<%BANNER%>

Life History, Population Dynamics, and Fishery Management of the Golden Tilefish, Lopholatilus Chamaeleonticeps, from th...

Permanent Link: http://ufdc.ufl.edu/UFE0044580/00001

Material Information

Title: Life History, Population Dynamics, and Fishery Management of the Golden Tilefish, Lopholatilus Chamaeleonticeps, from the Southeast Atlantic and Gulf of Mexico
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Lombardi, Linda Anne
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: age -- assessment -- reproduction -- tilefish -- validation
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There is a growing concern over the lack of life history information for many deepwater fisheries, including golden tilefish, Lopholatilus chamaeleonticeps.   Basic age and growth estimates are essential for proper stock assessment and management decisions. My first objective was to use lead-radium dating to validate the timing of growth increments.  Radiometric ages closely agreed with age estimates from traditional age estimations (growth increment counts in thin sagittal otolith sections) for females and unknown sex fish.  However, radiometric ages did not agree with traditional age estimates for males.  This difference may be attributed to differing growth rates and increment formation by gender or the transition of gender.  Golden tilefish longevity of 26 ± 6 yrs was confirmed using both methods.  My second objective was to describe the reproductive parameters and strategy for golden tilefish along the east coast of Florida and the northern Gulf of Mexico.  Male golden tilefish reached maturity at a younger age (male, 1 yr; female, 2.5 yr) and smaller size (male, 150 FL mm; female, 331 FL mm) than females. Golden tilefish gonads are described as being intersexual, identified as having non-functional (opposite) sex tissue (functional males 71% had multiple stages of atretic oocytes; functional females 26% had self-contained male tubules).  Male gonads also contained a cavity that originated from ovarian lumen and sperm sinuses.  Golden tilefish gonads exhibit intersexual tendencies and I found evidence supporting a protogynous hermaphroditic reproductive strategy.  My third objective was to apply the resulting life history characteristics (objectives 1 and 2) to two statistical age structure stock assessment models.  Each of these models represented the ends of the spectrum of model complexity.  The available data for golden tilefish stock assessment was typical for many species with irregular sampling for length and age composition data and highly variable indices of abundance.  I completed a stock assessment for golden tilefish using the simple Stochastic Stock Reduction Analysis model and the more complex Stock Synthesis model.  The two models agreed in stock status of golden tilefish (not overfished and not undergoing overfishing), as well as, in historical annual biomass and exploitation rates.
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 Linda Anne Lombardi.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Allen, Micheal S.
Local: Co-adviser: Pine, William.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044580:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044580/00001

Material Information

Title: Life History, Population Dynamics, and Fishery Management of the Golden Tilefish, Lopholatilus Chamaeleonticeps, from the Southeast Atlantic and Gulf of Mexico
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Lombardi, Linda Anne
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: age -- assessment -- reproduction -- tilefish -- validation
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: There is a growing concern over the lack of life history information for many deepwater fisheries, including golden tilefish, Lopholatilus chamaeleonticeps.   Basic age and growth estimates are essential for proper stock assessment and management decisions. My first objective was to use lead-radium dating to validate the timing of growth increments.  Radiometric ages closely agreed with age estimates from traditional age estimations (growth increment counts in thin sagittal otolith sections) for females and unknown sex fish.  However, radiometric ages did not agree with traditional age estimates for males.  This difference may be attributed to differing growth rates and increment formation by gender or the transition of gender.  Golden tilefish longevity of 26 ± 6 yrs was confirmed using both methods.  My second objective was to describe the reproductive parameters and strategy for golden tilefish along the east coast of Florida and the northern Gulf of Mexico.  Male golden tilefish reached maturity at a younger age (male, 1 yr; female, 2.5 yr) and smaller size (male, 150 FL mm; female, 331 FL mm) than females. Golden tilefish gonads are described as being intersexual, identified as having non-functional (opposite) sex tissue (functional males 71% had multiple stages of atretic oocytes; functional females 26% had self-contained male tubules).  Male gonads also contained a cavity that originated from ovarian lumen and sperm sinuses.  Golden tilefish gonads exhibit intersexual tendencies and I found evidence supporting a protogynous hermaphroditic reproductive strategy.  My third objective was to apply the resulting life history characteristics (objectives 1 and 2) to two statistical age structure stock assessment models.  Each of these models represented the ends of the spectrum of model complexity.  The available data for golden tilefish stock assessment was typical for many species with irregular sampling for length and age composition data and highly variable indices of abundance.  I completed a stock assessment for golden tilefish using the simple Stochastic Stock Reduction Analysis model and the more complex Stock Synthesis model.  The two models agreed in stock status of golden tilefish (not overfished and not undergoing overfishing), as well as, in historical annual biomass and exploitation rates.
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 Linda Anne Lombardi.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Allen, Micheal S.
Local: Co-adviser: Pine, William.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044580:00001


This item has the following downloads:


Full Text

PAGE 1

1 LIFE HISTORY, POPULATION DYNAMICS, AND FISHERY MANAGEMENT OF THE GOLDEN TILEFISH, LOPHOLATILUS CHAMAELEONTICEPS FROM THE SOUTHEAST ATLANTIC AND GULF OF MEXICO By LINDA ANNE LOMBARDI CARLSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Linda Anne Lombardi Carlson

PAGE 3

3 To my Father

PAGE 4

4 ACKNOWLEDGMENTS This dissertation was made possible through the financial support of the U.S. Department of Commerce (DOC) N ational Oceanic and Atmospheric Administration (NOAA) Fisheries Service. Two proposals funded by the DOC/NOAA, Marine Fisheries Initiative Program ( 07MFIH007 ) and Cooperative Research Project ( 08CRP009) provided the necessary means to complete this dissertation. Financial assistance was also provided the DOC /NOAA, Advanced Studies Program which provided tuition assistance. Due to the extensive f ield sampling, this dissertation was successful given the assistance of NOAA, Trip Interview Program port agents (C. Dennis, D. Fable, L. Bullock, M. Gamby, K. Roberts, and B. Bourgeois), the crew of the NOAA R/V OREGON II and to all the scientific staff, especially M. Grace, L. Jones and T. Driggers, NOAA, Shark Bottom Longline Observer Program (observers S. Gulak and J. Combs), commercial captain C. Summers, and P. Antosh from the Alabama Marine Resources. Similarly, the success of this dissertation was al so due to assistance at NOAA, Southeast Fisheries Science Center (SEFSC) Panama City, FL laboratory by L. Goetz Thorton (sectioning otoliths), B. Walling (blocking gonads), and H. Lyon (expertise in gonad histological). I appreciate the assistance of A Andrews of NOAA, Pacific Islands Fisheries Science Center Honolulu, HI (formerly of Moss Landing Marine Laboratory) for his expertise in lead radium dating In addition to, H. Hawk at Moss Landing Marine Laboratories for assistance with sample processing and C. Lundstrom at University of Illinois Urbana Champaign for ICPMS processing of the purified radium samples. I am grateful to stock assessment biologist s B. Linton and J. Walter (NOAA, S outheast Fisheries Science Center Miami, FL) for their assistance with stock synthesis. I especially appreciate the support provided to me by my co chairs, Dr. Mich eal Allen and Dr. William Pine. Dr. Allen and Dr. Pine recognized t he drive and passion I have for fisheries

PAGE 5

5 science, and provided the last push I needed to start and complete my PhD. I also appreciate the confidence, guidance, and the use of the stochastic stock reduction analysis model from Dr. Carl Walters My gratitude extends to the additional members of my PhD committee, Dr. Debra Murie, Dr. Alan Bolt e n (University of Florida, Department of Biology) and Dr. Clay Porch (special member, NOAA Southeast Fisheries Science Center Miami, FL)

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .........................................................................................................................9 ABSTRACT ...................................................................................................................................11 CHAPTER 1 GENERAL INTRODUCTION ..............................................................................................13 2 AGE ESTIMATION AND L EADRADIUM DATING OF GOL DEN TILEFISH, LOPHOLATILUS CHAMAEL EONTICEPS ...........................................................................16 Introduction .............................................................................................................................16 Materials and Method s ...........................................................................................................20 Sample Collection ...........................................................................................................20 Traditional Age Estimation .............................................................................................20 Otolith Growth Increment Deposition .............................................................................21 Lead radium Dating Age Groups ....................................................................................22 Le ad radium Dating Analysis ..........................................................................................23 Radiometric Age Estimation ...........................................................................................25 Radiometric versus Tradition Age Estimation ................................................................26 Results .....................................................................................................................................26 Sample Collection ...........................................................................................................26 Traditional Age Estimation .............................................................................................26 Otolith Growth Increment Deposition .............................................................................27 Lead radium Dating Age Groups ....................................................................................28 Lead radium Dating Analysis ..........................................................................................29 Radiometric versus Tradition Age Estimation ................................................................30 Discussion ...............................................................................................................................30 Conclusion ..............................................................................................................................34 3 EVIDENCE FOR HERMAPHRODITISM IN GOLDEN T ILEFISH ( LOPHOLATILUS CHAMAELEONTICEPS ) .......................................................................................................47 Introduction .............................................................................................................................47 Materials and Method s ...........................................................................................................49 Sample Collection ...........................................................................................................49 Otolith Processing and Assigning Age ............................................................................49 Gonad Processing ............................................................................................................50 Gonad Development ........................................................................................................50 Estimates of Maturity ......................................................................................................52

PAGE 7

7 Spawning Seasonality ......................................................................................................52 Other Criter ia for Hermaphroditism ................................................................................52 Breeding Strategy ............................................................................................................53 Results .....................................................................................................................................53 Sample Collection ...........................................................................................................53 Otolith Processing and Assigning Age ............................................................................53 Gonad Processing ............................................................................................................54 Gonad Development ........................................................................................................54 Sexual Maturity ...............................................................................................................54 Histological Evidence for Hermaphroditism ...................................................................55 Spawning Seasonality ......................................................................................................55 Other Criteria for Hermaphroditism ................................................................................56 Breeding Strategy ............................................................................................................56 Discussion ...............................................................................................................................56 Conclusion ..............................................................................................................................62 4 MODEL CHOICE AND QUANTITY OF INFORMATIVE DATA: A CASE STUDY OF THE GOLDEN TILEFI SH ASSESSMENT IN THE GULF OF MEXICO ....................87 Introduction .............................................................................................................................87 Materials and Methods ...........................................................................................................89 Model Descriptions .........................................................................................................89 Gulf of Mexico Golden Tilefish Case Study ...................................................................91 Uncertainty in Stock Status .............................................................................................95 Results .....................................................................................................................................96 Stochastic Stock Reduction Analysis (SRA) ...................................................................96 Stock Synthesis (SS) ........................................................................................................97 Model Comparison ..........................................................................................................98 Uncertainty in Stock Status .............................................................................................99 Disc ussion .............................................................................................................................100 Conclusion ............................................................................................................................107 5 CONCLUSION .....................................................................................................................132 Importance of Sampling .......................................................................................................132 Fishery Management ............................................................................................................134 LIST OF REFERENCES .............................................................................................................136 BIOGRAPHICAL SKETCH .......................................................................................................149

PAGE 8

8 LIST OF TABLES Table page 21 Summary of estimated age with fish and otolith characteristics for golden tilefish agegroups from the east coast of Florida processed in this study. ...................................35 22 Coring and radiometric results for golden tilefish age groups from the east coast of Florida processed in this study ...........................................................................................36 23 Comparison of estimated age and radiometric age for golden tilefish from the east coast of Florida processed in this study .............................................................................37 31 Descriptions and photomicrographs of histological sections of male golden tilefish leading gamete stages ........................................................................................................64 32 Descriptions and photomicrographs of histological sections of female golden tilefish leading gamete stages ........................................................................................................66 33 Descriptions and photomicrographs of histological sections of male golden tilefish reproductive phases ............................................................................................................68 34 Descriptions and photomicrographs of histological sections of female golden tilefish reproductive phases ............................................................................................................71 35 The number of golden tilefish collected by month ............................................................74 36 Summary statistics by sex and maturity for golden tilefish ...............................................75 37 Results of the logistic regressions for size and age at maturity for male and female golden tilefish .....................................................................................................................76 41 Commercial landings for the northern Gulf of Mexico golden tilefish ...........................109 42 Annual sample sizes of length composition data for golden tilefish from the northern Gulf of Mexico .................................................................................................................111 43 Annual sample sizes of age composition data for golden tilefish from the northern Gulf of Mexico .................................................................................................................112 44 Life history parameters for golden tilefish from the northern Gulf of Mexico by region and gender .............................................................................................................113 45 Negative log likelihoods for the major components of the stock synthesis model for golden tilefish from the northern Gulf of Mexico ...........................................................114 46 Summary statistics and results of diagnostic tests for convergence for MCMC iterations f or parameters from each model that determined stock status for golden tilefish from the northern Gulf of Mexico .......................................................................115

PAGE 9

9 LIST OF FIGURES Figure page 21 Map of collection area for golden tilefish ..........................................................................38 22 Images of thin sectioned golden tilefish sagittal otoliths from previous ageing studies ...39 23 Image of the juvenile golden tilefish otolith ......................................................................40 24 Length frequency of golden tilefish ...................................................................................41 25 Thin sectioned sagittal otoliths of golden tilefish ..............................................................42 26 Age frequency of golden tilefish........................................................................................43 27 Otolith increment counts and band increment measurements by gender for golden tilefish from the east coast of Florida ................................................................................44 28 Examples of thin sectioned male golden tilefish sagittal otoliths .....................................45 29 Measured 210Pb:226Ra ratios with respect to average traditional age for golden tilefish samples processed in this study .........................................................................................46 31 Golden tilefish proportion by reproductive stage by month and gender ...........................77 32 Golden tilefish proportion mature and immature by gender ..............................................78 33 Photomicrographs of golden tilefish histological sections showing membrane lined cavity originating from ovarian lumen in both a female and male gonads ........................79 34 Photomicrograph of golden tilefish histological sections showing an immatur e female containing male tissue ............................................................................................80 35 Photomicrographs of golden tilefish histological sections showing the placement of oocyt es within functional male gonads ..............................................................................81 36 Photomicrograph of golden tilefish histological section showing a spawning male containing vitellogenic and late hydrated oocytes .............................................................82 37 Photomicrograph of golden tilefish histological section showing sperm sinuses within the gonad wall of a spawning male .........................................................................83 38 Golden tilefish gonadsomatic indices by month and by gender .......................................84 39 Golden tilefish age and length data by gender ...................................................................85 310 Golden tilefish adipose flap height by length, maturity stage and gender .........................86

PAGE 10

10 41 Northern Gulf of Mexico spatial map displaying the National Marine Fisheries Service (NMFS ) statistical grids ......................................................................................116 42 Standardized indices of abundance for golden tilefish from the northern Gulf of Mexico by region and by source ......................................................................................117 43 Annual age composition for golden tilefish from the northern Gulf of Mexico ..............118 44 Observed annual length composition data for golden tilefish from the northeastern Gulf of Mexico .................................................................................................................119 45 Observed annual length composition data for golden tilefish from the northwestern Gulf of Mexico .................................................................................................................120 46 Stock synthesis size selectivities for golden tilefish from the northern Gulf of Mexico ............................................................................................................................121 47 Age selectivities for each assessment model for golden tilefish from the northern Gulf of Mexico .................................................................................................................122 48 D istribution of maximum sustainable yield given the distribution of exploitation at maximum sustainable yield for golden tilefish from the northern Gulf of Mexico from Stochastic SRA ........................................................................................................123 49 Residuals for fits of length composition for golden tilefish from the northern Gulf of Mexico of unknown gender .............................................................................................124 410 Residuals for fits of age composition for golden tilefish from the northern Gulf of Mexico of unknown gender .............................................................................................125 411 Predicted historical biomass from both assessment models for golden tilefish from the northern Gulf of Mexico ............................................................................................126 412 Predicted historical exploitation rates from both assessment models for golden tilefish from the northern Gulf of Mexico .......................................................................127 413 Trace plots of the parameters from Stochastic SRA assessment model that determined the stock status of golden tilefish from the northern Gulf of Mexico ...........128 414 Trace plots of the parameters from Stock Synthesis assessment model that determined the stock status for golden tilefish from the northern Gulf of Mexico .........129 415 Fried egg and beer plot for the parameters from Stochastic SRA assessment model that determined the stock status for golden tilefish from the northern Gulf of Mexico. .130 416 Fried egg and beer plot for the parameters from Stock Synthesis assessment model that determined the stock status for golden tilefish from the northern Gulf of Mexico. .131

PAGE 11

11 Abstract of Disser tation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LIFE HISTORY, POPULATION DYNAMICS, AND FISHERY MANAGEMENT OF THE GOLDEN TILEFISH, LOPHOLATILUS CHAMAELEONTICEPS FROM THE SOUTHEAST ATLANTIC AND GULF OF MEXICO By Linda Anne Lombardi Carlson August 2012 Chair: Mich eal Allen Cochair: William Pine Major: Fisheries and Aquatic Sciences There is a growing concern over the lack of life history information for many deepwater fisheries species including golden tilefish, Lopholatilus chamaeleonticeps. Basic age and growth estimates are essential for proper stock assessment and management decisions M y first objective wa s to use lead radiu m dating to validate the timing of growth increments Radiometric ages closely agreed with age estimates from traditional age estimations (growth increment counts in thin sagittal otolith sections) for females and unknown sex fish. However, radiometric a ges did not agree with traditional age estimates for males. This difference may be attributed to differing growth rates and increment formation by gender or the transition of gender. Golden tilefish longevity of 26 6 yrs was confirmed using both methods. My second objective was to describe the reproductive parameters and strategy for golden tilefish along the east coast of Florida and the northern Gulf of Mexico. Male golden tilefish reached maturity at a younger age (male, <1 yr; female, 2.5 yr) and smaller size (male, 150 fork length [ FL ] mm; female, 331 FL mm) than females. Golden tilefish gonads are described as being intersexual identified as having nonfunctional (opposite) sex tissue ( functional males

PAGE 12

12 71% had multiple stages of atretic oocytes ; func tional females 26% had self contained male tubules ) Male gonads also contained a cavity that originated from ovarian lumen and sperm sinuses. Golden tilefish gonads exhibit ed intersexual tendencies and I found evidence support ing a protogynous he rmaphroditic reproductive strategy. My third objective wa s to apply the resulting life history characteristics (objectives 1 and 2) to two statistical age structure stock assessment models. Each of these models represented the ends of the spectrum of model complexity. The available data for golden tilefish stock assessment was typical for many species with irregular sampling for length and age composition data and highly variable indices of abundance. I completed a stock assessment for golden tilefish using the simple Stochastic Stock Reduction Analysis model and the more complex Stock Synthesis model T he two models agreed in stock status of golden tilefish (not overfished and not undergoing overfishing), as well as in historical annual biomass and e xploitation rates.

PAGE 13

13 CHAPTER 1 GENERAL INTRODUCTION Golden tilefish, Lopholatilus chamaeleonticeps (Goode and Bean, 1880), is a deepwater demersal fish found in the Atlantic Ocean from Nova Scotia to the Gulf of Mexico (Dooley, 1978). Golden tilefish are managed by three fishery management councils (Northeast Atlantic, South Atlantic, and Gulf of Mexico). In 2004, the first fishery management regulation in the Gulf of Mexico was established as an annual quota for golden tilefish of 200 metric tons. This regulation was initiated as part of the red grouper Epinephelus morio rebuilding plan to guard against commercial long liners shift to deep water groupers and tilefish fisheries. The regulation was not based on any biological basis ( NOAA, 2004). The lack of basic life history information and golden tilefishs unknown population status in the Gulf of Mexico motivated this research. Age, growth, mortality, and reproduction are a fe w of the most valuable life history data necessary for stock assessment. Earlier research on golden tilefish from the waters of southern New England, South and North Carolina and Georgia concluded that golden tilefish are a long lived fish reaching maximu m ages of up to 40 yrs, have a slow growth rate, display sexual dimorphic growth and mature at least by age 5 (Turner et al., 1983; Harris and Grossman, 1985; Grimes et al., 1986; Palmer et al., 2004). These studies provided initial data as a foundation f or my research. In my first chapter, I validated the timing of band deposition in sagittal otoliths using the natural decay of 210Pb and 226Ra and establish an accurate age estimation criteria for golden tilefish For some teleost s establishing the timing of band deposition can be imprecise given the effect of environmental conditions on the rate of absorption of calcium carbonate and other trace elements into the otolith matrix. Golden tilefish reside in a very specific habitat p referring a soft, but malleable sediment along the continental shelf in water depths 80400 m (Freeman and

PAGE 14

14 Turner, 1977), inhabit burrows (Able et al., 1982), and within a specific thermal cline (914oC; Grimes et al., 1986). My first chapter will be subm itted to a journal that publishes original studies on the ecology of fishes and their relationship with the environment. In t he second chapter I describe the reproductive strategy of golden tilefish. Teleosts exhibit a large diversity of reproductive st rategies from fish remaining the same sex throughout their life time (gonochoristic) to fish changing sex during their lifetime (hermaphroditic). This chapter will be submitted to a journal that focuses on general fish biology in particular, reproductive studies involving the classification of reproductive phases based on histology. Finally, in my third chapter, I used the age, growth, and reproductive information obtained in my research along with data collected from the golden tilefish fishery to deter mine the status of the stock in the Gulf of Mexico. Stock status was predicted using two statistical age structure d stock assessment models. I plan to publish this chapter in a journal that features original research affecting marine fisheries. My disse rtation provides the first description of life history and stock status of golden tilefish from the northern Gulf of Mexico. My research describes the methods to appropriately assign age to golden tilefish in the Gulf of Mexico and south Atlantic through validating the timing of band deposition in sagittal otoliths. This work also disputes golden tilefishs gonochoristic reproductive strategy by providing a comprehensive description of the histological classification of the maturity phases of males and females, the occurrence of non functional tissue in both mature, functional males and females, and histological evidence to describe golden tilefish as protogynous hermaphrodites. And finally, my discoveries of basic life history information of golden tilef ish are applied to two age structure stock assessment models to

PAGE 15

15 determine if the golden tilefish stock in the Gulf of Mexico is overfished or undergoing overfishing

PAGE 16

16 CHAPTER 2 AGE ESTIMATION AND L EADRADIUM DATING OF GOL DEN TILEFISH, LOPHOLATILUS CHAMAE LEONTICEPS1Introduction Age structured stock assessment models that rely on accurate estimates of fish age can provide misleading results leading to erroneous management decisions if fish ag e ing is imprecise or in validated. As an example, estimates of sizeat age are required for growth estimation, which is critical for diagnosing growth overfishing (Parma and Deriso, 1990), and estimates of longevity play an important role in calculating natural mortality (Hoeni g, 1983) and lifetime fecundity (Rochet, 2000). Thus, accurate age and growth estimates and age estimation criteri a are essential for a proper stock assessment. Validating the timing of band deposition i n otoliths is critical in determining longevity, a s well as an accurate rate of growth. Otoliths form rings but proving whether or not the ring (or band) is laid down annually or bi annually, due to some physiological or environmental factor can be difficult, especially for deep water species (Cailliet et al., 2001). There are several methods to validate the timing of band deposition such as marginal increment analysis, tag and recapture of known age d fish, chemically marking fish with recapture, and radiochemical dating (Campana, 2001). The focus of my dissertation is the golden tilefish, Lopholatilus chamaeleonticeps which is a deep water demersal fish found in the Atlantic from Nova Scotia to the Gulf of Mexico (Dooley, 1978). Previous research used both sagittal otoliths (Turner et al., 1983; Harr is et al., 2001; Palmer et al., 2004) and anal fin rays (Harris and Grossman, 1985) to assign an annual age 1The research in this chapter was completed in collaboration with A. Andrews (NOAA Fisheries Service, Pacific Fisheries Science Center, Honolulu HI). A. Andrews provided technical expertise on analytical methodology. I conceived the original project, aided in securing funding, drafted methods, assisted with sample collection, and interpretation of analytical results Research was funded by NOAA Fisheries Service, Marine Initiative Research Program (Andrews, 2009a).

