

J-S estimates of absolute abundance of trap catches for 1999 and 2000 for six rivers.

[We did not include Little Wood River because the ﬁsh in that river were unusually

abundant and highly clumped (Fig. 2B); such a situation produced a low CPUE rela-

tive to absolute abundance because traps saturated in high density areas around the

river mouth]. e regression model provided an eﬃcient means of estimating abun-

dance outside of the TTI system and in areas where capture frequencies were too low

for direct use of the J-S method.

Estimating the abundance over all sizes (and therefore all gears) required accept-

ing three assumptions: (1) that all sizes were present in river and island systems, (2)

that year-class strength was constant among sampling years, and (3) that mortality

was uniform across all age groups. e ﬁrst assumption is justiﬁed because in our

collections, ﬁsh sizes ranged from 20 to 800 mm TL in rivers and 20 to 930 mm TL

Figure 4. Percent frequency of size classes of juvenile goliath grouper Epinephelus itajara cap-

tured by each of the three gear types in the Ten Thousand Islands. TL = total length.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

576

Table 2. Mean estimates of absolute abundance (n), survival probability (PHI), and density (D) of

juvenile goliath grouper, Epinephelus itajara, sampled in mangrove shoreline of the Ten Thou-

sand Islands (TTI), Everglades National Park (ENP), and Florida Bay (FB), southwest Florida,

USA, 1998–2001. Density estimates include all sh 20–1,000 mm SL (ages 0–5). S = number

of sampling times. α = maximum proportion of population marked. n

J-S

= age-specic absolute

abundance estimate based on Jolly-Seber model. PHI = nite probability of survival standardized

to 30 d. n

= absolute abundance based on all ages, with annual mortality (M) of 0.05. H = area

sampled in kilometers of mangrove shoreline.

System S α n

J-S

(95% CL)

PHI

(95% CL)

(M = 0.05)

(km)

(n km

−1

)

TTI rivers

Little Wood River 6 0.39 202 (163–242) 0.84 (0.48–1.0) 571 13.1 44

Palm River 6 0.58 70 (58–81) 0.69 (0.42–0.93) 196 9.2 21

Blackwater River 6 0.67 42 (1–83) 0.64 (0.28–0.87) 118 9.2 13

92 Canal East 6 0.79 41 (28–55) 0.86 (0.60–1.0) 116 10.4 11

Pumpkin River 5 0.43 38 (14–62) 0.81 (0.52–1.0) 106 8.5 12

Faka Union Canal 5 0.26 12 (2–22) 0.32 (0.16–0.60) 24 10.6 2

92 Canal West 5 0.72 10 (8–36) 0.64 (0.23–0.88) 27 10.4 3

Wood River* 5 – 5 (0–20) – 14 5.4 3

Whitney River* 6 – 3 (0–20) – 8 8.2 1

Ferguson River* 3 – 3 (0–20) – 8 10.7 1

ENP rivers

Harney River* 2 – 26 (18–36) – 74 15.9 5

Broad River* 2 – 18 (7–29) – 52 15.8 3

Shark River* 2 – 3 (0–20)

– 8 12.2 1

Roberts River* 2 – 0 – 0 13.7 0

Lostmans River* 2 – 0 – 0 7.8 0

North Prong* 2 – 0 – 0 12.1 0

TTI mangrove islands

Remuda Ranch Pass 4 0.42 31 (13–158) 1.0 (0.57–1.0) 82 2.0 41

Russell Key Pass 7 0.80 31 (15–117) 0.78 (0.32–1.0) 82 3.0 27

Fakahatchee Pass 5 0.75 18 (12–58) 0.61 (0.22–0.93) 49 3.2 15

Rabbit Key Pass 7 0.86 23 (13–70) 0.95 (0.53–0.97) 60 4.0 15

FB mangrove islands

Blackwater Sound* 2 – 3 (0–20) – 8 3.6 2

Little Buttonwood S* 2 – 0 – 0 4.3 0

E. Bob Allen Key* 2 – 0 – 0 2.3 0

Black Betsey Key* 2 – 0 – 0 5.3 0

*Absolute abundance estimate based on catch-per-unit-effort linear-regression model (Y = −7.9 + 5.1X).

around islands. at, combined with strong site ﬁdelity (see below), suggests that

juveniles remained near settlement sites in both systems. e second assumption

is justiﬁed because between sampling years at each site, we observed similar J-S ab-

solute abundances, suggesting a similar year-class size among years, especially in

rivers where highly selective crab-trap gear was used. For rivers, the mean absolute

abundances for 1999 and 2000 were: 217 and 193 for Littlewood; 64 and 73 for Palm;

33 and 42 for Blackwater; 37 and 39 for Pumpkin; and 48 and 37 for 92 Canal East.

