Regional Stock Structure of Greater Amberjack in the Southeastern United States Using Otolith Shape Analysis

MISSING IMAGE

Material Information

Title:
Regional Stock Structure of Greater Amberjack in the Southeastern United States Using Otolith Shape Analysis
Physical Description:
1 online resource (61 p.)
Language:
english
Creator:
Campbell, Chelsey A
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Fisheries and Aquatic Sciences, Forest Resources and Conservation
Committee Chair:
Parkyn, Daryl C
Committee Members:
Murie, Debra J
Devries, Douglas

Subjects

Subjects / Keywords:
amberjack -- dumerili -- gulf -- otolith -- seriola -- shape -- stock -- structure
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre:
Fisheries and Aquatic Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Otolith shape analysis was used to examine the current management hypothesis that greater amberjack (Seriola dumerili) in the Gulf of Mexico belong to a single stock. Shape of the sagittae was quantified using a combination of the shape indices rectangularity, ellipticity, roundness, and form factor, along with elliptical Fourier analysis for 444 otoliths collected from Louisiana, North Florida, and Central Florida. A smaller quantity of Atlantic stock otoliths from fish collected north of the Florida/Georgia border was included to test the validity of the technique for distinguishing Atlantic stock fish from Gulf of Mexico greater amberjack.   No significant differences were detected between left and right otoliths or between male and female otoliths from amberjack in the Gulf of Mexico. However, sex-based differences were significant in the otoliths from Atlantic stock amberjack. Principal Component Analysis showed evidence of grouping in otolith shape between the Gulf and Atlantic stocks, although there was overlap in shape between the two regions; the Gulf showed less clear signs of regional grouping, with no differences in shape visible between North and South Florida. Discriminant analysis had a 75% classification success rate between otoliths from the Gulf of Mexico and Atlantic stock. Similarly, a 75% classification rate was attained for fish collected in Louisiana, while only a 25% and 40% classification success rate was attained for otoliths from North and Central Florida respectively. This suggests that while some differentiation between Florida and Louisiana is present, overall the analysis supports the current one-stock management of greater amberjack within the Gulf of Mexico.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Chelsey A Campbell.
Thesis:
Thesis (M.S.)--University of Florida, 2012.
Local:
Adviser: Parkyn, Daryl C.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-06-30

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044788:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 REGIONAL STOCK STRUCTURE OF GREATER AMBERJACK IN THE SOUTHEASTERN UNITED STATES USING OTOLITH SHAPE ANALYSIS By CHELSEY ADELE CAMPBELL A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Chelsey Adele Campbell

PAGE 3

3 To my parents, who always enco uraged me to pursue my passions

PAGE 4

4 ACKNOWLEDGMENTS I would like to express my deepest gratitude to my committee chair, Dr. Daryl Parkyn, as well as to the other members of my supervisory committee, Dr. Debra Murie and Dr. Doug DeVries, for your help, support and guidance throughout this process. I would also like to thank all the members of the Murie/Parkyn lab for your help and advice. I would like to a cknowledge the NOAA Panama City Laboratory, the Gulf States Fisheries Commission the NOAA Beaufort City Laboratory, and the Murie/Parkyn lab for collecting and providing the otolith samples used in this research. Also many thanks to Fabien Morat of the Un iversit de la Mditerrane for assistance in interpreting the SHAPE software outputs. Finally, I thank the School of Forest Resources and through a m atching graduate assistantship, and the MARFIN program for providing funding.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ 4 LIST OF TABLES ................................ ................................ ................................ ........... 7 LIST OF FIGURES ................................ ................................ ................................ ........ 9 ABSTRACT ................................ ................................ ................................ .................. 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ... 12 Fisheries Background ................................ ................................ ............................ 12 Stock Identification ................................ ................................ ................................ 13 Introduction to the Stock Concept ................................ ................................ ... 13 Otolith Shape An alysis ................................ ................................ .................... 14 Greater Amberjack ................................ ................................ ................................ 15 Background ................................ ................................ ................................ ..... 15 GAJ Fishery ................................ ................................ ................................ .... 16 GAJ Stock Assessment ................................ ................................ ................... 18 GAJ Stock Structure ................................ ................................ ........................ 19 Study Objectives ................................ ................................ ................................ .... 21 2 METHODS ................................ ................................ ................................ ............ 23 Sample Collection ................................ ................................ ................................ .. 23 Otolith Morphology ................................ ................................ ................................ 23 Rostrum Exclusion ................................ ................................ .......................... 23 Shape Quantification ................................ ................................ ....................... 24 Shape indices ................................ ................................ ........................... 24 Fourier analysis ................................ ................................ ........................ 25 Data Analysis ................................ ................................ ................................ ........ 26 Comparison of Left and Right Otoliths ................................ ................................ ... 27 Comparison of Male and Female Otoliths ................................ .............................. 28 Regional Comparisons ................................ ................................ .......................... 28 3 RESULTS ................................ ................................ ................................ .............. 34 Sample Collection ................................ ................................ ................................ .. 34 Data Analy sis ................................ ................................ ................................ ........ 34 Comparison of Left and Right Otoliths ................................ ................................ ... 34 Comparison of Male and Female Otoliths ................................ .............................. 35 Regional Comparisons ................................ ................................ .......................... 35 Data Exp loration ................................ ................................ .............................. 35

PAGE 6

6 Analysis of Variance ................................ ................................ ........................ 35 Discriminant Analysis ................................ ................................ ...................... 36 4 DISCUSSION ................................ ................................ ................................ ........ 48 LIST OF REFERENCES ................................ ................................ .............................. 53 BIOGRAPHICAL SKETCH ................................ ................................ ........................... 59

PAGE 7

7 LIST OF TABLES Table page 2 1 Shape indices of greater amberjack ( Seriola dumerili ) sagittae calculated following Tuset et al. (2003), with ML corresponding to maximum length and MH to maximum height of the greatest enclosing rectangle. .............................. 31 3 1 Otolith sample sizes of greater amberjack ( Seriola dumerili ) sagittae by region used in the present study. ................................ ................................ ....... 37 3 2 Analysis of Covariance (ANCOVA) of shape indices of greater amberjack ( Seriola dumerili ) sagittae, with forklength as the covariate and region as a factor. ................................ ................................ ................................ ................ 37 3 3 moment correlation coefficients resulting from analysis of shape indices of greater amberjack ( Seriola dumerili ) sagittae. ......................... 37 3 4 Paired t test comparing shape indices of greater amberjack ( Seriola dumerili ) sagittae between left and right otoliths from the same individual. ...................... 37 3 5 Comparison of male and female greater amberjack ( Seriola dumerili ) otolith shape indices in the Gulf of Mexico stock using Analysis of Variance (ANOVA). ................................ ................................ ................................ .......... 38 3 6 Comparison of male and female greater amberjack ( Seriola dumerili ) otolith shape indices in the Atlantic stock using Analysis of Variance (ANOVA). .......... 38 3 7 Eigenvalues from principal component analysis (PCA) comparing otolith shape between Gulf and Atlantic stocks of greater amberjack ( Seriola dumerili ) ................................ ................................ ................................ ............ 39 3 8 Eigenvalues from principal component analysis (PCA) comparing otolith shape among central Florida, north Florida, and Louisiana samples of greater amberjack ( Seriola dumerili ) ................................ ................................ .............. 40 3 9 Comparison of Atlantic and Gulf of Mexico stock greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance (ANOVA). ................ 41 3 10 Comparison of central Florida, north Florida, and Louisiana greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance (ANOVA). ................................ ................................ ................................ .......... 41 3 11 Comparison of central Florida, north Florida, and Louisiana age 3 greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance ( ANOVA). ................................ ................................ ................................ .......... 41

PAGE 8

8 3 12 Discriminant analysis comparing otolith shape indices between Gulf and Atlantic samples of greater amberjack ( Seriola dumerili ). ................................ .. 42 3 13 Discriminant analysis comparing greater amberjack ( Seriola dumerili ) otolith shape among three regions in the Gulf of Mexico, with CF= Central Florida, NF=North Florida, and LA=Louisiana. ................................ ............................... 42

PAGE 9

9 LIST OF FIGURES Figure page 1 1. Trends in commercial catch of greater amberjack, Seriola dumerili over time in round (whole) weight by state in the Gulf of Mexico (data from SEDAR 2011). ................................ ................................ ................................ ................ 22 1 2. Commercial and recreational landings by weight for Gulf of Mexico and Atlantic greater amberjack, Seriola dumerili, fisheries (data from SEDAR 2008, 2011), with inclusion of Gulf of Mexico management regimes. ................ 22 2 1. Gulf of Mexico regions comp ared in this study, as delineated by the dotted lines.. ................................ ................................ ................................ ................. 32 2 2. Whole left sagitta otolith of a greater amberjack, Seriola dumerili ..................... 32 2 3. The maximum height (MH) and maximum length (ML) of the greatest enclosing rectangle, excluding the rostrum, wer e measured in each greater amberjack ( Seriola dumerili ) otolith; area and perimeter of the otolith posterior to the rostrum were also calculated. ................................ ................... 33 2 4. Maximum ventral length (MVL) of greater amberjack ( Seriola dumerili) otoliths across ages and forklengths. ................................ ................................ 33 3 1. Forklength (mm) of greater amberjack ( Seriola dumerili ) samples organized by collection region. ................................ ................................ ........................... 43 3 2. Age in years of Gulf of Mexico greater amberjack ( Seriola dumerili ) samples organized by collection region. ................................ ................................ .......... 44 3 3. Power analysis on elliptical Fourier descriptors showed that 13 harmonics described 99% of the shape of greater amberjack ( Seriola dumerili ) otoliths. .... 44 3 4. Values of shape indices of greater amberjack ( Seriola dumerili ) otoliths plotted across forklength show no relationship between size of fish and shape in dex value. ................................ ................................ ............................ 45 3 5. Two dimensional Principal Component Analysis projection comparing otolith shape between males from the Gulf and Atlantic stocks of greater amberjack (Seriola dumerili), with maximum convex polygons enclosing the regions. ........ 46 3 6. Two dimensi onal Principal Component Analysis projection comparing otolith shape of greater amberjack ( Seriola dumerili ) among regions in the Gulf of Mexico, with maximum convex polygons enclosing the regions. ....................... 47

