nursery delineation, habitat utilization, movements, and

W&M ScholarWorks W&M ScholarWorks

Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects

2001

Nursery delineation, habitat utilization, movements, and migration Nursery delineation, habitat utilization, movements, and migration

of juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia, of juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia,

United States of America United States of America

R. Dean. Grubbs College of William and Mary - Virginia Institute of Marine Science

Follow this and additional works at: https://scholarworks.wm.edu/etd

Part of the Ecology and Evolutionary Biology Commons, Fresh Water Studies Commons,

Oceanography Commons, and the Zoology Commons

Recommended Citation Recommended Citation Grubbs, R. Dean., "Nursery delineation, habitat utilization, movements, and migration of juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia, United States of America" (2001). Dissertations, Theses, and Masters Projects. Paper 1539616675. https://dx.doi.org/doi:10.25773/v5-vdqp-sc21

This Dissertation is brought to you for free and open access by the Theses, Dissertations, & Master Projects at W&M ScholarWorks. It has been accepted for inclusion in Dissertations, Theses, and Masters Projects by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected].

https://scholarworks.wm.edu/

https://scholarworks.wm.edu/etd

https://scholarworks.wm.edu/etds

https://scholarworks.wm.edu/etd?utm_source=scholarworks.wm.edu%2Fetd%2F1539616675&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/14?utm_source=scholarworks.wm.edu%2Fetd%2F1539616675&utm_medium=PDF&utm_campaign=PDFCoverPages




https://dx.doi.org/doi:10.25773/v5-vdqp-sc21

mailto:[email protected]

INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films

the text directly from the original or copy submitted. Thus, some thesis and

dissertation copies are in typewriter face, while others may be from any type of

computer printer.

The quality of this reproduction is dependent upon the quality of the

copy submitted. Broken or indistinct print, colored or poor quality illustrations

and photographs, print bleedthrough, substandard margins, and improper

alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript

and there are missing pages, these will be noted. Also, if unauthorized

copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by

sectioning the original, beginning at the upper left-hand comer and continuing

from left to right in equal sections with small overlaps.

Photographs included in the original manuscript have been reproduced

xerographically in this copy. Higher quality 6” x 9” black and white

photographic prints are available for any photographs or illustrations appearing

in this copy for an additional charge. Contact UMI directly to order.

Bell & Howell Information and Learning 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA

800-521-0600

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

NURSERY DELINEATION, HABITAT UTILIZATION, MOVEMENTS, AND

MIGRATION OF JUVENILE CARCHARHINUS PLUMBEUS IN CHESAPEAKE

BAY, VIRGINIA, USA

A Dissertation

Presented to

The Faculty of the School of Marine Science

The College of William and Mary in Virginia

In Partial Fulfillment

Of the Requirements for the Degree of

Doctor of Philosophy

by

R. Dean Grubbs

2001


UMI Number: 3003491

__ ®

UMIUMI Microform 3003491

Copyright 2001 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against

unauthorized copying under Title 17, United States Code.

Bell & Howell Information and Learning Company 300 North Zeeb Road

P.O. Box 1346 Ann Arbor, Ml 48106-1346


APPROVAL SHEET

This dissertation is submitted in partial fulfillment of

the requirements for the degree of

Doctor of Philosophy

R. Dean Grubbs II

Approved April 9, 2001

V jf lb ir A ' Musick, Fm.D.Committee Chairman/ Advisor

Herbert M. Au^in, Ph.D.

D3vrtfA \tvans, PI

Thomas B. Hoff, Ph.D. / /Mid-Atlantic Fisheries Management Council Dover, Delaware

JofyfF. Morrissey, P h , E L / Hofstra University Hempstead, New York


DEDICATION

I dedicate the completion of my doctoral dissertation to the memory of Captain Julian Anthony (Tony) Penello and the FA/ Anthony Anne. Captain Tony commercially fished the waters of the western North Atlantic for more than half a century. In an effort to give something back to the sea that had provided him so much, he began providing his vessel, a solid-built wooden trawler of ninety-eight feet, to researchers from the Virginia Institute of Marine Science in the mid- 1980’s. For the use of the Anthony Anne, he charged only the cost of operation and the lifetime of sea-going experiences and knowledge he offered free of charge. Captain Tony lost his cherished vessel, the Anthony Anne, in the winter of 1995. Despite this tragedy, he continued to provide his time, insight, and knowledge to a few students and their projects. This dissertation represents the completion of one of those projects. For me, Captain Tony became an irreplaceable mentor, advisor, confidante, and friend. Unfortunately, Captain Tony lost a battle with cancer in June 2000. For his efforts and undying loyalty I dedicate this body of research to his memory. To have met Captain Tony Penello was a privilege and to have known him was an honor.