For island sites, the mean absolute abundances for 1998 and 1999 were: 28 and 33 for

Remuda Ranch; 18 and 18 for Fakahatchee; 34 and 17 for Rabbit Key; and 36 and 28

for Russell Key. e third assumption is probably unjustiﬁed because young (small)

ﬁsh likely experience higher mortality rates than older (larger) ﬁsh, which is a general

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

577

pattern (Roﬀ, 1992). Because we have no way of estimating size-speciﬁc mortality, we

used the Kaplan-Meier estimate of annual mortality, M = 0.05, which was observed

on telemetered ﬁsh from about 400 to 1000 mm TL.

Survival.—e proportion (α) of marked juvenile goliath grouper increased over

the course of the study to a maximum that usually exceeded 0.5 (Table 2), suggesting

high survival of marked ﬁsh, which was further supported by monitoring teleme-

tered ﬁsh. e survival probability of TTI mangrove island ﬁsh (mean = 0.835) and

river ﬁsh (mean = 0.727) did not diﬀer signiﬁcantly in J-S 30-d survival probability

(ANOVA, P > 0.05, Table 2).

Two of the transmitters implanted in ﬁsh malfunctioned (could not be detected

upon release), and three others were never detected after the initial check. e re-

maining 17 ﬁsh were monitored for 34 mo, and Kaplan-Meier survival was estimated

on those ﬁsh. e 30-d survival rate was 0.995 (95% CL = 0.985–1.00); annual sur-

vival rate was 0.947 (95% CL = 0.834–1.00).

Habitat Association, Home Range, and Density.—Observations of telemetered ju-

veniles showed that they were strongly associated with structure (mangrove under-

cuts and root curtains, submerged mangrove branches or trees, and rocky outcrops).

ey also showed patchy distributions in both rivers and around islands (mean CD

for islands = 5.3 and rivers = 3.0). Because structure is associated with shoreline, we

used a linear measure (number km

–1

shoreline) as the basis for density estimation.

Recaptured juveniles showed high site ﬁdelity; over 90% moved no farther than

0.5 km during the entire sampling period. Straight-line home ranges for telemetered

ﬁsh in rivers averaged 586 m (radius = 293 m) and those around mangrove islands,

170 m (radius = 85 m; Table 3). Home range estimates served as the basis for density

estimation.

Mean densities around mangrove islands (25 km

–1

, SE = 6.2, CV = 0.5) were sig-

niﬁcantly greater and less variable than those in rivers (11 km

–1

, SE = 4.2, CV = 1.2)

(Table 2).

Figure 5. Linear regression of absolute abundance on catch per unit effort (CPUE; individuals

per trap set) for trap-caught goliath grouper, Epinephelus itajara, in rivers of the Ten Thousand

Islands: Blackwater River, Palm River, Pumpkin River, Faka Union Canal, 92 Canal West, and

92 Canal East. Mean values for 1999 and 2000. Regression model: Y = −7.9 + 5.1X (R

= 0.89).

Dashed lines represent 95% condence boundaries.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

578

Water Condition.—Fish densities in all TTI rivers and mangrove islands were neg-

atively correlated with low salinity (Correlation, P < 0.05) and with low salinity in

concert with low DO (Correlation, P < 0.01) (Fig. 6), but not with low DO alone (P =

0.07). Low densities in FB were not correlated with low DO or salinity (Table 4).