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science REGIONAL STOCK STRUCTURE OF G REATER AMBERJACK IN THE SOUTHEASTERN UNITED STATES USING OTOLITH SHAPE ANALYSIS By Chelsey Adele Campbell December 2012 Chair: Daryl C. Parkyn Major: Fisheries and Aquatic Sciences Otolith shape analysis was used to examine the current management hypothesis that greater amberjack ( Seriola dumerili ) in the Gulf of Mexico belong to a single stock. Shape of the sagittae was quantified using a combination of the shape indices rectangularity, ellipticity, roundness, and form factor, a long with elliptical Fourier analysis for 379 otoliths collected from Louisiana, North Florida, and Central Florida. A smaller quantity (n=69) of Atlantic stock otoliths from fish collected north of the Florida/Georgia border was included to test the valid ity of the technique for distinguishing Atlantic stock fish from Gulf of Mexico greater amberjack. No significant differences were detected between left and right otoliths or between male and female otoliths from greater amberjack in the Gulf of Mexico or the Atlantic Principal Component Analysis showed evidence of grouping in otolith shape between the Gulf and Atlantic stocks, although there was overlap in shape between the two regions; the Gulf showed less clear signs of regional grouping, with no diffe rences in shape visible between North and Central Florida Discriminant analysis had a 70 % classification success rate between otoliths from the Gulf of Me xico and Atlantic stock. In addition, the shape indices form factor and circularity were found to dif fer significantly

PAGE 11

11 between the stocks. Similarly, a 75% classification rate was attained for fish collected in Louisiana, while only a 25% and 40% classification success rate was attained for otoliths from North and Central Florida respectively. No signific ant differences were seen in shape indices amongst the three regions; however, when a subset of age 3 fish was tested, significant differences in the shape index rectangularity were present between the Louisiana and Florida samples. This suggests that some differentiation between Florida and Louisiana is present, but overall the analysis supports the current one stock management of greater amberjack within the Gulf of Mexico.

PAGE 12

12 CHAPTER 1 INTRODUCTION Fisheries Background The demand for marine fish is growing, with the annual global catch of fish more than tripling in the past 50 years (Pauly et al. 2002; Walters and Ahrens 2009). The fisheries are either full y or over exploited (Pauly et al. 2002; Hilborn et al. 2003; Worm et al. 2006) and an estimated 14 29% of fish stocks had collapsed by the early 2000 s (Mullon et al. 2005; Worm et al. 2006; Worm et al. 2009). Management rebuilding efforts have yet to reverse the ov erall trend of depletion in many stocks, and fisheries management has often been unsuccessful (Cunningham and Bostock 2005; Beddington et al. 2007; Worm et al. 2009) This results from challenges faced by management, including the difficulties associated w ith balancing a variety of stakeholder interests (Lackey 2005; Beddington et al. 2007; Hilborn 2007) with the biological limitations intrinsic to a species. Notably, insufficient biological knowledge (e.g., fecundity, sex ratio, age at first reproduction, and frequency of reproduction) has made it difficult for management to determine appropriate regulations for some species (Beddington et al. 2007, Smith 2011). Management of a fishery requires a good scientific understanding of the biology and behavior of a species (Beddington et al. 2007); therefore it is important to future management to improve our scientific knowledge. One area fundamental to proper management of fisheries is the appropriate delineation of fish stocks (MacLean and Evans 1981; Begg and W aldman 1999).

PAGE 13

13 Stock Identification Introduction to the Stock Concept The concept of stock, which originated in the field of animal husbandry, was first solidified as a basis for fisheries management at the end of the 19 th century, when A.C. Anderson and l ater John Peace Babcock noted that the salmon stocks in British Columbia rivers were discrete and varied in abundance from river to river (McDonald 1981). Anderson and Babcock recognized the relationship between abundance of spawners and the number of recr uits produced in each stock, and concluded that catch should be regulated differently among rivers (McDonald 1981; Ebbin 1996). In this way, the stock concept became established very early as a basis for managing salmon fisheries (McDonald 1981). Over tim e, numerous definitions have arisen to capture the essence of the stock concept, varying from delineations based on harvest to genetically based distinctions (Ihssen et al. 1981; MacLean and Evans 1981; Carvalho and Hauser 1994; Ebbin 1996; Coyle 1997; Boo temporal, spatial, genetic, and/or morphological differences separating them from each other (Ebbin 1996). Stocks are typically reproductively isolated, often with differing patterns of growth and recruitment, and respond independently to exploitation (McDonald 1981; Carvalho and Hauser 1994; Coyle 1997). This can lead to differential status of stocks within a species, with some stocks sustainably fished and others considered overfish ed or even collapsed. For example, the Atlantic stock of greater amberjack, Seriola dumerili has recovered from overfishing and is currently considered sustainable, while the Gulf of Mexico stock has been classified as overfished and undergoing overfishin g despite an ongoing rebuilding effort (SEDAR 2008, 2011).

PAGE 14

14 Numerous techniques have been employed to identify fish stocks, ranging from simple qualitative observations (such as the differential timing of spawning runs) to highly tech nical quantitative tec hniques (lh ssen et al. 1981; MacLean and Evans 1981; Ebbin 1996; Coyle 1997; Begg and Waldman 1999). In a review of methodologies, Begg and Waldman (1999) list catch data, tag recoveries, meristics, morphometrics, scale morphology, parasites, cytogenetics, protein electrophoresis, immunogenetics, mitochondrial (mtDNA) and nuclear DNA, heart tissue fatty acids, otolith elemental composition, osteological interdigitation features, stable isotope measurements, and thermal marking as some of the techniques empl oyed to define stock structure in fisheries species. An additional tool recently found useful in stock delineation is otolith shape analysis (Campana and Casselman 1993). Otolith Shape Analysis calcium carbonate stru ctures found in the otic capsules of bony fish es that are involved in hearing and balance. Most f ish have three sets of paired otoliths: the sagittae, lapilli and asteriscii. Otoliths are thought to grow continuously throughout the life of a fish in propo and are likely not subject to resor p tion ( Ricker 1975; Campana and Thorrold 2001 ). Otolith shape is typically species specific and is influenced by both genetic and environmental parameters (Campana and Casselman 1993 ; C ardinale et al. 2004; Vignon and Morat 2010 ). Because of these influences, o tolith shape has been observed to differ between stocks of the same species, including king mackerel, Scomberomorus cavalla (DeVries et al. 2002; Shep ard et al. 2010), Pacific sard ine, Sardinops sagax (Felix Uraga et al. 2005), haddock, Melanogrammus aeglefinus (Begg et al. 2001 ), sole, Solea solea (Merigot et al. 2007), Atlantic saury, Scomberesox saurus saurus (Aguera

PAGE 15

15 and Brophy 2011), common coral trout, Plectropomus leopardus (B ergenius et al. 2006) yellowstripe goatfish, Mulliodichthys flavolineatus (Pothin et al. 2006), monkfish Lophius piscatorius (Canas et al. 2012), mulloway, Argyrosomus japonicus (Ferguson et al. 2011), and Atlantic cod, Gadus morhua (Campana and Casselma n 1993; Jonsdottir et al. 2006; Petursdottir et al. 2006). Otolith shape is less variable than fish growth, and the otolith typical of sensory structures, remains unaffected by short term changes in fish condition such as starvation that might confound bo dy morphometrics (Campana and Casselman 1993 ; Pankhurst and Montgomery 1994 ). While studies of otolith shape cannot distinguish between environmental and genetic influences, the contributing differences in these factors are likely to influence otolith shap e among populations that remain at least partially segregated (Campana and Casselman 1993). Greater Amberjack Background G reater amberjack are widely distributed in marine tropical and subtropical waters. Individuals can attain lengths greater than 1500 mm forklength making them the largest members of the jack family (Carangidae). In the Western Atlantic t he species ranges along the coast from Massachusetts to southeastern Brazil, including the Gul f of Mexico (Robins et al. 1986; McEach ern and Fechhelm 2005). Greater amberjack (GAJ) are predatory, consuming invertebrates (including shrimp, crabs and squid) a s well as ray finned fishes of variou s species including members of the Clupeidae, Bothidae, and Sparidae (Burch 1979 ; Manooch and Haimovici 1983 ). I n the Gulf of Mexico, GAJ are currently managed as two distinct stocks with one stock along the Southeast Atlantic coast, including the Florid a Keys and the other residing in the Gulf of Mexico. Variation in mtDNA and microsatellite DNA is consistent

PAGE 16

16 with the two stock hypothesis (Gold and Richardson 1998 ; Murie et al. 2011 ). Tagging studies also support t he current two stock management, with M c C lellan and Cummings (1997) reporting an exchange rate from the Atlantic to the Gulf of Mexico of only 1.3%, with a 1.6% exchange rate from the Gulf of Mexico to the Atlantic. In the Gulf of Mexico, GAJ are assumed to spawn offshore, and pelagic Sargassu m mats have been observed to provide nursery habitat for young of year GAJ (Wells and Rooker 2004). As adults, GAJ are typically found schooling around reefs or other structure (such as wrecks and oil rigs) in 10 100 m of water. Because of this, GAJ are co nsidered part of the Reef Fish Management Unit, which includes species such as red snapper Lutjanus campechanus and gag grouper Mycteroperca microlepis Currently little is known about GAJ movement in the Gulf, although it has been noted that they are often only seasonally abundant in certain parts of their range and may associate with a variety of different habitats or areas each year (SEDAR 2011). From a fisheri es perspective, these seasonal movements likely affect local abundance and access to the resource, especially by recreational fishers due to added costs of fuel and distance. GAJ Fishery Greater amberjack are fished both commercially and recreationally; they are a popular sport fish (likely due to their aggressive fighting behavior) and important recreational fisheries for this spec ies have existed since the 1950 s (Cummings and McClellan 2000). Ho s that the species became co mmercially popular, possibly spurred by the decline of other commercial species such as red snapper and red drum Sciaenops ocellatus (Cummings and McClellan 2000).

