Griffin 2003 Gm Ex Sci Distribution, Habitat Partitioning, and Abundance of Atlantic Spotted Dolphins, Bottlenose Dolphins, and Loggerhead Sea Turtles on the Eastern Gulf of Mexico

7/28/2019 Griffin 2003 Gm Ex Sci Distribution, Habitat Partitioning, and Abundance of Atlantic Spotted Dolphins, Bottlenose D

1/12

2003 by the Marine Environmental Sciences Consortium of Alabama

Gulf of Mexico Science, 2003(1), pp. 2334

Distribution, Habitat Partitioning, and Abundance of Atlantic SpottedDolphins, Bottlenose Dolphins, and Loggerhead Sea Turtles on the

Eastern Gulf of Mexico Continental Shelf

ROBERT B. GRIFFIN AND NANCY J. GRIFFIN

We surveyed cetaceans and marine turtles from Nov. 1998 to Nov. 2000 along a

series of prescribed transects between Tampa Bay and Charlotte Harbor, Florida,

and between the coast and the 180-m isobath. Vertical profiles of temperature,

salinity, and chlorophyll concentration were collected at 65 stations, and contin-

uous surface data on these variables and transmittance were collected while un-

derway. Habitat partitioning among Atlantic spotted dolphins (Stenella frontalis),

bottlenose dolphins (Tursiops truncatus), and loggerhead sea turtles (Caretta car-

etta) was examined by canonical correspondence analyses of environmental char-

acteristics at sighting locations. Environmental characteristics and primary pro-

ductivity of S. frontalis and T. truncatus habitat on the eastern Gulf of Mexico

continental shelf significantly differed. In shelf waters shallower than 20 m, T.truncatus were the dominant cetacean species, whereas S. frontalis were the most

common shelf species at depths of 20180 m. Environmental preferences of C.

caretta were intermediate between the two dolphin species and showed no appar-

ent relationship with depth. The continental shelf in the eastern Gulf of Mexico

is broad, with distances from coast to slope as great as 200 km. Although S.

frontalis habitat has elsewhere been described as ubiquitous over the shelf, our

data suggest that S. frontalisin the eastern Gulf of Mexico prefer midshelf habitat.

Two delphinid species that predominate onthe Gulf of Mexico continental shelf arethe bottlenose dolphin (Tursiops truncatus) andAtlantic spotted dolphin (Stenella frontalis)(Mills and Rademacher, 1996; Jefferson andSchiro, 1997). Among species of marine tur-tles, the loggerhead sea turtle (Caretta caretta)is the most abundant in the Gulf of Mexico(Henwood, 1987). Research in the Gulf ofMexico has focused primarily on abundance ofthese species, and little work has comparedhabitat-use patterns.

Current population estimates (using aerialsurveys) for T. truncatus in the U.S. Gulf of

Mexico suggest that approximately 50,000 dol-phins live on the outer continental shelf (fromapproximately 9 km seaward of the 18-m iso-bath to the continental slope and from theUnited StatesMexico border to the FloridaKeys) and 17,600 dolphins live in coastal andinner shelf waters (from shore to the outershelf boundary) (Waring et al., 1997). Abun-dance of T. truncatus within 37 km of the Gulfof Mexico coast (estimated using aircraft striptransects) was 16,000 (Mullin et al., 1990).

Population estimates for S. frontalis in the

Gulf of Mexico are incomplete, with an esti-mate of 3,200 dolphins in the northern Gulfof Mexico (from approximately the 200-m iso-bath along the U.S. coast to the seaward extentof the U.S. Exclusive Economic Zone) (Waring

et al., 1997). This is considered a partial stockestimate because continental shelf areas were

generally not covered. Yet, data from 7 yr ofopportunistic effort on the continental shelf inthe northern and eastern Gulf of Mexicoshowed that the primary depth range for S.

frontalis was between 15 and 100 m (Mills andRademacher, 1996), with highest sighting rateseast of the Mississippi River. Beyond the con-tinental shelf, this species is sighted exclusivelyalong the upper continental slope (Mullin andHansen, 1999).

A shipboard survey along the continentalslope in the north-central and western Gulf of

Mexico from the FloridaAlabama border(87.5W) to the TexasMexico border (26.0N)and between the 100- and 2,000-m isobathsfound that habitat partitioning of these twospecies was best explained by bottom depth(Davis et al., 1998, 2002; Baumgartner et al.,2001). Stenella frontalis were consistently foundon the continental shelf and shelf break,

whereas T. truncatus were found primarily indeeper waters along the upper slope. AlthoughT. truncatus are also found on the Gulf of Mex-ico shelf, these surveys were limited to shelf-

break regions and did not examine habitat par-titioning between these species on the conti-nental shelf.