PAGE 17

17 to golden tilefish caught in the Middle Atlantic Southern New England region and the U.S. South Atlantic waters off the coasts of North and South Ca rolina and Georgia. These studies revealed golden tilefish had longevities of 40 years (Palmer et al., 2004) with females more prevalent in the older age classes (Turner et al., 1983; Harris et al., 2001). These longevities were based on the results of m arginal increment analysis in thin sectioned sagittal otoliths (Turner et al., 1983) and thin sectioned anal fin rays (Harris and Grossman, 1985). Marginal increment analysis (MIA) is used widely in age and growth studies and involves recording, either digitally or manually, multiple measurements from age ing structure (otoliths, fin ray) measurements (e.g. otolith core to otolith edge, otolith core to each growth increment). But to properly validate the timing of band deposition through MIA several cri teria must be met, such as a random selection of fish collected over a long time period (e.g. including more than one deposition cycle), age structure measurements must be exact and not subjective given the instrumentation used to measure ( manually or dig itally ) and results should be considered age specific (Campana, 2001). Validating g olden tilefish band deposition in sagittal otoliths using MIA was obscured by low sample sizes (n < 5) per month per age group (only 8 age groups; ages 13, 4, 5, 7, 8, 9, 1017), and large confidence intervals among mean values (Turner et al., 1983). Nevertheless it was still concluded that golden tilefish had an annuli band formation with most opaque zone of annulus formation completed by June, even though younger and older fish exhibited different timing of band formation (June versus March, respectively Turner et al., 1983). The MIA study using golden tilefish anal fin rays also concluded an annuli band formation although the increment measurements were imprecise given skin still attached to the external ray, and in some ray sections the translucent bands along the margin may not ha ve represented a completed increment (Harris and Grossman, 1985).

PAGE 18

18 Due to the inconclusiveness of the MIA, radiochemical validation methods were applied to golden tilefish. One such method is bomb radiocarbon which relies on C14 levels emitted into the atmosphere during nuclear testing in the late 1950s (Campana, 1997). The increase in C14 levels in the ocean and subsequent C14 levels in corporated into carbonate structures (i.e., otoliths) can provide an independent estimate of age (Kalish, 1995). This method validated the timing of band deposition successfully for a variety of species inhabiting shallow water reefs red snapper, Lutjanus campechanus (Baker and Wilson, 2001), benthopelagic red fish Centroberyx affinis (Kalish, 1995), and continental shelf quillback rockfish Sebastes maliger (Kerr et al., 2005). However, r adioca rbon dating is a ffected by the amount of C14 available in the water column, which decreases below depths of 100 m (Stuiver and stlund, 1980). Earlier efforts to validate golden tilefish band deposition using bomb radiocarbon resulted in lower amounts of radiocarbon in otolith cores compared to haddock ( Melanogramm us aeglefinus ) and red snapper ( Lutjanus campechanus ) delta C14 reference curves (Harris, 2005). One possible reason for the inconclusiveness of the radiocarbon study is the depth at which golden tilefish reside and the absorption of atmospheric C14 at these depths (Stuiver and stlund, 1980) There is very little information regarding the golden tilefish early life history (larval and juvenile stages) but of the few larval golden tilefish that have been identified these were collected at depths > 1 00 m (pers. comm., D. Drass, NOAA Fisheries Service, Pascagoula, MS) a nd most adult tilefish reside at a minimum of 100 m (SEDAR, 2011a ). Similarly, lower radiocarbon levels detected in yellowedge grouper ( Epinephelus flavolimbatus ) compared to red snappe r radiocarbon levels were contributed to the depths preferred by these fish (Cook et al., 2009). This suggests that use of C14 for golden tilefish may not be successful.

PAGE 19

19 Therefore, in my dissertation a second method of radiochemical dating using the natural decay of lead (210Pb) and radium (226Ra) was applied. For fish, leadradium dating depends on the incorporation of radium 226 from the environment, where it is locked into the otolith matrix and subsequently decays to lead210. T he lead radium da ting can be used to provide an independent estimate of age, as well as a form of age validation for traditional age estimation methodologies (Smith et al., 1991; Panfili et al., 2002). Lead radium dating is a geochronological technique that has been used to date recent geological formations, such as accretionary carbonates (e.g. Condomines and Rihs, 2006). Use of this system as a chronometer relies on the decay of the relatively long lived radioisotope radium 226 (226Ra) to the relatively short lived gr anddaughter product lead210 (210Pb). Because the half life of radium 226 is much greater (~1600 years) than lead210 (22.26 years) the disequilibrium of the lead radium system can function as a natural chronometer as lead 210 builds into equilibrium with radium 226. The otolith leadradium system can be used to provide a radiometric estimate of age based on the disequilibrium within the first year or few years of growth (otolith core). This tool is unique because it is strictly regulated by the passage of time. Given a measured lead radium activity ratio from otolith core material, an age can be estimated within the margin of uncertainty from the measured quantities occurring from the natural decay of radium 226 to lead 210 (Smith et al., 1991; Panfili et al., 2002). This approach is an improvement over radiocarbon because it is not dependent on atmospheric C14 deposition, which may not be reliable at the depths tilefish are found. This kind of information can serve as a form of age validation for other age estimation methods (e.g., traditional growth zone counts; Andrews et al., 2009).

PAGE 20

20 My objectives were to establish an accurate traditional age estimation methodology using sagittal otoliths and to validate growth increments in golden tilefish sagittal otoliths using innovative radiochemical dating methods. It is expected that by validating the timing of band deposition in golden tilefish sagittal otoliths an accurate age assignment will provide new insight into golden tilefish life history such as age at maturity, fecundity, growth, and mortality. This is information necessary in the development of age based stock assessment models (as discussed in chapter 4) to make informed fishery management regulations. Errors in age estimation could lead to erron eous conclusions related to the stock status of golden tilefish that could have serious ecological or economical impacts. Materials and Methods Sample C ollection Federal port agents collected otoliths from golden tilefish harvested by commercial longline gear in 2007 along the east coast of Florida (Fig. 2 1). Port agents targeted twenty fish (for both male and female) per 100 mm size bins (300 1000 mm, fork length). Port agents provided field measurements of fish lengths (fork or total length, to 1.0 mm), weights (whole or gutted, to 0.1 kg), and removed both sagittal otoliths and gonads. Port agents also gathered information describing catch location (latitude, longitude, depth, or NMFS statistical shrimp grid) during dockside interviews. Sex was assigned using macroscopic description of whole gonads (female oval in shape, pinkish to red in color; male thin, taper to a point, normally white in color). Traditional A ge E stimation In order to compare the results of lead radium dating, actual counts of pairs of opaque and translucent growth bands (referred to as traditional age) were interpreted from thin sectioned sagittal otoliths. One otolith from each pair of golden tilefish sagittal otoliths was sectioned by taking a transverse section (dorso ventral plane) through the focus using a Hillquist petrographic -

PAGE 21

21 type saw (Cowen et al., 1995). A secondary reader and I aged the sectioned otoliths using transmitted lig ht with a stereo microscope. Th is method w as similar to th at of previous golden tilefish ageing studies (Fig. 2 2) (Turner et al., 1983; Palmer et al., 2004). Each reader made initial estimates of age by examining the otolith sections with a wide ranging magnification to fully investigate the quantifiable pairs of opaque and translucent growth bands. The secondary reader and I discussed our initial readings to establish traditional age estimation criteria The otolith sections were re read by each reader using the established criteri a Indices of precision (aver age percent error and percent agreement 1 and 2 bands ) ( Campana, 2001) were calculated to determine the accuracy of the ageing criteria between readers. Both readers agreed on the final age estimates that determined the otoliths used for lead radium dating. A third independent reader (experienced in ageing deep water grouper species) also aged the golden tilefish sections to determine if my ageing criteri a w ere consistent and repeatable. Otolith Growth Increment Deposition Growth increments in thinsectioned sagittal otoliths were described both q ualitative ly and quantitative ly to determine sexual dimorphic patterns of growth. I quali tatively characterized growth increment deposition as either horizontal (increments deposited horizontally along the ve ntral sulcus) or vertical (increments deposited vertically along the ventral axis). Each thin sectioned sagittal otolith was captured digitally using a digital camera ( Hitachi, Model KP D50 color ) connected to a desktop computer equipped with Image Tool s oftware (University of Texas Health Science Center in San Antonio: UTHSCA, Version 3.0). The distance between the core and the mid line of the completed growth increment (a pair of translucent and opaque bands) was measured along the ventral axis to quant itatively describe growth increment deposition.

PAGE 22

22 Lead radium D ating A ge G roups There are four major limitations of lead radium dating that must be considered before forming age groups. These limitations are: 1) core size and potential age of juvenile otoliths; 2) individual and collective otolith sample mass availability for juvenile and cored adult otoliths; 3) potential radium 226 activity; and 4) total sample age (estimated age plus time since capture). The approach in this study was similar to mos t other studies performed in that initial sample masses were chosen to provide a good indication of lead210 and radium 226 activity, given a best guess at the lowest case scenario. Typical radium 226 activity in otolith material is 0.03 to 0.05 dpmg1. Based on this estimate, a minimum of 0.5 grams of core material for each lead radium age group was targeted to collect sufficient counts at the alpha ( ) spectrometer, with a more optimal target of more than 1.0 g for each age group. Because of its rarit y in collections, only one juvenile otolith was available as a reference for adult otolith coring. Readers agreed the juvenile tilefish was 2 years of age and based on the dimensions and whole otolith weight of this otolith, the target of core extractions was approximately 8 mm x 4 mm x 1 mm (length x width x thickness) and a weight of 0.05 grams (g). Hence, the number of otoliths required for an adequate sample size of 0.5 g was ~10, with more than 10 providing more optimal conditions for measuring leadradium activities. After otoliths were chosen for each of leadradium age groups (based on traditional age estimation) each otolith was cored to the first 2 years of growth by: 1) grinding on a lapidary wheel; and 2) comparing the extracted core microscopically, as well as macroscopically, to the reference juvenile otolith mentioned above. It was necessary to core each otolith to isolate the lead radium activity to the first few years of the otoliths growth. Alternate pairs of opaque and translucent g rowth bands visible in the otolith were used to verify the concentric structure of each core to the first few years of growth. Careful consideration to the shape of the juvenile

PAGE 23

23 otolith was taken as the extraction took place. Due to the sensitivity of le ad radium dating to the amount of otolith material, otolith coring was completed under the supervision of A. Andrews at Moss Marine Laboratory (Moss Landing, CA). It was noted that the orientation of the otolith slowly rotated upward on the anterior end w ithin the first few years of growth. In addition, the shape of the juvenile otolith was slightly concaved to the distal side (Fig. 2 3). Both of these morphological characteristics were approximated in the core extractions, where a combination of microsc opic examination and measurements provided the best indication of when the core extraction was finished. Dry (room temperature) otolith core weights (similar to the reference juvenile tilefish otolith) were used as the final determination that core extrac tion was complete. In most cases, the cores needed finetuning with minor grinding and more microscopic and measurement observations. Finally, otolith cores were pooled into respective leadradium age groups and prepared for leadradium dating analysis. Lead radium D ating A nalysis Given the sophisticated laboratory equipment and specialized training needed to complete the lead radium dating, this analysis was conducted by A. Andrews at Moss Marine Laboratory (Moss Landing, CA). A detailed protocol descri bing sample preparation, chromatographic separation of radium 226 from barium and calcium, and analysis of radium 226 using mass spectrometry is described elsewhere (Andrews et al., 1999). The procedures used in this study were an improvement of the estab lished protocol: 1) by shifting the collection interval on the recovery; and 2) by using an improved ICP MS (Inductively Coupled Plasma Mass Spectrometry) technique to purified radium samples. Other than these details, only an overview of the radium 226 procedures is provided with details on the determination of lead210 activity. Standardized trace metal clean procedures and equipment were used throughout sample

PAGE 24

24 prep aration, separation, and analysis, since the levels of radium 226 and lead210 typically found in otoliths were extremely low (femtograms (10 15 g) for radium 226 and attograms (1018 g) for lead 210). Also, ultra pure, double distilled (GFS Chemicals) a cids and dilutions used in this study were made using Millipore filtered MilliQ water (18 MW cm1). Dried and weighed samples were dissolved in TFE beakers on hot plates at 90C by adding 8N HNO3 in 12 mL aliquots. Alternation between 8N HNO3 and 6N H Cl, with an aqua regia transition, several times resulted in complete sample dissolution. The dried sample, after dissolution, formed yellowish precipitate. To reduce remaining organics (otolin), and to put the residue into the chloride form required for the lead 210 activity determination procedure, samples were redissolved in 1 mL 6N HCl and taken to dryness five times at 90120C. A sufficient amount of organics w as removed once a whitish residue formed. It was these samples that were used to determi ne lead 210 activity prior to ICP MS analysis. The alpha decay of polonium 210 as a daughter proxy for lead 210 was used to determine lead 210 activity in the otolith samples. Since it takes two years for the alpha decay of polonium 210 and to ensure the activity of polonium 210 was due solely to the ingrowth from lead 210, only the cores of adult fish were used in analysis. Samples were prepared for polonium 210 analysis by spiking them with polonium 208 (a yield tracer). The amount of polonium 208 adde d was based on observed radium 226 levels in other studies of deepwater fishes (Andrews, 2009b). This amount was adjusted to about 5 times the expected polonium 210 activity in the otolith sample to reduce error in the lead 210 activity determination. These spiked samples were redissolved in approximately 50 mL of 0.5N HCl on a hot plate at 90C covered with a watch glass. The polonium 210 and polonium 208tracer was autodeposited for 4 hours onto a silver planchet. The activities of these isotopes were determined using a spectrometry on

PAGE 25

25 the plated samples. Additional procedural and system details are described elsewhere (Andrews et al. 1999). The remaining solution after polonium plating was dried and saved for radium 226 analyses on a Nu Plasma HR ICP MS located in the Department of Geology at University of Illinois Urbana, Champagne. Radiometric A ge E stimation Radiometric age was determined from the measured lead 210 and radium 226 activities. Since the lead radium activities were measured using the same sample, these calculations were independent of sample mass. Radiometric age was calculated using an equation derived from Smith et al. (1991) to compensate for the ingrowth gradient of lead210radium 226 in the otolith core, tage = T T e 1 R 1 Ra A Pb A 1 lnT 0 226 210 where tage = the radiometric age at the time of analysis, A210Pb = the measured lead 210 activity at time of analysis and reported as dpmg1 (disintegrations per minute per gram), A226Ra = the radium 226 activity measured using ICPMS (report ed as dpmg1), R0 = the activity ratio of lead 210 and radium 226 initially incorporated, l = the decay constant for lead 210 (ln(2).26yr1), and T = the estimated core age based on the first few growth zones. An initial uptake ratio of R0 = 0.0 was used, based on the close agreement of the measured juvenile age group leadradium ratio with the expected ingrowth curve; however, other studies have accounted for what appeared to be exogenous lead210 with minor adjustments necessary

PAGE 26

26 (e.g. Kastelle et al., 2000; Stransky et al., 2005). A radiometric age range was calculated, based on the analytical uncertainty, for each sample by using error propagation through to the final age determinations (2 SE, standard error). Calculated error included the standard sources of error (i.e. pipetting, spike and calibration uncertainties, etc.), alpha counting statistics for lead 210, and the ICPMS analysis routine. Radiometric versus T radition A ge E stimation Traditional age estimation (determined from otoliths growth band counts) and radiometric age estimation (determined from lead radium dating) were compared by lead radium sample age groups These estimations were also compared to the expected 210Pb:226Ra ingrowth curve. Age agreement, or disagreem ent, between the two methods in terms of potential ageing bias (95% confidence intervals, 2 SE) was given consideration. Results Sample C ollection Port agents collected a total of 223 golden tilefish (female, n = 95; male, n = 101; unknown sex, n = 27). The sex of tilefish was determined by examining whole gonads The smallest fish was a male (389 mm FL) and the largest fish was of unknown sex (935 mm FL). Of the eight size bins, port agents sampled four size bins (400799 mm FL; Fig. 24) sufficiently (sample size > 20 per size bin). There were difficulties in collect ing small tilefish due to gear selectivity and a lack of larger fish (800+ mm FL ) caught by the fishery. Traditional A ge E stimation Readers determined golden tilefish thin sectioned sagittal otoliths were difficult to interpret given several different shapes of otolith sections and diverse patterns of band deposition (Fig. 2 5). Rea ders interpreted otolith sections using a magnification of 10 20x with a stereo microscope aided by transmitted light. Initial reader agreements resulted in 18% average percent

PAGE 27

27 error (APE), 8% reader agreement, and reader agreement 1 band 22% and 2 bands 26%. The traditional age estimation criteria involved 1) counting alternating pairs of thin opaque and translucent bands that were present both along the ventral axis and ventral sulcus, 2) interpreting otolith sections at no more than 20x magnifica tion, and 3) taking the shape (stocky or elongated) of the otolith section into consideration when interpreting the distance between bands. After the agreement of an ageing criteri a reader agreement increased to 28%, APE lowered to a more acceptable valu e of 6% (similar to other long lived fish aged at the NOAA Fisheries Service, Panama City Lab oratory, Panama City, FL ) with percent agreements 1 band and 2 bands increased to 78% and 95%, respectively. Final age estimates used for further analysis we re based on the agreed pattern of growth increment recognition between readers. Readers determined tilefish to be age 4 32 years old, with females occurring in most age groups (5 32 years; Fig. 2 6). Otolith interpretations were completed by an addit ional third reader and resulted in similar age estimates (female age 5 25, male age 4 12) and similar precision (APE 7.5%, percent agreement 1 band 25%, percent agreement 2 bands 63%). Otolith Growth Increment Deposition Golden tilefish otoliths s howed sexual dimorphic di fferences in otolith morphology. Adult females tended to deposit growth increments in a manner that thickened the otolith horizontally on the proximal side (sulcus) (Fig. 2 5A ) Males deposited growth bands in a more vertical, longitudinal manner (Fig. 25B ) Golden tilefish also showed sexual dimorphic growth increment deposition. Average growth increments were similar between genders until age 8, thereafter males deposited increments at larger distances (Fig. 2 7). Additionally, some male sagittal otoliths showed an unusual pattern of growth increment deposition; the first four bands composed of several thinner, darker increments followed by fainter increments deposited in more even groupings (Fig. 28).

PAGE 28

28 Lead radium D ating A ge G roups Twelve age groups were formed for the lead radium analysis of golden tilefish otoliths (Table 2 1). Estimated age for the series of groups ranged from 57 years to 2228 years with sexes separated for the first two age groups. Within the separatesex age groups, each was randomly split into sample replicates, resulting in 2 male and 2 female groups for both 57 year and 1013 year age groups. Randomization led to age ranges that differed slightly in range (i.e., GTL 5 7 M1 = 67 years and GT L 10 13 M1 = 10 12 years); however, the average estimated age was relatively consistent at 6.2 to 6.4 years for the 57 yr groups, and 10.6 to 11.4 years for the 1013 yr age groups. The older age groups could not be separated by sex because of a lack of reproductive tissue collected and the low sample availability. Two 1519 year groups were separated by otolith weight into a low (A) and high (B) weight classification. A similar situation was the result for the oldest age groups, where randomization led to slightly different age ranges, the low weight age group range from 2027 years and the higher otolith weight group ranged from 22 28 years. Fish length was lowest on average for the 5 7 yr female age groups and range up to more than 750 mm FL on ave rage for the oldest age groups ( Table 2 1) In general, the length of fish increased with estimated age, but some age groups included some rel atively large individuals (e.g., GTL 5 7 M1, maximum length = 860 mm FL ; Table 2 1). Whole otolith weight was al so lowest for the 57 yr female age groups at below 0.4 g on average. Otoliths for adults were massive and exceeded 1.7 g on average for the 22 28 yr age group. For the younger randomized age groups, average otolith weight was lower for females relative to males in all cases. Hence, relative to estimated age, females were typically smaller with lighter otoliths for the first two age groups (57 yr and 1013 yr). For the two older age groups, the heaviest 1519 yr group (GTL 1519 B) was similar on avera ge for both fish size and otolith weight to the 2027 yr age group.

PAGE 29

29 The oldest age group (GTL 2228) exceeded almost all groups in terms of otolith weight both on average and in range; only GTL 15 19 B included otoliths that overlapped in weight. Lead r adium D ating A nalysis Lead radium determinations were made for all 12 agegroup samples (Table 22). The number of otoliths that made up each age group ranged from 9 to 21 for total sample weights of just over 0.5 g to greater than 1.5 g. Mean core weig ht was slightly greater than the target weight of 0.05 g and ranged from approximately 0.063 g to 0.081 g per core. This was anticipated to an extent because the cleaning process usually removes ~5% of the external material, but in this study removed only 1 3%. To make a better approximation on the representation of core age in radiometric modeling, 3 years was used in lieu of 2 years as the core age since the juvenile otolith used for reference contained growth pass the end of the second annulus formation. Lead 210 activity increased as expected from the youngest to the oldest age groups by a factor of ~3 times. Samples ranged from 0.00419 13.8% dpmg1 for GTL 5 7 F1 to 0.01359 6.7% dpmg1 for GTL 22 28. Radium 226 activity was lower than expected for the region by a factor of about 2 to 5 times. The activity ranged from an average of 0.01413 22.6% dpmg1 for the GTL 10 13 M1 groups to 0.02573 16.1% dpmg1 for GTL 22 28. Radium 226 recovery was low for a few samples (the most massive samples may have overloaded the Sr column with barium causing radium recovery to suffer) and the runs were deemed unreliable (error greater than 30% and close to background counts on the ICP MS). To re cover the opportunity to age the samples, an average of all measured values with less than 25% error (13 23%) was used in place of the unreliable values (n = 6; 0.01989 15.7% dpmg1); this replacement occurred for GTL 5 7 M1, M2, and GTL 1013 F2. For the other randomly split age groups an average of the measured values for the pair was used as the radium 226 activity for the groups. For the older age groups that were not randomly split, sample specific radium 226 activities were used.

PAGE 30

30 Radiometric versus T radition A ge E stimation Both the total sample age (estimated age plus time since collection) and age corrected for time since collection were calculated to compare radiometric age with traditional age estimated from growth zone counting (Table 23). Agreement of the total sample age with the ingrowth curve was good for some sample groups and markedly different from what was expected for other groups (Fig. 2 9). Sample measurements to the left of the ingrowth curve were underestimated for age usin g growth increment counting, which was most apparent with the male groups especially GTL 10 13 M1 and M2. Young female and the older mixed age groups radiometric and traditional age estimates were largely in agreement, but the 15 19 yr groups age were sl ightly underestimated by a few years on average. No apparent trend related to otolith weight was discernible based on the similarity of radiometric age determinations for the oldest groups (GTL 1519 A and B, GTL 2027, and GTL 2228). Discussion Longevi ty, the maximum number of years of life, is related to natural mortality. If a fish lives a very long time (longevities in terms of multiple decades), the overall stock will be less productive. This phenomenon has been best documented in very long lived (centennial) fish such as orange roughy ( Hoplostethus atlanticus ) and numerous species of rockfish (Mace et al., 1990; Musick, 1999). I am very confident in my estimates of longevity for golden tilefish from the east coast of Florida, since lead radium da ting and traditional age estimations were in good agreement in estimating longevity (lead radium 26 6 yrs, traditional average age 25.2 3 yrs). My estimates of longevity were very similar to those published for golden tilefish from other ageing studie s (35 yr, Turner et al., 1983; 33 yr, Harris and Grossman, 1985; 34 yr, Harris et al., 2001; 40 yr, Palmer et al., 2004). Not only have I proven longevity using the natural decay of radium 206 to lead 210 but I have also provided a more precise method to interpret growth

PAGE 31

31 bands in thin sectioned sagittal otoliths. Even though radiometric and traditional age estimation methods were in agreement in the oldest age groups, these methods did not agree in the male age groups, with ages differing by 515 years. The differences in traditional age versus lead radium dating in male tilefish may be due to : 1) the uptake of 226Ra differed from the mean values used, 2) growth rates or 3) the pattern of growth increment deposition differed by gender. It is important to recall that lead radium dating depends on the level of radium 226 uptake. Radium 226 levels in the otolith cores were lower than expected (typically 0.3 to 0.5 dpmg1) and this wa s difficult to explain in the broader context of measured values from other lead radium dating fish studies to date (Andrews, 2009 b). The flux of radium 226 is typically greatest near continental margins and sea floors with low sedimentation rates (Broker and Peng, 1982), or from nearshore environments like the coastal waterwa ys of Florida (Fanning et al., 1982). However, seasonal fluctuations of radium isotopes along the southeastern U.S. continental shelf (Moore, 2007) and the decrease of 226Ra by depth and distance from shore (W. Moore pers. comm., University of South Carol ina, Columbia, SC) could account for these lower radium levels. In addition, little or no information exist for the early life stages of golden tilefish and commercial fisheries typically catch tilefish > 400 mm FL even though there is no size limit (Fre eman and Turner, 1977). The only exception on record was with the 210 mm FL juvenile that was recovered from the mouth of moray eel by an observer. Small fish may be avoiding the gear because most commercial long liners use large circle hooks (15/013/0) or may inhabit a different area than where fishing occurs ( depths > 200 m ) (Freeman and Turner, 1977; Hale, 2011). It is assumed male and female juveniles settle in similar habitats and are exposed to similar levels of radium; therefore, the intake of radium should be the same regardless of sex.