PAGE 17

17 Gulf of Mexico greater amberjack are currently assessed as both overfished and undergoing ov erfishing (SEDAR 2011). The commercial fishery for greater amberjack utilizes vertical handlines (bandit gear or electric reels), with a small proportion of catch coming from long lining and spearfishing (SEDAR 2006). Commercial landings for this species i n the Gulf are highest in Florida, followed by Louisiana and Texas (SEDAR 2006) (Figure 1 1). Commercial fishery data for GAJ in the Gulf begins in the 1960s for Florida and in the 1980 s for all other states (SEDAR 2011). Catch of GAJ in the Gulf exceeds t hat of the Atlantic (Figure 1 2 ). Commercial GAJ landings in the Gulf grew from between five to eight thousand pounds ( whole weight) in the early 1960 s to a peak of 2,055,639 pounds in 1988 (SEDAR 2011). After the 1988 peak, landings fell stabilizing arou nd 1.2 million pounds between 1994 and 1997 (SEDAR 2011). The decline in landings after 1991 corresponds to the implementation of management regulations (Cummings and McClellan 2000, Fig. 1 2). Following 1997, landings declined again and have remained belo w one million pounds since, reaching a low of 504,114 pounds in 2008 (SEDAR 2011). Recreational angling methods for GAJ include hook and line and spearfishing. The recreational fishery for GAJ (and other reef species) has high economic value in the Gulf o f Mexico. For example, in 2008 charter boats in the reef fish fishery earned approximately $88,000/vessel and headboats approximately $461,000/vessel (Rauch 2011); while there are no data on how many of these vessels target GAJ specifically, overall the la rgest proportion of recreational GAJ catch comes equally from charterboats and private anglers (SEDAR 2011).

PAGE 18

18 GAJ Stock Assessment The underlying principle in modern fisheries management is that every fish population produces a surplus, and that the large st surplus that can be harvested annually from that population (known as Maximum Sustainable Yield, or MSY) can be estimated through scientific analysis (i.e. stock assessments) (Ricker 1975; Lackey pressure and population biomass (F msy and B msy respectively), which wi ll maintain MSY (Beddington et al. 2007; Gabriel and Mace 1999). When fishing pressure (F) exceeds F msy (F current /F msy >1) the stock is considered to be unde rgoing overfishing, and is assessed as overfished when the biomass (B) is less than B msy (B current /B msy <1). In the 1990 s, the Gulf of Mexico Fishery Management Council received anecdotal information from eastern Gulf of Mexico fishermen that the GAJ stock was in decline, notable due to a decrease average size and abundance. I Southeast Fisheries Science Center (SEFSC) conducted a stock assessment in 1996 (Hood 2006). Though this assessment was deemed too imprecise to specify allowable cat ch limits, there was concern that the stock was in decline and so more stringent management regulations were implemented (Hood 2006). In 2000, the stock was reassessed, and the resulting stock assessment models showed an overfished condition (as of 1998 ) and suggested that overfishing had also occurred (Hood 2006). In Mexico Fishery Management Council tha t the stock had been overfished and, in response, the Council devel oped a 7 year rebuilding plan beginning in 2003 (SEDAR 2011). In 2006, a stock assessment was conducted through the Southeast Data Assessment and Review (SEDAR) process which found the stock to continue to be both

PAGE 19

19 overfished and undergoing overfishing as of 2004, estimating F 2004 /F msy at 1.017 and B 2004 /B msy at 0.706 (SEDAR 2006). The most recent assessment update (SEDAR 2011) estimated F 2009 /F msy at 1.83, with B 2009 /B msy at 0.31, indicating continued overexploitation. In addition, backwards projections u sing the updated models estimated conditions to have been worse in 2004 than estimated by the assessment, with F 2004 /F msy at 2.4 and B 2004 /B msy at 0.38. The models estimate d that the stock has been overfished continuously, reaching a low in 1997 with a B 19 97 /B msy of 0.28. In addition, it was estimated that overfishing has occurred since 1986, with F 1986 /F msy of 1.77, with the exception of 1990 (F 1990 /F msy =0.89). Furthermore, models based on current rates of exploitation predict that the stock will not rec over, nor will overfishing end, within the time frame of the rebuilding plan (SEDAR 2011). GAJ Stock Structure GAJ within the Gulf of Mexico are currently managed as one continuous stock, with presumed mixing across the entire region. Genetic analysis has yielded evidence of continuous gene flow in the northern Gulf, with no signs of regional differentiation (Gold and Richardson 1998; Murie et al. 2011). However, tagging studies in the Gulf have found that many individual GAJ exhibit little net movement bet ween release and recapture, suggesting that mixing across the region may not be continuous. For example, amberjack tagged off of a Louisiana oil platform for age validation in an age and growth study were observed to remain on site for at least 9 months af ter tagging (Beasley 1993). On a larger scale, McClellan and Cummings (1997) found 30% of fish tagged and recaptured in the Gulf showed zero net movement, with 58% of recaptures made within 25 nautical miles of the release site. Of the fish that moved, 90. 6% were

PAGE 20

20 recaptured within 100 nautical miles of the release site. Recapture times ranged from 1 6 years (1.2 on average), with 98% recaptured within 2 years. McClellan and Cummings (1997) also saw a negative relationship between movement and time at large, with the highest movement exhibited by fish at large for a short period of time. In addition, Ingram and Patterson (2001) found that 97% of GAJ tagged off of Pensacola were recaptured within the original release area (though no information was given on ti me at large), and found that fish tagged off of Panama City traveled an average distance of 10.8 km (with a mean time at large of 200 days). Furthermore, a study currently underway by Murie et al. (2011) has found an average distance traveled of 69.54 km f or tagged GAJ, with a median distance of 8.0 km, and no relationship between distance traveled, time at large, or size of fish. Tagging studies have found that some individual GAJ do undergo large scale movements in the Gulf. For example, Murie et al. (20 11) found one fish tagged in Apalachicola, FL was recaptured 11 months later in Tampico, Mexico, with another tagged near Madeira Beach, FL recaptured 10 months later near Port Maria, Jamaica. In addition, Ingram and Patterson (2001) saw one fish tagged of f of Panama City Beach, FL recaptured 396 days later off of Port Fourchain, LA. This suggests at least some degree of mixing in the region, which may make it difficult to elucidate regional structure using genetic data. Studies have shown that only a few m igrants per generation are sufficient to prevent development of genetic differentiation (Allendorf 1983; Carvalho and Hauser 1994; Coyle 1997). Therefore, a failure to detect genetic differentiation does not mean that functionally no separation exists (Coy le 1997).

PAGE 21

21 Study Objectives GAJ in the Gulf are both overfished and undergoing overfishing, and have failed to recover despite a rebuilding plan and increasingly stringent management regulations. A failure to accurately grasp the stock structure of GAJ may contribute to the ongoing decline in this species in the Gulf. If the Gulf of Mexico stock is not completely mixed, the disproportionate fishing effort could lead to localized overfishing of the species, particularly off the coast of Florida. In delineat ing stock structure, it is important to use a holistic approach that incorporates results from a variety of methodologies (Begg and Waldman 1999). The present study examines the utility of otolith shape analysis in regional comparisons of GAJ in the Gulf o f Mexico, and between the Gulf and the Atlantic stocks of greater amberjack. In conjunction with tagging and genetic data, this stock identification method may help form a holistic picture of the stock structure of GAJ in the southeastern United States Th is should aid in future management of the species in this region and assist in rebuilding efforts for the Gulf of Mexico stock by allowing managers to accurately understand the structure of the stock

PAGE 22

22 Figure 1 1. Trends in commercial catch of greater amberjack, Seriola dumerili over time in round (whole) weight by state in the Gulf of Mexico (data from SEDAR 2011). Figure 1 2. Commercial and recreational landings by weight for Gulf of Mexico and Atlantic greater amberjack, Seriola dumerili fisheries (data from SEDAR 2008, 2011), with inclusion of Gulf of Mexico management regimes 0 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1960 1970 1980 1990 2000 2010 Whole weight (pounds ) Year Texas Louisiana Mississippi/Alabama West Florida 28" rec, 3 fish daily limit 36" comm 1 fish daily limit comm March/May closure 30" rec 36" comm rec June/July closure TAC overage closure TAC overage closure

PAGE 23

23 CHAPTER 2 METHODS Sample Collection Th e present study used otoliths collected from port sampling and scientific sampling by the University of Florida, NOAA Fisheries and regional state agencies including the Gulf States Marine Fisheries Council ( GSMFC ) Otoliths utilized were limited to those collected between 20 02 and 2008 Samples were restricted t o fish between 700 105 0 mm forklength (FL) to limit possible size based variation and to standardize fish sam ples across regions. Sample regions included central Florida (Madeira Beach to Sarasota), North Florida (Apalachicola to Suwannee), and central and western Louisiana (Figure 2 1 ). In addition as otolith shape comparisons have never been conducted for GAJ, otoliths from the Atlantic stock were included for a separate shape comparison between stocks Atlantic stock otoliths were provided by NOAA Fisheries, and were restricted to otoliths collected north of the Florida/Georgia border to limit the chance that otoliths came from Gulf migrants. The study focused on saggitae, which have been seen to be the most informative for morphological studies (Campana and Casselman 1993) and have been utilized in other studies of this nature (Begg et al. 2001; DeVries et al. 2002; Felix Uraga et al. 2005 ; Bergenius et al. 2006; Jonsdottir et al. 2006 ; Petursdottir et al. 2006 ; Pothin et al. 2006 ; Merigot et al. 2007 ; Shep ard et al. 2010 ; Aguera and Brophy 2011; Ferguson et al. 2011; Canas et al. 2012). Otolith M orphology Rostrum E xclusion G reater amberjack otoliths are fragile and easily broken during removal, i n addition to being small in size relative to the b ody size of the fish (Figure 2 2 ) As a result of this, a