Little is known of sea turtle distributions andabundance in the Gulf of Mexico. Aerial sur-


2/12

24 GULF OF MEXICO SCIENCE, 2003, VOL. 21(1)

Fig. 1. Location of study area. Solid lines represent ECOHAB synoptic survey track line. Abundanceestimates refer to region contained within ECOHAB block (14,400 km2). Conductivitytemperaturedepthstation locations (filled circles) are shown.

veys of a 9,000-km2 area, 50 km south of Mo-bile, AL (Levenson et al., 1992), yielded a com-bined density estimate of 0.01 turtles km2 forthree turtle species (C. caretta; leatherback tur-tle, Dermochelys coriacea; and green turtle, Che-lonia mydas) during Nov. 1991April 1992. Car-etta caretta densities of 0.04 turtles km2 werereported for the northeastern Gulf of Mexico(Mullin and Hoggard, 2000). Satellite sea-sur-face temperature data and aerial survey data

were used to identify an upper (28 C) and low-

er (13.3 C) limit of preferred sea-surface tem-peratures for C. caretta (Coles and Musick,2000). The study suggests that sea turtles arenot randomly distributed geographically butstay within preferred temperature ranges thatare seasonally variable.

Partitioning of habitat between the primaryaquatic tetrapods on the west Florida continen-tal shelf, T. truncatus, S. frontalis, and C. caretta,has not been studied, and S. frontalis and C.caretta population densities have not been ex-amined in this region. We examined habitat

partitioning of T. truncatus and S. frontaliswithreference to physical and biotic oceanographicparameters, testing the hypothesis of minimalhabitat overlap between these species on thecontinental shelf, as found by others on the

continental slope (Davis et al., 1998). Habitatuse by these two closely related taxa was alsocompared with that of C. caretta.

METHODS

We gathered cetacean- and turtle-sightingdata from Nov. 1998 through Nov. 2000.Monthly shipboard oceanographic surveysaboard the R/V Suncoaster(Florida Institute ofOceanography) transected an area of the west

Florida continental shelf bounded by 8284.5W and 2628N (Fig. 1). General surveydesign was generated by the Ecology of Harm-ful Algal Blooms (ECOHAB) research group atthe University of South Florida, St. Petersburg,FL, for purposes of understanding physicaland biological mechanisms underlying bloomsof the toxic dinoflagellate, Karenia brevis. Sur-

veys included a series of repeatable transects,with 79 oceanographic stations, at 9-km inter-vals (Fig. 2). Two cross-shelf transects between10- and 50-m depths, as well as one cross-shelf

transect between 10- and 180-m depth, weresurveyed on a monthly basis throughout thestudy period. Surveys consisted of 34 d of ef-fort per month, covering approximately 100km/d. Surveys were completed each month


3/12

25GRIFFIN AND GRIFFINDOLPHIN AND TURTLE HABITAT AND ABUNDANCE

Fig. 2. Contour of cetacean sighting effort (months surveyed) along track, between Nov. 1998 and Nov.2000.

during the study, with the exception of Julyand Sep. 1999 and Oct. 2000. Other transectssurveyed during a part of the study period in-cluded 1) 10-m isobath coastal transect (Dec.1998June 2001; Nov. 2001); 2) 10- to 50-m-deep diagonal transect (Dec. 1998Aug. 1999;May, Sep., and Dec. 2001); 3) 50-m isobath

(Nov. 1998Nov. 1999; June 2001). During sur-veys, vertical profiles of temperature, salinity,chlorophyll concentration, and transmittance

were collected at oceanographic stations byconductivitytemperaturedepth (CTD) bathy-thermograph (Seabird SPE25 Sealogger).Fluorescence was measured as a proxy for chlo-rophyll using a Chelsea Instruments AQUA-tracka Mk III fluorometer. Continuous under-

way surface data on temperature, salinity, chlo-rophyll concentration, and transmittance werecollected using a Falmouth Scientific Instru-

ment Micro-CTD3 integrated with a SeapointChlorophyll fluorometer manufactured by Sea-point Sensors, Inc. (Kingston, NH), a Wet LabsC-Star transmissometer, and a Seapoint turbid-ity meter, mounted on the port deck in a plas-tic vessel through which near-surface seawater(2 m deep) passed continuously.

During surveys, observers were on watchduring transit between stations (approximately30 min) and then broke from effort for 1520min while data were gathered at oceanograph-ic stations. Surveys were conducted by three

observers, with two observers on effort duringduty rotations. Additional observers permittedduty rotation, enabling additional break time.Two observers maintained a watch from thebow while underway during daylight hours,

scanning with naked eye for the presence ofcetaceans and turtles. Biological and physicaldata within transect segments (9-km effortunit between oceanographic stations) were col-lected by observers to document conditions be-tween oceanographic stations. These data in-cluded observations of surface biological man-

ifestations (e.g., birds, flying fish, schoolingfish, cnidarians), descriptors of sea-state andsighting conditions, and number of cargo, fish-ing, and recreational vessels present. Hand-held binoculars (7 50) were used to sightand identify species when cues or animals werefound. When cetaceans or sea turtles were en-countered, data collected included time andlocation of sighting, bearing and distance toanimals when initially sighted, species, totalgroup size, and number of calves. Bearing wasestimated using a 360 course plotter. Distances

to animals when sighted were estimated by ob-servers with prior training and experience indistance approximation. Estimation skills wereperiodically tested by comparing estimated dis-tances to buoys with distances obtained byships radar. Calves were defined as dolphinshaving 75% the body length of associatedmaternal escort. Species identifications wereassigned by experienced observers. For somesightings, the vessel was diverted from track toallow for species identifications.