TABLE OF CONTENTS

ACKNOWLEDGEMENTS......................................................................................................vi

FUNDING............................................................................................................................... viii

LIST OF MAPS....................................................................................................................... ix

LIST OF FIGURES...............................................................................................................xiii

LIST OF TABLES...............................................................................................................xviii

ABSTRACT..........................................................................................................................xxii

CHAPTER 1: SPATIAL DELINEATION OF SUMMER NURSERY AREAS TO

DEFINE ESSENTIAL FISH HABITAT FOR JUVENILE CARCHARHINUS

PLUMBEUS IN CHESAPEAKE BAY, VIRGINIA................................................................2

ABSTRACT.......................................................................................................................... 3

INTRODUCTION................................................................................................................. 5

MATERIALS AND METHODS..........................................................................................10Sampling Gear..................................................................................................................10Sampling Design............................................................................................................. 10Data Processing.............................................................................................................. 12Data Analysis....................................................................................................................18

RESULTS...........................................................................................................................27/; Spearman’s Rank Correlation................................................................................... 29II: Classification and Regression Tree (CART) Modeling ......................................36III: Testing the Tree Models......................................................................................... 49

DISCUSSION.....................................................................................................................68

LITERATURE C ITED........................................................................................................75

CHAPTER 2: MIGRATORY MOVEMENTS, PHILOPATRY, AND TEMPORAL

DELINEATION OF SUMMER NURSERIES FOR JUVENILE CARCHARHINUS

PLUMBEUS IN THE CHESAPEAKE BAY REGION.........................................................80

ABSTRACT........................................................................................................................81

INTRODUCTION............................................................................................................... 83


TABLE OF CONTENTS (continued)

MATERIALS AND METHODS......................................................................................... 89Sampling Gear.................................................................................................................89Sampling Design.............................................................................................................90Data Processing..............................................................................................................90Survival Factors and Tagging...................................................................................... 92Data Analysis................................................................................................................... 93

RESULTS.......................................................................................................................... 95I: Temporal Nursery Delineation.................................................................................95II: Migration Patterns.................................................................................................. 112III: Philopatry................................................................................................................ 113

DISCUSSION.................................................................................................................. 125

LITERATURE CITED..................................................................................................... 129

CHAPTER 3: SHORT-TERM MOVEMENTS AND SWIMMING DEPTH OF

JUVENILE CARCHARHINUS PLUMBEUS IN CHESAPEAKE BAY ..........................134

ABSTRACT......................................................................................................................135

INTRODUCTION............................................................................................................. 137

MATERIALS AND METHODS....................................................................................... 141Study Area ......................................................................................................................141Telemetry Gear.............................................................................................................. 143Transmitter Attachment............................................................................................... 144Tracking Methodology and Experimental Design..................................................147Data Analysis................................................................................................................. 148

RESULTS.........................................................................................................................157

I Activity Space.............................................................................................................. 159II Depth Telemetry.........................................................................................................171III Tidal Current Correlation........................................................................................ 196

DISCUSSION.................................................................................................................. 206

LITERATURE CITED..................................................................................................... 219

APPENDICESAPPENDIX 1 Nursery Spatial Delineation Station Data............................................ A-6APPENDIX 2-A Nursery Temporal Delineation D ata...............................................A-12APPENDIX 2-B Tag Return Data...............................................................................A-16