Adult Abundance.—Oﬀshore adult abundance (from REEF surveys) was positively

correlated with the quantity of mangrove shoreline along the Florida coast (Logistic

Regression, P < 0.001, R

= 0.83; Table 5, Fig. 7). A stronger relationship (R

= 0.92)

resulted if the Florida Keys were excluded—exclusion was justiﬁed by the low pro-

portion of goliath-grouper sightings (4% of 14,386 observations) and the poor water

quality around the Keys (Kruczynski, 1999). Adult abundance was negatively related

to seagrass abundance (Logistic Regression, P < 0.001, R

=0.75).

D

Beck et al. (2001) reﬁned the deﬁnition of nursery habitat in such a way that dem-

onstrating nursery function for a particular species requires that researchers evaluate

density, survival, growth, and above all, the contribution of the habitat to the produc-

tion of the adult population. Other than this study, there are virtually no published

studies that have unequivocally demonstrated a nursery function of mangroves for

ﬁshes according to this deﬁnition (Faunce and Serafy, 2006).

is study clearly demonstrates that mangroves can only provide important nurs-

ery habitat for juvenile goliath grouper when local environmental conditions are

suitable. Rivers are more likely than mangrove islands to experience frequent hy-

poxic (DO < 3 mg L

–1

) or low-salinity (< 5) conditions because they drain upland ar-

Table 3. Home-range estimates of telemetered juvenile goliath grouper, Epinephelus itajara, in

the Ten Thousand Islands rivers and mangrove islands, southwest Florida, USA. Ranges are mea-

sured linearly, in kilometers of mangrove shoreline.

Location Fish size at

tagging (mm TL)

Times observed Observation period

(months)

Maximum distance

traversed (m)

Rivers

Little Wood R. 555 22 20 259

543 20 18 611

Palm R. 438 18 12 926

620 10 19 315

579 24 20 818

Mean home range 586

Mangrove islands

Rabbitt Key 610 6 26 340

640 9 16 350

Russell Key 909 4 10 197

825 2 9 86

580 4 16 100

Fakahatchee 695 9 25 50

805 4 15 67

580 5 25 86

Pumpkin Bay 508 8 13 100

559 16 31 320

Mean home range 170

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

579

eas, include man-made canals, and are impacted more acutely by upland freshwater

management projects that occur throughout this area (Ogden et al., 2005). us, it

is not surprising to ﬁnd in rivers that juveniles occur at lower densities, experience

slower growth rates, and have greater home ranges than island ﬁsh—keeping in mind

that density may be inﬂuenced by the availability of structure (Frias-Torres, 2006),

and growth rates may be aﬀected by food availability.

e near lack of juveniles in FB is best explained by the bay’s long-standing de-

graded state, characterized by periods of hypersaline and high-nutrient conditions;

periodic toxic phytoplankton blooms; and signiﬁcant seagrass, sponge, and man-

grove die-oﬀs (Boyer et al., 1999; Fourqurean and Robblee, 1999). ese conditions

and events probably aﬀected nursery function (Butler et al., 1995), although other

factors, such as limited larval delivery and lack of microhabitat features, cannot be

ruled out.

Nursery habitat is expected to promote survival in part by providing refuge from

predation. While mangrove habitat is rife with possible predators of juvenile goliath

grouper—including gray snapper (Lutjanus griseus Linnaeus, 1758), snook [Centro-

pomus undecimalis (Bloch, 1792)], red drum [Sciaenops ocellatus (Linnaeus, 1766)],

and gag [Mycteroperca microlepis (Goode and Bean, 1879)], as well as bull sharks

[Carcharhinus leucas (Müller and Henle, 1839)], and lemon sharks [Negaprion brevi-

rostris (Poey, 1868)]—it also is structurally complex, providing signiﬁcant refugia.

Measuring survival is problematic. We found that the J-S method routinely un-

derestimated survival in goliath grouper because ﬁsh loss due to emigration or es-

capement was indistinguishable from ﬁsh loss due to mortality. e Kaplan-Meier

method minimizes these biases, and therefore provides a more realistic estimate.