Abundances of S. frontalis, T. truncatus, and

C. caretta were estimated using the programDISTANCE (Thomas et al., 1998). Sightingsfrom all months were pooled for these analy-ses. Data were right truncated to exclude thegreatest 5% of perpendicular distances. Detec-


4/12


TABLE 1. Variables used in canonical correspondence analysis of Tursiops truncatus, Stenella frontalis, andhabitat use. Surface values of temperature, salinity, density, chlorophyll, and transmittance at cetacean lo-cations were extracted from the continuous underway surface data set. Water column properties at cetaceanlocations were calculated as means of CTD values at casts bracketing transect segments where sightings were

made.a

1 Surface temperature (C) at sighting location or at midsegment where no sighting was made2 Surface minus bottom temperature (C) in 9-km transect segment associated with sighting location3 Stratified (1) or nonstratified (0) water column defined as the presence or absence of a well-developed

thermocline in given transect segments4 Surface salinity (PSU) at sighting location or at midsegment where no sighting was made5 Mean surface minus bottom salinity (PSU) in 9-km transect segment6 Density (Sigma-T, kg m3) at sighting location or at midsegment where no sighting was made7 Mean surface minus bottom density (Sigma-T) in 9-km transect segment8 Surface transmittance (%) at sighting location or at midsegment where no sighting was made9 Maximum chlorophyll (g liter1) in the water column in 9-km transect segment

10 Surface chlorophyll (g liter1) at sighting location or at midsegment where no sighting was made11 Latitude of sighting

12 Longitude of sighting13 Closest distance of sighting from land (km)14 Depth (m) at sighting location or at midsegment where no sighting was made15 Month16 Year17 Sequential date (day of year, from 1 to 366)18 Cos of sequential date (cos and sine of sequential date analyzed to test for cyclical temporal variation)19 Sine of sequential date

a CTD, conductivitytemperaturedepth.

tion function and group size were estimated

globally by species, and analyses were poststra-tified by sighting-depth ranges: 010 m, 1020m, 2030 m, 3040 m, 4050 m, and 50 m.The 50-m stratum included waters between50 m and 180 m, the maximum depth in thesurvey area. Data were combined in this stra-tum because of relatively low sighting effort inindividual 10-m increments in depth. For S.

frontalis densities, effort within the 0- to 10-mstratum was not used for the density estimatebecause the minimum depth of sighting loca-tions for this species was 16 m. Three models

were tested (i.e., uniform+cosine, half-nor-mal+cosine, and half-normal+hermite polyno-mial), and Akaikes information criterion(Akaike, 1973) was used to select the most par-simonious model for each analysis. Regressionsof observed group size against distance werenot significant at an alpha level of 0.15; hence,mean group sizes were calculated as the meanof observed values.

Relationships of cetacean species and habi-tat use to the physical and biological environ-ment were analyzed by canonical correspon-

dence analyses (CCA) (ter Braak, 1986, 1995;ter Braak and Verdonschot, 1995) using theprogram CANOCO 3.10 (ter Braak, 1992).These analyses have been successful in under-standing cetacean distributions in the eastern

tropical Pacific (Fiedler and Reilly, 1994; Reilly

and Fiedler, 1994). Differences in habitat char-acteristics and temporal use patterns were test-ed by CCA of 19 environmental, spatial, andtemporal variables (Table 1). We included cosand sine transformations of sighting sequentialdates to test for the influence of cyclical annual

variation. Analyses were done by a forward se-lection process to minimize the number of var-iables used in ordination, and variables signif-icantly contributing to explaining species vari-ance (tested by Monte Carlo simulation with999 permutations) were retained. Addition of

variables to the ordination ended when thecontribution of the variable under consider-ation was insignificant (P 0.05).

Canonical correspondence analysis is an ei-genvector ordination technique, which relatescommunity composition to variation in the en-

vironment, using an iterative procedure to di-rectly relate species ordinations to environ-mental variables. In CCA, species are ordinat-ed along synthetic axes that are constrained tobe linear combinations of environmental vari-ables. Axes are generated subject to the restric-

tion that they be uncorrelated with previousaxes. Biplots of species ordinations and envi-ronmental vectors permit direct interpretationof relationships between species distributionsand the environment. In CCA ordination dia-


5/12


grams, species points are plotted at their op-tima locations (center of species curve) alongthe axes, representing a two-dimensional nichecenter. Environmental variables are plotted aseigenvector axes, leading away from the originin the direction of increasing value. Relativelengths of environmental vectors are propor-tional to the importance of the environmental

variable in explaining species distributions.Similarities in direction of environmental vec-tors are related to degree of correlation be-tween environmental parameters.