PAGE 24

24 large proportion of otoliths available to be used in th e present study were broken or had chipped rostra. Otolith shape was therefore standardized to exclude the rostral portion Following DeVries et al. (2002 ), a straight line was drawn from the dorsal surface of the antirostrum to the ventral edge of the oto lith (Figure 2 3 ), and the rostral portion anterior to this line was excluded from further analysis. Without this methodology, sample size would have been greatly reduced. Shape Q uantification Although t here are many ways to quantify otolith shape a combi nation of shape indices developed by Tuset et al. (2003) and elliptical Fourier analysis was chosen for this study as recommended by Tracey et al. (2006). O tolith images were captured and digitized using Motic Images (v 3.0) software (Motic Group North Am erica, Vancouver, Canada) on a Leica MZ50 dissecting scope using a Panasonic WV CP224 CCD Microscope Camera Linear measurements were taken using the Motic software and included maximum height (MH) and maximum length (ML) of the greatest enclosing rectangle of the posterior po rtion of the otolith (Figure 2 3 ), as well as area and perimeter of the posterior portion of the otolith; these measures allowed for the ca lculation of five shape indices. Shape indices Shape ind ic es were calculated following Tu set et al. (2003). Shape indices serve to quantify general shape characteristics and have been useful in other studies of otolith shape ( Ferguson et al. 2011 ; Jonsdottir et al. 2006 ; Pothin et al. 2006 ; Merigot et al. 2007 ; Shepard et al. 2010 ). The shape indices used in the present study include form factor, roundness, circularity, rectangularity and ellipticity (Table 2 1 ). The index f orm factor estimates edge irregularity, with a value of 1 representing a perfectly smooth

PAGE 25

25 edge and values < 1 when the edge is irregular. Roundness and circularity describe the similarity of certain features to a perfect circle, with respective minimum values of 1 and Rectangularity quantifies the variations in length and width with respe c t to area, with a value of 1 representing a perfect square. Finally, ellipticity examines whether the changes in the axes lengths are proportional. Fourier analysis Among the many classes of Fourie r analysis, Elliptical Fourier A nalysis (EFA) is considered the most powerful for otolith shape analysis (Tracey et al. 2006; Merigot et al. 2007). EFA describes the outline of a shape using a number of representative variables, termed harmonics. These harmonics are each characterized by four coefficients, or elli ptical Fourier descriptors, a b c and d which correspond to the projection of the outline on the x ( a and b ) and y ( c and d ) axes (Merigot et al. 2007; Canas et al. 2012). Elliptical Fourier analysis was conducted using the Shape program (Iwata and Ukai 2002). Shape inputs the digitized image and calculates the Fourier coefficients, then normalizes them based on Kuhl and G iardin a ( 1982) to correct for di fferences in size and orientation. A greater number of harmonics increases the accuracy of shape outline; however, too large a number can overcomplicate analyses. Fourier power analysis was therefore calculated to determine the appropriate number of harmo nics using E quation 2 1 PF n = 0.5( a 2 n + b 2 n + c 2 n + d 2 n ) (2 1) where PF n = power of the Fourier harmonic with a n b n c n and d n refer ring to the a, b, c and d coefficients of the n th harmonic ( Pothin et al. 2006; Merigot et al. 2007 ). The

PAGE 26

26 cumulative power percentage was then calculated using the sum of the previous PF n Th e goal was to reach a threshold cumulative power percentage of 99%; after this, little information would be added by additional harmonics ( Pothin et al. 2006; Merig ot et al. 2007 ) Power analysis was run on a randomly selected subsample of 30 otoliths, to determine a cumulative power of 99% Data Analysis Statistical analyses were run using the SAS and JMP software (SAS Institute Inc. 2008 ). All statistical tests were conducted at the = 0.05 criterion level unless otherwise stated Prior to analyses, the shape indices and Fourier coefficients were examined for agreement with statistical assumptions of normality and homoscedasticity using the Kolmogorov Smirnov test (Zar 1999), respectively Initial data exploration indicated a number of outliers which caused the data to vary significantly from normal. Each outlier was checked against its sample image, and it was determined that the outliers corresponded to deformed otoliths (i.e. otoliths with jagged, irregular posterior portions). A total of 18 outliers were therefore removed from the dataset for subsequent analyses. Following outlier removal, parameters still found to vary from normal included the shape index circularity and the harmonics A5, A6, A8, B2, and D8. Circularity was normalized using the square root transformation, and the harmonics were normalized using the log e transformation (Zar 1999) Following transformatio n, all parameters conformed to assumptions of normality and homogeneity of variance. Otoliths grow over the life of a fish, and it is poss i ble that shape varies with fish size Therefore, Analysis of Covariance (ANCOVA) was used to examine the effect of fi sh

PAGE 27

27 size (forklength) on each shape index, with sampling region included as a factor and fish length the covariate (Bergenius et al. 2006; Jonsdottir et al. 2006 ; Petursdottir et al. 2006 ). Fish length was chosen over age as a covariate because i nitial data exploration determined that otolith growth has a stronger relationship with fish length than fish age (Figure 2 4 ). An Analysis of covariance (ANCOVA) was used to examine if the shape indices varied with length Many of the shape indices were constructed u sing different combinations of the same parameters, and correlation was susp ected amongst the shape indices ; therefore moment correlation coefficient s w ere calculated as a precursor to analysis A ll indices wer e retained for further analy ses, but correlations were taken into account when choosing analysis methodology. Comparison of Left and Right Otoliths Often, only a single left or right otolith was available in the otolith collection. Consequently, a paired t test was used to compare le ft and right otoliths when available from the same individual to look for differences that might bias analyses Most species previously examined including Atlantic c od, Gadus morhua haddock, Melanogrammus aeglefinus saithe Pollachius virens, golden redf ish, Sebastes marinus, Atlantic herring, Clupea harengus, and Atlantic mackerel, Scomber scombrus, appear ed to show no significant statistical differences between left and right otoliths (Hunt 1992; Petursduttir et al. 2006) In contrast, common sole, Solea solea, (Merigot et al. 2007) have left and right otoliths that differ significantly within an individua l, likely a consequence of their side oriented benthic existence. A p aired t test was used to compare shape indices between left and right otoliths of 25 male GAJ individuals from the same region (Louisiana) to explore this possible source for error in shape analysis

PAGE 28

28 Comparison of Male and Female Otoliths Sex based differences in growth rates between male and female GAJ have been documented in the Atlantic (Harris et al. 2007); differences in growth rates of GAJ in the Gulf of Mexico are less pronounced (Murie and Parkyn 2008) but may still exist. Studies have found correlations between differences in growth rate and differences in otolith shape, a nd sexual dimorphism in otolith shape has been observed in other species (e.g., cod and haddock) (Campana and Casselman 1993; Begg et al. 2001). Similarly it is possible that male and female GAJ otoliths exhibit morphological differences. Shape indices of male and female GAJ sagittae collected from the same region (Centr al Florida) were compared using analysis of variance (ANOVA), with a Bonferroni correction for repeated testing, to look for sex specific differences in GAJ otolith shape. Individual ANOVAs were chosen over multipl e analysis of variance (MANOVA) as MANOVA is known to work best with moderately correlated data ( Salkind 2010 ) and most of the indices were found to have low or high correlations. The Bonferroni adjustment gave a significance crite rion of = 0.01. Regional Comparisons C omparisons of otolith shape, as quantified by shape indices and elliptical Fourier analysis, between the Gulf and Atlantic stocks and among three regions in the Gulf of Mexico (Central Florida, Northern Florida and L ouisiana) were first explored descriptively using Principal Component Analysis (PCA) (PC ORD v.6.0). PCA does not require a priori assumptions about group membership, and presents an unbiased indication of separation between regions (Begg et al. 2001; Merigot et al. 2007 ). The

PAGE 29

29 cross products matrix for the PCA was calculated using the variance/covariance method. Next, shape indices were compared between the Gulf and Atlantic stocks of GAJ and among the three Gulf regions using multiple individual analyses of variance (ANOVAs) to test for specific shape differences. ANOVA was corrected for multiple testing using t, giving a significance criterion of = 0.01. Though samples were restricted to a size range, they still contained fish of variable ages. Therefore, a complimentary analysis was run on a subset of the data. Shape indices of age 3 fish were compared among the three Gulf of Mexico regions using ANOVA, again corrected = 0.01. Analysis on the data subset was restricted to ANOVA and excluded from the other multivariate analyses due to the small sam ple size of age 3 fish. Linear Discriminant Analysis (DA) was then used as an a posteriori test to examine if otolith shape could distinguish among regions Discriminant analysis investigate s the integrity of pre defined groups (Pothin et al. 2006; Merigo t et al. 2007), and has been employed in several recent studies of otolith morpholo gy (Begg et al. 2001; DeVries et al. 2002 ; Felix Uraga et al. 2005 ; Merigot et al. 2007 ; Petursdottir et al. 2006; Pothin et al. 2006 ). Like PCA, discriminant analysis allow s for the comparison of multiple parameters across groups simultaneously; however, DA requires the sample size of the smallest group to exceed the number of predictor variables by a value of at least n 2 (with n = the sample size of the smallest group), though a greater discrepancy is thought to be preferable ( Meyers et al. 2006 ). Shape indices and Fourier analysis combined give a total of 53 parameters to describe GAJ otolith shape. This is

PAGE 30

30 sufficiently smaller than the sample size for each of the Gulf r egions; however, it is too large to allow appropriate testing of the Atlantic stock samples. Therefore, DA was run on shape indices alone when comparing the Gulf and Atlantic stock otoliths, but all parameters (shape indices and Fourier descriptors combine d ) were included for comparison among the three Gulf regions. Prior to DA, samples were randomly split into a model data set and a test data set. Discriminant analysis was then run on the model data sets, and the resulting discriminant functions were used to test the ability of otolith shape to predi region of origin. Performance of the DA statistic ( Fleiss 1981 ) which compares the discriminatory power of the analysis to what might be expected by random chanc e alone.