Indeed, the Kaplan-Meier survival estimate is quite high (annual rate = 0.947, 95%

Figure 6. Relationship between density (number km

–1

of mangrove shoreline) of juvenile goliath

grouper, Epinephelus itajara, and dissolved oxygen and salinity minima (percent of monthly

sampled dissolved oxygen and salinity measures collected top and bottom from 1997 to 2001,

total of about 120 samples, below 3.0 mg L

–1

and 5, respectively), in the Ten Thousand Islands and

Everglades National Park rivers and mangrove islands. Negative correlations for dissolved oxygen

and salinity minima are P = 0.07 and P = 0.03, respectively.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

580

CL = 0.834–1.00), and thus strengthens conﬁdence that the marked (= telemetered)

population adequately represents the unmarked population, an important consider-

ation in any mark-recapture study.

e most important criterion for determining the nursery status of a particular

habitat is its contribution to the adult population relative to that of other juvenile

habitats occupied by the particular species (Beck et al., 2001). e fact is that juve-

nile goliath grouper only rarely occur in any other habitats besides mangroves. No

juveniles were collected in intensive inshore surveys of seagrass beds (> 2500 trawl

tows and gillnet sets by Coleman et al., 1993, 1994; ayer et al., 1999; Fitzhugh et

al., 2005) and mud bottom (~500 trawl tows by Coleman et al., 1993, 1994) from the

Florida panhandle to Florida Bay. Only three of 143 juvenile sightings reported to the

FWC hotline occurred in habitat other than mangroves, including one oﬀshore, one

in seagrass about 100 m from a mangrove coastline, and one in concrete rubble near

a river mouth. Further, anecdotal information from interviews with long-time local

Table 4. Dissolved oxygen, salinity, and the density of juvenile goliath grouper, Epinephelus itajara,

in mangroves at Ten Thousand Islands (TTI), Everglades National Park (ENP), and Florida Bay (FB)

water-quality stations. Density estimated for all juveniles 20–1,000 mm TL (ages 0–5). DO min =

percentage of monthly dissolved oxygen observations below 3 mg L

–1

. Salinity min = percentage of

monthly salinity observations below 5; max = percentage of monthly salinity observations above 40.

Density was measured as individuals per km of shoreline.

Salinity

Location Station DO min Minimum Maximum Density

Rivers

Littlewood River TTI 72 2.5 0.0 3.4 44

Palm River TTI 75 15.0 2.6 0.0 21

Blackwater River TTI 75 15.0 2.6 0.0 13

Pumpkin River TTI 72 2.5 0.0 3.4 12

92 canal E TTI 75 15.0 2.6 0.0 11

Harney River FLAB 36 10.0 0.0 0.0 5

92 Canal W TTI 75 15.0 2.6 0.0 3

Wood River TTI 70 8.6 15.5 3.4 3

Faka Union Canal TTI 70 8.6 15.5 3.4 3

Broad River FLAB 35 5.1 13.0 0.0 3

Whitney River TTI 75 15.0 2.6 0.0 1

Ferguson River TTI 64 16.0 12.0 1.7 1

Shark River FLAB 39 28.0 37.0 0.0 1

Lostmans River FLAB 29 0.0 20.0 0.0 0

Roberts River FLAB 47 0.0 39.0 0.0 0

North Prong River FLAB 38 11.2 74.3 0.0 0

TTI mangrove islands

Remuda Ranch TTI 74 2.1 0.0 0.0 41

Russell Key TTI 65 0.0 0.0 2.5 27

Fakahatchee TTI 69 1.7 1.7 2.5 15

Rabbit Key TTI 52 0.0 0.0 2.6 15

FB mangrove islands

Blackwater Sound FLAB 05 0.0 0.0 0.0 2

E. Bob Allen Keys FLAB 21 0.0 0.0 17.1 0

Little Buttonwood Sound FLAB 05 0.0 0.0 0.0 0

Black Betsey Key FLAB 23 0.0 0.0 0.0 0

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

581

ﬁshermen revealed that mangrove-associated goliath grouper are exclusively juve-

niles (i.e., < 1000 mm TL) (L. Demere, Chokoloskee, FL ﬁshing guide and third gen-

eration ﬁsherman), and that oﬀshore reef-associated goliath grouper are virtually all

adults (i.e., > 1000 mm TL) (D. DeMaria, Key West, FL commercial spearﬁsher who

targeted goliath grouper before 1990; Porch and Eklund, 2004). Indeed, the cover

provided by habitat types other than mangroves is insuﬃcient for a ﬁsh that reaches

a meter in length.