For these analyses, effort and sighting datagathered where sighting conditions includedBeaufort sea states of 3 were used, andsites were defined as the 9-km transect seg-ments between oceanographic stations. Sight-

ing data were weighted in these analyses by nat-ural logarithms of the group size estimateswithin each sighting to minimize the effect oferrors in the estimates of the size of largergroups and to reduce the relative influence oflarger groups on these analyses. Althoughgroup size in delphinids may reflect availabilityof food sources, additional factors that are notrelated to suitability of habitat may influencegroup size (e.g., aggregation for mating, per-ceived risk of predation, or age and sex ofgroup members) (Evans, 1987).

Community ordination diagrams were con-structed using CCA results to relate cetaceanand turtle distributions to physical and biolog-ical variables making significant contributions.Species scores, or ordination coordinates, werecalculated as weighted mean sample scores inall tests. Interspecies ordination distances ap-proximate their chi-square distances when thisscaling is used. KruskalWallis test and theMannWhitney U-test (Sokal and Rohlf, 1981)

were used to test for differences in axes scoresbetween cetacean species and to examine dif-

ferences in species means of physical and bio-logical variables associated with species distri-butions.

RESULTS

Monthly sighting effort (Fig. 2) within thestudy area varied as a function of daylightlength and scientific operations aboard the ves-sel. The three cross-shelf transects were consis-tently surveyed for cetaceans, whereas the di-agonal transect received the least attention.

Over 7,000 km of survey effort was completedin the study area during the 2-yr period, with267 on-effort dolphin sightings [119 S. frontalissightings, 663 dolphins; 113 T. truncatus sight-ings, 316 dolphins; one rough-toothed dolphin

(Steno bredanensis) sighting, seven dolphins; 34unidentified dolphin sightings, 94 dolphins]for an overall sighting rate of 0.154 dolphinskm1. Mean (SD, median) group size was 2.8(2.27, 2) for T. truncatus and 5.6 (5.29, 4) forS. frontalis. Group sizes of S. frontalis rangedfrom 1 to 48 dolphins, whereas T. truncatusgroup sizes ranged from 1 to 15 dolphins. Ap-proximately 81% of S. frontalis groups sightedapproached the vessel to bow-ride, compared

with 42% of T. truncatus groups. This differ-ence was highly significant (chi-square test; 2

46.49, P 0.001). Three species of marineturtles were sighted, including 36 C. caretta,three D. coriacea, and one Kemps Ridley (Lep-idochelys kempi), along with 21 turtles not iden-tified to species.

Stenella frontalis sightings tended to be indeeper waters farther from the coast (Fig. 3)compared with T. truncatus sightings, whereasC. caretta were more often seen at mediandepths and distances. The minimum depth forS. frontalis sightings was 16 m, with only eightsightings at depths 20 m, whereas T. truncatusand C. carettawere found throughout the studyarea. Mean (SD, median) sighting depths forthe two dolphin species were 40 m (19.1, 37m) and 23 m (16.1, 13 m), respectively, where-as mean (SD, median) distances from coast

were 71 km (36.0, 68 km) and 37 km (39.3, 17km), respectively. Mean (SD, median) C. carettasighting depth was 30 m (17.5, 30), and meandistance from land was 55 km (42.6, 54 km).

Using Akaikes information criterion, thehalf-normal+cosine model was selected forabundance estimates of S. frontalis and T. trun-catus, whereas the uniform+cosine method wasselected for C. caretta estimates. The effectivestrip width (ESW) of S. frontalis was 202 m,compared with an ESW of 168 m for T. trun-catus. Pooled data showed an abundance of

3,703 S. frontalis(2,6355,202, 95% ConfidenceInterval (CI)) and 1,346 T. truncatus (9591,889, 95% CI) in the study area. Overall den-sity of S. frontalis was 0.260 dolphins km2,

whereas overall T. truncatusdensity throughoutall depth strata was 0.093 dolphins km2. Den-sity estimates stratified by sighting depth (Fig.4) indicate a primary depth range of 2050 mfor S. frontalis in this region, whereas T. trun-catus are more likely to be sighted from thecoast to 20-m depth.

Effective strip width for C. caretta was 182 m.

Estimated abundance of C. caretta within thestudy area was 181 (114286, 95% CI), with anoverall density of 0.013 turtles km2. No rela-tionship was apparent between C. caretta sight-ing densities and depth strata (Fig. 4).