PAGE 32

32 Because of the low values, and i n some cases poor radium recovery, the activity of radium 226 suffered for some samples. Counts at the ICP MS were close to background making the end result unreliable. This led to the use of a mean activity value, in lieu of the unreliable measured valu es to recover radiometric age determinations for a few groups. While this could be considered a weak point in the findings, the overall determination of radiometric age from the mean values was consistent with the relatively accurate lead 210 and radium 226 activity values, both in terms of activity and group age. Historically, mean radium 226 values were used in successful lead radium dating studies, typically providing lower precision and an additional assumption for radium 226 uptake (e.g., Bennett et al., 1982; Campana et al., 1990; Fenton et al., 1991). While it is desirable to acquire much greater age estimat ion accuracy, the relatively low precision from radium 226 uncertainty in the golden tilefish otolith cores did not preclude an age determinatio n within the reported margin of error. Hence, it is unlikely that the differences in age observed for the male age groups were not the result of analytical error and that there were actual differences in the age and growth of adult male fish. Golden tile fish otoliths had structural differences in otolith morphology that suggest variations in band deposition and growth rate by gender. Male golden tilefish do grow faster and obtain much larger sizes than female golden tilefish (see chapter 3). In the foll owing chapter, I provide evidence that golden tilefish are protogynous hermaphrodites. The differences in growth, differences in deposition of growth increments and morphology of sagittal otoliths between the genders may be due to some females transitioni ng into males. In each of the male sample groups used for leadradium dating, at least 70% of the male gonads contained female gametes of varying degree of development (see chapter 3). This transition may also be the reason for different patterns of grow th increment depositions in male sagittal. Nevertheless, I am

PAGE 33

33 confident that the lead radium analysis was not obscured by the differences in band deposition by gender since the extracted core material composed of the first t hree years of otolith growth, wh ich was similar by gender The comparison of traditional age estimates and radiometric age estimates relies on the accuracy and precision of the traditional ageing methodology. Accuracy refers to the similarity of the traditional age estimate and actual a ge (based on a validation method) and precision is the repeatability (or consistency) of the method of growth band interpretation (Campana, 2001) Golden tilefish thin sectioned sagittal otoliths showed a variety of different patterns in growth increments that were difficult to interpret. My traditional ageing methodologies were based on previous published methods for ageing golden tilefish sagittal otoliths (Turner et al., 1983; Palmer et al., 2004), along with the integration of methods used for interpre ting yelloweye rockfish ( Sebastes ruberrimus ; Andrews et al., 2002). The difficulty in the interpretation of growth increments in golden tilefish sagittal otoliths was also one of the reasons for the inconclusiveness of the radiocarbon validation study (H arris, 2005). Although my traditional age estimates were in good agreement for the female and unknown sex age groups, the disagreement in the male age groups is most likely due to the underageing and misinterpretation of the growth increments in male gold en tilefish rather than the results of lead radium analysis. My traditional age estimations may have been less accurate due to the different patterns of growth increments in primary (some males deriving from a male immature phase) and secondary (males tra nsitioning from females) male golden tilefish (Sadovy and Shapiro, 1987; Walker and McCormick, 2009). The application of leadradium radiometric dating relies on several assumptions 1) the otolith is a closed system with no loss or gain of 226Ra, 2) the update of exogenous 226Ra is negligible, and 3) the update of 226Ra is in constant proportion to the otolith mass growth rate

PAGE 34

34 (Kimura and Kastelle, 1995). My study addressed each of those assumptions by using an improved technique (Andrews et al., 1999) tha t used otolith cores (as well as the first years of growth) so that any difference in 226Ra intake given the changes in growth rate of the otoliths was eliminated. Also, the usage of advances in thermal ionization mass spectrometry (TIMS) were applied to reduce error and subsequently increasing analytical precision of small quantities of 226Ra isotopes through the use of improved ICP MS. And finally, in addition to these new techniques, precautions were made to minimize contamination during all sample processing (e.g. double distilled acids). Conclusion Determining sustainability in fish stocks relies on estimates of growth, age at maturity, longevity, natural mortality, and recruitment variability; all of which rely on an accurate estimate of age. In some fish, for example splitnose rockfish ( Sebastes diploproa; Bennett et al., 1982) and Patogonian toot hfish (Dissostichus eleginoides ; Andrews, 2009b) traditional age estimations are difficult even with successful validation studies. In situations where age estimates are both imprecise and bias ed an ageing error matrix can be incorporated into the modeli ng process (Punt et al., 2008; Gertseva and Cope, 2011; Candy et al., 2012). It is recommended that a reference collection of known ages be routinely read by multiple readers from multiple ageing facilities to fully capture the imprecision and bias associ ated with traditional ageing estimations into the ageing error matrix. The reference collection needs to include samples that fully represent the range of ages (especially the older fish) and with sufficient sample sizes per age class to enable appropriat e statistical analysis (Campana, 2001; Punt et al., 2008). Therefore, given the inconclusiveness of validating each of the age and gender groups for golden tilefish, I recommend the use of an ageing error matrix in assessment models to incorporate the unc ertainty in traditional age estimates (as is used in following chapter 4 in Stock Synthesis)

PAGE 35

35 Table 2 1. Summary of estimated age with fish and otolith characteristics for golden tilefish age groups from the east coast of Florida processed in this study. Estimated age composition, fork length (FL) and whole otolith weight are given. The sex composition for samples GTL 5 7 and 1013 was signified by an M or F in the sample number. Sexes were mixed or unavailable for the older age groups. Age group & sa mple number Age r ange (yr) Average age (yr) Average FL (mm) FL (mm) r ange (2 SE) Average otolith wt. (g) Otolith weight range (2 SE) GTL 5 7 M1 6 7 6.4 549 436 860 ( 39) 0.470 0.348 0.596 ( 0.034) GTL 5 7 M2 5 7 6.3 520 465 620 (17) 0.453 0.330 0.663 (0.030) GTL 5 7 F1 5 7 6.4 483 402 535 (28) 0.391 0.277 0.504 (0.046) GTL 5 7 F2 5 7 6.2 465 412 508 (18) 0.366 0.270 0.488 (0.039) GTL 10 13 M1 10 12 10.6 683 494 796 (37) 0.889 0.681 1.198 (0.079) GTL 10 13 M2 10 13 10.8 729 570 810 (40) 0.964 0.630 1.261 (0.095) GTL 10 13 F1 10 13 11.1 613 524 660 (22) 0.776 0.592 0.964 (0.071) GTL 10 13 F2 10 13 11.4 620 570 725 (24) 0.757 0.589 1.063 (0.073) GTL 15 19 A 15 19 16.8 699 620 780 (31) 1.024 0.899 1.124 (0.037) GTL 15 19 B 15 19 16.4 750 643 842 (36) 1.345 1.139 1.628 (0.095) GTL 20 27 20 27 22.5 731 680 824 (28) 1.287 1.073 1.521 (0.086) GTL 22 28 22 28 25.2 763 700 900 (43) 1.710 1.541 2.002 (0.096)

PAGE 36

36 Table 2 2. Coring and radiometric results for golden tilefish age groups from the east coast of Florida processed in this study. The tot al number of otoliths (n), age group weight and the average core weight are listed with the measured 210Pb and 226Ra activities fo r the samples ( 2 SE). Calculated activity ratios and their corresponding margin of error were used to calculate sample age and uncertainty (Table 2 3). Age group & sample number n Sample Weight (g)1 Average c ore weights (g)2 210 Pb (dpm g 1 ) % error3 226 Ra (dpm g 1 ) % error3 210 Pb: 226 Ra (2 SE) GTL 5 7 M1 20 1.52521 0.07626 0.00491 7.8% 0.01989 15.7% 4 0.246 86 (0.043 19 ) GTL 5 7 M2 21 1.57928 0.07520 0.00589 6.1% 0.01989 15.7% 4 0.296 13 (0.049 72 ) GTL 5 7 F1 9 0.57104 0.06345 0.00419 13.8% 0.02478 18.2% 5 0.169 22 (0.031 87 ) GTL 5 7 F2 10 0.70606 0.07061 0.00564 9.9% 0.02478 18.2% 5 0.227 58 (0.036 90 ) GTL 10 13 M1 15 1.02671 0.06845 0.00783 6.6% 0.01413 22.6% 5 0.553 85 (0.09596 ) GTL 10 13 M2 14 1.03334 0.07381 0.00737 6.8% 0.01413 22.6% 5 0.521 66 (0.090 78 ) GTL 10 13 F1 11 0.75149 0.06832 0.00593 8.9% 0.02083 13.1% 6 0.284 65 (0.045 07 ) GTL 10 13 F2 11 0.89425 0.08130 0.00766 7.0% 0.01989 15.7% 6 0.385 18 (0.06 599 ) GTL 15 19 A 12 0.79502 0.06625 0.00866 7.1% 0.01758 23.2% 6 0. 49260 (0. 04396 ) GTL 15 19 B 10 0.67835 0.06167 0.00862 8.2% 0.01917 14.1% 6 0. 44981 (0. 04785 ) GTL 20 27 10 0.67733 0.06773 0.00886 7.8% 0.01788 20.1% 6 0.4955 1 (0.1067 2 ) GTL 22 28 10 0.68193 0.06819 0.01359 6.7% 0.02573 16.1% 6 0.5283 2 (0.092 15 ) 1 Cleaned and dried sample weight prior to processing. 2 Extracted otolith cores after cleaning. 3 Calculation based on propagation of 2 SE using the delta method (Knoll 1989) and the ICPMS analysis routine ( 2 SE). 4 Poor radi um recover from original sample led to use of an average radium activity for all sample specific measurements below 25% error. 5 Average radium 226 for sample replicates. 6 Sample specific radium 226.

PAGE 37

37 Table 2 3. Comparison of estimated age and radiometric age for golden tilefish from the east coast of Florida processed in this study Total sample age is given as the average age plus the time since capture for direct comparison with radiometric age. Ra diometric age was calculated from the measured 210Pb:226Ra activity ratios and age range was based on the analytical uncertainty and error propagation ( 2 SE). Corrected age for time since capture date was calculated for direct comparison with the estima ted average age of the age group. Age group & sample number Average age (yr) Time since capture (yr) Total sample age (yr) Radiometric age (range; yr) Corrected age (range; yr) GTL 5 7 M1 6.4 1. 3 7.7 10.6 (8.8 12.5) 9.3 (7.6 11.2) GTL 5 7 M2 6.3 1.3 7.6 12.8 (10.6 15.1) 11.5 (9.3 13.8) GTL 5 7 F1 6.4 1. 4 7.8 7.5 (6.3 8.7) 6.1 (4.9 7.4) GTL 5 7 F2 6.2 1. 2 7.4 9.8 (8.3 11.4) 8.6 (7.1 10.1) GTL 10 13 M1 10.6 1.4 12.0 27.4 (21.2 35.2) 26.0 (19.8 33.8) GTL 10 13 M2 10.8 1.4 12.2 25.2 (19.6 32.0) 23.8 (18.2 30.5) GTL 10 13 F1 11.1 1. 4 12.5 12.3 (10.3 14.4) 10.9 (8.9 13.0) GTL 10 13 F2 11.4 1.4 12.8 17.1 (13.9 20.8) 15.7 (12.4 19.4) GTL 15 19 A 16.8 1. 5 18.3 23.3 (16.5 31.9) 21.8 (15.0 30.4) GTL 15 19 B 16.4 1. 5 17.9 20.7 (16.7 25.3) 19.2 (15.0 23.8) GTL 20 27 22.5 1. 5 24.0 23.5 (17.3 31.1) 22.0 (15.9 29.7) GTL 22 28 25.2 1.4 26.6 25.6 (19.9 32.6) 24.2 (18.5 31.2)

PAGE 38

38 Figure 2 1. Map of collection area for golden tilefish. The shaded area indicates where golden tilefish otolith samples were collected and used in the lead radium analyses. Golden tilefish were intercepted from commercial long line vessels fishing along the 200 m contour off the east coast of Florida.

PAGE 39

39 ( A ) ( B ) Figure 2 2. Images of thin sectioned golden tilefish sagittal otoliths from previous ageing studies Displayed are the respective age estimations ( A ) 8 yr old male golden tilefish (77 cm fork length ) (Turner et al., 1983) and ( B ) 7 yr old golden tilefish (Palmer et al., 2004). An increment consisted of one translucent and one opaque zone (reflected light, Turner et al., 1983; transmitted light, Palmer et al., 2004) Photographs copied from citations.

PAGE 40

40 ( A ) ( B ) ( C ) Figure 2 3. Image of the juvenile golden tilefish otolith ( A ) whole, ( B ) thin sectioned (transmitted light) and ( C ) extracted core. 1 mm 25x 1 mm 15x

PAGE 41

41 Figure 2 4. Length frequency of golden tilefish. Fish were collected from commercial long line fishery in 2007 from the east coast of Florida (male, black bars; female, white bars; unknown sex, gray bars). 0 5 10 15 20 25 30 35 40 45 300 399 400 499 500 599 600 699 700 799 800 899 900 999 Number Fork length (mm) male female unknown

PAGE 42

42 ( A ) ( B ) Figure 2 5. Thinsectioned sagittal otoliths of golden tilefish Otolith sections were viewed with transmitted light (15x) ( A ) female (age 26 yrs, 734 mm FL) and ( B ) male (age 10 yrs, 758 mm FL) Traditional age was determined by interpreting opaque increments along ventral axis (solid lines) and along ventral sulcus (d ashed lines)

PAGE 43

43 Figure 2 6. Age frequency of golden tilefish. Fish were collected from commercial long line fishery in 2007 from the east coast of Florida T raditional age estimation (counting otolith growth zones) for male (black bars ), female ( white bars ) and unknown sex ( gray bars). 0 5 10 15 20 25 30 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 Number Traditional age (yr) male female unknown

PAGE 44

44 Figure 2 7 Otolith increment counts and band increment measurements by gender for golden tilefish from the east coast of Florida Band increments were measured from the core to the mid line of the increment fo r males (solid black circles) and females (solid white circles) and reported as mean standard deviation. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Band Increment (mm) Increment count male female

PAGE 45

45 Figure 2 8. Examples of thinsectioned male golden tilefish sagittal otoliths. ( A ) age 6 yr, 530 mm FL and ( B ) age 11 yr, 810 mm FL depicting changes in growth increment deposition possibly due to the transition of gender. Otolith sections were viewed using a stereo microscope with transmitted light (15x). Traditional age was determined by interpreting opaque i ncrements along ventral axis (solid lines) and along ventral sulcus (dashed lines) A B

PAGE 46

46 Figure 2 9. M easured 210Pb:226Ra ratios with respect to average traditional age for golden tilefish samples processed in this study M ales ( solid black circles ), females ( solid white circles ) and unknown sex ( solid gray circles), plotted with the expected 210Pb:226R ingrowth curve (solid line) and 2 year ingrowth compensation (dotted line). Horizontal error bars represent 2 standard errors (SE) around the mean of the sample growth zone age. The vertical error bars represent the analytical uncertainty associated with measuring the 210Pb:226Ra ratio (2 SE). 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 10 20 30 40 50 60 70 80 90 100 110 120 Lead 210:radium 226 activity ratio Traditional Age (yr) male Female unknown

PAGE 47

47 CHAPTER 3 EVIDENCE FOR HERMAPHRODITISM IN GOLDEN TILEFISH ( LOPHOLATILUS CHAMAELEONTICEPS ) Intro duction Teleosts can be described as being gonochoristic or hermaphroditic ( Sadovy de Mitcheson and Liu, 2008). Gonochores are individuals that remain the same sex their entire lives and hermaphrodites spend some proportion of their lives as either sex. Hermaphrodites can be further described as either simultaneous (contain both functional and mature ovarian and testicular tissue) or sequential. There are two types of sequential hermaphrodites, protogynous (functional female tissue replaced by functiona l male tissue) or protandrous (functional male tissue replaced by functional female tissue). All hermaphrodites can also be described as intersexual, having both male and female germinal tissue (Atz, 1964). However, just describing the presence of non fu nctional (opposite) sex tissue within an individuals gonad does not necessarily mean that individual is a functional hermaphrodite. It is necessary to prove, histologically, that the individual functioned as both sexes during its lifetime to determine a hermaphroditic reproductive strategy ( Sadovy de Mitcheson and Liu, 2008 ; Lowerre Barbieri et al ., 2011). The focus of this study is the golden tilefish, Lopholatilus chamaeleonticeps (Goode and Bean, 1880). Tilefishes are classified into two families (Branchiostegidae, Malacanthidae) and five genera ( Caulolatilus, Lopholatilus, Branchiostegus, Malacanthus, Hoplolatilus ) (Dooley, 1978). The tilefish families, like most teleosts (Atz 1964), exhibit a large diversity of reproductive strategies. T ilefish families are described as being gonochoristic ( Caulolatilus priceps ; Elorduy Garay and Ramirez Luna, 1994) to having intersexual gonads ( Caulolatilus microps Ross and Merriner, 1983; Branchiostegus japonicas Watanabe & Suzuki, 1996) and

PAGE 48

48 protogynous hermaphrodites with polygyny mating systems ( Malacanthus plumieri Baird, 1988). Previous reproductive studies have classified golden tilefish as gonochoristic but these studies also detected a small percentage of male gonads contained previtel l ogenic oocytes (1%, Erickson and Grossman, 1986; <0.1%, Grimes et al., 1988; < 0.01 %, Palmer et al., 2004). The presence of early developed oocytes in testes is not uncommon in fish with a g onochoristic reproductive strategy, especially if the species exhibits a bisexual (non functioning) juvenile stage (Sadovy and Shapiro, 1987). Furthermore, golden tilefish male gonads did not contain oocytes in a more advanced stage (or atretic, degeneratin g oocytes) nor did male gonads contain an ovarian lumen, characteristics that suggest a male gonad originated from a female gonad ( Erickson and Grossman, 1986; Grimes et al., 1988; Palmer et al., 2004) T here are several criteria to successfully prove a protogynous hermaphroditic reproductive strategy (Sadovy and Shapiro, 1987). These criteria rely on the histological analysis of gonadal tissue from functional male s and females Three criteria are specific to functional males: 1) a lumen (or membrane lined cavity) remains unused in testes, 2) late stage (cortical alveolar, vitellogenic, hydrated) and atretic (or degenerating) oocytes are present in testes, and 3) sperm sinuses a re present along the testes wall. The final criterion is the identification of gonads in a transitional stage, which is a gonad containing degenerative tissue of one sex with developing functional tissue of the opposite sex Hermaphroditism is difficult to prove and can be misdiagnosed. Therefore, it is necessary to not only follow the criteria above but also to collect fish from a large range of sizes and from throughout the year (Sadovy and Sharpiro, 1987; Sadovy and Domeier, 2005).

PAGE 49

49 My objectives of t his chapter are: 1) describe the reproductive phases by sex, 2) determine the reproductive seasonality, 3) estimate the age and size at maturity and 4) apply the criteria described by Sadovy and Shapiro (1987) to classify the reproductive strategy of gold en tilefish collected along the east coast of Florida and throughout the northern Gulf of Mexico. This is information necessary in the development of age based stock assessment models (as discussed in chapter 4) to make informed fishery management regulat ions (SEDAR, 2011a ) Misidentifying the reproductive strategy of a species can be troublesome, especially if the measure of spawning stock biomass does not consider sperm limitation with the removal of the largest, typically male fish from the population (Alonzo and Mangel, 2004). Materials and Methods Sample C ollection Golden tilefish were collected by NOAA Fisheries Service, Southeast Fisheries Science Center (SEFSC) sampling programs throughout the northern Gulf of Mexico and east coast of Florida (2 000 2009). Since most golden tilefish landed by the commercial fishery are thoroughly gutted at sea, Trip Interview Program port agents made special requests to willing commercial captains to collect whole fish to obtain gonad tissue. At sea observers on board commercial long line vessels also provided biological samples (otoliths and gonads) Two fishery independent surveys provided additional golden tilefish biological samples (NOAA/SEFSC Pascagoula, MS; Cooperative Research Project 08CRP009). Both of these longline surveys used a standardized sampling design with sets chosen randomly based on depth (see Grace et al., 2 004 for complete description). Otolith Processing and Assigning Age The sagittal otolith was used as the primary ageing structure ( se e chapter 2). Sagittal otoliths were sectioned using a Hillquist thin sectioning saw, and sections were viewed using a

PAGE 50

50 stereomicroscope with reflected light. The traditional age estimation criterions, as described in chapter 2, were applied to the otoliths associated with fish collected with gonads. Two readers (myself the primary reader and a secondary reader) interpreted the otolith s and indices of precisi on (Average Percent Error, Percent Agreement, Coefficient of Variation) were calculated ( Campana 2001). All fish were assigned an annual age equal to the annulus count by convention. Gonad P rocessing Golden tilefish gonads were weighed to the nearest 0.1 g and fixed in 10% neutral buffered formalin for a minimum of two weeks. Preserved gonads were subsampled along the posterior anterior axes of the gonad and a small subsample (1 cm3) was removed and placed in a cassette for histological processing. His tological processing of golden tilefish gonads were prepared by the Louisiana State University School of Veterinary Medicine, Histopathology Laboratory in Baton Rouge, LA. Tissues were embedded in paraffin, sectioned to a thickness of 4mounted on g lass slides and stained with hematoxylin 1 and eosinY following standard histological procedures. Gonad D evelopment Golden tilefish histological slides were viewed using a compound microscope at 40 to 400x magnification to determine the functional sex, r eproductive phase, and the reproductive strategy (gonochoristic or hermaphroditic)1 1 Histological slides were interpreted by H. Lyon (NOAA Fisheries Service, Panama City, FL), given her expertise in interpreting both gonochoristic and hermaphroditic gonads. Functional sex is defined by the presence and prevalence of leading (most advanced) gamete stage identified in the gonad, following Sadovy de Mitcheson and Liu 2008 (Tabl e 31 and 32) Male and female gonads were classified as immature or mature given the dominate stage of spermatogenesis or oocyte development,

PAGE 51

51 respectively (Table 33 and 34). The structural and germinal features of male and female gonads were describe d to meet the criteria, as described by Sadovy and Shapiro (1987), to determine whether or not golden tilefish exhibit characteristics of hermaphrodites. Female gonads were staged using oocyte developmental characteristics (Wallace and Selman, 1981; Hunter and Macewicz, 1985; Tyler and Sumpter, 1996; McMillan, 2007) and assigned to a phase of reproduction (Table 3 4) based on the leading gamete stage, indicators of prior spawning and short term atresia. Evidence of prior spawning is described by the number of brown bodies (macrophages), presence of old hydrated oocytes, stage of atresia, the condition of the muscle bundles, presence of connective tissue, appearance of lamellae in the gonad tissue, and thickness of gonad wall. Golden tilefish females with d eveloping, active, spawning or resting gonads were considered sexually mature. Females that possessed cortical alveolar oocytes were considered mature only if indicators of prior spawning were present (Rideout et al., 2000; Rhodes and Sadovy, 2002). As w ith females, male maturation was based on the dominant gamete stage. Golden tilefish males with all stages of spermatogenesis were considered sexually mature (Table 3 3) Male and female reproductive phases were based on standardized terminology used in teleost reproductive literature (Brown Peterson et al., 2011). The criteria described by Sadovy and Shapiro (1987) were used to classify the reproductive strategy (gonochoristic, hermaphroditic) of golden tilefish. These criteria are based on the configu ration of germinal tissue within functional male and female gonads: 1) determine the existence of a cavity within the germinal tissue and if the cavity is membrane lined in both male and female gonads, 2) determine the existence of transitional individua ls where gonads contain degenerative tissue of one sex and developing tissue of the opposite sex, 3) describe the present or absence and stage of development of nonfunctional (opposite) sex tissue and classify the

PAGE 52

52 placement of the non functional tissue ( floater, embedded, attached, in duct, or a combination of placements), and 4) describe the presence or absence of sperm sinuses and their location within male germinal tissue. Estimates of M aturity Size and age at maturity were calculated using a logisti c regression model : Yi = ( exp ( a + (b Xi)/(1 + exp (a + (b Xi)), where Yi = proportion mature at length or age (Xi), a = intercept, and b = slope of the logistic regression. The logistic regression model was fit to binomial maturity data (immature=0, mature=1) and was used to determine the size and age at which 50% of females or males in the population reached sexual maturity The slope and intercept of the logistic regression were determined using a linear model with the binomial family and logistic option in R (R Development Core Team 2011). Spawning S easonality The timing of peak spawning was assigned using gonad somatic ind ices (GSI ) for males and females using the following formula: GSI = (GW/(TW GW)) 100; where GW = total gonad weight (g) and TW = total fish weight (g). Monthly mean GSI values were calculated to estimate seasonal reproductive patterns. Other C riteria for H ermaphroditism There are two population level life history traits which support the occurrence of hermaphroditism (Sad ovy and Shapiro, 1987) Male and female length at age data was compared to determine if growth rates differed between the sexes and if one sex dominated the larger and subsequent older age classes. An unpaired Students t test with unequal variances was used to determine if mean length at age differed significantly between the sexes at each age. The comparison of size at age data was restricted to age classes with sample sizes ratio of golden tilefish was also calculated to determine if the sex ratio skewed from 1:1.