PAGE 31

31 Table 2 1. Shape indices of greater amberjack ( Seriola dumerili ) sagittae calculated following Tuset et al. (2003), with ML corresponding to maximum length and MH to maximum height of the greatest enclosing rectangle. Shape index Equation Form factor 2 Roundness 2 Circularity Perimeter 2 /Area Rectangularity Area/(ML*MH) Ellipticity (ML MH)/(ML + MH)

PAGE 32

32 Figure 2 1 Gulf of Mexico regions compared in this study, as delineated by the dotted lines. Region abbreviations correspond to CF: Central Florida (Madeira Beach to Sarasota), NF: North Florida (Apalachicola to Cedar Key, FL) and LA: Louisiana. Figure 2 2 Whole left sagitta otolith of a greater amberjack, Seriola dumerili LA NF CF Rostrum Antirostrum Sulcus Dorsal Ventral

PAGE 33

33 Figure 2 3 The maximum height (MH) and maximum length (ML) of the greatest enclosing rectangle, excluding the rostrum, were measured in each greater amberjack ( Seriola dumerili ) otolith; area and perimeter of the otolith posterior to the rostrum were also calculated. Figure 2 4 Maximum ventral length (MVL) of greater amberjack ( Seriola dumerili) otoliths across ages and forklengths. y = 1.19x + 6.72 r = 0.60 4 6 8 10 12 14 0 2 4 6 MVL (mm) Age (years) y = 0.01x + 3.32 r = 0.86 4 6 8 10 12 14 200 700 1200 1700 MVL (mm) Forklength (mm) MH ML

PAGE 34

34 CHAPTER 3 RESULTS Sample Collection In total, 455 samples were analyz ed using shape analysis (Table 3 1 ). A ll Gulf regio ns had greater than 100 samples; t he Atlantic stock had f ewer samples, with a tot al of 69 analyzed for the region. Final size and age frequencies across regions are depicte d in Figures 3 1 and 3 2; most fish were around 3 years of age. Data Analysis Power analysis determined that the a priori criterion level of 99% was achieved within 13 harmonic calculations (Figure 3 3 ), indicating that GAJ otolith shape could be summarized by 13 harmonics (or 52 Fourier coefficients). However, the first harmonic was excluded from the analyses because the outline reconstructed by the first coefficients is a simple ellipse (Merigot et al. 2007). Ther efore, a final of 12 harmonics, and thus a total of 48 Fourier coefficients were retained for each otolith. The ANCOVA demonstrated that none of the shape indices varied signifi cantly with fish length (Table 3 2). In addition, shape indices as a function of fish length sh owed no visible trends (Figure 3 4 ), therefore all indices were retained for further analyses. All shape ind ices were correlated, with mo st correlations low; however, circularity and form factor were highly correlated (0.99), as were ellip tici ty and roundness (0.89) (Table 3 3). Comparison of Left and Right Otoliths The paired t test found no significant differences in shape indices between le ft and right otoliths (Tabl e 3 4 ). Therefore, when the right otolith was absent, the mi rror image o f the left (a digital manipulation using the Motic imaging software) was used in the analysis.

PAGE 35

35 Comparison of Male and Female Otoliths The ANOVAs showed no significant differences in any shape index between male and female otoliths in the Gulf of Mexico (T able 3 5 ) or Atlantic (Table 3 6 ) stocks of GAJ. Therefore, the sexes were pooled for subsequent analyses. Regional Comparisons Data E xploration Principal component analysis (PCA) showed that 98 % of the variance could be explained by the first two principal component ax es (Tables 3 7 and 3 8 ) therefore the results were examined on a two dimensional plane. Principal components analysis showed evidence of regional grouping of otolith shape between the Gulf and Atl antic otoliths (Figure 3 5 ) indicati ng differences in otolith shape between these two stocks, although there was overlap in otolith shape between the two regions. PCA projections comparing otolith shape among Gulf regions showed less clea r signs of grouping (Figure 3 6 ), with north and centr al Florida displaying a high degree of overlap S ome differences in otolith shape were evident between Louisiana and the two Florida regions, although grouping was less clear than between the Atlantic and Gulf stocks. Analysis of Variance Analysis of Varia nce (ANOVA) found significant differences in the shape indices form factor and circularity between the Gulf and Atlantic stocks of GAJ (p values of 0.007 and 0.0012 respectively) (Table 3 9 ) No significant differences were seen among the three Gulf region s (Table 3 10) However, when limited to the age 3 fish data subset, significant differences were seen in the shape index rectangularity between the two Florida regions and the Louisiana region ( Table 3 11 ).

PAGE 36

36 Discriminant Analysis The Gulf regions model data set consisted of 270 randomly selected samples (90 per region), with a test data set of 60 randomly selected samples (20 from each region); for comparison between the Atlantic and Gulf, the model data set consisted of 98 randoml y select ed samples (49 from each stock) with a test data set of 54 randomly selected samples (27 per stock) Discriminant analysis (DA) between the Gulf and Atlantic stocks of GAJ conducted on shape indices alone showed a 70 % classification success, sugge sting that shape differences are present in otoliths from GAJ in these two regions (Table 3 12 ). value of = 0.40 indicated a 40% improvement over random chance. Within the Gulf, however otolith shape was less able to correctly predicted region, with only a 47% classification success among the three regions overall predicting only a 20% improvement over random chance (Table 3 13 ). DA had the highest success assigning otoliths from Louisiana (75% classification success) but did a poor job assigning otoliths from North Florida (40% classification success) and Central Florida (25% classification success).

PAGE 37

37 Table 3 1. Otolith sample sizes of greater amberjack ( Seriola dumerili ) sagittae by region used in the pres ent study. Region n Central Florida 143 North Florida 115 Louisiana 121 Atlantic 69 Table 3 2. Analysis of Covariance (ANCOVA) of shape indices of greater amberjack ( Seriola dumerili ) sagittae, with forklength as the covariate and region as a factor. Parameter df p value F n Form factor 1 0.4691 0.53 442 Roundness 1 0.7944 0.07 442 Circularity 1 0.4471 0.58 442 Rectangularity 1 0.2294 1.46 442 Ellipticity 1 0.6719 0.18 442 Table 3 moment correlation coefficients resulting from analysis of shape indices of greater amberjack ( Seriola dumerili ) sagittae. Index Roundness Circularity Rectangularity Ellipticity Form factor 0.32 0.99 0.20 0.24 Roundness 0.32 0.40 0.89 Circularity 0.21 0.24 Rectangularity 0.06 Table 3 4 Paired t test comparing shape indices of greater amberjack ( Seriola dumerili ) sagittae between left and right otoliths from the same individual. Index df t statistic p value n Form f actor 24 2.064 0.617 25 Circularity 24 2.064 0.525 25 Roundness 24 2.064 0.986 25 Rectangularity 24 2.069 0.227 25 Ellipticity 24 2.055 0.393 25

PAGE 38

38 Table 3 5 Comparison of male and female greater amberjack ( Seriola dumerili ) otolith shape indices in the Gulf of Mexico stock using Analysis of Variance (ANOVA). Index df F p value n Form f actor 1 0.0325 0.8573 80 Roundness 1 5.1945 0.0255 80 Circularity 1 0.0188 0.8914 80 Rectangularity 1 0.7295 0.3957 80 Ellipticity 1 3.9309 0.051 80 Table 3 6 Comparison of male and female greater amberjack ( Seriola dumerili ) otolith shape indices in the Atlantic stock using Analysis of Variance (ANOVA). Index df F p value n Form f actor 1 2.9686 0.0911 52 Roundness 1 0.1868 0.6674 52 Circularity 1 3.4152 0.0705 52 Rectangularity 1 1.4023 0.2419 52 Ellipticity 1 0.0027 0.9588 52

PAGE 39

39 Table 3 7 Eigenvalues from principal component analysis (PCA) comparing otolith shape between Gulf and Atlantic stocks of greater amberjack ( Seriola dumerili ) Number Eigenvalue Percent Cum ulative Percent 1 12406886.1624 92.0623 92.0623 2 744966.7353 5.5278 97.5901 3 304486.5123 2.2594 99.8495 4 6447.4660 0.0478 99.8974 5 3679.5648 0.0273 99.9247 6 3091.2776 0.0229 99.9476 7 2313.6054 0.0172 99.9648 8 983.8244 0.0073 99.9721 9 927.5728 0.0069 99.9789 10 598.0848 0.0044 99.9834 11 457.2442 0.0034 99.9868 12 342.2807 0.0025 99.9893 13 246.8228 0.0018 99.9911 14 186.0984 0.0014 99.9925 15 135.0716 0.0010 99.9935 16 123.9071 0.0009 99.9945 17 110.1558 0.0008 99.9953 18 106.3880 0.0008 99.9961 19 99.3859 0.0007 99.9968 20 62.6120 0.0005 99.9973 21 57.8768 0.0004 99.9977 22 49.2333 0.0004 99.9981 23 47.0896 0.0003 99.9984 24 37.9093 0.0003 99.9987 25 30.5891 0.0002 99.9989 26 25.0799 0.0002 99.9991 27 20.9755 0.0002 99.9993 28 19.5557 0.0001 99.9994 29 16.5765 0.0001 99.9995 30 11.9904 0.0001 99.9996