6/12


Fig. 3. Group sightings per kilometer in transect segments during study period.

Canonical correspondence analyses.Of 19 physi-cal and biological variables used in CCA, fourmade significant contributions to explaining

variance in cetacean habitat characteristics:

transmittance, surface temperature, surface sa-linity, and difference between surface and bot-tom salinity. Correlation values (Table 2) sug-gest that canonical axis 1 represented variationin transmittance and surface minus bottom sa-

linity, whereas canonical axis 2 represented var-iation in surface temperature. Variation in sur-face salinity contributed to both axes. For S.

frontalisand T. truncatus, CCA explained 29%

and 25% of the species variation (Table 3),respectively, whereas 27% of C. caretta varia-tion was explained. Axis 1 was more importantin explaining S. frontalis variation. Axis 2 wasmore important in explaining characteristics


7/12


Fig. 4. Estimated densities (animals per square kilometer) ofStenella frontalis(Sf), Tursiops truncatus(Tt),and Caretta caretta (Cc) by depth stratum (m) for 2-yr pooled data.

TABLE 2. Correlations of canonical axes with signif-icantly (P 0.05) contributing variable (n 605).

Variable Axis 1 Axis 2

SB salinityaTemperatureSalinityTransmittance

0.3750.014

0.1210.154

0.0270.3850.1640.050

a SB indicates surface minus bottom.

TABLE 3. Percentage of variation explained by ca-nonical axes, by species.

Species

Axes

1 2 Total

Stenella frontalis

Tursiops truncatus

Caretta caretta

27.6013.85

9.50

1.1011.6317.18

28.7025.4826.68

of C. caretta habitat, whereas both axes wereimportant in explaining T. truncatus habitat.

Speciesenvironment ordination biplots in-dicated environmental similarities and differ-ences in optimal habitat (represented by axeslocation) of the three species. Axis 1 (Fig. 5)separated the S. frontalis habitat characteristicsfrom those ofT. truncatus and C. caretta, where-as axis 2 separated the C. carettaenvironmentalconditions from the two dolphin species habi-tats. Canonical ordination suggests that S. fron-

talis are likely to be found in waters with greatersurface salinity, lower or negative surface mi-nus bottom salinity values, and greater trans-mittance (corresponding with lower chloro-phyll values) compared with C. caretta or T.truncatus. Caretta caretta are more likely foundin warmer waters than S. frontalis or T. trunca-tus.

Plotting of weighted species CCA standarddeviations along each axis (Fig. 6) provided ameasure of niche breadth (Carnes and Slade,1982) and permitted an examination of niche

separation between species, as described by en-vironmental characteristics. Standard deviationellipses about estimated optimum environ-ments overlap for all species combinations butdo not coincide. Although variation in environ-mental conditions at sighting locations was

large, MannWhitney U-test shows that thesespecies significantly differed in location in ca-nonical space (Table 4). Stenella frontalisand T.truncatus ordinations significantly differedalong axis 1, which was most important in ex-plaining variation between species. Axis 1 (sa-linity and transmittance) separated S. frontalisfrom T. truncatus and C. caretta, whereas axis 2(temperature and salinity) separated C. carettafrom the two dolphin species.

Mean values of many of the environmental

variables tested by CCA significantly differed byspecies (Tables 5, 6), providing further evi-dence of differences in habitat conditions in-dicated by canonical ordination. Transmit-tance of light through water was greater andchlorophyll content was lower in waters whereS. frontalis were found than in waters where T.truncatus were sighted. Tursiops truncatus weresighted in water with significantly less salinityand smaller water column temperature gradi-ent compared with C. caretta and S. frontalis.

Caretta caretta were sighted in waters with a mi-nor water column salinity gradient. All speciesdiffered in mean sighting depth and mean dis-tance from shore, with C. caretta intermediatebetween the two dolphin species.


8/12


Fig. 5. Ordination for model with abundance logarithmically transformed. All variables significantly

contributed (P 0.05). Arrows point in direction of variable increase, and crosses represent variable grandmeans. Species ordinations: Sf, Stenella frontalis; Tt, Tursiops truncatus; Cc, Caretta caretta. S B Sal, mean

value for surface salinity bottom salinity; Trans, surface transmittance; Salinity, surface salinity; Temp,surface temperature.

DISCUSSION

We found that densities of S. frontalis(0.260dolphins km2) in the eastern Gulf of Mexico

were greater than densities of T. truncatus(0.093 dolphins km2). Aerial surveys in the

northeastern Gulf of Mexico (Mullin and Hog-gard, 2000) reported a greater density of T.truncatus (0.148 dolphins km2) on that areaof the shelf (waters 100 m in depth) and alower density of S. frontalis (0.089 dolphinskm2). Differences in survey methodologymake comparisons of our results with earlier

work difficult. It is not known whether the ap-parent dissimilarity among studies on relativedensity of these two species between the east-ern and northeastern Gulf of Mexico is an ar-tifact of methodology or represent true region-

al differences. Observed differences betweenthe two regions suggest ecological variation be-tween broad-shelf habitat in the eastern Gulfof Mexico and narrow-shelf habitat in thenorth.