Demonstrating connectivity between nursery and adult habitats for a protected

species such as goliath grouper is more diﬃcult in the absence of a regular monitor-

ing program. In this study, we opted to compare regional abundance of adult goliath

grouper oﬀshore with the quantity of mangrove shoreline immediately inshore of

sampled populations. (We assumed that the local population, still being in the re-

covery phase, had not reached the carrying capacity that would force population

expansion to neighboring lower-density areas.) e relationship between adult abun-

dance and quantity of mangrove shoreline is strongly positive and corroborates the

contention that southwest Florida is the center of goliath grouper abundance in the

southeastern United States (see Sadovy and Eklund, 1999; Porch and Eklund, 2004).

Emigration of juvenile goliath grouper occurred in 5- to 6-yr-olds at about 1000

mm TL, just under the size of maturity estimated by Bullock et al. (1992): 1100–1150

mm TL for males; 1200–1350 mm TL for females. Emigration is more likely to be

size rather than age dependent, given that size-at-age became increasingly more vari-

able over time and that ﬁsh left the estuary at about one meter TL regardless of age

(Table 1). A seasonal cue such as temperature is unlikely to trigger egress in goliath

grouper as it appears to in gag (see, e.g., Koenig and Coleman, 1998) because juvenile

goliath grouper remain in mangroves for years, whereas gag leave their nursery habi-

tat during the ﬁrst fall season of residency.

Mangrove habitat suitable as goliath grouper nursery in the South Florida Ecosys-

tem was probably much more abundant and widespread a century ago than it is now,

extending at least from Indian River Lagoon on the east coast to the Caloosahatchee

estuary on the west coast, including Biscayne and Florida bays (Ogden et al., 2005).

Human activity has had an enormous impact, including that caused by water man-

agement, agriculture, industrial and housing developments, roadways, and mosquito

abatement, resulting in major changes in the timing, quantity, and quality of fresh-

water ﬂows into the estuary. Goliath grouper is only one of a diverse array of species

Table 5. Distributions of adult goliath grouper Epinephelus itajara sightings offshore (1993–2004)

and the extent of mangrove shoreline and seagrass area inshore in all Florida coastal regions, from

Reef Environmental Education Foundation surveys. % = (no. sightings/no. surveys) × 100.

Location (as marked in Fig. 1A) No.

sightings

No.

surveys

% Mangrove

shoreline (km)

Seagrass

(km

)

Gulf of Mexico Coast

Northwest Florida 36 322 11 48 2,778

Central West Florida 105 332 32 2,778 307

Southwest Florida (with Florida Bay) 365 535 68 8,752 1,538

Florida Keys 640 14,386 4 2,584 2,862

Atlantic Coast

NE Florida to Canaveral 33 170 19 1,065 217

Canaveral to Jupiter 61 765 8 545 71

Jupiter to Biscayne 170 4,498 4 532 2,018

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

582

aﬀected by these hydrologic and water quality alterations (Sklar and Browder, 1998;

Ogden et al., 2005).

In mangrove habitat, goliath grouper appear to be intolerant of large and/or highly

variable ﬂows of fresh water, but conditions of low DO, other water quality factors,

and episodic events such as red tide, possibly stimulated by enhanced nutrient loads,

undoubtedly aﬀect the suitability of the mangrove nursery. Additional studies are

needed to determine the productivity of other South Florida mangrove habitat for

goliath grouper (and other species) and to deﬁne the various factors aﬀecting density,

growth, and home range including interactions of water quality, food availability,

microhabitat features.

In addition to reduced mangrove habitat quality in the South Florida system, di-

rect mangrove loss is occurring. Current mangrove coverage (~2174 km

based on

FWRI data) represents a signiﬁcant reduction from coverage that existed 100 yrs ago,

before signiﬁcant human-induced changes in the system (Johnson, 1974; Gilmore

and Snedaker, 1993; Sklar and Browder, 1998; Ogden et al., 2005). e most recent

losses (between 1986 and 1997) include 3.3% (from 366 to 354 km

) for FB and 28%

(from 259 km

to 186 km

) for TTI (Fig. 8). Worldwide, mangrove ecosystems have

declined by ~35% since the 1980s, primarily because of mariculture in third-world

countries (Mendelssohn and Mckee, 2000; Valiela et al., 2001).