PAGE 40

40 Table 3 8 Eigenvalues from principal component analysis (PCA) comparing otolith shape among central Florida, north Florida, and Louisiana samples of greater amberjack ( Seriola dumerili ) PCA axis Eigenvalue Percent Cum ulative Percent 1 1467.6461 92.5836 92.5836 2 80.6105 5.0852 97.6687 3 34.4548 2.1735 99.8422 4 0.7838 0.0494 99.8917 5 0.5184 0.0327 99.9244 6 0.3714 0.0234 99.9478 7 0.1652 0.0104 99.9582 8 0.1338 0.0084 99.9667 9 0.1090 0.0069 99.9736 10 0.0881 0.0056 99.9791 11 0.0549 0.0035 99.9826 12 0.0482 0.0030 99.9856 13 0.0340 0.0021 99.9878 14 0.0265 0.0017 99.9894 15 0.0220 0.0014 99.9908 16 0.0197 0.0012 99.9921 17 0.0171 0.0011 99.9931 18 0.0154 0.0010 99.9941 19 0.0130 0.0008 99.9949 20 0.0119 0.0007 99.9957 21 0.0087 0.0005 99.9962 22 0.0085 0.0005 99.9968 23 0.0072 0.0005 99.9972 24 0.0053 0.0003 99.9976 25 0.0051 0.0003 99.9979 26 0.0047 0.0003 99.9982 27 0.0038 0.0002 99.9984 28 0.0035 0.0002 99.9986 29 0.0028 0.0002 99.9988 30 0.0027 0.0002 99.9990

PAGE 41

41 Table 3 9 Comparison of Atlantic and Gulf of Mexico stock greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance (ANOVA). Index df F p value n Form f actor 1 12.153 0.0007 98 Roundness 1 0.2831 0.5959 98 Circularity 1 11.1997 0.0012 98 Rectangularity 1 6.6579 0.0114 98 Ellipticity 1 0.4276 0.5197 98 Table 3 10 Comparison of c entral Florida, north Florida, and Louisiana greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance (ANOVA). Index df F p value n Form f actor 2 2.3232 0.0999 270 Roundness 2 0.2649 0.7675 270 Circularity 2 2.4169 0.0911 270 Rectangularity 2 1.7278 0.1797 270 Ellipticity 2 0.0838 0.9196 270 Table 3 11 Comparison of central Florida, north Florida, and Louisiana age 3 greater amberjack ( Seriola dumerili ) otolith shape indices using Analysis of Variance (ANOVA). Index df F p value n Form factor 2 3.7449 0.0267 114 Roundness 2 2.3634 0.0989 114 Circularity 2 3.983 0.0214 114 Rectangularity 2 7.1636 0.0012 114 Ellipticity 2 0.5467 0.5804 114

PAGE 42

42 Table 3 12 Discriminant analysis comparing otolith shape indices between Gulf and Atlantic samples of greater amberjack ( Seriola dumerili ). Region Model n Test n Gulf Atlantic correctly classified (%) Gulf 49 20 14 6 7 0 Atlantic 49 20 6 14 7 0 Total 98 40 20 20 70 0.40 0 Table 3 13 Discriminant analysis comparing greater amberjack ( Seriola dumerili ) otolith shape among three regions in the Gulf of Mexico, with CF= Central Florida, NF=North Florida, and LA=Louisiana. Region Model n Test n CF NF LA correctly classified (%) CF 90 20 5 7 8 25 NF 90 20 7 8 5 40 LA 90 20 3 2 15 75 Total 270 60 15 17 28 47 0.199

PAGE 43

43 Figure 3 1. Forklength (mm) of greater amberjack ( Seriola dumerili ) samples organized by collection region. 0 10 20 30 700 750 800 850 900 950 1000 1050 Frequency Forklength (mm) West Florida 0 5 10 15 20 700 750 800 850 900 950 1000 1050 Frequency Forklength (mm) North Florida 0 5 10 15 20 700 750 800 850 900 950 1000 1050 Frequency Forklength (mm) Louisiana 0 5 10 15 20 700 750 800 850 900 950 1000 1050 Frequency Forklength (mm) Atlantic Central Florida

PAGE 44

44 Figure 3 2. Age in years of Gulf of Mexico greater amberjack ( Seriola dumerili ) samples organized by collection region. Figure 3 3 P ower analysis on ellipt ical Fourier descriptors showed that 13 harmonics described 99% of the shape of greater amberjack ( Seriola dumerili ) otoliths. 0 50 100 150 1 2 3 4 5 6 7 8 9 10 Frequency Age (years) Central Florida 0 20 40 60 1 2 3 4 5 6 7 8 9 10 Frequency Age (years) North Florida 0 10 20 30 40 1 2 3 4 5 6 7 8 9 10 Frequency Age (years) Louisiana 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 14 16 18 20 Percentage of shape described Number of descriptors

PAGE 45

45 Figure 3 4 Values of shape indices of greater amberjack ( Seriola dumerili ) otoliths plotted across forklength show no relationship between size of fis h and shape index value. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 700 750 800 850 900 950 1000 1050 1100 Circularity value Shape index value Forklength (mm) Form Factor Roundness Rectangularity Ellipticity Cicularity

PAGE 46

46 Figure 3 5 Two dimensional Principal Component Analysis projection comparing otolith shape between males from the Gulf and Atlantic stoc k s of greater amberjack (Seriola dumerili), with maximum convex polygons enclosing the regions. Atlantic Gulf of Mexico

PAGE 47

47 Figure 3 6 Two dimensional Principal Component Analysis projection comparing otolith shape of greater amberjack ( Seriola dumerili ) among regions in the Gulf of Mexico, with maximum convex polygons enclosing the regions. Central Florida North Florida Louisiana

PAGE 48

48 CHAPTER 4 DISCUSSION There is no consensus in otolith shape analysis as to what constitutes a classification success informative to management, and studies have reported variable levels of success. Jonsdottir et al. (2006) compared cod otoliths from locations in northern and s outhern Iceland, and found that only 0 44% classified correctly to region based on otolith shape. However, misclassified otoliths were most often classified to adjacent locations, and a high percentage of cod south of Iceland were classified to other south ern locations (66 72%) and north of Iceland to other northern locations (61 67%). The authors considered the results to successfully discriminate a northern and southern spawning group for Icelandic cod, which previously had been managed as a single manage ment unit, and suggested that the current single stock management of Icelandic cod be reconsidered. DeVries et al. (2002) compared otolith shape among Gulf of Mexico and Atlantic king mackerel and found shape correctly classified 80% of Atlantic and 86% of Gulf king mackerel, which was considered high enough to use otolith shape to discern between the two stocks in mixing zones. A later study on king mackerel (Shepard et al. 2011) found classification success rates from 60 73%. Tuset et al. (2003) reported 68.8% classification accuracy in otolith shape between Atlantic and Mediterranean comber; though differences were slight between the two stocks, the authors considered it better than would be expected by chance and therefore reported otolith shape capable of separating the two regions. Campana and Casselman (1993) compared otolith shape among northwestern Atlantic cod and found classification success ranged from 20 80% depending upon location and scale of classification (i.e.

PAGE 49

49 classification was more accurat e to region than to specific location), and interpreted this In th e present study, otolith shape wa s found to correctly classify 70 % of Atlantic and Gulf of Mexico oto liths, which is interpreted as a high classification succ ess based upon previous studies. In addition, analysis of variance showed significant differences in the shape indices form factor and circularity between the two stocks, and PCA projections showed e two stocks of GAJ, validating the tool for this species providing further support for the uniqueness of the Gulf and Atlantic stocks and their separate management. Using otol ith shape to discriminate among Louisiana, north Florida and central Florida GAJ showed variable success. Overall, discriminant analysis had only a 47% classification success, with = 0.199 indicating little improvement over random chance alone, while PCA projections also revealed a high degree of overlap in otolith shape among regions. Similarly, analysis of variance showed no significant differences in otolith shape overall among the three regions. This supports both the current management assumption of one continuous Gulf of Mexico stock a s well as the current genetic data indicating mixing within the Gulf of Mexico stock (Gold and Richardson 1998; Murie et al. 2011). However, when restricted by age, analysis of variance did show significant differences in rectangularity between Florida and Louisiana samples, and despite an overall low classification success, discriminant analysis was able to correctly classify 75% of Louisiana samples, which suggests some otolith shape characteristics local to the subreg ion. Despite this success many Florida samples (40% for central Florida and 25% for northern Florida) were incorrectly classified to Louisiana,

PAGE 50

50 which supported the the PCA results that suggest ed a high degree of overlap among the three regions. Though s amples were restricted to a size range of 700 1050 mm in an effort to standardize amongst regions, Louisiana fish were larger and older on average than fish from the two Florida regions. It is possible that this size and age discrepancy could have contribu ted to the differences seen between Florida and Louisiana otoliths. However, analysis of covariance showed no relationships between size and otolith shape, and differences were still apparent in the age restricted subset of the data. Therefore it is assume d that the differences seen were due to regional distinctions in otolith shape and not to differences in size or age among the regions. Otoliths used in this study had been previously collected for age and growth analyses, and not specifically for regiona l shape comparisons and otolith collections ranged across years and seasons. However, preliminary data exploration showed no differences in shape between breeding and non breeding seasons, and so time of collection was not taken into account for regional comparisons. However, it is possible that regional differences may be more discerni ble in a future targeted study of otoliths collected from a single year It is also reflect the st ructure of the Gulf of Mexico stock of GAJ. While tagging data show most individuals recaptured close to the original release site, some tags have been recovered over greater distances (Beasley 1993; McClellan and Cummings 1997; Ingram and Patterson 2001; Murie et al. 2011). This could reflect a stock consisting of both migratory and resident individuals, which has been shown in other species. Tagging

PAGE 51

51 data of Gulf of Mexico cobia, Rachycentron canadum for example, have found that while most individuals mig rate from the northern Gulf to south Florida to overwinter, some individuals have been found to remain in the northern Gulf year round, suggesting separate migratory and non migratory groups (Hendon and Franks 2010). A stock structure consisting of migrato ry and resident sub populations would confound morphological differences in otolith shape, making regional differences among resident individuals difficult to discern. It may therefore be worthwhile to examine this stock using isotope analysis, which could be able to distinguish between migratory and resident individuals and thereby supplement the current genetic and tagging data. In summary, GAJ otoliths in the Gulf of Mexico did not exhibit clear differences in shape among regions sampled. While there is evidence that Louisiana samples differ, with age three individuals significantly differing in rectangularity and an overall high classification success, there was overlap among the regions and northern and central Florida regions were indistinguishable. However, this does not necessarily mean that the stock is completely mixed. Otolith shape analysis is a novel approach for exploring stock structure in this species, and while it was able to distinguish between the Gulf and Atlantic stocks it showed a hig h degree of variation Th erefore this form of analysis may not accurately reflect stock structure in the Gulf, and studies should continue to elucidate the structure of this species. If the stock is not continuously mixed, as is strongly suggested by the low movement rates of GAJ in the Gulf of Mexico observed in the tagging data the disproportionately high fishing effort off of Florida could lead to local ized overfishing of the species. Stock delineation is vital to the appropriate

PAGE 52

52 management of fisheries. Therefore, it is important to determine with certainty as the Gulf of Mexico GAJ stock enters the next phase in its rebuilding efforts.