The importance of S. frontalis habitat ofgreater than 20-m depth agrees with earlierfindings indicating that S. frontalis principallyoccupy waters 15100 m in depth (Mills andRademacher, 1996). In that study, S. frontalisdistribution on the entire Gulf of Mexico con-tinental shelf was examined using opportunis-tic data gathered from various National MarineFisheries Service resource surveys.

Because C. caretta spend 90% of their timesubmerged during any given season (Renaudand Carpenter, 1994), with average submer-gence times as great as 171 min., abundancesfor this species are probably underestimated.In addition, unidentified turtles that could po-tentially increase C. caretta density estimates

were not included in these analyses. Mean sea-

surface temperature (26.3 C) associated withour C. caretta sightings was in agreement withmean sea-surface temperature reported else-

where for C. caretta distributions (13.328 C;Coles and Musick, 2000).


9/12


Fig. 6. Ellipses of uncertainty (95% CI) aboutspecies ordinations on the first and second canoni-cal axes, taken from canonical correspondence anal-

ysis of environmental data. Ellipses are 1 SD aboutthe estimated optimal location for each species onthe first and second canonical axes. Sf, Stenella fron-talis; Tt, Tursiops truncatus; Cc, Caretta caretta.

TABLE 4. P values for MannWhitney U-test, com-paring canonical axes scores by species.

Species

Stenella frontalis

Axis 1 Axis 2

Tursiopes truncatus

Axis 1 Axis 2

T. truncatus

Axis 1Axis 2Axis 1Axis 2

0.003

0.0010.14

0.0050.15

0.001

Some assumptions of line transect theorywere violated in this study. It is not likely thatall animals on track line were seen. Further, itis likely that dolphins were often aware of the

vessels approach before they were detected.Some initial sightings were made while dol-phins approached the vessel, which may re-duce calculated ESW, and lead to an inflatedabundance estimate (Turnock and Quinn,1991). Although some bow-riding groups mayhave initially been on the track line, a higherproportion of S. frontalis bow-riders suggests

that S. frontalis may be more likely to approachthe vessel than T. truncatus, potentially leadingto an artificial increase in relative abundanceof this species.

The greater number ofS. frontalis(663) thanT. truncatus(316) seen by observers during thisstudy may reflect relative densities of dolphinspecies. This could also have resulted fromgreater visibility and the differential attractionof S. frontalis to the research vessel. Work hasshown that these two species show 0% avoid-ance reaction toward ships (Wursig et al.,

1998); yet, no work has been done to examinerelative detectability of these two species as afunction of response to vessel. Stenella frontalisapproaching the vessel to bow-ride tended todisplay exhibitory behaviors (pers. obs.)

(e.g., porpoising, leaping, splashing, andbreaches), whereas T. truncatus seldom dis-played these behaviors. Such behaviors may en-able observer detection of groups at a greater

relative distance. The greater ESW reported inthis study for S. frontalis supports this hypoth-esis of early detection for this species. Al-though abundance estimates reported in thisstudy may be positively biased, they can be use-ful for detection of seasonal and interannualtrends within species.

The four variables significantly contributingto CCA represent parameters that reflect near-shore vs offshore regions (e.g., greater salinityand blue water at greater distances from thecoast). The eastern Gulf of Mexico exhibits en-

vironmental variability between nearshore andoffshore waters, with consistent differences inprimary productivity, temperature, and salinity.Nearshore chlorophyll concentrations are rel-atively high, and chlorophyll concentrationsrapidly decline beyond 10 km from the coast.Nearshore waters are often well mixed, where-as offshore waters may be thermally stratified.Greater transmittance with distance from thecoast, as in S. frontalis optimum habitat, resultsfrom lower primary productivity in offshore

waters. High gradients in surface to bottom sa-

linity can result from 1) less mixing in the wa-ter column, 2) input of higher-salinity waterfrom offshore regions, or 3) high freshwateroutflow from estuaries such as Tampa Bay andCharlotte Harbor.

Salinity and transmittance of water (a proxyfor primary production) were important in de-scribing variation in species habitat use andmay reflect differences in water masses and as-sociated productivity. Salinity is a conservativecharacteristic, useful for identification of watermasses. Salinity levels in the region are elevat-

ed by intrusion of Loop Current filaments,whereas freshwater flow from coastal bays andestuaries results in a relatively strong salinitygradient of fresher water. Thermal fronts wereoften located at boundaries between well-


10/12


11/12


TABLE6.

Means(SD)ofbioticandabioticvariablesassociatedwithdolphina

ndturtlesightings.