Our data suggest that availability of suitable mangrove habitat is the primary bot-

tleneck to goliath grouper production, making suitable mangrove habitat not only

essential, as deﬁned by the Sustainable Fisheries Act, but critical in the sense deﬁned

by the U.S. Endangered Species Act (16 U.S.C. Section 1531 et seq.). Given that nearly

75% of Florida’s human population lives in coastal counties, that the population is

expected to increase by 9 million people in the next 25 yrs (Florida Oﬃce of Eco-

nomic and Demographic Research, http://edr.state.ﬂ.us/population.htm), and that

mangrove losses of 20% are expected in the next 50 yrs (J.U., unpubl. data), we con-

tend that to be eﬀective, any recovery plan for this species should include protection

and restoration of mangrove nursery habitat, which in turn depends in large part

on restoration of “deﬁning characteristics” (Ogden et al., 2005) of the South Florida

ecosystem.

Figure 7. Logistic regression of proportion of REEF surveys with adult goliath grouper (proxy for

offshore adult abundance) on kilometers of mangrove shoreline for coastal regions of Florida (P

< 0.001, R

= 0.83).

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

583

A

We thank L. Demere (Chokoloskee, FL), J. and B. White (Chokoloskee, FL), and T. Schmidt

(National Park Service) for habitat information. M. Finn (Goodland, FL), D. DeMaria (Key

West, FL), T. Bevis (Florida State University), J. Brusher (NMFS, Panama City) provided tech-

nical assistance; H. Norris and N. Morton (FWC, FWRI) provided data on mangrove distribu-

tion. J. Dodrill, B. Horn, and K. Mille (FWC) provided data on goliath grouper abundance. e

FWC Florida Wildlife Research Institute supported the tagging hotline. e Ocean Conser-

vancy provided posters. We acknowledge grants from the NMFS Oﬃce of Protected Species,

the NMFS Oﬃce of Habitat, the Curtis and Edith Munson Foundation, the National Fish and

Wildlife Foundation, and e Pew Fellows Program in Marine Conservation (to F.C.C.). We

also thank A. B. istle for formatting and editing the manuscript. is research was con-

ducted under the guidelines of the Florida State University Animal Care and Use Commit-

tee and under permits from the Florida Fish and Wildlife Conservation Commission. is is

Contribution No. 1109 of the Florida State University Coastal and Marine Laboratory.

L C

Atwood, H. L., S. P. Young, and J. R. J. Tomasso. 2001. Salinity and temperature tolerance of

black seabass juveniles. N. Am. J. Fish. Manage. 63: 285–288.

Beck, M. W. 1995. Size-speciﬁc shelter limitation in stone crabs: a test of the demographic

bottleneck hypothesis. Ecology 76: 968–980.

Figure 8. Satellite images of Florida Bay and Ten Thousand Islands mangrove coverage. Upper

panels, before mangrove die-back; lower panels, after mangrove die-back. Black = mangrove,

gray = land and other wetlands; white = water. Data are from Landsat TM imagery through an

ISO data classication at 30-m resolution. Fish and Wildlife Research Institute (Florida Fish and

Wildlife Conservation Commission), St. Petersburg, FL, USA.

BULLETIN OF MARINE SCIENCE, VOL. 80, NO. 3, 2007

584

_________, K. L. Heck, K. W. Able, D. L. Childers, D. B. Eggleston, B. M. Gillanders, B. Halpern,

C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. Weinstein. 2001. e

identiﬁcation, conservation, and management of estuarine and marine nurseries for ﬁsh

and invertebrates. BioScience 51: 633–641.

Boyer, J. N., J. W. Fourqurean, and R. D. Jones. 1999. Seasonal and long-term trends in the water

quality of Florida Bay (1989–1997). Estuaries 22: 417–430.

Bullock, L. H., M. D. Murphy, M. F. Godcharles, and M. E. Mitchell. 1992. Age, growth, and

reproduction of jewﬁsh Epinephelus itajara in the eastern Gulf of Mexico. Fish. Bull. 90:

243–249.

Butler, M. J., J. H. Hunt, W. F. Herrnkind, M. J. Childress, R. Bertelsen, W. Sharp, T. Mathews, J.