PAGE 53

53 Breeding Strategy Previous research documented the importa nce of the predorsal adipose flap in defining alternative breeding strategies for golden tilefish males (Grimes et al., 1988) The height of the adipose flap has also been used to classify golden tilefish as male or female. T herefore, the height of the adipose flap was measured from the base of the head to the m aximum vertical height of the adipose flap The relationship between the heigh t of the adipose flap by gender and maturity stage was evaluated Results Sample C ollection A total of 1 ,859 golden tilefish were provided by multiple NOAA Fisheries Service sampling programs throughout the Gulf of Mexico and east coast of Florida (Table 35 ). A majority of samples were collected in 2007 2009 (east coast of Florida, 93%; Gulf of Mexico, 73%). A decrease in the number of third and fourth quarter annual samples from the east coast of Florida occurred due to regional fishery closures. The golden tilefish fishery in the Gulf of Mexico also experienced seasonal time closures, but sampling through fishery independent surveys provided continuous monthly samples. Otolith Processing and Assigning Age Golden tilefish thin sectioned otoliths are difficult to interpret given several different shapes of otolith sections and diverse patterns of growth deposition ( see chapter 2 ). Given the difficulty in determining accurate age estimates, I had the assistance of a secondary reader (D. Berrane, NOAA Fisheries Service Beau fort, NC ) to quantify the imprecision in age estimation Indices of precision were calculated from those otoliths read by both readers (n = 200) with an average percent error of 7%, with percent agreement of 32% increasing to 95% 3 years

PAGE 54

54 Gonad P rocessi ng In a majority (66%) of males the entire gonad was blocked, due to the small size of male gonads. Female gonads were much larger and were subsampled before blocking, germinal tissue was removed from the posterior portion (53%), both the posterior and a nterior portions (14%), and in 12% of females the entire gonad was blocked. Gonad D evelopment Of the 1825 gonads histologically classified (98% of gonads collected), 773 were functional females and 1052 were functional males. Although sampling was extens ive and a large range of lengths of golden tilefish were collected (280 1040 mm FL ), only 16 ( 12 males, 4 females) golden tilefish were classified as immature ( Table 36) The i mmature males were collected in August, September, and December, with mature, spawning males capable of spawning collected in most months (Fig. 31A). The majority of mature female gonads were in spawning capable reproductive stage from January June (Fig. 3 1 B ). In ten gonad samples maturity stage could not be determined due to inconsistencies in gonad preservation. Golden tilefish gonads contained both structural and germinal features to support hermaphroditism. Sexual M aturity It wa s difficult to determine the difference between an immature male and resting male during nonspawning season. Therefore, the classification of maturity in males wa s subjective to the interpretation of the stages of spermatogenesis. Fish with undetermined maturity stage were not used in this analysis (male, n = 4; female, n = 6). Immature males were evident from a larger distribution of lengths and ages compared to immature females (Table 3 6) Male golden tilefish were estimated to reach maturity at a younger age (male, < 1 yr; female, 2.5 yr) and s maller length (male, 150 mm FL ; female 331 mm FL ) than females ( Table 3 7, Fig 32 ) Only four

PAGE 55

55 immature females were collected and these fish were less than 400 mm in length and six years of age. The smallest mature female was 338 mm FL and age 3 years. Histological E vidence for H ermaphroditism Both functional male and functional female gonads contained a membrane lined cavity (Fig. 33A ): criterion 1) in approximately 15% of the males this cavity was qualified as originating from ova rian lumen and remained unused for sperm transportation (Fig. 33, B and C ). Non functional (opposite) sex tissue was present in both functional males and females: criterion 2) 26% of the functional females had tubules containing all stages of spermatogen esis and 71% of the functional males had multiple stages of oocyte development. The relative placement of non functional tissue was a combination of locations: functional females had self contained tubules of male tissue (consisting of several stages of s permatogenesis ) scattered throughout the lumen (Fig. 34 ) and functional males contained oocytes attached, floating, embedded, in the ducts or a combination of locations within the germinal tissue (Fig 35, A, B, C, and D ). Functional male gonads contain ed oocytes of all stages in testes; criterion 3) primary growth was the dominate stage (88%), followed by vitellogenic (29%), cortical alveolar (28%), late hydrated (6%), and hydrated (4%) (Fig. 36) And the final criterion, male sperm sinuses were present and located within the gonad wall (Fig. 37 ). Spawning S easonality Gonadsomatic index (GSI ) suggests golden tilefish have a prolonged spawning season. Golden tilefish GSI values were elevated from January through June with the peak in April for both males and females (Fig. 3 8). Reproductive seasonality did not differ between the fish collected off the east coast of Florida and the northern Gulf of Mexico, therefore data was combined.

PAGE 56

56 Other C riteria for H ermaphroditism Golden tilefish exhibited sexual ly dimorphic growth. Male golden tilefish were more prevalent in the larger size classes (Fig. 39A ). However, male and female golden tilefish occurred within the same range of ages (age 3 33 yrs, Fig. 39 B ). Males were significantly larger in all but one (age 4) of the age classes with sample sizes 16, Fig. 3 9C ). Finally, the overall sex ratio of golden tilefish was slightly skewed (1.3:1.0) with males dominating the population. B reeding Strategy Measurements of the height of the adipose flap proved to be sexual dimorphic. Large mature males had larger adipose flaps than females especially for fish > 500 mm FL (Fig. 3 10 ). Most females adipose flaps measured < 10 mm in height. It was difficult to distinguish an immature male and immature or mature female given just the height of the adipose flap. Discussion Through detailed histological photomicrographs I documented the simultaneous occurrence of mature testicular and ovarian tissues in golden tilefish from the U.S. South Atlantic waters off the east coast of Florida and the northern Gulf of Mexico. T his is strong evidence that golden tilefish are hermaphroditic, not gonochoristic ( Sadovy de Mitcheson and Liu, 2008) Golden tilefish gonads qualify as being protogynous hermaphrodites by satisfying each of the criteria of Sadovy and Shapiro (1987); male gonads contained a membrane lined cavity, originating from ovarian lu men, which remains unused for sperm transportation and female gonads also contained a similar membrane lined lumen, male gonads contained follicles of all stages in testes and sperm sinuses were present within the testes wall. The most difficult criteri on to satisfy is identifying fish undergoing transition since not all sex changing fish exhibit the same structural changes in germinal tissue when undergoing

PAGE 57

57 transition (Sadovy and Shapiro, 1987; Sadovy de Mitcheson and Liu, 2008). For example, in the Fa mily Serranidae there are three types of configurations of the germinal tissue during transition: Serranus has complete separation of ovarian and testicular tissue by connective tissue; Ryiticus Anthias has testicular tissue remains isolated along the ov arian wall with no connective tissue, and Epinephelus has mixed ovarian and testicular tissue (Smith, 1965). In order to apply these configurations to other sex changing fish, Sadovy and Shapiro (1987) relabeled these as delimited, unlimited, and mixed. Given these definitions, I have classified golden tilefish with a mixed configuration of germinal tissue since the nonfunctional sex tissue was located throughout the gonad and was adjacent to f unctional sex tissue. In functional males, nonfunctional f emale oocytes were located in a variety of placements within the male germinal tissue (attached, embedded, floating, and in the ducts) and in functional females, nonfunctional male tubules were found scattered within the female germinal tissue and were self contained within their own germinal epithelium. In comparison, the germinal configuration of golden tilefish females undergoing transition to male were different than that of protogynous hermaphrodites Epi ne phelus morio (red grouper; Moe, 1969) and Mycteroperca phenax (scamp; Lombardi Carlson et al., 2012), given the nonproliferation (in terms of the amount) of male tissue throughout the female gonad. In both functional males and females golden tilefish the nonfunctional (opposite sex) tissue amounted to only a small percentage (<1%) of the entire gonad. A majority (49%) of the golden tilefish sampled had nonfunctional tissue The functional males with newly developed oocytes (primary growth, cortical alveolar, and vitellogenic) were considered to have transitioned from mature, active females that had not recently spawn ed whereas males with early or late hydrated oocytes tra nsitioned from recently

PAGE 58

58 spawned or spent females These lat t er males would be considered truly transitional fish since these functional male gonads contained degenerating female oocytes and were undergoing proliferating spermatogenesis (Shapiro and Sadovy, 1987) For functional females with male tubules, these females were considered to be preparing for sex change. Similar structural features have been documented in protogynous gobies ( Eviota sp.; Cole, 1990) and coral sea trout ( Plectropomus sp.; Adam s, 2003) where testicular tissue remains dormant in functional females until sex change. Functional males and functional females with opposite sex tissue were collected mainly during the spawning season (males, 69%; females, 85%), alluding to a very high ly social mating behavior (as first prop osed by Grimes et al., 1988) Golden tilefish have likely minimized the cost of changing sex and increased their overall reproductive fitness by minimizing the time not reproducing given the occurrence of dormant testicular tissue in females and transitional fish during the spawning season (Hoffman et al. 1985; Adams, 2003) It is possible that I have misconstrued fish in transition with golden tilefish exhibiting a juvenile bisexual pha se. This phase is defined in terms of gonad morphology and not gonad function (Sadovy and Domeier, 2005) Typically a juvenile bisexual phase is temporary and restricted to a certain size, age, or gamete stage ( Sadovy de Mitcheson and Liu, 2008) but gol den tilefish males and females with opposite sex tissue were identified from the same distribution of lengths In addition, there were not any differences in the lengths of functional male golden tilefish characterized as having primary oocytes to more developed (vitellogenic, hydrated) oocytes within the testes. Therefore, it is unlikely that golden tilefish exhibit a juvenile, bisexual phase. After thorough examination of male histological slides, evidence supports that male golden tilefish have a dia ndric mode of maturation. Males with a diandric mode of maturation are

PAGE 59

59 described as some males deriving from an immature phase (primary) and other males transitioning from females (secondary). This was evident given the collection of small, immature male s, as well as the occurrence of remnant atretic oocytes in male testes. This greater flexibility in male development has been attributed to variations in social and environmental conditions (Adams, 2003). Further evidence to support diandry is that mal es were observed at similar ages as the female age at first maturity and there was a complete overlap of ages for males and females (Fennessy and Sadovy, 2002). Typically, gonadal investment differs between primary males (those derived from immature males ) and secondary males (those derived from females) (Choat et al., 1996 ; Drillings and Grober, 2005) but there were not any difference in gonado somatic index between primary and secondary males. Previous reproductive research on golden tilefish also reported having two classes of males (spawning and non spawning, Grimes et al., 1988). They based their conclusions on the differentiation of males by a secondary sexual characteristic (height of dorsal adipose flap), immature males at smaller sizes, and th e maintenance of reproductive territories by males in habitat limited environments. Golden tilefish from the Gulf of Mexico and along the east coast of Florida have a secondary sexual characteristic that is sexual l y dimorphic with large mature males havi ng larger adipose flaps and females having smaller adipose flaps. However, the average height of the adipose flap did not vary between male types. Grimes et al 1988 concluded that the larger percentage of small, immature males represented non spawning m ales but this reproductive phase was based only on visual inspection of male gonads. In addition, it is not clear if these males were sampled during the spawning season, as it is difficult to distinguish immature from resting, mature males during non spaw ning seasons. Golden tilefish do exist in a very specific habitat type of soft, but malleable sediment along the continental shelf in water

PAGE 60

60 depths 80400 m (Freeman and Turner, 1977), inhabit burrows (Able et al., 1982), and reside in a specific thermal cline (9 14oC; Grimes et al., 1986). Previous submersible observations and still camera images of golden tilefish in and around their burrows have revealed great insight on the behavior of these elusive fish (Able et al., 1982; Grimes et al ., 1986; Jones et al., 1989; Able et al., 1993 ) but until a combination of visual observations of tilefish in their natural environment along with histological confirmation of sex and maturity of those observed individuals can be conducted, any conclusions on the mating behavior and morphological development of males and females is speculative. Theoretically, from a population ecological perspective, it can be advantageous to be hermaphroditic (Tomlinson, 1966). There is a higher chance of successful br eeding if individuals can change sex when populations are at low densities. In addition, there is a 50/50 chance an individual will come into contact with the opposite sex especially if individuals exist in relative isolation given a limited habitat and a re less mobile. Golden tilefish do exist in relative isolation within burrows and densities may fluctuate given exploitation rates and seasonal changes in water temperature, making it beneficial to be able to change sex. Reproductive strategy can vary geo graphically, especially if sexual development is influenced by social factors and habitat availability ( e.g. Teleost families: Pomacentridae [damselfishes ] ; Sadovy de Mitcheson and Liu, 2008). The an e m onefish ( Amphiprion clarkii ) relies heavily on sea an emone for shelter and spawning and the distribution of anemone s dictates the anemonefish reproductive strategy and mating behavior (Moyer, 1980; Ochi, 1989). In tropical areas with low density of sea anemones, female anemonefish are more likely to change sex (functional hermaphrodites) when the larger male anemonefish are removed (Moyer, 1980) but sex change is less likely (functional gonochor e s) in temperature areas with high densities of

PAGE 61

61 anemone and subsequent higher densities of male anemonefish (Ochi, 1989). Similarly, in obligate coral dwelling damselfishes ( Dascyllus aruanus ) the area coverage and distance between coral colonies governs the need to change sex (Cole, 2002; Asoh, 2003). Therefore, it is plausible that the reproductive strategy of gol den tilefish varies along its distribution given the availability of suitable habitat (see description above). Teleosts can display a large diversity of reproductive strategies from remaining the same sex their entire lifetime (gonochores) to changing sex during their lifetime (hermaphrodites) to some fish gonads containing nonfunctional, opposite sex tissue (intersex) (Atz, 1964) It is not uncommon for gonochoristic teleosts to contain immature, oocyte stage (primary oocytes) in a male gonad s ( Lower re Barbieri et al ., 2011), but it is uncommon for male gonads of a gonochoristic teleosts to contain mature, oocyte stages (c ortical a lveolar vitellogenic, early hydrated, late hydrated) as detected in golden tilefish. Simultaneous hermaphroditic teleost s are characterized by a 1:1 sex ratio, an undelimited germinal tissue configuration (no mixing of opposite sex tissue), and both male and female germinal tissues are clearly functional (Sadovy and Shapiro, 1987). Functional male and female golden tilefis h did contain opposite sex germinal tissue but this tissue amount to < 1% of the entire gonad and was assumed nonfunctional; therefore, lessening the possibility of simultaneous hermaphroditism. Intersex is mostly prevalent in freshwater teleosts and is typically associated with a hormonal imbalance in male teleosts due to endocrine disruptors in the environment (Hinck et al., 2009). Unlike the male gonads of golden tilefish, intersex gonads contain only early staged oocytes (previtellogenic and vitellog enic) and compose of a small proportion (< 3%) of the population ( Bateman et al., 2004; Hinck et al., 2009). Protandr ous hermaphroditism is difficult to prove since male tissue seldom remains in mature, female gonad s T he best evidence for protandry is

PAGE 62

62 t he occurrence of degenerative male tissue in a gonad with developing female tissue ( Sadovy and Shapiro, 1987). Female golden tilefish gonads did not contain degenerative male tissue but instead tubules containing several stages of spermatogenesis. Theref ore, based on histological evidence supporting each of the above criteria and population level life history, golden tilefish from U.S. South Atlantic waters off the east coast of Florida and the northern Gulf of Mexico are most likely sequential protogynous hermaphrodites. Previous literature on the reproductive strategy of golden tilefish concluded this fish is gonochoristic (Erickson and Grossman, 1986; Grimes et al., 1988; Palmer et al., 2004) not hermaphroditic as I have concluded. There are diffi culties to scientifically prove the factors contributing to the occurrence of hermaphroditism in golden tilefish along the east coast of Florida and throughout the northern Gulf of Mexico. The best evidence could be obtained through c ontrolled scientific experiments on golden tilefish in their natural environment using underwater remotely operated vehicles or manned submersibles These experiments could help identify the triggering factors, environmental or social conditions changes in densities, control s by the central nervous system or variations in hormones (Baroiller et al., 1999) or mating systems (St. Mary, 2000) that are contributing to this alternative reproductive strategy. Conclusion Given the prem ise that fishing is size selective (removing the larger, older fish), the reproductive strategy (gonochoristic or hermaphroditic) of the species can influence the response to overfishing (Bannerot et al., 1987; Armsworth, 2001). As in the case in protogynous hermaphrodites, the larger, older fish are predominately males and these populations are more likely to feel the effects of size selective fishing through sperm limitation, shifts in behavior and skewed sex ratios (Bannerot et al., 1987; Armsworth, 2001; Alonzo and Mangel, 2004; Heppell et al., 2006). Since it is difficult to observe how a population compensates for sperm limitations,

PAGE 63

63 especially for deep water fish, stock assessment models can attempt to incorporate the uncertainty of how a fish with a protogynous hermaphroditic reproductive strategy responds to the impact of fishing. In the next chapter, I explored how understanding and modeling the protogynous hermaphroditic life history can influence stock assessments, and compared those results to a simpler model that ignores hermaphrod itism.

PAGE 64

64 Table 31. Descriptions and photomicrographs of histological sections of male golden tilefish leading gamete stages. Hematoxylin1 and eosinY stain magnification s and scales on photomicrographs Photomic rographs Leading Gamete Stage Description of male leading gamete stage Spermatogonium (SG) Stained light purple, fuzzy in appearance, large in diameter but not located in spermatocysts. Primary Spermatocyte (PS) Darkly stained, mostly non spherical in shape, relatively large diameter and present in spermatocysts. Secondary Spermatocyte (SS) Cells more spherical and darker than primary spermatocytes, smaller in diameter and present in spermatocysts. 0.05 mm, 400x 0.05 mm, 200x 0.05 mm, 200x

PAGE 65

65 Table 3 1. Continued Photomicrographs Leading Gamete Stage Description of male leading gamete stage Spermatid (ST) Black staining, spherical and smaller than previous stages, in spermatocysts. Tails not present. Spermatozoa (SZ) Black staining, pink tails present, spherical and similar in size to spermatids. Spermatocysts appear to merge together (due to extensive discontinuous germinal epithelium) forming large pools. 0.05 mm, 200x 0.05 mm, 40x

PAGE 66

66 Table 32. Descriptions and photomicrographs of histological sections of female golden tilefish leading gamete stages. Hematoxylin 1 and eosinY stain magnification s and scales on photomicrographs Photomicrographs Leading Gamete Stage Description of female leading gamete stage Primary Growth (PG) Definitive follicle is formed, follicular epithelium envelopes oocyte, germinal vesicle present, large clear cytoplasmic area Cortical Alveolar (CA) Three main components: cortical granules (appear white) thin band or zona radiate (appears as a red ring) surrounds the oocyte, and small droplets or lipids appear. Germinal vesicle present Vitellogenic (V) Yolked oocytes are mostly spherical. Yolk is first seen as small, red, spherical granules around the germinal vesicle and cortical alveolar granules (appear white) move to t he periphery of oocyte 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 100x

PAGE 67

67 Table 3 2. Continue d Photomicrographs Leading Gamete Stage Description of female leading gamete stage Early Hydrated (EH) Yolked oocytes less spherical in shape in conjunction with lipids coalescing into one or two clear spheres. Some attributes of the maturation stage may have begun including migration of the germinal vesicle Start of final oocyte maturation. Late Hydrated (LH) Oocytes reach their maximum volume through hydration and are amoeboid in shape. Oocyte becomes more translucent as lipid and protein yolk droplets coalesce, germinal vesicle disappeared 0.05 mm, 40x 0.05 mm, 40x

PAGE 68

68 Table 3 3. Descriptions and photomicrographs of histological sections of male golden tilefish reproductive phases. The sample size (n) for each phase is provided. Hematoxylin 1 and eosinY stain magnification s and scales on photomicrographs Photomicrographs Reproductive Phase n Description of male reproductive phase Immature, inactive 8 Includes males with spermatogonia and no evidence of spermatogenesis. Inactive, uncertain 11 Difficult to distinguish from regressed, except for lack of developed tissue. Developing virgin 1 Spermatogenesis begins, spermatocytes present and no prior indicator of maturity. 0.05 mm, 200x 0.05 mm, 200x 0.05 mm, 100x

PAGE 69

69 Table 3 3. Continued Photomicrographs Reproductive Phase n Description of male reproductive phase Developing 203 Spermatogenesis and the formation of spermatocytes begins. Little or no spermatozoa. Active, mature 0 This class is not used for males, essential the same as developing Spawning, capable 591 All stages of spermatogenesis may be present. Spermatozoa evident and fill ing lobules and sperm ducts Spent, p ost spawn 2 Spermatogenesis ceasing, some residual spermatozoa present, spermatogonia proliferation common. 0.05 mm, 100x 0.05 mm, 100x 0.05 mm, 40x

PAGE 70

70 Table 3 3. Continued Photomicrographs Reproductive Phase n Description of male reproductive phase Regressed, inactive, mature 234 Spermatogonia dominate, no active spermatogenesis, and some residual spermatozoa may be present. 0.05 mm, 40x

PAGE 71

71 Table 3 4. Descriptions and photomicrographs of histological sections of female golden tilefish reproductive phases. The sample size (n) for each phase is provided. Hematoxylin 1 and eosinY stain magnification and scale on photomicrographs Photomicrographs Reproductive Phase n Description of f emale reproductive phase Immature, inactive 1 Primary growth oocytes only, no evidence of prior spawning. Inactive, uncertain 4 Only primary growth oocytes present, not capable of spawning in distant future and no evidence of prior spawning. Developing virgin 0 Cortical alveolar oocytes dominate and no prior indicators of maturity No female tilefish were assigned this phase Developing 12 Cortical alveolar oocytes present. Prior spawning indicators confirm maturity. 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 40x

PAGE 72

72 Table 3 4. Continued Photomicrographs Reproductive Phase n Description of female reproductive phase Active, mature 100 Vitellogenic oocytes present and fish should spawn within days or weeks. Spawning, capable 327 Fish is reproductively active and capable of spawning. Early and late hydrated oocytes, as well as vitellogenic oocytes. Postovulatory follicles (old or new) may be present. Post ovulatory, spent 7 All oocytes stages may be present, majority of oocytes (>50%) experiencing atresia. 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 40x