PAGE 53

53 LIST OF REFERENCES Allendorf, F.W. 1983. Isolation, gene flow, and genetic differentiation among populations. Great Lakes Fishery Commi s sion 1983 Project Completion Report. Aguera, A., and D. Brophy. 2011. Use of saggital otolith shape analysis to discriminate Northeast Atlantic and Western Mediterranean stocks of Atlantic saury, Scomberesox saurus saurus (Walbaum). Fisheries Research 110: 465 471. Beasley M.L. 1993. Age and growth of greater amberjack, Seriola dumerili from the Beddington, J.R., D.J. Agnew, and C.W. Clark. 2007. Current problems in the management of marine fisheries. Science 316: 1713 1716. Begg, G.A., and J.R. Waldman. 1999. An holistic approach to fish stock identification. Fisheries Research 43:35 44. Begg, G.A., W.J. Overholtz, and N.J. Munroe. 2001. T he use of internal otolith morphometrics for identification of haddock ( Melanogrammus aeglefinus ) stocks on Georges Bank. Fisheries Bulletin 99: 1 14. Bergenius, M.A.J., G.A. Begg, and B.D. Mapstone. 2006. The use of oto lith morphology to indicate the stock structure of common coral trout ( Plectropomus leopardus ) on the Great Barrier Reef, Australia. Fishery Bulletin 104(4): 498 511. Booke, H.E. 1999. The stock concept revisited: perspectives on its history in fisheries. Fisheries Research 43: 9 11. Burch, R.K. 1979. The greater amberjack, Seriola dumerili : its biology and fishery off Campana, S.E. and J.M. Casselman. 1993. Stock disc rimination using otolith sha pe analysis. Canadian Journal of Fisheries and Aquatic Sciences 50: 1062 1083. Campana, S.E., and S.R. Thorrold. 2001. Otoliths, increments, and elements: keys to a comprehensive understanding of fish populations? Canadian Journal of Fisheries and Aquatic Sciences 58: 30 38. Canas, L., C. Stransky, J. Schlickeisen, M.P. Sampedro, and A.C. Farina. 2012. Use of otolith shape analysis in stock identification of anglerfish ( Lophius piscatorius ) in the Northeast Atlantic. ICES Journal of Marine Science 69(2): 25 0 256. Carvalho, G.R., and L. Hauser. 1994. Molecular genetics and the stock concept in fisheries. Reviews in Fish Biology and Fisheries 4: 326 350.

PAGE 54

54 Cardinale, M., P. Doering Arjes, M. Kastowsky, and H. Mosegaard. 2004. Effects of sex, stock, and environme nt on the shape of known age Atlantic cod ( Gadus morhua ) otoliths. Canadian Journal of Fisheries and Aquatic Sciences 61(2): 158 167. Coyle, T. 1997. Stock identification and fisheries management: the importance of using several methods in a stock identification study. Proceedings of the Joint Workshop of the Australian Society for Fish Biology and the Fish and Aquatic Resource Management Association of Australasia: 173 182. Cummings, N.J. and D.B. McClellan. 2000. Trends in the Gu lf of Mexico great er amberjack fishery through 1998: Commercial landings, recreational catches, observed length frequencies, estimates of landed and discarded catch at age, and selectivity at age. U.S. Dept of Commerce, National Oceanographic and Atmospheric Administratio n, National Marine Fisheries Service, Sustainable Fisheries Division. Cunningham, S. and T. Bostock. 2005. Successful Fisheries Management: Issues, Case Studies and Perspectives. SIFAR/ World Bank Study of Good Management Practice in Fisheries. Eburon Ac ademic Publishers. Delft, Netherlands. DeVries, D.A., C.B. Grimes, and M.H. Prager. 2002. U sing otolith shape analysis to distinguish eastern Gulf of Mexico and Atlantic Ocean stocks of king mackerel. Fisheries Research 57 (2002): 51 62. Ebbin, S.A. 1996. The stock concept: constructing tools for Pacific salmon management. Coastal Management 24(4): 355 364. Felix Uraga, R., V.M. Gomez Munoz, C. Quinonez Vela zquez, F.N. Melo Barrera, K.T. Hill, W. Garcia Franco. 2005. Pacific sardine ( Sardinops sagax ) stock discrimination off the west coast of Baja California and southern California using otolith morphometry. California Cooperative O ceanic Fisheries Investigations Report 46: 113 121. Ferguson, G.J., T.M. Ward, and B.M. Gillanders. 2011. Otolith shape and el emental composition: complementary tools for stock discrimination of mulloway ( Argyrosomus japonicus ) in southern Australia. Fisheries Research 110(1): 75 83. Fleiss, J.L. 1981 Statistical methods for rates and proportions, 2 nd edition. John Wiley and Son s, New York, New York. Gabriel, W.L., and P.M. Mace. 1999. A review of biological reference points in the context of the precautionary approach. Proceedings of the 5 th Annual NMFS Stock Assessment Workshop: 34 45. Gold, J.R. and L.R. Richardson. 1998. Population structure in greater amberjack, Seriola dumerili from the Gulf of Mexico and the w estern Atlantic Ocean. Fishery Bulletin 96(4): 767 778.

PAGE 55

55 Harris, P.J., D.M. Wyanski, D.B. White, and P.P. Mikell. 2007. Age, growth, and reproduction of greater amberjack off the sou theastern U.S. Atlantic coast. Transactions of the American Fisheries Society 136: 1534 1545. Hendo n, J.R., and J.S. Franks. 2010. Sport fish tag and release in Mississippi coastal waters and the adj acent Gulf of Mexico. SEDAR28 RD23. Hilborn, R., T.A. Branch, B. Ernst, A. Magnusson, C.V. Minte Vera, M.D. Scheuerell, Environmental Resources 28: 359 399. Hilborn, R. 2007. Defining success in fisheries and conflicts in objectives. Marine Policy 31:153 158. Hood, P. 2006. History of vermillion snapper, greater amberjack, and gray triggerfish management in federal waters of the U.S. Gulf of Mexico 1984 2005. SEDAR 9: DW1. Hunt, J.J 1992. Morphological characteristics of otoliths for selected fish in the northwest Atlantic. Journal of Northwest Atlantic Fishery Science 13: 63 75. Ihssen, P.E., H.E. Booke, J.M. Casselman, J.M. McGlade, N.R. Payne, and F.M. Utter. 1981. Stock identifi cation: materials and methods. Canadian Journal of Fisheries and Aquatic Sciences 38: 1838 1855. Ingram, G.W. Jr. and W.F. Patterson III. 2001. Mo vement patterns of red snapper ( Lutjanus campechanus ), greater amberjack ( Seriola dumerili ), and gray triggerf ish ( Balistes capriscus ) in the Gulf of Me xico and the utility of marine reserves as management tools. Proceedings of the 52 nd Gulf and Caribbean Fisheries Institute 52: 686 699. Iwata, H. and Y. Ukai. 2002. SHAPE: a computer pr ogram package for quantitati ve evaluation of biological shapes based on ellip tical Fourier descriptors. The Journal of Heredity 93(5): 384 385. Jonsdottir, I.G., S.E. Campana, and G. Marteinsdottir. 20 06. Otolith shape and temporal stability of spawning groups of Icelandic cod ( Gadus morhua L.). International Council for the Exploration of the Sea Journal of Marine Science 63: 1501 1512. Kuhl, F.P. and C.R. Giardina. 1982. Elliptic Fourier features of a closed contour. Computer Graphics Image Processing 18 : 236 258. Lackey, R.T. 2005. Fisheries: history, science, management. Pages 121 129 i n J.H. Lehr and J. Keely, editors. Water Encyclopedia: Surface and Agricultural Water. John Wiley and Sons Incorporated, New York, New York.