M. Field, and H. G. Marshall. 1995. Cascading disturbances in Florida Bay, USA: cyanobac-

teria blooms, sponge mortality, and implications for juvenile spiny lobsters Panulirus argus.

Mar. Ecol. Prog. Ser. 129: 119–125.

Coleman, F. C., C. C. Koenig, and W. F. Herrnkind. 1993. Annual report: survey of Florida in-

shore shrimp trawling by-catch and preliminary tests of by-catch reduction devices. 1992.

Florida Department of Natural Resources, Tallahassee, Florida. 28 p + appendices.

____________, ___________, and _____________. 1994. Second annual report: survey of Florida

inshore shrimp trawlers bycatch. DNR Contract #C-7779. Florida Department of Natural

Resources, Florida Marine Research Institute, Tallahassee, Florida. 60 p.

Crowder, L. B., S. J. Lyman, W. F. Figueria, and J. Priddy. 2000. Source-sink population dynam-

ics and the problem of siting marine reserves. Bull. Mar. Sci. 66: 799–820.

Eklund, A. M. and J. Schull. 2001. A stepwise approach to investigating the movement patterns

and habitat utilization of Goliath grouper, Epinephelus itajara, using conventional tagging,

acoustic telemetry, and satellite tracking. Pages 189–216 in J. R. Sibert and J. L. Nielsen, eds.

Electronic tagging and tracking in marine ﬁsheries. Kluwer, Dordrecht.

Fitzhugh, G. R., C. C. Koenig, F. C. Coleman, C. B. Grimes, and W. A. Sturges. 2005. Spatial and

temporal patterns in fertilization and settlement of young gag (Mycteroperca microlepis)

along the West Florida Shelf. Bull. Mar. Sci. 77: 377–396.

Fourqurean, J. W. and M. B. Robblee. 1999. Florida Bay: a history of recent ecological changes.

Estuaries 66: 345–357.

Frias-Torres, S. 2006. Habitat use of juvenile goliath grouper (Epinephelus itajara) in the Flor-

ida Keys, USA. Endang. Species Res. 1: 1–6.

Gilmore, R. G. and S. C. Snedaker. 1993. Mangrove forests. Pages 165–198 in W. H. Martin, S.

G. Boyce, and A. C. Echternacht, eds. Biodiversity of the southeastern United States: low-

land terrestrial communities. Wiley, New York.

GMFMC (Gulf of Mexico Fishery Management Council). 1990. Amendment 2 to the ﬁshery

management plan for reef ﬁsh, includes regulatory impact review, regulatory ﬂexibility

analysis and environmental assessment. Gulf of Mexico Fishery Management Council,

Tampa, Florida. 20 p.

Faunce, C. H. and J. E. Serafy. 2006. Mangroves as ﬁsh habitat: 50 years of ﬁeld studies. Mar.

Ecol. Prog. Ser. 318: 1–18.

Halpern, B. S. 2004a. Are mangroves a limiting resource for two coral reef ﬁshes? Mar. Ecol.

Prog. Ser. 272: 93–98.

__________. 2004b. Habitat bottlenecks in stage-structured species: hermit crabs as a model

system. Mar. Ecol. Prog. Ser. 276: 197–207.

Heemstra, P. C. and J. E. Randall. 1993. FAO species catalogue. Vol. 16. Groupers of the world.

(Family Serranidae, Subfamily Epinephelinae). An annotated and illustrated catalogue of

the grouper, rockcod, hind, coral grouper and lyretail species known to date. Food and

Agriculture Organization of the United Nations, Rome.

Johnson, L. 1974. Beyond the fourth generation. University Presses of Florida, Gainesville. 230 p.

Kapetsky, J. M. 1985. Mangroves, ﬁsheries, and aquaculture. Pages 17–36 in FAO Fish. Rep. No.

338 (Suppl.): 87 p. (Doc FIRD/R338 Suppl.).

KOENIG ET AL.: JUVENILE GOLIATH GROUPER HABITAT

585

Koenig, C. C. and F. C. Coleman. 1998. Absolute abundance and survival of juvenile gags in sea

grass beds of the northeastern Gulf of Mexico. Trans. Am. Fish. Soc. 127: 44–55.