PAGE 73

73 Table 3 4. Continued Photomicrographs Reproductive Phase n Description of female reproductive phase Regressed, inactive, mature 3 21 Primary growth oocytes only, evidence of sexual maturity and recentl spawned. 0.05 mm, 40x

PAGE 74

74 Table 35. The number of golden tilefish collected by month. Fish were collected by multiple sources throughout the east coast of Florida and Gulf of M exico (years combined: 20002009). Month east coast of Florida Gulf of Mexico Total January 168 82 250 February 196 47 243 March 115 73 188 April 235 39 274 May 115 55 170 June 48 49 97 July 0 11 11 August 0 147 147 September 28 167 195 October 45 20 65 November 0 152 152 December 0 67 67 Total 950 909 1859

PAGE 75

75 Table 3 6. Summary statistics by sex and maturity for golden tilefish Data was combined for fish collected from the east coast of Florida and Gulf of Mexico. F unctional sex (with and without opposite sex tissue) maturity stage (immature, mature (immature, mature) showing sample size (n), range of fork length and mean length standard error (SE ). F ish where maturity stage could not be accurately determined were not included (male, n = 4; female, n = 6). Functional sex Maturity stage Opposite sex tissue n Range Mean SE Male immature No Yes 3 9 471 499 432 748 487 8 5 59 37 Male mature No Yes 341 695 375 1040 389970 667 6 647 5 Female immature No Yes 3 1 290 394 336 357 3 3 na Female mature No Yes 562 201 338 824 390 800 531 4 565 6

PAGE 76

76 Table 3 7. Results of the logistic regressions for size and age at maturity for male and female golden tilefish. Data was combined for fish collected from the east coast of Florida and Gulf of Mexico. Parameters of the logistic regression (intercept, slope) were calculated using a general linear model with the binomial family and logistic option in R (R Development Core Team 2011). Intercept Slope Estimate Male FL (mm) Age (yr) 1.509 0.864 0.010 0.442 150 > 1 Female FL (mm) Age (yr) 18.175 3.087 0.055 1.256 331 2.46

PAGE 77

77 Figure 3 1. Golden tilefish proportion by reproductive stage by month and gender. See t ables 3.3 and 3.4 for complete description of reproductive stages for ( A ) males and ( B ) females respectively Data was combined for fish collected from the east coast of Florida and Gulf of Mexic o. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Percent Month Regressed, inactive Spent Spawning capable Active, mature Developing Developing virgin Inactive, uncertain Immature, inactive 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Percent Month Regressed, inactive Spent Spawning capable Active, mature Developing Developing virgin Inactive, uncertain Immature, Inactive A B

PAGE 78

78 Figure 3 2. Golden tilefish p roportion mature and immature by gender. Maturity estimates for male s ( mature: solid black line immature: black circles ) and females ( mature: dotted black line immature: white circles ) by ( A ) length and ( B ) age. Data was combined for fish collected from the east coast of Florida and Gulf of Mexico Proportion mature predicted by logistic regression. P roportion immature based on observed data at length and age Solid gray line indicates 50% mature. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 200 400 600 800 1000 Percent Frequency Fork length (mm) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Percent Frequency Age (yr) A B mature, male immature male mature female immature female

PAGE 79

79 Figure 33. Photomicrographs of golden tilefish histological sections showing membrane lined cavity originating from ovarian lumen in both a female and male gonads. ( A ) Active mature female (490 mm FL, age 9) collected in January 2007 in the Gulf of Mexico, showing primary growth, cortical alveolar and vitellogenic oocytes. This cavity remains unused in males for sperm transportation as shown in a ( B ) spawning male (570 mm FL, age 12) collected in March 2009 in the Gulf of Mexico, showing all stages of spermatogenesis, spermatozoa evident and f illing lobules and sperm ducts and a ( C ) developing male (524 mm FL, age 10) collected in January 2008 off the east coast of Florida, showing formation of spermatocytes beginning. Hematoxylin1 and eosin Y stain m agnification s and scales on photomicrographs B C 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 40x A

PAGE 80

80 Figure 34. Photomicrograph of golden tilefish histological sections showing an immature female containing male tissue The female ( 336 mm FL, age 4) collected in March 2009 in the Gulf of Mexico contained male tubules (several stages of spermatogenesis) at ( A ) 40x magnification and ( B ) 200x magnifications Hematoxylin 1 and eosin Y stain magnifications and scales on photomicrographs. A B 0.05 mm, 40x 0.05 mm, 200x

PAGE 81

81 Figure 3 5. Photomicrographs of golden tilefish histological sections showing the placement of oocytes within functional male gonads ( A ) A ttached vitellogenic oocyte in a spawning male (620 mm FL, age 10) collected in June 2009 from the Gulf of Mexico, ( B ) floater cortical alveolar oocytes in a spawning male (507 mm FL, age 6) collected in February 2007 off the east coast of Florida ( C ) e mbedded vitellogenic oocyte in a spawning male (484 mm FL, age 6) collected in April 2008 off the east coast of Flori da and ( D ) primary growth, cortical alveolar, and vitellogenic oocytes in the ducts of a spawning male (670 mm FL, age 11) collected in February 2007 off the east coast of Florida. Hematoxylin 1 and eosinY stain, magnification s and scales on photomicrogr aphs. A B C D 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 40x 0.05 mm, 40x

PAGE 82

82 Figure 36. Photomicrograph of golden tilefish histological section showing a spawning male containing vitellogenic and late hydrated oocytes. The male (646 mm FL, age 8) collected in June 2006 in the northern Gulf of Mexico Hematoxylin 1 and eosinY stain magnification and scale on photomicrograph. 0.05 mm, 40x

PAGE 83

83 Figure 37. Photomicrograph of golden tilefish histological section showing sperm sinuses within the gonad wall of a spawning male. The male (748 mm FL, age 9) collected in April 2008 off the east coast of Florida Hematoxylin 1 and eosinY stain, magnification and scale on photomicrograph. 0.05 mm, 10x

PAGE 84

84 Figure 38. Golden tilefish g onad somatic i ndices by month and by gender. ( A ) Male and ( B ) female golden tilefish Gonadsomatic index reported as mean standard error Reproductive seasonality did not differ between the fish collected off the east coast of Florida and the northern Gulf of Mexico, therefore data was combined. Sample sizes appear above error bars. 116 102 98 160 81 45 4 76 109 15 29 15 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Gonad somatic Index Month 94 81 64 55 53 32 1 44 64 44 74 48 0.00 1.00 2.00 3.00 4.00 5.00 6.00 Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Gonad somatic Index Month A B

PAGE 85

85 Figure 39. Golden tilefish a ge and length data by gender. ( A ) F ork length ( mm) and ( B ) a ge distributions by sex males ( black bars) and females ( white bars ) and ( C ) mean size standard error at age by sex (males, black circles; females, white circles). *indicates significant differences between the sexes determined through an unpaired Students t test with unequal variances, p Data was combined for fish col lected from the east coast of Florida and Gulf of Mexico 0% 5% 10% 15% 20% 25% 30% 250 350 450 550 650 750 850 950 Percent Frequency Fork length (mm) 0% 4% 8% 12% 16% 20% 3 5 7 9 11 13 15 17 19 21 23 25 27 Percent Frequency Age (yr) 300 400 500 600 700 800 900 4 5* 6* 7* 8* 9* 10* 11* 12* 13* 14* 15* 16* 17* 18* 19 20* Fork length (mm) Age (yr) B C A male female male female male female

PAGE 86

86 Figure 3 10. Golden tilefish adipose flap height by length, maturity stage and gender Sample sizes are above the male data points (immature, n = 2, gray triangles; mature, n = 282, black triangles) and below the female data points (immature, n = 2, gray circles; mature, n = 235, white circles). Maturity determined through histology. 1 1 1 10 50 58 43 34 13 15 6 5 2 5 11 20 27 36 46 29 36 40 20 10 1 0 10 20 30 40 50 60 0 100 200 300 400 500 600 700 800 900 1000 Adipose flap height (mm) Fork length (mm) mature, male immature, male mature, female immature, female

PAGE 87

87 CHAPTER 4 MODEL CHOICE AND QUANTITY OF INFORMATIVE DATA: A CASE STUDY OF THE GOLDEN TILEFISH ASSE SSMENT IN THE GULF O F M EXICO Introduction A stock assessment is a n analysis in which the dynamics of a fish population are modeled, historical trends of harvest (biomass, exploitation rate) are estimated and alter native management scenarios are tested (Hilborn and Walters, 1992). There are several choices of model types used in stock assessment analyses ranging from simple, surplus production models to complex, statistical catch at age models that incorporate auxi liary information (e.g. environmental and stock mixing parameters). The complexity of the model increases as the quantity of input data and the amount of observed data predicted increases. There are generally two schools of thought in regards to the use of simple versus complex assessment models 1) use a simple model and add complexity given the quantity of informative data ( Richards and Schnute, 1998; Schnute and Richards, 2001; Walters and Martell, 2004) or 2) use a complex model and remove model detai l until the observed data are sufficiently modeled (Methot, 2009a ; Cope, in press ). A compromise (and a recommendation) is to use multiple models in assessments ( Hilborn and Walters, 1992; National Research Council, 1998; Walters and Martel l 2004). Fe deral assessments for fish stocks in the Gulf of Mexico have included a multitude of model configurations, from simple, surplus production models, virtual population analysis, to complex statistical catch at age models. As an example, a surplus production which incorporates covariates (ASPIC; Prager, 1992, 1994) is a bi omass dynamic model used by NOAA Fisheries Service and is based on the Graham Schaeffer logistic production model This model is easily modified by the user and eliminates the requirement of other assessment approaches to use catch per unit effort data but cannot make the use of any available age and length data therefore

PAGE 88

88 requires less data. Alternatively, s tatistical catch at age models can incorporate a larger quantity of da ta such as age and size specific information and are also used by NOAA Fisheries Service, for example, agestructured assessment program (ASAP; Legault and Restrepo, 1999), and numerous modified virtual population analysis (VPAs ; Goodyear, 1997; Porch, 1999) (e.g., ADAPT; Conser and Powers, 1990; Powers and Restrepo, 1992). Most recently, fish stocks in the Gulf of Mexico are being modeled using Stock Synthesis (Methot, 2000, 2009b) This statistical catch at age model uses likelihood methods to determine the goodness of fit between the observed data and simulated population models and allows an even larger quantity of data to model size and age selectivities, spawner recruitment functions, hermaphrodism, time varying catchability, environmental variabili ty, and even models the potential biases in the observation process. Stock Synthesis is likely the most highly parameterized model routinely used as part of the stock assessment process by US federal fisheries assessments. Regardless of the simplicity or the complexity of the stock assessment model used to assess a fish population, most models have a similar overall structure (Walters and Martell, 2004) Assessment models consist of a state dynamic model ( the submodel including historical disturbances and other unknown processes in the actual system ) and an observation model (the submodel how observa tional quantities are related to the unobserved system state ) (Walters, 1986). These models also include a model fitting criteri on ( e.g. maximum likelihoo d estimation ) to find what combination of state dynamics and observation model best predicts the observed data Similarly stock assessment model s have the same objective to estimate the current state of the fished population. Each type of model has dif ferent assumptions related to the parameters in the model and how these parameters interact to create the prediction model. These assumptions are often

PAGE 89

89 unique to each type of model and ultimately are fundamental to creating the prediction model and in tur n generating parameter estimates of interest such as fishing mortality rate used to define the stock status. Because these assumptions are fundamental to the model outcomes, and ultimately the management decisions based on these model outcomes, it is highly recommend that different types of models, with different assumptions, be used in assessing fish stocks ( Hilborn and Walters 1992; National Research Council, 1998; Walters and Martel, 2004). That way if two types of models give different predictions about stock status, then managers should explore the differences in model assumptions to determine why the results diverge and whether additional research on specific assumptions could resolve uncertainties about the stock status by improving the available d ata (Hilborn and Walters, 1992). Therefore, my objective wa s to compare two age structure models that differ in complexity a relatively simple stochastic Stock Reduction Analysis (SRA) a ge structure model and a more complex model Stock Synthesis (SS) using Gulf of Mexico golden tilefish as my case study1Materials and Methods Multiple model variations for SRA and SS will be presented. The comparison between models will be conducted on one of the SS model variation for golden tilefish (indices emphasized an d age and length composition data deemphasized) and the Gulf wide stochastic Stock Reduction Analysis model (age selectivity fitted through SS and without age composition) Th ese model formulation s allow a straightforward comparison of the model s results Model D escriptions Stochastic Stock Reduction Analysis (SRA) is a stochastic age structured population model with BevertonHolt stock recruitment function that simulates populations forward in time 1 The research presented in this chapter incorporates data, model inputs, and results evaluated during the stock assessment of golden tilefish in the Gulf of Mexico (SEDAR, 2011a).

PAGE 90

90 starting before fishing began (Wa lters et al., 2006). This model is parameterized by taking management objectives, annual exploitation rate (U) producing maximum sustainable yield at equilibrium (UMSY) and maximum sustainable yield (MSY) as leading parameters, then calculates the Beverto nHolt stock recruit parameters from these parameters and from per recruit fished and unfished eggs and vulnerable biomasses ( Forrest et al., 2008; Martell et al., 200 8). In SRA, recruitment is assumed to have had log normally distributed annual anomalies The SRA model produces a distribution of trajectories of vulnerable biomass over time (due to recruit ment anomalies) as well as, probability distributions of MSY, Umsy, Uyear, Goodyears Compensation Ratio (recK), and stock status. There are four main assumptions in SRA : 1) the population is at virgin conditions the first year data are provided; 2) there is stochastic variation in recruitment for all years ; 3) historical recruitment have arisen from variation around a single, stationary stock recruitme nt curve; and 4 ) survival rate is constant for all years The main premise in stochastic SRA is to try to best explain the fluctuations in stock size given historical catches and recruitment (Walters et al., 2006). In order to model changes in the fisher y given changes in targeted species, gear used, and size and/or age regulations (i.e., size limits), age selectivities (or vulnerabilities) can vary over time. Stochastic SRA is written in visual basic, uses a Windows based interface for user inputs, and Bayes posterior distributions for biomass, exploitation rate, and leading parameters are calculated using Markov Chain Monte Car l o (MCMC) integration The likelihood function used in the MCMC search algorithm can include terms for relative abundance indi ces that change over time, deviations of current exploitation rates, and multiple length and/or age composition. S tochastic Stock Reduction Analysis has been applied to several Gulf of Mexico species including red snapper ( Lutjanus campechanus ; SEDAR, 2005), red

PAGE 91

91 grouper (Epinephelus morio; SEDAR, 2006), and yellowedge grouper ( Epinephelus flavolimbatus ; SEDAR, 2011b). Stock Synthesis (SS) is a statistical catch at age model that incorporates a variety of biological, population level, and fisheries data (Methot, 2000). Stock Synthesis can simultaneously analyze age and length composition data, growth, mortality, age and length selectivities, catch histories, indices of abundance, fishery effort, recruitment deviations, etc., with each parameter having its own degree of uncertainty. Stock Synthesis has been termed an integrated analysis model because of its flexibility a nd its ability to handle missing or incomplete data series, multiple stock sub areas, and timevarying parameters (Methot, 2009a ) Stock synthesis can model recruitment using either a BevertonHolt or Ricker stock recruitment functions. Recruits are assumed to be age 0 fish and, just as in SRA, recruitment variation is modeled with a lognormal distribution. The premise of SS is to bring the model to the data and model the potential biases in the observed data (Methot, 2000) Stock Synthesis is compile d using ADMB (Auto Differentiation Model Builder). Stock Synthesis has mainly been applied to groundfish fisheries along the United States w est c oast (Pacific Fishery Management Council 2000) but within the la st few years, SS (version 3.03a; Method, 2009b) has been applied to fish stocks in the Gulf of Mexico assessments (SEDAR, 2011a, 2011b). Gulf of Mexico G olden T ilefish C ase S tudy Golden tilefish, Lopholatilus chamaeleonticeps is a deep water demersal fish found in the Atlantic from Nova Scotia to the Gulf of Mexico (Dooley, 1978). Golden tilefish are managed by three fishery management councils (Northeast Atlantic, South Atlantic, and Gulf of Mexico). In the Gulf of Mexico, golden tilefish are managed by a total allowable catch of 200 metric ton s which was implemented in 2004. This regulation was initiated as part of the red

PAGE 92

92 grouper rebuilding plan to guard against commercial long liners shifting to deep water groupers and tilefish fisheries ( NOAA 2004). As I reported in previous chapters, g olden tilefish use very specific habitat types, commonly soft, but malleable sediment along the continental shelf in water depths 80400 m (Freeman and Turner, 1977). Golden tilefish inhabit burrows (Able et al., 1982), reside in a specific thermal cline (9 14oC; Grimes et al., 1986) and have limited movement (Grimes et al., 1983). Therefore, the golden tilefish stock in the Gulf of Mexico was modeled within regions of the northern Gulf of Mexico (east and west of the Mississippi drainage, Fig. 41). Thi s separation accounted for spatial differences in fishing effort (Table 41), region specific growth ( see below ), and subsequent recruitment variability. Golden tilefish a ge and length composition data were available for the fishery dependent and fishery independent sources. Data was delineated by region providing the capture location information (latitude, longitude, or NMFS statistical grid). A majority of the age and length composition data w ere obtained from the commercial long line fishery in the e astern region (19842009; Table 42 and 43, Fig 43, 44, and 45). Given the difficulty in validating the age estimates of golden tilefish (chapter 2), an ageing error matrix was incorporated into Stock Synthesis. S ex specific age and length composit ion data were also available for the last nine years of the time series (200 12009, Table 43 ). Golden tilefish ages and total lengths from the entire time series (1997 2009) were used to obtain gulf wide and region specific growth parameters. The data was fit to a von Bertalanffy growth model and growth param eters were estimated by non linear regression (Solver, Microsoft Excel) (Table 4 4) I also determined golden tilefish to have sexual ly dimorphic growth (see chapter 3). G olden tilefish were also modeled (in SS only) with sexual dimorphic growth, with

PAGE 93

93 males capable of obtaining larger sizes at faster rates than females (Table 4 4). Gender and region specific growth parameters were estimated within SS, due to the lack of observational data (Table 4 3). In addition, S S modeled golden tilefish as protogynous hermaphrodites (given the evidence in chapter 3) with only 57% of population remaining female and estimating size at transition through a logistic regression. Using a logistic regression I estimated that female tilefish reach maturity at a young age (2 yrs) and small size (344 mm TL) (see chapter 3 for details), even though golden tilefish are long lived and reach lengths of over a meter Golden tilefish f ecundity was not modeled directl y but as a proxy, through a regression of body weight and gonad weight (Table 44 ). Only females with vitellogenic or hydrated oocytes collected during the spawning months (January through June) were used to estimate regressions Each assessment model us ed a different regression as a proxy for fecundity. Stochastic SRA used a linear relationship between body weight and the average weight at maturity. There are three possible regressions to use as proxies for fecundity in Stock Synthesis (Methot, 2000) an d given the data, I used a nonlinear relationship between gonad weight and body weight. Estimates for natural mortality for golden tilefish were calculated using n umerous regressions. These regressions were applied to a variety of golden tilefish dat asets (e.g., males, females, all data; von Bertalanffy growth model unconstrained, and with to constrained at zero), resulting in a distribution of natural mortality estimates (SEDAR, 2011 a ) After removing unrealistically high natural mortality values (M > 0.25), the mode of the distribution was 0.14. This estimate of natural mortality was the same as was estimated using Hoenig (1983) regression model for teleosts for all data based on longevity (30 yrs). Therefore, n atural mortality was based on longevity for both models (Table 44), with SS also incorporating an age specific mortality (reference age 4 ; Lorenzen, 2005).

PAGE 94

94 Golden tilefish catch histories began in 1965 with the primary source of harvest the commercial hand line fishery (Table 41). Commercial bottom longline gear was introduced in the Gulf of Mexico in the late 1970s, and by the 1980s golden tilefish were primarily being harvest by this gear type. Since golden tilefish are not managed by a size limit and are caught at great depths, discards are minimal and any reported discards were added to the respective fishery landings (SEDAR, 2011a ) There were two sources of indices of abundance for golden tilefish: fishery dependent, commercial long line fishery (CMLL) and fishery independent, NOAA/ National Marine Fisheries Service (NMFS) bottom long line survey (LL Survey) (SEDAR, 2011a ) The fishery dependent index was based on NOAA Fisheries Service, Southeast Fisheries Science Center Coastal F isheries Logbook Program for self reported commercial bottom long line catch per unit effort data (19922009). Since commercial long line trips are not identifiable by the target fishery, the Stephens and MacCall 2004 method was used to restrict the logbook data to trips most likely to catch golden tilefish (e.g. NMFS Statistical Grids, species associations). The standardized index of abundance for commercial long line fishery was constructed using a delta lognormal approach with three factors (subregion, days at sea, year) and three factor interactions (subregion*year, days at sea*year, subregion*days at seas). The same modeling approach was used for the fishery independent bottom long line survey (factors: water depth, survey area, y ear). The fishery independent index began in 2000 and only used data from stations that fished in the depths golden tilefish inhabit (125365 m). Both standardized indices were calculated by region and corresponding coefficients of variations (CV) were i ncorporated in both assessment models (Fi g. 4 2, A, B, C, and D ).