Stenellafrontalis

n

Tursiopstruncatus

n

Carettacaretta

n

Depth(m)

Distance(km

)

Temperature

(C)

Salinity(psu)

Sigma-Ta

Chlorophyll

(gliter1)

Transmittance(%)

SBbtemper

ature(C)

SBbsalinity

(psu)

SBbdensity

(kgm3)

40.7

(

18.14)

76.7

(

35.63)

24.6

(

3.15)

35.850(

0.5369)

24.097(

0.09540)

0.241(

0.1912)

39.1

(

26.33)

1.55(

2.100)

0.189(

0.3

751)

0.596(

0.7111)

151

151

131

131

131

131

131

129

129

129

17.7

(16.0

7)

26.4

(38.86)

24.0

(3.72)

34.885(1.7629)

23.512(1.1543)

0.639(0.8

718)

31.5

(24.51)

1.06

(2.099)

0.128(0.3035)

0.389(0.5259)

188188135135135135132919191

29.5

(17.38)

54.9

(42.50)

26.3

(3.1

7)

35.696(0.8433)

23.507(0.8441)

0.342(0.3654)

34.6

(25.98)

1.68

(2.582)

0.003(0.2069)

0.518(0.7437)

40

40

33

33

33

33

31

32

32

32

aSigma-T

de

nsity(kgm3)minus1000.

bSBindicates

surfaceminusbottom.

in part by an anonymous donation to Mote Ma-rine Laboratory, in support of the postdoctoral

work of RBG.

LITERATURE CITED

AKAIKE, H. 1973. Information theory and an exten-sion of the maximum likelihood principle, p. 267281. In: International symposium on informationtheory. 2d ed. B. N. Petran and F. Csaaki (eds.).

Akadeecemiai Kiaki, Budapest, Hungary.BAUMGARTNER, M. F., K. D. MULLIN, L. N. MAY, AND

T. D. LEMING. 2001. Cetacean habitats in thenorthern Gulf of Mexico. Fish. Bull. 99:219239.

BERO, D. 2001. Population structure of the Atlanticspotted dolphin (Stenella frontalis) in the Gulf ofMexico and western North Atlantic. M.S. thesis.Univ. of Charleston, Charleston, SC.

CARNES, B. A., AND N. A. SLADE. 1982. Some com-ments on niche analysis in canonical space. Ecol-ogy 63:888893.

COLES, W. C., AND J. A. MUSICK. 2000. Satellite seasurface temperature analysis and correlation withsea turtle distribution off North Carolina. Copeia2000:551554.

DAVIS, R. W., G. S. FARGION, N. MAY, T. D. LEMING,M. BAUMGARTNER, W. E. EVANS, L. J. HANSEN, ANDK. MULLIN. 1998. Physical habitat of cetaceansalong the continental slope in the north-centraland western Gulf of Mexico. Mar. Mamm. Sci. 14:490507.

, J. G. ORTEGA-ORTIZ, C. A. RIBIC, W. W. EVANS,D. C. BIGGS, P. H. RESSLER, R. B. CADY, R. R. LEBEN,K. D. MULLIN, AND B. WURSIG. 2002. Cetacean hab-itat in the northern oceanic Gulf of Mexico. Deep-Sea Res. Part I 49:121142.

ESHER, R. J., C. LEVENSON, AND T. DRUMMER. 1992.Aerial surveys of endangered and protected spe-cies in the Empress II ship trial operating area inthe Gulf of Mexico. Naval Research LaboratoryNRL/MR/717492-7002. Stennis Space Center,MS. 55 p.

EVANS, P. G. H. 1987. The natural history of whalesand dolphins. Facts on File, New York.

EVANS, W. E., B. WU

RSIG, J. ORTEGA-ORTIZ, AND R. W.DAVIS. 2000. Cetacean habitat associations in thenorthern Gulf of Mexico: an overview, p. 289292.In: Proceedings: eighteenth annual Gulf of Mexi-co information transfer meeting; Dec. 1998. U.S.Department of the Interior, Minerals ManagementService, New Orleans, LA.

FIEDLER, P. C., AND S. B. REILLY. 1994. Interannualvariability of dolphin habitats in the eastern trop-ical Pacific. II: effects on abundances estimatedfrom tuna vessel sightings, 19751990. Fish. Bull.92:451463.

GRIFFIN, R. B. 1999. Sperm whale distributions and

community ecology associated with a warm-corering of Georges Bank. Mar. Mamm. Sci. 15:3351.HENWOOD, T. A. 1987. Sea turtles of the southeastern

United States, with emphasis on the life historyand population dynamics of the turtle, Caretta car-etta. Ph.D. diss., Auburn Univ., Auburn, AL.


12/12


JEFFERSON, T. A., AND A. J. SCHIRO. 1997. Distributionof cetaceans in the offshore Gulf of Mexico.Mamm. Rev. 27:2750.