PAGE 56

56 MacLean, J.A., and D.O. Evans. 1981. The stock concept, di screteness of fish stocks, and fisheries management. Canadian Journal of Fisheries and Aquatic Sciences 38: 1889 1898. Manooch, C.S., and M. Haimovici. 1983. Foods of greater amberjack, Seriola dumerili and almaco jack, Seriola rivolian a (Pisces Carangidae) from the South Atlantic Bight. Journal of the Elisha Mitchell Scientific Society 99(1): 1 9. McClellan, D.B and N.J. Cummings. 1997. Preliminary anal ysis of tag and recapture data of the greater amberjack, Seriola dumerili in t he s outheastern United States. Proceedings of the 49 th Gulf and Caribbean Fisheries Institute 49: 25 45. fisheries. Canadian Journal of Fisheries and Aquatic Sciences 38(12): 1657 1664. McEach ern J.D. and J.D. Fechhelm. 2005. Fishes o f the Gulf of Mexico Volume 2: Scorpaeniformes to Tetraodontiformes. University of Texas Press, Austin, Texas. Merigot, B., Y. Letourneur, and R. Lecomte Finiger. 2 007. Characterization of local populations of the common sole Solea solea (Pisces, Soleidae) in the NW Mediterranean through otolith morphometr ics and shape analysis. Marine Biology 151: 997 1008. Mullon, C., P. Freon, and P. Cury. 2005. The dynamics of collapse in world fisheries. Fis h and Fisheries 6:111 120. Murie, D.J. and D.C. Parkyn. 2008. Age, growth and sexual maturity of greater amberjack ( Seriola dumerili ) in the Gulf of Mexico. MARFIN final report NA05NMF4331071. Murie, D., D. Parkyn, and J. Austin. 2011. Seasonal movement and mixing rates of greater amberjack in the Gulf of Mexico and assessment of exchange within the south Atlantic spawning stock. NOAA Fisherie s Cooperative Research Program Final Report NA07NMF4540076 Myers, L.S., G.C. Gamst, and A.J. Guarino. 2006. Applied multivariate research: design and interpretation. Sage Publications Incorporated, Thousand Oaks, California. NMFS. 2011. Establishment of a June and July closed season for the recreational harvest of greater amberjack in the Gulf of Mexico. Southea st Fishery Bulletin FB11 040 Pankhurst, N.W. and J.C. Montgomery. 1994. Uncoupling of visual and somatic growth in the rainbow trout Oncorhynchus mykiss Brain, Behavior, and Evolution 44(3):149 55.

PAGE 57

57 Paul y, D., V. Christensen, S. Guenette, T.J. Pitcher, U. Rashid Sumaila, C.J. Walters, R. Watson, and D. Zeller. 2002. Towards sustainability in world fisheries. Nature 418: 689 695. Petursdottir, G., G.A. Begg, and G. Marteinsdottir 2006. Discrimination between Icelandic cod ( Gadus morhua L.) populations fr om adjacent spawning areas based on otolith shape and growth. Fisheries Research 80(2006): 182 189. Pothin, K., C. Gonzalez Salas, P. Chabanet and R. Leco mte Finiger. 2006. Distinction between Mulloidichthyes flavolineatus juv eniles from Reunion Island and Mauritius Island (south west Indian Ocean) b ased on otolith morphometrics. Journal of Fish Biology 68: 1 16. Rauch, S.D. III. 2011. Fisheries of the Caribbean, Gulf of Mexico, and south Atlantic; reef fish fishery of the Gulf of Mexico: greater amberjack management measures. Federal Register 76(15): 4084 4087. Ricker, W.E. 1975. Computation and Interpretation of Biological Statistics of Fish Populations. Bulletin 181. Fisheries Research Board of Canada. Ottawa, Canada. Robins, C.R., G.C. Ray, J. Douglass and R. Freund. 1986. A field guide to Atlantic coast fishes of North America. Houghton Mifflin Co mpany, New York, New York. Salkind, N.J. 2010. Encyclopedia of research design: volume 1. Sage Publications Incorporated, Thousand Oaks, California. Carolina. SEDAR 2006. SEDAR9 SAR: Gulf of Mexico Grea ter Amberjack Stock Assessment Report. SEDAR. 2008. SEDAR15: Stock assessment report 2 (SAR 2) South Atlantic greater amberjack. SEDAR. 2011. SEDAR9 SAUR: Gulf of Mexico Greater Am berjack Stock Assessment Update Report. Sm ith, G. H., jr. 2011. Field Based Non lethal Sex Determination and Effects of Sex Ratio on Population Dynamics of Greater Amberjack, Seriola dumerili. Masters of Science Thesis, University of Florida Shepard, K.E., W.F. Patterson and D.A. DeVries. 2010. Tren ds in Atlantic contribution to mixed stock king mackerel landings in south Flori da inferred from otolith shape analysis. Marine and Coastal Fisheries: Dynam ics, Management, and Ecosystem Science 2: 195 204.

PAGE 58

58 Thompson, B.A., M. Beasley, and C.A. Wilson. 1998. Age distribution and growth of greater amberjack, Seriola dumerili from the north central Gulf of Mexico. Fishery Bulletin 97(2): 362 371. Tracey, S.R., J.M. Lyle and G. Duhamel. 2006. Application of elliptical Fourier analysis of otolith form as a tool for stock identification. Fisheries Research 77 (2006): 138 147. Tuset, V.M., I.J. Lozano, J.A. Gonzalez, J.F. Pert usa, and M.M. Garci Diaz. 2003. Shape indices to identify regional differences in otolith morphology of comber, Serranus cabrilla (L., 1758). Journal of Applied Ichthyology 19(2): 88 93. Vignon, M., and F. Morat. 2010. Environmental and genetic determinant of otolith shape revealed by a non indigenous tropical fish. Marine Ecology Progress Series 411: 231 241. Walters, C. J. and R. Ahrens 2009. Oceans and Estuaries: Managing the Commons. Pages 221 240 i n C. Folke, G.P. Kofinas, and S. Capin editors. Principles of Ecosystem Stewardship: Resiliance Based Natural Resource Management in a Changing World. Springer New York New York Wells, R .J. D. and J.R. Rooker. 2004. Distribution, age, a nd growth of young of the year greater amberjack ( Seriola dumerili ) associated with pelagic Sargassum Fisheries Bulletin 102: 545 554. Worm, B., E.B. Barbier, N. Beaumont, J.E. Duffy, C. Folke, B.S. Halpern, J.B.C. Jackson, H.K. Lotze, F. Micheli, S.R. Palumbi, E. Sala, K.A. Selkoe, J.J. Stachowicz, and R. Watson. 2006. Impacts on biodiversity loss on ocean ecosystem services. Science 314: 787 780. Worm, B., R. Hilborn, J.K. Baum, T.A. Branch, J.S Col lie, C. Costello, M.J. Fogarty, E.A. Fulton, J.A. Hutchings, S. Jennings, O.P. Jensen, H.K. Lotze, P.M. Mace, T.R. McClanahan, C. Minto, S.R. Palumbi, A.M. Parma, D. Ricard, A.A. Rosenberg, R. Watson, D. Zeller. 2009. Rebuilding global fisheries. Science 3 25: 578 585. Zar, J.H. 1999. Biostatistical Analysis Prentice Hall Incorporated, Englewood Cliffs, New Jersey.

PAGE 59

59 BIOGRAPHICAL SKETCH Chelsey was born in Tampa, Florida, where she School and then H.B. Plant High School. While in Tampa, Chelsey taught Tae Kwon Do for several years, attained the rank of 3 rd degree blackbelt and competed on the national level. In 2004 she graduated from high school an d attended the University of Florida, where she majored in Zoology. While an undergraduate at UF, Chelsey volunteered on a variety of research projects. In 2005 she worked for Dr. Iske Larkin conducting manatee surveys in Kin gs Bay, Florida. The next year she was hired to assist graduate student Elise Hoover and Dr. Shirley Baker on a study of tripl oid clam survival. That fall, she began working in Dr. terrestrial snakes. In addition, she assisted on work with cotton mouths o n the island of Seahorse Key, FL and also worked with then graduate student Leslie Babonis researching marine snake osmoregu lation. In the summer of 2006, Chelsey worked in the Florida Keys as an intern for the Seakeys program; while there, she also volunteered for the Sea Turtle Hosp ital on Marathon Key. In 2006, she was awarded a University Scholars grant to study the relative abundance of invasive rats on the island of Seahorse Key. Chelsey also began volunteering with the Florida Program for S hark Research that year, where she worked on shark age and growth and reproduction studies. In 2007 her work with the Lillywhite lab allowed her to join a research expedition to Orchid Island, Taiwan, where they collected and studied several species of sea krait. That same summer she assisted the Florida Program for Shark Research in the field, tagging cownose rays in the Mosquit o Lagoon, FL. Later that year, Chelsey began

PAGE 60

60 volunteering with the Invertebrate Division of the Florida Museum of Natural Histo ry, where she was soon hir ed as a collections assistant. Chelsey also assisted the Paulay lab with their research on species diversity of various sea cucumber groups. During her time at UF, Chelsey also worked with numerous elementary and midd le school gro ups. In addition, she led the students on field trips to Cedar Key, Florida, where the collected and identified marine organisms. Chelsey greatly enjoyed this outreach experience. Chelsey graduated from the University of Florida with a Bache lor in Science degree in May, 2008. She was then hired as an intern by Dr. Kenneth Emberton on a study of biodiversity in was able to travel to Madagascar to conduct field research, and remained there a total of 6 weeks; the rest of my internship was spent analyzing the specimens collected in Gainesville, FL at the Florida Museum of Natural History. In the fall of 2008, she attained my Science Diver certification, and was invited to assist then graduate student Dr. Mary Hart in field c ollections of chalk bass in Carrie Bowe Cay, Belize. In 2009, Chelsey was hired as the field technician for the Florida Pro gram for Shark Research, where she tagged numerous shark species as well as conducted abundance surveys for the Sawfish Recovery Prog ram. Chelsey also joined the Mur ie/Parkyn lab in 2009 and 2010 In 2010, Chelsey began work as a M.S. graduate assistant in the University of atic Sciences with Dr. Daryl Parkyn. Her work embodied this thesis on otolith shape as a tool to examine stock structure in Gulf of Mexico greater amberjack. Her research was recognized with a best poster award at the Florida Chapter of the American Fisher ies Society, and the Roger Rottmann

PAGE 61

61 Scho larship as the top Master of Science student in the Florida Chapter of the American Fisheries Society for 2012. Upon completion of her Master s of Science degree, Chelsey will begin a Ph.D. in Interdisciplinary Ecolog y at the University of Florida.