PAGE 95

95 Age and size selectivities were constructed based on both fishery dependent and fishery independent age and size composition. There were multiple age and size selectivities used in Stock S ynthesis (SS) for each specific combination of source (commercial hand line, commercial long line and bottom long line survey), gender (female, male), and region (east and west). The best fit for both length and age data were through logistic regressions (SEDAR, 2011a ) Size selectivities were fitted using a double logistic regression with normal distribution and constrained to be asymptotic ( Fig. 4 6). Age selectivities (in SS) were fitted based on the logistic regression ( Fig. 4 7A ). In Stochastic Stock Reduction Analysis (SRA) selectivities are modeled as a matrix of age vulnerabilities by year. Initial age vulnerabilities in SRA were modeled using a virtual population analysis but final model runs (for comparison to SS) used the resulting average (across genders and regions) commercial long line age selectivities fitted by SS (Fig 47B ). Uncertainty in S tock S tatus I evaluated the uncertainty in the parameters that determined the status of golden tilefish stock in the Gulf of Mexico. S tock assessments conducted by NOAA Fisheries Service follow regional fishery management and scientific guidelines when determining stock status. The probabilities of being overfished and/or undergoing overfishing to determine stock status were based on Spawning St ock Biomass (SSB) and Exploitation (U) at Maximum Sustainable Yield (MSY) for Stochastic SRA and based on Spawning Stock Biomass (SSB ) and Fishing mortality (F) at a Spawning Potential Ratio (SPR) of 30% for the Stock Synthesis model. Therefore, the uncer tainty in the ratios of SSB2009/SSBMSY and U2009/UMSY for Stochastic SRA and the ratios of SSB2009/SSBSPR 30% and F2009/FSPR 30% for Stock Synthesis were examined through the resulting MCMC chains. I tested the convergence of the MCMC chain for each paramet er using two

PAGE 96

96 diagnostic tests (Geweke and Hiedelberger and Welch convergence tests) and trace and correlation plots (coda and PBS modeling packages; R Development Core Team, 2011). Results Stochastic Stock Reduction Analysis (SRA) In order to make predic tions about the status of golden tilefish populations east and west of the Mississippi drainage, stochastic SRA model was executed separately by region and for the entire Gulf of Mexico stock For each region specific model, I applied the default values for recruitment anomalies (1.0 ) and standard deviation of recruitment (0.5) and recruitment was modeled without autocorrelation. Next, I ran SRA using region specific life history parameters (Ta ble 42), catch histor ies (Table 4 1) and commercial long line indices with varying degrees of uncertainty (coefficient of variation) (Fig. 4 2) I compared the two age vulnerability schedules by running each model (gulf wide, East, West) twice once us ing the age vulnerability back calculated by virtual population analysis and again using the age vulnerabilities estimated by SS through a logistic regression (Fig. 4 7B). Finally, I incorporated the age composition data into the SRA model for the gulf wi de and Eastern Gulf of Mexico model runs. A ge composition data was not sufficient to model the Western Gulf of Mexico with age data. Stochastic SRA model convergence is based on the acceptance rate of the Markov Chain Monte Carlo (MCMC) sampling procedure. Stochastic SRA experienced difficulties in fitting the eastern region and Gulf wide models for golden tilefish; therefore, the default paramet er s controlling the movement along parameter space for the MCMC were adjusted. All models were manually ceased after several million iterations (Gulf wide, 4.0 x 106, 16% acceptance rate; East, 4.2 x 106, 23% acceptance rate; West, 4.4 x 106, 30% acceptance rate). The eastern region was predicted to have a high probability that recent catches have been at or above MSY and in

PAGE 97

97 the western region recent catches have been at MSY by SRA, given the sample distribution of maximum sustainable yield (MSY) and exp loitation at MSY (Umsy) (Fig. 4 8). Stock Synthesis (SS) Stock synthesis was also applied to region speci fic data to predict whether or not the golden tilefish population in the Gulf of Mexico was overfished or undergoing overfishing. In SS only one mod el was compiled r ecruitment was distributed evenly between the regions. A single BevertonHolt stock recruitment function was estimated in SS with recruitment deviations forced to sum to zero with a fixed value of 0.15 for the standard deviation of recruitment. Bias adjustments were applied to the calculation of recruitment (Methot and Taylor, 2011). The level of adjustment was based on the amount of data informing recruitment (19651983 no bias adjustment; 19841997 linear increase in bias adjustment; 19972006 full bias adjustment; 20072009 no bias adjustment). Stock Synthesis can process a larger capacity and complexity of data therefore, region and gender specific growth patterns and mortality functions (Table 42) were used. In addition, a ge selectivities by gender and source ( Fig. 46 and 4 7A ) were applied in SS but due to the lack of sufficient gender data collected with length frequency data, size selectivities were only constructed by source and region. Finally, golden tilefish in the Gulf of Mexico w ere modeled using region specific catch history by commercial gear type (Table 41) and commercial long line indices and bottom long line survey indices with varying degrees of uncertainty (coefficient of variation) ( Fig. 4 2). The fit of Stock Synthesis (SS) is measured in terms of the value of negative log likeliho od (NLL). The total negative log likelihood in SS (for the models ran for golden tilefish) contained six major components (catch histories, indices of abundance, length composition, age composition, recruitment, and priors for estimated parameters). I chose to discuss two (model 1 and model 12) of the fifteen model variations (sensitivities), which were presented during the

PAGE 98

98 Southeast Data, Review, and Assessment Workshops for golden tilefish from the Gulf of Mexico (SEDAR, 2011a ) and are thus of greatest interest to fishery managers These models (1 and 12) varied by the amount of uncertainty incorporated into the calculation of the total NLL. In Stock component to c alculate the total NLL to incorporate uncertainty into the model (Methot, 2009b). In Model 1, no lambdas were applied to the data components and a large NLL was calculated with most of the NLL pertaining to the age and length compositions (Table 4.5). By evaluating (visually) the residuals of the age and length composition (Fig. 4.9 and 4.10), it was evident that SS had difficulties predicting the age and length composition especially in the western region Therefore, to test the sensitivity of SS to the uncertainty of the age and length composition, Model C omparison Models were compared in terms of predicted biomass, exploitation rates, and stock status. Both SRA (gulf wide) and SS (model 12) predicted the eastern region of the Gulf of Mexico yielded a higher carrying capacity of golden tilefish compared to the western region given the historical catches ( Fig. 4 11). In addition, both models predicted the same trend in historical biomass levels by region, SRA biomass trajectories appear lower than SS because SRA reports the vulnerable, not the total biomass as in SS. Gulf wide exploitation rates were nearly identical between the models until 1993, at which SRA predicted slightly higher exploitation ( Fig. 4 12 ). Stochastic SRA used total egg production as a proxy for spawning stock biomass (SSB). The probability of being overfished is the resultant of unique MCMC iterations in which the ratio of SSB2009/SSBMSY is less than 1.0 and the probability of overfishing comes from the number of unique MCMC runs in which the ratio of U2009/UMSY is greater than 1.0. SRA predi cted that

PAGE 99

99 golden tilefish in the Gulf of Mexico are not experiencing overfishing (probability overfishing: 1.26%) and are not overfished (probability overfished: 0.11%). In SS, the probabilities of being overfished and/or undergoing overfishing are based on spawning potential ratio (SPR) and not MSY therefore, Gulf of Mexico golden tilefish stock status between the models cannot be compared directly. Nevertheless, the final model runs in SS determined tilefish was not undergoing overfishing (based on SPR3 0%) and was not overfished. The comparison of SRA model runs using varying age vulnerability schedules resulted in similar stock status (not overfished and not undergoing overfishing) for each region and for the entire Gulf of Mexico golden tilefish stock Predicted historical vulnerable biomass trajectories had similar trajectories but SRA model runs using the age vulnerability schedule back calculated through virtual population analysis (VPA) resulted in lower initial (eastern region, 30%; western regio n, 8%) and final biomass (eastern region, 10%; western region, 6%), compared to SRA model runs using the age vulnerabilities estimated through SS (Fig. 4 11) SRA model runs with the VPA age vulnerabilities predicted 1620% higher historical exploitation, during peak exploitation years (1988, 1995, and 2005) (Fig. 4 12) Age composition data w ere added to the eastern and Gulf wide models to help better inform the model for estimating mortality rates and recruitment. Stochastic SRA model runs with age com position data provided similar historical biomass trajectories (Fig. 4 11) and exploitation rates (Fig. 4 12) as well as decreased probabilities of overfishing (with age composition, 0.35%) and overfished (with age composition, 0.05%). Uncertainty in S tock S tatus For both models, original MCMC chains were thinned prior to tests of convergence. The MCMC chains tested for convergence from the Stochastic SRA (SRA) model consisted of 5,048 iterations that originated from the 4.0 x 106 MCMC iterations, the first 200 iterations were part of

PAGE 100

100 the burnin and removed, and the remaining chain was thinned every 800th iteration. The MCMC chains tested for convergence from the Stock Synthesis (SS) model consisted of 4,000 iterations. These chains were constructed f rom MCMCs ran for 1.0 x 106 iterations, the first 1,000 iterations were part of the burnin and removed, with every 200th iteration saved. The resulting MCMC chains for the ratios of SSB2009/SSBMSY and U2009/UMSY for Stochastic SRA and the ratios of SSB2 009/SSBSPR and F2009/FSPR for Stock Synthesis passed both tests of convergence (Table 4 6). The Gewekes test of convergence did not find any differences in the mean of the first 10% of the chain compared to the mean of the last 50% of the chain, concludi ng all chains resulted from a stationary distribution. Similarly, the Heidelberger and Welch's (stationary compared samples along the entire chain and half width compared half the width of the 95% confidence intervals of the mean with the mean) tests of convergence concluded the samples of each chain came from stationary distribution and from a chain of sufficient length. Trace plots for each ratio also provided evidence that MCMC sampling for each chain were from stationary distributions (Fig. 4 13 a nd 414). The stock status parameters from Stochastic SRA and Stock Synthesis had similar correlations (SRA, 68; SS, 62) and uncertainties (Fig. 4 15 and 416). By presenting the uncertainty in the MCMC chains for those parameters that determine stock status, better confidence in the models results can be concluded. Discussion These two age structure models provide an example of comparing a simple (stochastic Stock Reduction Analysis, SRA) and a complex (Stock Synthesis, SS) model structure. On one side of the spectrum, SRA, a simple age structure model uses basic biological in puts and catch data (conditioned on catch removal process) to provide an estimate of the current stock size and status, based on the historical catch history. The complex agestructure mo del (SS) modeled region and gender specific growth, mortality, selec tivities, indices of abundance, and estimated

PAGE 101

101 62 parameters (e.g. von Bertalanffy growth parameters, virgin recruitment, stockrecruitment steepness parameter, recruitment deviations). Using similar data, but different model structure, the re s ults of these models suggest similar conclusions concerning the stock status of golden tilefish in the northern Gulf of Mexico. S pecifically this stock is not overfished and currently not undergoing overfishing. A main aim of any stock assessment model is to determine the present abundance of fish given the historical removals. Since assessment scientists cannot simply count the number of fish present in a stock, a vector of removals is constructed, an index of abundance is calculated, and information describing the age, growth, reproduction, and mortality of a population are used to predict the present abundance (Schnute and Richards, 2001). In each of the main inputs of assessment models (removals, indices, life history), both process (error associated with unpredictability of biological processes) and measurement (error associated with observational data, e.g. data reporting) errors are present (Schnute and Richards, 2001). Assessment models also contain structural errors associated with programmed assumptions and constraints (Hilborn, 2003). Each of these errors make up the uncertainties associated with stock assessments. In complex models, like Stock Synthesis, as more functional components are incorporated into the model structure, basic parameter int eractions are more difficult to interpret and the assessment process is less transparent to managers and stake holders (Walters, 1986; Hilborn, 2003). Although complex models can have less process and measurement error an increase in modeling structural e rror can be just as problematic (Mohn, 2009) Therefore, to conduct a successful stock assessment Schnute and Richards (2001) recommend the following: 1) use a suite of models because no model can fully capture all the intermingled processes of a fish sto ck; 2) do not only rely on the trends of the fishery (i.e., landings) 3) investigate other fields (e.g.,

PAGE 102

102 economics, oceanography); 4) start with a simple model and add complexity given the observed data, and 5) use computer simulation to test management s cenarios. A quote attributed to the statistician G eorge E. P Box states that Remember that all models are wrong; the practical question is how wrong they have to be to not be useful and this certainly applies to fisheries assessment model s In the case of golden tilefish, my challenge was to develop and compare two types of population assessment models using the same data, knowing that the data on golden tilefish biology and the fishery landings were sparse. We can attempt to model the un certainty in the observational model but a key question is can we truly model the biological and environmental complexity of fisheries (Schnute and Richards, 2001) and if not, does this impact our ability to develop sustainable management plans In the ex ample of golden tilefish from the northern Gulf of Mexico, there are three trends evident from the data inputs that have very large influences on my ability to determine stock status 1) t here was a lack of consistency in sampling the fishery for age and l ength composition, 2) t here was a lack of signal from either the fishery dependent or the fishery independent indices of abundance and 3) t here was a lack of duration (length of coverage for the time series) for the most of the data (except for catch). The golden tilefish fishery in the Gulf of Mexico consists of a limited number of vessels with a low total allowable catch (TAC, 200 metric tons ) in comparison to the larger reef fish fisheries (red snapper, TAC = 1 ,610 metric tons ; red grouper, TAC = 2,61 0 metric tons ; GMFMC 2011). This small number of vessels creates opportunities and also adds complexity to filling the data gaps. Ideally with a small number of vessels, fewer port agents would be required to canvas the landings for biological samples. However, port agents are primarily dedicated to sampling much larger and more economically important fisheries such as those from

PAGE 103

103 reef fish. A dditional there is difficulty in obtaining biological samples from port agents since most golden tilefish are gutted atsea and most fish are iced and boxed onboard fishing vessels, with delivery trucks waiting dockside to quickly ship the catch to fish markets in the northeastern United States (pers. comm., P. Antosh state of Alaba ma port agent). This requires biological samples to be taken from onboard observer programs which are more expensive and dictated by vessel selection. In the Gulf of Mexico, biological and landings data have been collect ed from the commercial sect or since the early 1990s through a NOAA Fisheries Service dock side sampling program called the Trip Interview Program (TIP). This program has port agents stationed at major commercial ports along the shoreline of the entire Gulf of Mexico from Key West, FL to Brownsville, TX. Federal stock assessments are based on the data collected by these port agents, many of which have collected data since the inception of this program. In general, the goal of TIP is to conduct representative sampling (age and lengt h data) across time, location, gears and trips (Hoenig, 2007). This type of on site sampling is the most reliable but requires an appropriate statistical design to be maintained (National Research Council, 1998). These port agents are field statisticians attempting to purposely sample landings of multiple species from one or many fishing vessels at one time. A recent independent review of TIP sampling design recommended that a systematic sampling scheme (sampling the jth fish) could provide less bias in data collection then the current purposely sampling (Hoenig, 2007). As a result of these shortcomings from the sampling program, both stock assessment models were affected by the lack of biological samples, as well as, the limited number (<10 yr) of year s of consecutive data that was available for modeling.

PAGE 104

104 The data requirements for stock assessment models increase exponentially with model complexity (Hall, 2003). Regardless of the complexity of the model only quality, informative data should be used in models, otherwise noisy, insufficient data can lead to bad management decisions (Adkison, 2009). As in the case of golden tilefish, data inputs consisted of short time series of consecutive data, small annual sample sizes for age and length composition ( especially by gender and region) and highly variable indices of abundance. The only long term time series data, which was considered collected with some confidence and both models assumed to be collected without error, was the historical catch data. The total catch history is one of the most vital fishery based data (Shepherd, 1984; Pope, 1988; Hall, 2003). The next most important fishery based data is an unbiased estimate of abundance ( Hilborn and Walters, 1992; Schnute and Richards 2001; Methot, 2009a ) The two indices of abundance used to model golden tilefish were highly uncertain and did not provide much signal for either model. An index of abundance that does not truly reflect the changes in abundance or indices that are conflicting can result in misleading stock status (National Research Council, 1998), as was suggested in the case in the northern cod (Walters and Maguire, 1996). In my opinion, the data used to construct both the fishery dependent and independent indices were inadequate The fishery dependent index was based on data subsetting from commercial log book data using the Stephens and MacCall (2004) method. This method was first tested using species associated with the bocaccio rockfish recreational fishery off the coast of Northern California. This data sub sampling relies on the accurate reporting by fishers of species associated with golden tilefish but if a co species does not have a commercial value or a fisher does not have permit, than the cospecies is discarded, not landed and most likely not recorded in the logbook. Yellowedge grouper and other deepwater groupers had the highest associations

PAGE 105

105 with the golden tilefish from the analysis of commercial logbook data for both the subareas (SEDAR, 2011a ). However, the species most associated with golden tilefish during fishery independent surveys were gulf hake ( Urophycis cirrata) and Cuban dogfish ( Squalus cubensis ) neither species appeared in the species association list for data sub setting nor are of much value commer cially. Therefore, there needs to be a more appropriate manner to query the logbook database. My recommendation is to add a target species data field to the logbook database to allow for more direct sampling of positive trips by species. Similarly, it is as important to have a fishery independent survey to estimate the trend in relative abundance. This survey should be conducted through a stratified random, sampling design to calculate a true independent estimate of abundance not just for the exploite d stock but for the mature, spawning fish as well as recruits (Hall, 2003) An index of abundance is considered vital information for stock assessments (Shepherd, 1984). However, the data used to construct the fishery independent index of abundance is ba sed on a bottom long line survey that only spends 10% of the annual sampling in depths (183366 m) associated with golden tilefish. This annual survey is conducted for eight weeks in the Gulf of Mexico covering depths of 9366 m with most effort in the sh allowest depths (SEDAR, 2011a ). This bottom long line survey provides the data for calculating indices of abundance for numerous shallow water reef fish (groupers, snappers; SEDAR, 2005, 2006), as well as small coastal sharks (Highly Migratory Species, 20 07). I propose that an annual survey, statistically designed for deepwater, demersal species be conducted throughout the Gulf of Mexico with effort s solely at the deeper depths (183366 m). This survey would complement the current survey but be designed to provide better quality and quantity of informative data on abundance for numerous poorly known commercially harvested species (e.g. deep water groupers).

PAGE 106

106 Additionally, in developing a stock assessment it is as important to obtained length frequency data (Shepherd, 1984). Typically, length frequency data can provide insights into recruitment trends, growth, and mortality (Anderson and Neumann, 1996) assuming that length data has been collected systematically, in a randomized strategy that encompasses the entire range of lengths and at sufficient annual sample sizes (for length based assessment, sample size > 1000, Erzini, 1990; for age based assessment, samp le size > 100, Anderson and Neumann, 1996). However, golden tilefish length frequency and age samples were not collected consecutively throughout the time series of catch and were collected in insufficient sample sizes. The number of consecutive years o f data and the sample size of the age composition data were determined to heavily influence the results of Stock Synthesis especially in terms of biomass and predictability of the catch (Yin and Sampson, 2004). Therefore, one must be as explicit and straig htforward about the uncertainty of the data and the assumptions of the model, regardless of the models complexity, during a stock assessment (Walters and Maguire, 1996). The assessment for golden tilefish was heavily influenced by the lack of quality an d especially the quantity of data and was judged to be a datapoor species (SEDAR, 2011a ). I have presented two assessment models that vary in complexity, a simple, age structure depletion based model stochastic Stock Reduction Analysis model and a complex, integrated analysis model Stock Synthesis but both agreed in stock status, historical biomass and exploitation. Stock Synthesis had difficulty in predicting observational data (e.g., age and length frequencies, indices of abundance) and model per formance increased as age and length data streams were deemphasized. In stochastic Stock Reduction Analysis, model results were not sensitive to the additional age composition data and the status of the golden tilefish stock predicted historical biomass trajectories and exploitation w ere similar with and without the age composition data

PAGE 107

107 Although the stock status of golden tilefish in the northern Gulf of Mexico was determined to be in good standing, this species has characteristics that makes it highly susceptibility to capture and to overfishing (Grimes and Turner, 1999). Golden tilefish have a unique habitat preference and borrowing behavior that increases this species vulnerability to fishing, especially if fishers overharvest at one particular site In addition, golden tilefish are long lived, slow growing, and have a complex breeding and reproductive strategy that is not completely understood. The previous assessments for golden tilefish from the Northeast Atlantic (NEFSC, 2005) and South Atlanti c (SEDAR, 2004) management areas have resulted in overfished and overfishing conditions and providing stricter management regulations golden tilefish in both of these areas are now in rebuilding stages (NEFSC, 2009; SEDAR, 2011c). The assessments in these areas also suffered from insufficient life history data and unbiased estimates of abundance, same as the Gulf of Mexico assessment models presented here. Therefore, efforts need to continue to collect annual, biological samples from fishery dependent sources, as well as, developing species or at least complex level (e.g., deep water) fishery independent surveys. Conclusion Federal stock assessment scientists, as well as independent experts agree that t he quality and complexity of a particular assessm ent should be directly related to the quality and quantity of data available ( Ludwig and Walters, 1985; Hilborn and Walters, 1992; Schnute and Richards, 2001; Mace et al. 2001; Walters and Martell, 2004 ). I recommend that data should be scored and ranked accordingly and the complexity of the model should reflect the quality and quantity of informative data. Data should be ranked based on the number of consecutive years of data (e.g., 10 years; Yin and Sampson, 2004), the number of i ndividual samples that generated the data (e.g. for length composition at least 10x the number of length bins; Gerritsen and McGrath, 2007), how the data w ere collected (e.g., randomly, fishery independent), and the spatial

PAGE 108

108 coverage of the data (e.g., the number of vessels interviewed, the number of fish houses/docks ) Additionally, I recommend meta data on the behavior of the industry should be documented, since a majority of data used in stock assessment models rely on the cooperation of the fishers (Co tter et al., 2004). A s it is recommended by many assessment scientists, data inputs should always incorporat e some degree of uncertainty ( such as standard deviations, probability distributions) and the assumptions of assessment models should be discussed openly for the assessment process to remain transparent to managers and stakeholders.

PAGE 109

109 Table 4 1. Commercial landings for the northern Gulf of Mexico golden tilefish. Landings ( gutted metric tons) are presented by commercial hand line (CMHL) and long line (CMLL) fishery by region, east (E) and west (W), delineated by the Mississippi River drainage. Year CMHL E CMHL W CMLL E CMLL W Total 1965 2.82 0.00 0.00 0.00 2.82 1966 0.81 0.00 0.00 0.00 0.81 1967 0.44 0.00 0.00 0.00 0.44 1968 0.60 0.00 0.00 0.00 0.60 1969 0.13 0.00 0.00 0.00 0.13 1970 0.00 0.00 0.00 0.00 0.00 1971 1.33 0.00 0.00 0.00 1.33 1972 0.45 0.00 0.00 0.00 0.45 1973 1.62 0.00 0.00 0.00 1.62 1974 1.70 0.00 0.00 0.00 1.70 1975 6.00 0.00 0.00 0.00 6.00 1976 9.98 0.00 0.00 0.00 9.98 1977 14.74 0.00 0.00 0.00 14.74 1978 9.57 0.24 0.00 0.00 9.81 1979 11.88 0.00 2.32 0.48 14.68 1980 7.79 0.00 2.79 0.73 11.31 1981 52.51 0.00 36.43 11.30 100.24 1982 24.68 0.00 56.14 41.51 122.33 1983 5.78 0.21 56.11 32.29 94.39 1984 5.32 0.85 74.53 41.62 122.32 1985 4.05 4.98 37.18 66.61 112.82 1986 19.69 3.98 59.82 51.22 134.72 1987 30.90 8.04 69.41 109.16 217.50 1988 35.88 19.16 104.74 225.05 384.82 1989 16.51 25.73 41.61 100.27 184.12 1990 27.01 1.37 50.71 65.01 144.11 1991 7.48 9.98 48.32 20.98 86.76 1992 5.43 5.98 38.82 35.73 85.96 1993 6.33 2.96 61.26 44.67 115.22 1994 5.09 0.66 107.99 43.56 157.30 1995 0.95 3.25 66.93 120.71 191.84 1996 0.59 1.20 48.99 34.50 85.28 1997 1.05 0.23 116.58 18.72 136.59 1998 0.55 0.65 90.93 28.21 120.34 1999 2.53 1.78 88.82 58.51 151.64 2000 1.72 2.24 109.25 80.96 194.17 2001 6.49 0.12 136.73 52.81 196.15 2002 3.92 0.64 99.93 114.59 219.08 2003 1.40 0.88 95.45 64.46 162.19 2004 1.24 0.25 114.88 72.88 189.25

PAGE 110

110 Table 4 1 Continued. Year CMHL E CMHL W CMLL E CMLL W Total 2005 1.63 1.74 138.60 96.57 238.54 2006 2.37 0.08 100.18 23.41 126.04 2007 0.42 0.84 118.05 9.78 129.10 2008 0.05 0.13 117.47 23.52 141.18 2009 0.54 0.03 142.19 23.61 166.38 Total 341.95 98.23 2433.17 1713.45 4586.80

PAGE 111

111 Table 4 2. Annual sample sizes of length composition data for golden tilefish from the northern Gulf of Mexico Data reported by commercial gear (hand line, CM HL; long line, CM LL) and region (east, E; west, W) northern Gulf of Mexico divided by Mississippi River drainage For those years with higher sampling effort, an effective sample size of 200 was applied. Numbers in parenthesis reflect the number of sex specific length records which were modeled in Stock Synthesis. Year CMHL E CMHL W CMLL E CMLL W 1984 106 100 1985 1986 2 26 1987 1 25 126 1988 1 48 175 1989 2 82 1990 3 153 128 1991 14 (4) 15 35 415 1992 1 95 23 395 1993 22 (1) 12 14 (1) 162 1994 2 47 642 (6) 295 1995 2 7 245 185 (5) 1996 30 (1) 1 (2) 316 62 1997 20 655 20 1998 19 411 13 1999 34 5 560 2000 24 (6) 795 35 2001 46 (108) 2 777 2002 100 (3) 197 2003 2 19 610 (7) 18 2004 10 1685 40 2005 174(2) 20 1559 (6) 34 2006 1 (2) 1,125 15 (2) 2007 (1) 9 905 93 (2) 2008 2 (2) 52 473 (18) 410 2009 11 38 (10) 580 (91) 577 Total 641 342 12,069 3415