Krebs, C. J. 1999. Ecological methodology. Benjamin/Cummings, Menlo Park, California. 620 p.

Kruczynski, W. L. 1999. Water quality concerns in the Florida Keys: sources, eﬀects, and so-

lutions. Report to the FKNMS Water Quality Protection Program. Florida Keys National

Marine Sanctuary, Marathon, Florida. 65 p.

Laegdsgaard, P. and C. Johnson. 2001. Why do juvenile ﬁsh utilize mangrove habitats? J. Exp.

Mar. Biol. Ecol. 257: 229–253.

Loftus, W. F. 2000. Inventory of ﬁshes of Everglades National Park. Fla. Sci. 63: 29–26.

Mendelssohn, I. A. and K. L. Mckee. 2000. Saltmarshes and mangroves. Pages 501–536 in M.

G. Barbour and W. D. Billings, eds. North American terrestrial vegetation, 2nd ed. Cam-

bridge University Press, Cambridge.

Nagelkerken, I., C. M. Roberts, G. van der Velde, M. Dorenbosch, M. C. van Riel, E. Cocheret

del la Morinière, and P. H. Hienhuis. 2002. How important are mangroves and seagrass

beds for coral-reef ﬁsh? e nursery hypothesis tested on an island scale. Mar. Ecol. Prog.

Ser. 244: 299–305.

NMFS (National Marine Fisheries Service). 2004. Annual Report to Congress on the status of

U.S. Fisheries—2003. U.S. Department of Commerce, NOAA, National Marine Fisheries

Service, Silver Spring, Maryland. 24 p.

Ogden, J. C., S. M. Davis, T. K. Barnes, K. J. Jacobs, and J. H. Gentile. 2005. Total system con-

ceptual ecological model. Wetlands 25: 955–979.

Porch, C. E. and A. M. Eklund. 2004. Standardized visual counts of goliath grouper oﬀ south

Florida and their possible use as indices of abundance. Gulf Mexico Sci. 2: 155–163.

Robins, C. R., G. C. Ray, J. Douglass, and R. Freund. 1986. A ﬁeld guide to Atlantic coast ﬁshes

of North America. Houghton Miﬄin, Boston. 354 p.

Roﬀ, D. A. 1992. e evolution of life histories, theory and analysis. Chapman & Hall, New

York. 535 p.

Sadovy, Y. and A. M. Eklund. 1999. Synopsis of biological information on the Nassau grouper,

Epinephelus striatus (Bloch 1792), and the jewﬁsh, E. itajara (Lichtenstein 1822). NOAA

Tech. Rep. NMFS 146. 65 p.

Seber, G. A. F. 1982. e estimation of animal abundance, 2nd ed. Griﬃn, London. 654 p.

Sklar, F. H. and J. A. Browder. 1998. Coastal environmental impacts brought about by altera-

tions to freshwater ﬂow in the Gulf of Mexico. Environ. Manage. 23: 347–358.

ayer, G. W., A. B. Powell, and D. E. Hoss. 1999. Composition of larval, juvenile, and small

adult ﬁshes relative to changes in environmental conditions in Florida Bay. Estuaries 22:

518–533.

orrold, S. R., C. M. Jones, P. K. Swart, and T. E. Targett. 1998. Accurate classiﬁcation of juve-

nile weakﬁsh Cynoscion regalis to estuarine nursery areas based on chemical signatures in

otoliths. Mar. Ecol. Prog. Ser. 173: 253–265.

Valiela, I., J. L. Bowen, and J. K. York. 2001. Mangrove forests: one of the world’s threatened

major tropical environments. BioScience 51: 807–815.

White, G. C. and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press,

New York. 383 p.

A: (C.C.K., F.C.C.) Florida State University Coastal and Marine Laboratory, St. Te-

resa, Florida 32858-2702. (A.M.E., J.S.) NOAA-Fisheries, Southeast Fisheries Science Center,

Division of Protected Resources and Biodiversity, 75 Virginia Beach Drive, Miami, Florida

33149. (J.U.) Department of Geography, Clippinger Lab 122, Ohio University, Athens, Ohio

45701. C A: (C.C.K.): E-mail: <koenig@bio.fsu.edu>.