MILLS, L. R., AND K. R. RADEMACHER. 1996. Atlanticspotted dolphins (Stenella frontalis) in the Gulf ofMexico. Gulf Mex. Sci. 14:114120.

MULLIN, K. D., and L. J. HANSEN, 1999. Marine mam-mals of the northern Gulf of Mexico, p. 269277.In: The Gulf of Mexico large marine ecosystem:assessment, sustainability and management. H.Kumpf, K. Steidinger, and K. Sherman (eds.).Blackwell Science, Inc. Malden, MA.

MULLIN, K. D., AND W. HOGGARD. 2000. Visual surveysof cetaceans and sea turtles from aircraft andships, p. 111171. In: Cetaceans, sea turtles andseabirds in the northern Gulf of Mexico: Distri-bution, abundance and habitat associations. Vol-ume II: Technical Report. R. W. Davis, W. E.Evans, and B. Wursig (eds.). Prepared by Texas

A&M Univ. at Galveston and the National MarineFisheries Service. U.S. Department of the Interior,Geological Survey, Biological Resources Division,USGS/BRD/CR-1999-0006 and Minerals Manage-ment Service, Gulf of Mexico OCS Region, NewOrleans, LA. OCS Study MMS 2000-003.

, R. R. LOHOEFENER, W. HOGGARD, C. L. RO-DEN, AND C. M. ROGERS. 1990. Abundance of bot-tlenose dolphins, Tursiops truncatus, in the coastalGulf of Mexico. Northeast Gulf Sci. 11:113122.

REILLY, S. B., AND P. C. FIEDLER. 1994. Interannualvariability of dolphin habitats in the eastern trop-ical Pacific. I: research vessel survey, 19861990.

Fish. Bull. 92:434450.RENAUD, M. L., AND J. A. CARPENTER. 1994. Move-ments and submergence patterns of turtles (Car-etta caretta) in the Gulf of Mexico determinedthrough satellite telemetry. Bull. Mar. Sci. 55:115.

ROBERTS, H. H., R. A. MCBRIDE, AND J. M. COLEMAN.1999. Outer shelf and slope geology of the Gulfof Mexico: an overview, p. 93112. In: The Gulf ofMexico large marine ecosystem: assessment, sus-tainability, and management. H. Kumpf, K. Stei-dinger, and K. Sherman (eds.). Blackwell Science,Malden, MA.

SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. W. H.Freeman and Company, San Francisco, CA.

TER BRAAK, C. J. F. 1986. Canonical correspondenceanalysis: a new eigenvector technique for multi-

variate direct gradient analysis. Ecology 67:11671179.

. 1992. CANOCOa FORTRAN program forCanonical Community Ordination. Microcomput-er Power, Ithaca, NY.

. 1995. Ordination, p. 91173. In: Data anal-ysis in community and landscape ecology. R. H. G.Jongman, C. J. F. ter Braak, and O. F. R. van Ton-geren (eds.). Univ. of Cambridge Press, Cam-bridge, U.K.

, AND P. F. M. VERDONSCHOT. 1995. Canonicalcorrespondence analysis and related multivariatemethods in aquatic ecology. Aquat. Sci. 57:255289.

THOMAS, L., J. L. LAAKE, J. F. DERRY, S. T. BUCKLAND,

D. L. BORCHERS, D. R. ANDERSON, K. P. BURNHAM,S. STRINDBERG, S . L . HEDLEY, M . L . BURT, F.MARQUES, J. H. POLLARD, AND R. M. FEWSTER. 1998.Distance 3.5. Research Unit for Wildlife Popula-tion Assessment. Univ. of St. Andrews, St. Andrews,Scotland.

TURNOCK, B. J., AND T. J. QUINN II. 1991. The effectof responsive movement on abundance estimationusing line transect sampling. Biometrics 47:701715.

WARING, G. T., D. L. PALKA, K. D. MULLIN, J. H. W.HAIN, L. J. HANSEN, AND K. D. BISACK. 1997. U.S.

Atlantic and Gulf of Mexico marine mammal stock

assessments1996. NOAA Technical Memoran-dum NMFS-NE-114. Woods Hole, MA.WURSIG, B., S. K. LYNN, T. A. JEFFERSON, AND K. D.

MULLIN. 1998. Behaviour of cetaceans in thenorthern Gulf of Mexico relative to survey shipsand aircraft. Aquat. Mamm. 24:110.

OFFSHORE CETACEAN ECOLOGY PROGRAM, MOTEMARINE LABORATORY, CENTER FOR MARINEMAMMAL AND SEA TURTLE RESEARCH, 1600KEN THOMPSON PARKWAY, SARASOTA, FLORIDA34236. Date accepted: February 20, 2003.

Documents

Griffin 2003 Gm Ex Sci Distribution, Habitat Partitioning, and Abundance of Atlantic Spotted Dolphins, Bottlenose Dolphins, and Loggerhead Sea Turtles on the Eastern Gulf of Mexico