PAGE 112

112 Table 4 3. Annual sample sizes of age composition data for golden tilefish from the northern Gulf of Mexico. Data reported by fishery dependent (commercial hand line, CMHL; commercial long line, CMLL) and fishery independent (National Marine Fisheries Service bottom long line survey, LL Survey) sources Sex was determined through the histological examination of gonad tissue. Year 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 CMHL Female Male Unknown 5 24 34 2 44 3 10 55 6 10 43 CMLL Female Male Unknown 43 4 11 44 74 273 3 4 500 8 51 491 29 208 8 53 276 58 54 580 171 156 996 LL Survey Female Male Unknown 3 1 2 19 18 10 7 8 30 6 30 1 8 12 11 15 5 15 41 2 5 21 2 13 31 Total 43 4 0 22 91 143 307 578 614 271 405 775 1426

PAGE 113

113 Table 4 4. Life history parameters for golden tilefish from the northern Gulf of Mexico by region and gender. Growth parameters were estimated through nonlinear regression. Length at maturity was estimated using logistic regression. Proxies for fecundity were based on spawning female body weight and gonad weight regressions. Natural mortality was estimated using Hoenig (1983) regression model for teleosts for all data based on longevity (30 yrs). See text for further details. *these values represent the estimated parameters used in S tock Synthesis (total length (cm), weight (kg)). Parameter Applied to both regions East Sexes combined East Male East Female West Sexes combined West Male West Female Von Bertalanffy asymptotic length growth coefficient 83 0.14 88 0.11 92 0.13 82 0.11 77 0.17 87 0.16 66 0.19 Length at Maturity 34 Fecundity Stochastic SRA Stock Synthesis proxies linear regression non linear regression Length Weight alpha beta 7.52 x 106 3.08 Natural Mortality 0.137

PAGE 114

114 Table 4 5. Negative log l ikelihoods for the major components of the s tock s ynthesis model for golden tilefish from the northern Gulf of Mexico. S tock S ynthesis (SS) model 1 was the base model and model 12 emphasized the indices of abundances ( = 25) and deemphasized the age and length compositions ( = 0.05) These values represent the Negative log likelihoods ( NLL ) multiplied by the lambdas ( ) for all sources and regions combined. Component NLL : model 1 NLL : model 12 Catch 3.05 x 10 6 3.26 x 10 6 Indices 4.76 494.64 Length composition 2397.15 141.82 Age composition 4966.29 303.12 Recruitment 175.21 140.94 Parameter priors 35.59 35.69 Total 7579 .00 126.93

PAGE 115

115 Table 4 6. Summary statistics and results of diagnostic tests for convergence for MCMC iterations for parameters from each model that determined stock status for golden tilefish from the northern Gulf of Mexico Summary statistics for Stochastic Stock Reduction Analysis (SRA) and Stock Synthesis (SS) include sample size (n), median (2.5% and 97.5% quantiles), Geweke Z score, Hiedelberger and Welch pass or fail stationary test (H and W, stationary) and Hiedelberger and Welch pass or f ail half width test (H and W, half width) Parameters that determined stock status (SRA: Spawning Stock Biomass (SSB) in 2009/ Spawning Stock Biomass (SSB) at Maximum Sustainable Yield (MSY) and Exploitation (U) in 2009/ Exploitation (U) at Maximum Sustainable Yield (MSY); SS: Spawning Stock Biomass (SSB) in 2009/ Spawning Stock Biomass (SSB) at Spawning Potential Ratio (SPR) of 30% and (B) Fishing mortality (F) in 2009/Fishing mortality (F) at Spawning Potential Ratio (SPR) of 30%) Model Parameter n Median (2.5%, 97%) Geweke Z score H and W Stationary H and W Half width SRA SSB 2009 /SSB MSY U2009/UMSY 5028 5028 1.92 (1.46,2.35) 0.28 (0.16, 0.59) 4.12 2.38 pass pass pass pass SS SSB 2009 /SSB SPR F2009/FSPR 4000 4000 1.99 (1.76, 2.26) 0.49 (0.39, 0.62) 1.39 1.78 pass pass pass pass

PAGE 116

116 Figure 4 1. Northern Gulf of Mexico spatial map displaying the National Marine Fisheries Service (NMFS) statistical g rids. Golden tilefish inhabit depths 100400m. Golden tilefish were modeled with two subareas (east, west) of the Mississippi drainage (East: NMFS grids 1 11; West: NMFS grids 1223 ). D epth contours in meters

PAGE 117

117 Figure 4 2. Standardized indices of abundance for golden tilefish from the northern Gulf of Mexico by region and by source. ( A ) Commercial long line east, ( B ) commercial long line west, ( C ) bottom long line survey east, and ( D ) bottom longline survey west. Indices of abundance (observed, solid black line; predicted in Stock Synthesis, solid gray line) and coefficient of variation (CV, dotted lines) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.5 1.0 1.5 2.0 2.5 1992 1994 1996 1998 2000 2002 2004 2006 2008 CV Index of abundance Year A 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.5 1.0 1.5 2.0 2.5 1992 1994 1996 1998 2000 2002 2004 2006 2008 CV Index of abundance Year C 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.5 1.0 1.5 2.0 2.5 1992 1994 1996 1998 2000 2002 2004 2006 2008 CV Index of abundance Year B 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.0 0.5 1.0 1.5 2.0 2.5 1992 1994 1996 1998 2000 2002 2004 2006 2008 CV Index of abundance Year D observed predicted CV

PAGE 118

118 Figure 4 3. Annual age composition for golden tilefish from the northern Gulf of Mexico. D ata reported from commercial long line fishery by region ( A ) east and ( B ) west of the Mississippi River drainage Each circle is proportion to the sample size at age in a given year. The largest circle size represents 96 fish and the smallest circle 1 fish. A B

PAGE 119

119 Figure 4 4. Observed annual length composition data for golden tilefish from the northeastern Gulf of Mexico Data reported from commercial long line fishery. The trend lines indicate the normal distribution. A B

PAGE 120

120 Figure 4 5. Observed annual length composition data for golden tilefish from the northwestern Gulf of Mexico Data reported from comm ercial long line fishery. The trend lines indicate the normal distribution.

PAGE 121

121 Figure 4 6. Stock s ynthesis s ize selectivities for golden tilefish from the northern Gulf of Mexico Selectiv i ties were fit in s tock s ynthesis using a double logistic regression for commercial long line (solid black line), commercial hand line (solid gray line) and NMFS bottom long line survey (dotted black line). 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 11 15 19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 Selectivity Total length (cm) commercial long line commercial hand line NMFS bottom longline survey

PAGE 122

122 Figure 4 7. Age selectivities for each assessment model for golden tilefish from the northern Gulf of Mexico ( A ) Stock s ynthesis selectivities based on logistic regression for commercial long line fishery by region and gender ( female age selectivity same for both regions, solid black line; male east, solid gray line; male west, dotted gray line) and ( B ) stoc hastic Stock Reduction Analysis vulnerabilities fit within s tock s ynthesis (averaged across region and gender from commercial long line, solid black line), calculated through VPA and fixed to logistic curve (s olid gray line) and observed age vulnerabilities back calculated through VPA (dotted gray line). 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Age vulnerabilities Age (yr) 0.0 0.2 0.4 0.6 0.8 1.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Age selectivities Age (yr) A B female, east and west male, east male, west fit by stock synthesis estimated by VPA, fixed logistic estimated by VPA

PAGE 123

123 Figure 4 8. D istribution of maximum sustainable yield given the distribution of exploitation at maximum sustainable yield for golden tilefish from the northern Gulf of Mexico from Stochastic SRA. ( A ) East and ( B ) west of the Mississippi River drainage The s olid line indicates the average catch for the given time series for either region. Smooth scatter plot (R Develop ment Core Team 2011) color symbolizes density of points ( grey is lowest and black is highest density). A A B

PAGE 124

124 Figure 4 9. R esiduals for fits of le ngth composition for golden tilefish from the northern Gulf of Mexico of unknown gender Data reported from the commercial long line fishery ( A ) east and ( B ) west of the Mississippi River drainage from Stock Synthesis. Solid circles are positive residual s (i.e., observed greater than predicted) and open circles are negative residuals (i.e., predicted greater than observed). The size of the bubble indicates the sum of residuals at length bin by year (maximum residual sum; east = 0.27, west = 0.93) 0 20 40 60 80 100 120 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Total length (cm) Year 0 20 40 60 80 100 120 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Total length (cm) Year A B

PAGE 125

125 Figure 4 10. R esiduals for fits of age composition for golden tilefish from the northern Gulf of Mexico of unknown gender Data reported from the commercial long line fishery ( A ) east and ( B ) west of the Mississippi River drainage from Stock Synthe sis. Solid circles are positive residuals (i.e., observed greater than predicted) and open circles are negative residuals (i.e., predicted greater than observed). The size of the circle represents the sum of residuals at age by year (maximum residual sum; east = 0.27, west = 0.93) 0 5 10 15 20 25 30 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Age (yr) Year 0 5 10 15 20 25 30 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Age (yr) Year A B

PAGE 126

126 Figure 4 11. Predicted historical biomass from both assessment models for golden tilefish from the northern Gulf of Mexico. Biomass reported by region ( A ) east and ( B ) west of the Mississippi River drainage for s tock s ynthesis (SS) (model 1, solid gray line and model 12, dotted gray line ) and stochastic stock reduction analysis (SRA) model ( age vulnerabilities fitted through SS, solid black line; age vulnerabilities calculated throug h VPA, dotted black lines ; ag e composition data included, dashed black line ) Biomass from SS is reported as total biomass. Biomass from SRA is reported as vulnerable biomass. 0 500 1000 1500 2000 2500 3000 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 Biomass (metric tons) Year 0 500 1000 1500 2000 2500 3000 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 Biomass (metric tons) Year A B SS, Model 1 SS, Model 12 SRA, age vulnerabilities fit by SS SRA, age vulnerabilities VPA SRA, age vulnerabilities VPA with age composition SS, Model 1 SS, Model 12 SRA, age vulnerabilities fit by SS SRA, age vulnerabilities VPA

PAGE 127

127 Figure 4 12. Predicted historical exploitation rates from both assessment models for golden tilefish from the northern Gulf of Mexico. S tock synthesis (SS) (model 1, solid gray line and model 12, dotted gray line) and stochastic stock reduction analysis (SRA) model (age vulnerabilities fitted through SS, solid black line; age vulnerabilities calculated throu g h VPA, dotted black lines ; age vulnerabilities calculated through VPA with age composition data, dashed black line ). 0.00 0.05 0.10 0.15 0.20 0.25 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 Exploitation rate Year SS, Model 1 SRA age vulnerabilities fit by SS SS, Model 12 SRA age vulnerabilities VPA SRA age vulnerabilities VPA with age composition

PAGE 128

128 Figure 4 13. Trace plots of the parameters f rom Stochastic SRA assessment model that determin ed the stock status of golden tilefish from the northern Gulf of Mexico. (A) Spawning Stock Biomass (SSB) in 2009/ Spawning Stock Biomass (SSB) at Maximum Sustainable Yield (MSY) and (B) Exploitation (U) in 2009/ Exploitation (U) at Maximum Sustainable Yield (MSY) for values from resulting MCMC chains (coda package; R Development Core Team, 2011). Horizontal line represents the median value for each ratio. A B

PAGE 129

129 Figure 4 14. Trace plots of the parameters from Stock Synthesis assessment model that determin ed the stock status for golden tilefish from the northern Gulf of Mexico. ( A ) Spawning Stock Biomass (SSB) in 2009/ Spawning Stock Biomass (SSB) at Spawning Potential Ratio (SPR) of 30% and ( B ) Fishing mortality (F) in 2009/ Fishing mortality (F) at Spawni ng Pot ential Ratio (SPR) of 30% for values from resulting MCMC chains (coda package; R Development Core Team, 2011). Horizontal line represents the median value for each ratio. A B

PAGE 130

130 Figure 4 15. Fried egg and beer plot for the parameters from Stochastic SRA assessment model that determin ed the stock status for golden tilefish from the northern Gulf of Mexico. SSB2009/SSBMSY and U2009/UMSY values from resulting MCMC chains (PBS modelling package; R Development Core Team, 2011). The lower left half is the fried egg (density contours ) the upper right half is the glass filled to the correlation point, and the diagonals show densities of the parameters. SSB 2009 /SSB MSY U 2009 /U MSY

PAGE 131

131 Figure 4 16. Fried egg and beer plot for the parameters from Stock Synthesis assessment model that determin ed the stock status for golden tilefish from the northern Gulf of Mexico. SSB2009/SSBSPR 30 and F2009/FSPR 30 values from resulting MCMC chains (PBS modelling package; R Development Core Team, 2011) The lower left half is the fried egg (density contours), the upper right half is the glass filled to the correlation point, and the diagonals show densities of the parameters. SSB 2009 /SSB SPR30 F 2009 /F SPR30

PAGE 132

132 CHAPTER 5 CONCLUSION Importance of S ampling One of the key messages from this dissertation (and a reason for seeking my degree) is that for stock assessments to provide meaningful results to inform management policies, these stock assessments must use the best available data and assessment scientist s should be knowledgeable on data limitations. It is very important to understand the methodologies of data collection. My doctoral research relied on federally funded long term data collection and routine biological sampling. My dissertation included t hree main objectives: 1) the validation of the timing of band deposition in sagittal otoliths, 2) the description of the reproductive strategy, and 3) the prediction of the stock status of golden tilefish from the Gulf of Mexico. For each of these objectiv es, sampling from long term fishery dependent monitoring programs for biological data (lengths) and samples (e.g., otoliths, gonads), along with effort and landings data and sampling from fishery independent surveys for biological samples and abundance dat a were critical to my research results. There are caveats in the collection of both fishery dependent and independent data that need to be understood to use these data sets appropriately. Fishery dependent data can be advantageous in that it is more generally available for more species (all that have a management plan), inexpensive, and often routinely collected covering a broad geographic area (Begg, 2005). I relied on the longterm fishery dependent program the NOAA Fisheries Service, Southeast Fisherie s Science Center (SEFSC) Trip Interview Program (TIP) for collect ing biological data and samples and the NOAA Fisheries Service, SEFSC Coastal Fisheries Logbook Program (CFLP) for relative abundance data. One caveat of these data is the sizeselective nature of fisheries. A fishery can be size selective due to a variety of fishery regulations of minimum size limit, an upper slot limit, gear restriction

PAGE 133

133 (e.g., hook size, bait type), area closure s or depth restrictions Another caveat of fishery dependent data is how TIP port agents collect biological samples from the landed catch. The TIP port agents guidelines for biological sampling are to attempt to purposely sample landings of multiple species from one or many fishing vessels at one time, which has lead to some species being under sampled while other more economically important species are oversampled. A third caveat of fishery dependent data is the behavior of commercial fishers. The CFLP is a self reporting program relying on the cooperation of fishers to accurately report where they fish (e.g., NMFS statistical shrimp grid), what they catch (e.g., type of fish landed, bycatch, discarded), and how they fish (e.g., number of days, number of hooks, length of mainline), and interpreting trends in abundance data may not be straightforward. Abundance data are highly influenced by fishers behavior s A fishers behavior can be affected by the current economics (i.e., cost per pound, fuel price, boat slip price), as well as technological advances (i.e., vessel electronics, changes in gear), which can alter the species being fished and the location of fishing. I was able to address several of these limitations common in fishery dependent data through collaborating with TIP port agents. The TIP port agent s are the primary contact for NOAA Fisheries Service with commercial fishers and are responsible for the collection of biological samples and in some Gulf of Mexico states monitor the reporting of landings data. My research objectives required specific bi ological sampling (e.g., both sagittal otoliths, gonads), but during routine TIP port agent biological sampling only one sagittal otolith is obtained and gonad tissue is not collected given that a majority of commercially harvested fish are gutted at sea. I worked with TIP port agents by first explaining the necessity for specific biological samples and second by describing the type of biological information that would be gained from these samples. The TIP port agents then contacted willing commercial capt ains that agreed to

PAGE 134

134 return to port with whole fish and allowed TIP port agents to have more time to process fish for both sagittal otoliths and gonad tissue. Through the efforts of both TIP port agents and cooperative commercial captains supplemental biol ogical samples were collected and contributed to the sampling needs for my first two research objectives. Fishery independent surveys provide an opportunity to collect data without the influences of the dynamics of a fishery (see caveats above). My research relied on two fishery independent surveys for biological samples and abundance data. However, abundance data collected during the longterm fishery independent survey was limited since only 10% of the surveys time is spent in depths associated with golden tilefish. I designed the second fishery independent survey with two objectives: 1) to estimate abundances of tilefish and deepwater grouper and 2) to characterize the spawning season of tilefish and deepwater grouper. This survey had two sampling methodologies to meet each objective, 1) randomized and stratified by depth (100 400 m) bottom long line survey (2 nm mainline, 200 4 m gangions with #15 circle hooks) and 2) conduct site specific monthly sampling using bottom long line gear. This one year project also contributed to the monthly gonad samples necessary to describe the reproductive seasonality and strategy of golden tilefish. Although, this project was only conducted for one year, its abundance data can be used as a baseline for futur e longterm bottom long line survey, specifically for tilefish and deepwater groupers. Fishery Management In the United States, fishery management is governed by the MagnusonStevens Act ( NOAA, 2009). This act provides the framework to prevent overfishi ng by permitting fishing to occur within three levels of fishing (annual catch limit, acceptable biological catch, fishing at overfishing level). The status of a fish stock can be described as: 1) overfished if the current level of biomass has declined or 2) undergoing overfishing if the current level of fishing

PAGE 135

135 mortality has declined below a level capable of producing a maximum sustainable yield NOAA, 2009). In 2004, a total allowable catch (TAC) of 200 metric tons for golden tilefish in the Gulf of was established to prevent overfishing. This regulation was based, not on biological information, but on the average of landings for the previous five years of fishing (prior to 2004). My dissertation provided the much needed biological data to better infor m management decisions for golden tilefish. In my first objective, I provided evidence for the longevity of golden tilefish, as well as, the uncertainty associated with traditional age estimates. This information is of particular importance, especially w hen predicting the productivity of a stock. I also provided evidence of golden tilefish to be classified as protogynous hermaphrodites. This type of reproductive strategy can have confounding affects to a population given the size selective nature of fis hing (removing the larger, older fish, predominately male). I applied this biological information to two assessment models that varied in complexity. These models agreed that golden tilefish stock in the Gulf of Mexico was not overfished nor was the stock undergoing overfishing. Thus, I conclude that the current fishery management regulation of total allowable catch of 200 metric tons is adequate in maintaining the golden tilefish population biomass at an optimal level.

PAGE 136

136 LIST OF REFERENCES Able, K. W., C. B. Grimes, R. A. Cooper, and J. R. Uzmann. 1982. Burrow construction and behavior of tilefish, Lopholatilus chamaeleonticeps in the Hudson Submarine Canyon. Environ. Biol. Fishes 7:199205. Able, K. W., C. B. Grimes, R. S Jones, and D. C. Twichell. 1993. Temporal and spatial variation in habitat characteristics of tilefish ( Lopholatilus chamaeleonticeps ) off the east coast of Florida. Bull. Mar. Sci. 53:10131026. Adams, S. 2003. Morphological ontogeny of the gonad of three plectropomid species through sex differentiation and transition. J. Fish Biol. 63:2236. Adkison, M. D. 2009. Drawbacks of complex models in frequentist and Bayesian approaches to natural resource management. Ecol. Appl. 19:198205. Alonzo, S. H., and M. Mangel. 2004. The effects of size selective fisheries on the stock dynamics of and sperm limitation in sex changing fish. Fish. Bull. 102:113. Anderson, R. O., and R. M. Neumann 1996. Length, weight, and associated structural indices. In Fisheries Techniques, 2nd edition. (B. R. Murphy, and D. W. Willis, eds.), p. 447482. American Fisheries Society, Bethesda, Maryland. Andrews, A. H. 2009a. Final Report. Leadradium dating of golden tilefish ( Lopholatilus chamaeleonticeps ). MARFIN. 07MFIH007. 2009b. Lead radium dating of two deepwater fishes from the southern hemisphere, Patagonian toothfish ( Dissostichus eleginoides ) and orange roughy ( Hoplostethus atlanticus ). Ph.D. Dissertation, 192 p. Department of Ichthyology and Fisheries Science, Rhodes University, Grahamstown, South Africa. Andrews, A. H., G. M. Caillet, K. H. Coale, K. M. Munk, M. M. Mahoney, and V. M. O'Connell. 2002. Radiometric age validation of the yelloweye rockfish ( Sebastes ruberrimus ) from southeaste rn Alaska. Aust. J. Mar. Freshw. Res. 53:139146. Andrews, A. H., K. H. Coale, J. L. Nowicki, C. Lundstrom, Z. Palacz, E. J. Burton, and G. M. Cailliet. 1999. Application of an ionexchange separation technique and thermal ionization mass spectrometry to 226Ra determination in otoliths for radiometric age determination of long lived fishes. Can. J. Fish. Aquat. Sci. 56:13291338.

PAGE 137

137 Andrews, A. H., D. M. Tracey, and M. R. Dunn. 2009. Leadradium dating of orange roughy ( Hoplostethus atlanticus ): vali dation of a centenarian life span. Can. J. Fish. Aquat. Sci. 66:11301140. Armsworth, P. R. 2001. Effects of fishing on a protogynous hermaphrodite. Can. J. Fish. Aquat. Sci. 58:568578. Asoh, K. 2003. Gonadal development and infrequent sex change in a population of the humbug damselfish, Dascyllus aruanus in continuous coral cover habitat. Mar. Biol. 142:12071218. Atz, J. W. 1964. Intersexuality in fishes. In Intersexuality in Vertebrates Including Man (C. N. Armstrong, and A. J. Marshall, eds .), p. 145232. Academic Press, London. Baird, T. A. 1988. Female and male territoriality and mating system of the sand tilefish, Malacanthus plumieri Environ. Biol. Fishes 22:101116. Baker, J., M. S., and C. A. Wilson. 2001. Use of bomb radiocarbon to validate otolith section ages of red snapper Lutjanus campechanus from the northern Gulf of Mexico. Limnol. Oceanogr. 46:18191824. Bannerot, S., W. F. Fox, and J. E. Powers 1987. Reproductive strategies and the managem ent of snappers and groupers in the Gulf of Mexico and Caribbean. In Tropical Snappers and Groupers: Biology and Fisheries Management (J. J. Polovina, and S. Ralston, eds.), p. 561603. Westview Press, Boulder. Baroiller, J. F., Y. Guiguen, and A. Fostier 1999. Endocrine and environmental aspects of sex differentiation in fish. Cell. Mol. Life Sci. 55:910931. Bateman, K. S., G. D. Stentiford, and S. W. Feist. 2004. A ranking system for the evaluation of intersex condition in European flounder ( Plat ichthys flesus ). Environ. Toxicol Chem 23:28312836. Begg, G. A. 2005. Life history parameters. In Stock Identification Methods, Applications in Fishery Science (S. X. Cadrin, K. D. Friedland, and J. R. Waldman, eds.), p. 119150. Elsevier Academic Press, Amsterdam. Bennett, J. T., G. W. Boehlert, and K. K. Turekian. 1982. Confirmation of longevity in Sebastes diploproa (Pisces: Scorpaenidae) from 210 Pb/226Ra measurements in otoliths. Mar. Biol. 71:209215.

PAGE 138

138 Broker, W. S., and T. S. Peng. 1982. Tracers in the Sea, Lamont Doherty Geological Observatory, Columbia University, Palisades, New York. B rown Peterson, N. J., D. M. Wyanski, F. SaboridoRey, B. J. Macewicz, and S. K. Lowerre Barbieri. 2011. A standardized terminology for describing reproductive development in fishes. Mar. Coast. Fish. 3:5270. Cailliet, G. M., A. H. Andrews, E. J. Burton, D. L. Watters, D. E. Kline, and L. A. Ferry Graham. 2001. Age determination and validation studies of marine fishes: do deepdwellers live longer? Experimental Gerontology 36:739764. Campana, S. E. 1997. Use of radiocarbon from nuclear fallout as a dated marker in the otoliths of haddock Melanogrammus aeglefinus Mar. Ecol. Prog. Ser. 150:4956. 2001. Accuracy, precision and quality control in age determination, including a review of the use and abuse of age validation methods. J. Fish Biol. 59:197242. Campana, S. E., K. C. T. Zwanenburg, and J. N. Smith. 1990. 210Pb/226Ra determination of longevity in redfish. Can. J. Fish. Aquat. Sci. 47:163165. Candy, S. G., G. B. Nowara, D. C. Welsford, and J. P. McKinlay. 2012. Estimating an a geing error matrix for Patagonian toothfish ( Dissostichus eleginoides ) otoliths using betweenreader integer errors, readability scores, and continuation ratio models. Fish. Res. 115116:1423. Choat, J. H., L. M. Axe, and D. C. Lou. 1996. Growth and l ongevity in fishes of the Family Scaridae. Mar. Ecol. Prog. Ser. 145:3341. Cole, K. S. 1990. Patterns of gonad structure in hermaphroditic gobies (Teleostei Gobiidae). Environ. Biol. Fishes 28:125142. 2002. Gonad morphology, sexual development, and colony composition in the obligate coral dwelling damselfish Dascyllus aruanus Mar. Biol. 140:151163. Condomines, M., and S. Rihs. 2006. First 226Ra -210Pb dating of a young speleothem. Earth Planet. Sc. Lett. 250:410. Conser, R. J., and J. E. Powers. 1990. Extension of the ADAPT VPA tuning method designed to facilitate assessment work on tuna and swordfish stocks. ICCAT Collect. Vol. Sci. Pap. 32:461468.

PAGE 139

139 Cook, M., G. R. Fitzhugh, and J. S. Franks. 2009. Validation of yellowedge grouper, Epinephelus flavolimbatus age using nuclear bombproduced radiocarbon. Environ. Biol. Fishes 86:461472. Cope, J.M. In press Implementing a statistical catch at age model (Stock Synthesis) as a tool for deriving overfishing limits in data limited situ ations. Fish. Res Cotter, A. J. R., L. Burt, C. G. M. Paxton, C. Fernandez, S. T. Buckland, and J. X. Pan. 2004. Are stock assessment methods too complicated? Fish Fish. 5:235254. Cowan, J., J. H., R. L. Shipp, I. Bailey, H. K., and D. W. Haywick. 1995. Procedure for rapid processing of large otoliths. Trans. Am. Fish. Soc. 124:280282. Dooley, J. K. 1978. Systematics and biology of the tilefishes (Perciformes: Branchiostegidae and Malacanthidae), with description of two new species. NOAA Tec h. Rep. 411:178. Drillings, C. C., and M. S. Grober. 2005. An initial description of alternative male reproductive phenotypes in the bluebanded goby, Lythrypnus dalli (Teleostei, Gobiidae). Environ. Biol. Fishes 72:361372. Elorduy Garay, J. F., and S. Ramrez Luna. 1994. Gonadal development and spawning of female ocean whitefish, Caulolatilus princeps (Pisces: Branchiostegidae) in the Bay of La Paz, B.C.S., Mexico. J. Fish Biol. 44:553566. Erickson, D. L., and G. D. Grossman. 1986. Reproductive demography of tilefish from the South Atlantic Bight with a test for the presence of protogynous hermaphroditism. Trans. Am. Fish. Soc. 115:279285. Erzini, K. 1990. Sample size and grouping of data for lengthfrequency analysis. Fish. Res. 9:355366. Fanning, K. A., J. A. Breland, and R. H. Byrne. 1982. Radium 226 and Radon222 in the coastal waters of west Florida: high concentrations and atmospheric degassing. Science 215:667670. Fennessy, S. T., and Y. Sadovy. 2002. Reproduc tive biology of a diandric protogynous hermaphrodite, the serranid Epinephelus andersoni Aust. J. Mar. Freshw. Res. 53:147158.

PAGE 140

140 Fenton, G. E., S. A. Short, and D. A. Ritz. 1991. Age determination of orange roughy, Hoplostethus atlanticus (Pisces, Trachichthyidae) using 210Pb:226Ra disequilibria. Mar. Biol. 109:197202. Forrest, R. E., S. J. D. Martell, M. C. Melnychuk, and C. J. Walters. 2008. An age structured model with leading management parameters, incorporating agespecific selectivity and maturity. Can. J. Fish. Aquat. Sci. 65:286296. Freeman, B. L., and S. C. Turner. 1977. Biological and fisheries data on tilefish, Lopholatilus chamaeleonticeps Goode and Bean. NOAA Tech. Ser. Rep. NMFS NEFSC 5:1 41. Gerritsen, H. D., and D. McGrath. 2007. Precision estimates and suggested sample sizes for lengthfrequency data. Fish. Bull. 105:116120. Gertseva, V. V., and J. M. Cope. 2011. Population dynamics of splitnose rockfish ( Sebastes diploproa) in the Northeast Pacific Ocean. Ecol. Model. 222:973981. Goode, G. B., and T. H. Bean. 1880. Description of a new genius and species of fish, Lopholatilus chamaeleonticeps from the south coast of New England Nineteenth Proceedings of the United States National Museum 2:205209. Goodyear, C. P. 1997. Fish age determined from length: an evaluation of three methods using simulated red snapper data. Fish. Bull. 95:3946. Grace, M., D. E. de Anda Fuentes, and J. L. CastilloGniz. 2004. Biological surveys to assess the relative abundance and distribution of coastal sharks and teleosts of the Mexican Gulf of Mexico; 1997, 1998, 2001, and 2002. Proc. Gulf Caribb. Fish. Inst. 55:271279. Grimes, C. B., K. W. Able, and R. S. Jones. 1986. Tilefish, Lopholatilus chamaeleonticeps h abitat, behavior and community structure in Mid Atlantic and southern New England waters. Environ. Biol. Fishes 15:273292. Grimes, C. B., C. F. Idelberger, K. W. Able, and S. C. Turner. 1988. The reproductive biology of tilefish, Lopholatilus chamaele onticeps Goode and Bean, from the United States Mid Atlantic Bight, and the effects of fishing on the breeding system. Fish. Bull. 86:745762. Grimes, C. B., S. C. Turner, and K. W. Able. 1983. A technique for tagging deepwater fish. Fish. Bull. 81:663666.

PAGE 141

141 Grimes, C. B., and S. C. Turner 1999. The complex life history of tilefish Lopholatilus chamaeleonticeps and vulnerability to exploitation. In Life in the Slow Lane, Ecology and Conservation of Long Lived Marine Animals (J. A. Musick, ed.), p. 1726. American Fisheries Society, Bethesda, MD. GMFMC ( Gulf of Mexico Fisheries Management Council ) 2011. Commercial Fishing Regulations for Gulf of Mexico Federal Waters. Tampa, FL. Hale, L. 2011. Bottom longline fishery bycatch of golden tilefish from observer data. 6 p. SEDAR25 DW11. Southeast, Data, Assessment, and Review, North Charleston, SC. Hall, N. G. 2003. Data requirements of multispecies, spatial, and ecosystem models. Population dynamics for fisheries management 106116 p. Australian Society for Fish Biology Proceedings. Harris, M. J., and G. D. Grossman. 1985. Growth, mortality, and age composition of a lightly exploited tilefish substock off Georgia. Trans. Am. Fish. Soc. 114:837846. Harris, P. J. 2005. Validation of ages for species of the deepwater snapper/grouper complex off the southeastern coast of the United States. 12 p. MARFIN NA17FF2870. Harris, P. J., S. M. Padgett, and P. T. Powers. 2001. Exploitation relat ed changes in the growth and reproduction of tilefish and the implications for the management of deepwater fisheries. AFS Symposium 25:155210. Heppell, S. S., S. A. Heppell, F. C. Coleman, and C. C. Koenig. 2006. Models to compare management options f or a protogynous fish. Ecol. Appl. 16:238249. Highly Migratory Species. 2007. Stock assessment report small coastal shark complex: Atlantic Sharpnose, Blacknose, Bonnethead, and Finetooth Shark. 395 p. NOAA/NMFS Highly Migratory Species Management Di vision, Silver Spring, MD. Hilborn, R. 2003. The state of the art in stock assessment: where we are and where we are going. Sci. Mar. 67:1520.

PAGE 142

142 Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics and uncertainty, Chapman and Hall, London. Hinck, J. E., V. S. Blazer, C. J. Schmitt, D. M. Papoulias, and D. E. Tillitt. 2009. Widespread occurrence of intersex in black basses ( Micropterus spp.) from US rivers, 19952004. Aquat. Toxicol. 95:6070. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mortality rates. Fish. Bull. 81:898903. 2007. Evaluation of TIP sampling design, field protocols, and data analysis procedures. 15 p. Semi Annual Progress Report for the period 4/1/07 9/30/07 for NOAA/NMFS Grant # NA05NMF4741247 with NMFS SERO. Hoffman, S. G., M. P. Schildhauer, and R. R. Warner. 1985. The costs of changing sex and the ontogeny of males under contest competition for mates. Evolution 39:915927. Hunter, J. R., and B. J. Mac ewicz. 1985. Measurement of spawning frequency in multiple spawning fishes. NOAA Tech. Rep. 36:7994. Jones, R. S., E. J. Gutherz, W. R. Nelson, and G. C. Matlock. 1989. Burrow utilization by yellowedge grouper, Epinephelus flavolimbatus in the nort hwestern Gulf of Mexico. Environ. Biol. Fishes 26:277284. Kalish, J. M. 1995. Application of the bomb radiocarbon chronometer to the validation of redfish Centroberyx affinis age. Can. J. Fish. Aquat. Sci. 52:13991405. Kastelle, C. R., D. K. Kimura and S. R. Jay. 2000. Using 210Pb/226Ra disequilibrium to validate conventional ages in Scorpaenids (genera Sebastes and Sebastolobus ). Fish. Res. 46:299312. Kerr, L. A., A. H. Andrews, K. Munk, K. H. Coale, B. R. Frantz, G. M. Cailliet, and T. A. Br own. 2005. Age validation of quillback rockfish ( Sebastes maliger ) using bomb radiocarbon. Fish. Bull. 103:97107. Kimura, D. K., and C. R. Kastelle. 1995. Perspectives on the relationship between otolith growth and the conversion of isotope activity ratios to fish ages. Can. J. Fish. Aquat. Sci. 52:22962303. Knoll, G. F. 1989. Radiation Detection and Measurement, Wiley, New York.

PAGE 143

143 Legault, C. M., and V. R. Restrepo. 1999. A flexible forward age structured assessment program. 49:246253. Lombardi Carlson, L. A., M. Cook, H. Lyon, B. A. Barnett, and L. Bullock. 2012. A description of age, growth, and reproductive life history traits of scamp ( Mycteroperca phenax) from the northern Gulf of Mexico. Mar. Coast. Fish. 4:129144. Lorenzen, K. 2005. Population dynamics and potential of fisheries stock enhancement: practical theory for assessment and policy analysis. Philos Trans R Soc Lond. 360:171189. Lowerre Barbieri, S. K., N. J. BrownPeterson, H. Murua, J. Tomkiewicz, D. M. Wyanski, and F. SaboridoRey. 2011. Emerging issues and methodological advances in fisheries reproductive biology. Mar. Coast. Fish. 3:3251. Ludwig, D., and C J. Walters. 1985. Are age structured models appropriate for catch effort data? Can. J. Fish. Aquat. Sci. 42:10661072. Mace, P. M., N. W. Bartoo, A. B. Hollowed, P. Kleiber, R. D. Methot, S. A. Murawski, J. E. Powers, and G. P. Scott. 2001. Marine fisheries stock assessment improvement plan. Report of the National Marine Fisheries Service National Task Force for Improving Fish Stock Assessments. 175 p. US Department of Commerce, National Oceanic and Atmospheric Administration. Mace, P. M., J. M. Fe naughty, R. P. Coburn, and I. J. Doonan. 1990. Growth and productivity of orange roughy ( Hoplostethus atlanticus ) on the north Chatham Rise. N. Z. J. Mar. Freshwat. Res. 24:105119. Martell, S. J. D., I. Pine, W. E., and C. J. Walters. 2008. Parameterizing age structured models from a fisheries management perspective. Can. J. Fish. Aquat. Sci. 65:15861600. McMillan, D. B. 2007. Fish Histology, Female Reproductive Systems. Springer, Netherlands. Methot, R. D. 2000. Technical description of the Stock Synthesis assessment program. 56 p. NOAA Tech. Memo. NMFS NWFSC 43. 2009a. Stock assessment: operational models in support of fisheries management. In The Future of Fisheries Science in North America (R. J. Beamish, and B. J. Roths child, eds.), p. 137165. Springer. 2009b. User manual for stock synthesis model version 3.03a. 143 p. NOAA Fisheries Service, Seattle, WA.

PAGE 144

144 Methot, R.D. and I. G. Taylor 2011. Adjusting for bias due to variability of estimated recruitments in fishery assessment models. Can. J. Fish. Aquat. Sci. 68:17441760. Moe, M. A. 1969. Biology of the red grouper Epinephelus morio ( Valenciennes) from the eastern Gulf of Mexico. 95 p. Florida Department of Natural Resources, Marine Research Laboratory, St. Petersburg, FL. Mohn, R. 2009. The uncertain future of assessment uncertainty. In The Future of Fisheries Science in North America (R. J. Beamish, and B. J. Rothschild, eds.), p. 495504. Springer Moore, W. S. 2007. Seasonal distribution and flux of radium isotopes on the southeastern U.S. continental shelf. J. Geophys. Res. Atmos. 112:116. Moyer, J. T. 1980. Influence of temperate waters on the behavior of the tropical anemonefish Amphiprion clarkii at Miyake Jima, Japan. Bull. Mar. Sci. 30:261272. Musick, J. A. 1999. Criteria to define extinction risk in marine fishes: the American Fisheries Society initiative. Fisheries 24:6 14. NOAA ( N ational Oceanic and Atmospheric Administration ) 2004. Fisheries of the Caribbean, Gulf of Mexi co, and South Atlantic; Reef Fish Fishery of the Gulf of Mexico; Red Grouper Rebuilding Plan. Federal Register 69:78987903. 2009. Magnuson Stevens Act Provisions; Annual Catch Limits; National Standard Guidelines; Final Rule. Federal Register 74: 317 83213. N ational Research Council 1998. Improving fish stock assessments, National Academy Press, Washington, DC. NEFSC (Northeast Fisheries Science Center). 2005. A Report of the 41th Northeast Regional Stock Assessment Workshop (41th SAW). 246 p. Northeast Fisheries Science Center Reference Document 05 14, Woods Hole, MA. 2009. A Report of the 48th Northeast Regional Stock Assessment Workshop (48th SAW). 842 p. Northeast Fisheries Science Center Reference Document 09 15, Woods Hole, MA. Ochi, H. 1989. Mating behavior and sex change of the anemonefish, Amphiprion clarkii in the temperate waters of southern Japan. Environ. Biol. Fishes 26:257275.

PAGE 145

145 Pacific Fisheries Management Council. 2000. Status of the Pacific Coast groundfish fishery through 2000 and recommended acceptable biological catches for 2001. Pacific Fishery Management Council, Portland, Oregon. Palmer, S. M., P. J. Harris, and P. T. Powers. 2004. Age, growth and reproduction of tilefish, Lopholatilus chamaeleonticeps along the southeast Atlantic coast of the United States, 19801987 and 19961998. 34 p. Southeast Data, Assessment, and Review. South Atlantic Fisheries Management Council, Charleston, SC. Panfili, J., H. de Pont ual, H. Troadec, and P. J. Wright. 2002. Manual of Fish Sclerochronology, Ifremer IRD, Brest, France. Parma, A. M., and R. B. Deriso. 1990. Dynamics of age and size composition in a population subject to size selective mortality: effects of phenotypic variability in growth. Can. J. Fish. Aquat. Sci. 47:274289. Pope, J. G. 1988. Collecting fisheries assessment data. In Fish Population Dynamics, 2nd edition (J. A. Gulland, ed.), p. 6382. John Wily & Sons Ltd. Porch, C. E. 1999. A Bayesian VPA w ith randomly walking parameters. ICCAT Collect. Vol. Sci. Pap. 49:314326. Powers, J. E., and V. R. Restrepo. 1992. Additional options for age sequenced analysis. ICCAT Collect. Vol. Sci. Pap. 39:540553. Prager, M. H. 1992. ASPIC A surplus product ion model incorporating covariates. ICCAT Collect. Vol. Sci. Pap. 28:218229. 1994. A suite of extensions to a nonequilibrium surplus production model. ICCAT Collect. Vol. Sci. Pap. 92:374389. Punt, A. E., D. C. Smith, K. KrusicGolub, and S. Robertson. 2008. Quantifying age reading error for use in fisheries stock assessments, with application to species in Australias southern and eastern scalefish and shark fishery. Can. J. Fish. Aquat. Sci. 65:19912005. R Development Core Team 2011. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria

PAGE 146

146 Rhodes, K. L., and Y. Sadovy. 2002. Reproduction in the camouflage grouper (Pisces : Serranidae) in Pohnpei, federated states of Micronesia. Bull. Mar. Sci. 70:851869. Richards, L. J., and J. T. Schnute. 1998. Model complexity and catch age analysis. Can. J. Fish. Aquat. Sci. 55:949957. Rideout, R. M., M. P. M. Burton, and G. A. Rose. 2000. Observations on mass atresia and skipped s pawning in northern Atlantic cod, from Smith Sound, Newfoundland. J. Fish Biol. 57:14291440. Rochet, M. J. 2000. A comparative approach to life history strategies and tactics among four orders of teleost fish. ICES J. Mar. Sci. 57:228239. Ross, J. L., and J. V. Merriner. 1983. Reproductive biology of the blueline tilefish, Caulolatilus microps off North Carolina and South Carolina. Fish. Bull. 81:553568. Sadovy de Mitcheson, Y., and M. Liu. 2008. Functional hermaphroditism in teleosts. Fish Fish. 9:143. Sadovy, Y., and M. L. Domeier. 2005. Perplexing problems of sexual patterns in the fish genus Paralabrax (Serranidae, Serraninae). J. Zool. 267:121133. Sadovy, Y., and D. Y. Shapiro. 1987. Criteria for the diagnosis of hermaphroditis m in fishes. Copeia 136156. Schnute, J. T., and L. J. Richards. 2001. Use and abuse of fishery models. Can. J. Fish. Aquat. Sci. 58:1017. SEDAR (Southeast, Data, Assessment, and Review) 2004. SEDAR 4. Stock assessment of the deepwater snapper grouper complex in the South Atlantic. 594 p,. Charleston, SC. 2005. SEDAR 7. Stock assessment report of Gulf of Mexico red snapper. 480 p., North Charleston, SC. 2006. SEDAR 12. Stock assessment report Gulf of Mexico red grouper. 358 p. North Charleston, SC. 2011a SEDAR 22. Stock assessment report Gulf of Mexico tilefish. 467 p. North Charleston, SC. 2011b. SEDAR 22. Stock assessment report for yellowedge grouper from the Gulf of Mexico. 423 p., North Charleston, SC. 2011c SEDAR 2 5. Stoc k assessment report South Atlantic tilefish. 330 p. North Charleston, SC.

PAGE 147

147 Shepherd, J. G. 1984. The availability and information content of fisheries data. In Exploitation of Marine Communities (R. M. May, ed.), p. 95109. Springer Verlag, Berlin. Smith, C. L. 1965. The patterns of sexuality and the classification of serranid fishes. American Museum Novitates 2207:120. Smith, J. N., R. Nelson, and S. E. Campana 1991. The use of Pb210/Ra 226 and Th 228/Ra 228 disequilibria in the ageing of ot oliths of marine fish. In Radionuclides in the study of marine processes ( P. J. Kershaw and D. S. Woodhead, ed s .), p. 350359. Elsevier, New York. St. Mary, C. M. 2000. Sex allocation in Lythrypnus (Gobiidae): variations on a hermaphroditic theme. Environ. Biol. Fishes 58:321333. Stephens, A., and A. MacCall. 2004. A multispecies approach to subsetting logbook data for purposes of estimating CPUE. Fish. Res. 70:299310. Stransky, C., G. Kanis ch, A. Kruger, and S. Purkl. 2005. Radiometric age validation of golden redfish ( Sebastes marinus ) and deep sea redfish (S. mentella) in the Northeast Atlantic. Fish. Res. 74:186197. Stuiver, M. and H. G. Ostlund. 1980. GEOSECS Atlantic Radiocarbon. Radiocarbon 22: 124. Tomlinson, J. 1966. The advantages of hermaphroditism and parthenogenesis. J. Theor. Biol. 11:5458. Turner, S. C., C. B. Grimes, and K. W. Able. 1983. Growth, mortality, and age size structure of the fisheries for tilefish, Lopholatilus chamaeleonticeps in the Middle Atlantic Southern New England Region. Fish. Bull. 81:751763. Tyler, C. R., and J. P. Sumpter. 1996. Oocyte growth and development in teleosts. Rev. Fish Biol. Fish. 6:287318. Walker, S. P. W., and M. I. McCormick. 2009. Fish ears are sensitive to sex change. Biol. Lett. 5:7376. Wallace, R. A., and K. Selman. 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:325343.

PAGE 148

148 Walters, C. J. 1986. Adaptive management of renewable natural resources, The Blackburn Press, Caldwell, New Jersey. Walters, C., and J. J. Maguire. 1996. Lessons for stock assessment from the northern cod collapse. Rev. Fish Biol. Fish. 6:125137. Walters, C. J., and S. J. D. Martell. 2004. Fisheries ecology and management, Princeton University Press, Princeton. Walters, C. J., S. J. D. Martell, and J. Korman. 2006. A stochastic approach to stock reduction analysis. Can. J. Fish. Aquat. Sci. 63:212223. Watanabe, K., and N. Suzuki. 1996. Sex differentiation, sexual maturity and the spawning season of the red tilefish Branchiostegus japonicus on the Pacific coast of Tokushima prefecture. Nippon Suisan Gakkaishi 62:406413. Yin, Y., and D. B. Sampson. 2004. Bias and precision of estim ates from an age structured stock assessment program in relation to stock and data characteristics. N. Am. J. Fish. Manage. 24:865879.

PAGE 149

149 BIOGRAPHICAL SKETCH Linda Anne Lombardi Carlson was born in New Bern, North Carolina to Irene and to the late Fred Lombardi i n 1974. Lindas fathers employment with the U.S. Marine Corp permitted the family to move quite extensively during Lindas primary school years. Linda attended elementary schools in Kailua, Oahu, Hawaii and Stafford, Virginia. Lindas father was an avid fisher and although his catch per unit effort was extremely low, it was these exp eriences that enticed Lindas interest in the ocean Linda spent he r senior year of high school in Sydney Australia. While in Australia, Linda completed a direct independent study with the University of Sydney. Linda returned to the states and graduated from North Stafford High School Stafford, VA in 1993. Linda beg an her collegiate education at Northern Virginia Community College in Woodbridge, Virginia, transferred to the University of North Carolina in Wilmington (UNCW) and graduated with a Bachelor of Science degree in Marine Biology in December 1996. Linda worked for the National Undersea Research Center in Wilmington, NC and completed a senior project under the direct of Dr. David Lindquist. During Lindas undergraduate school years, she also spent a summer abroad with the School for Field Studies in the T urks and Caicos Islands studying the resource management of coral reefs. Following her graduation from the UNCW, Linda was an intern at Mote Marine Laboratory, Center for Shark Research. In 1998, Linda was employed by the NOAA Fisheries Service, Southeast Fisheries Science Center (SEFSC), Panama City Laboratory, Panama City, Florida and began to learn about Gul f of Mexico reef fish life history and the complexities reef fish fisheries management. In 1999, Linda began her Master of Science degree through a cooperative agreement between the National Marine Fisheries Service and the University of Mississippi, Department of Biology, under the direction of Dr. Glenn Parsons. Lindas graduate

PAGE 150

150 research focused on the ecology and population dynamics of the bonnet head shark. Linda continued her employment as a fisheries biologist at NOAA/SEFSC Panama City after her graduation at the University of Mississippi in 2001. Linda began taking graduate courses at the University of Florida, Fisheries and Aquatic Sciences in the Fall of 2006 to decide on continuing her education in quantitative fisheries. She officially began her PhD in the Fall of 2008 and defended in the summer of 2012. L inda is presently employed as a Research Fisheries Biologist at the NOAA/SEFSC labo ratory in Panama City, Florida.