VITA


ACKNOWLEDGEMENTS

I am indebted to many people who aided in the completion of my doctoral degree and this dissertation. I thank my committee members for their support, patience, and advice, especially Drs. John Morrissey and Thomas Hoff for their critical reviews of this dissertation. Three academics have been especially influential in my professional development. My major advisor, Dr. Jack Musick, heightened my enthusiasm for the natural sciences and taught me that acquired knowledge need not be arbitrarily restricted to microcosmic subjects as has become the norm in biology. He taught me as much about subjects that were far outside my dissertation topic as those within. Dr. John Morrissey broadened my intellect through his highly critical and skeptical approach to science. His deep knowledge and dynamic teaching methods were also very influential in the development of my own style of instruction. Dr. Sonny Gruber hired me as a college freshman to work as a laboratory assistant, and has provided many professional opportunities over the years since that time. I am particularly grateful for the opportunities to develop my true professional love, that of teaching, through the many undergraduate and graduate courses I taught at his laboratory in the Bahamas.

Two men also served as personal mentors and confidantes during my degree tenure. They were the extremely competent captains of the large vessels used for the longline survey. Their importance to the success of this research can not be understated, however, I am more grateful to Captain Durand Ward and the late Captain Tony Penello for their friendship, tolerance, and advice.

I thank VIMS Vessel Operations for their logistical support and professionalism. They are an under-appreciated yet invaluable resource to the institute. I am especially grateful to George Pongonis for providing a dedicated vessel for the telemetry work, Mike Spangler for equipment fabrication, Robert Hudgins for emergency mechanical assistance, Raymond Forrest for vessel preparation, and Susan Rollins for scheduling logistics.

I thank the following volunteers and interns that participated in the monotonous yet grueling tracking cruises: S. Nichols, M. Hardt, K. MacDonald,R. Pemberton, J. Romine, S. Payne, C. Jones, J. Gelslecihter, K. Goldman, B. Watkins, E. Flores, and D. Ha. I am especially indebted to Christina Conrath who participated in the majority of the tracking cruises. These are the unsung heroes of projects such as this. I also thank the VIMS longline crew (Jim Gelsleichter, Rich Kraus, Ken Goldman, Jason Romine, and Christina Conrath) as well as the many volunteers that endured long hours and harsh conditions during the delineation and tagging cruises. I also thank Pat Geer for GIS advice and for providing environmental data from the trawl survey, and Rich Kraus for GIS and statistical advice.

vi


I am grateful for the many friendships made while at VIMS. I especially thank Rich Kraus for the fishing and herping adventures, Tripp MacDonald for enduring as my roommate for five years yet remaining a friend, Wolf Lange for helping maintain my sanity through snowboarding, and Roy Pemberton and Brett Falterman just for being themselves.

I owe tremendous gratitude to my family for their love and encouragement. I especially thank my mother, Audrey, for her undying support, my father, Ralph, for introducing me to the outdoors, my brothers, Derek, Austin, and Garrett, for enduring the trials of having me as an older brother, and my stepmother, Teresa, for her patience. I am also indebted to my in-laws, Dwight and Debbie Bullock, for their love, friendship, and support throughout this degree. Finally, I thank my wonderful wife, Audra. Her love and kindness were often neglected, yet her generosity, unselfishness and confidence never wavered in calm or angry seas.

vii


FUNDING

Financial support for this research was provided by the Virginia Marine Resources Commission - Salt Water License Fund, Wallop-Breaux Funds through the U.S. Fish & Wildlife Service, the National Marine Fisheries Service’s Apex Predator Investigations, and the Environmental Protection Agency’s STAR Fellowship Program.

viii


LIST OF MAPS

Map 1-1: Locations of standard stations sampled by the VIMS longline survey from 1973-1999 including stations K (Kiptopeke) and M (Middleground) in the lower Chesapeake Bay........................................................................................... 14

Map 1-2: Distribution of longline stations used to spatially delineate nursery for Carcharhinus plumbeus in Chesapeake Bay. Darker circles represent higher CPUE (sharks per 100 hooks).......................................................................... 15

Map 1-3: VIMS Trawl Survey stations sampled in August 1999 and interpolated monthly grids of variables used to assign values to associated longline stations. a) VIMS Trawl Survey Stations, b-d) August 1999 longline stations with b) bottom salinity grid (practical salinity units), c) bottom temperature grid (°C), d) bottom dissolved oxygen grid (parts per million) 16

Map 1-4: Interpolated grid of Distance to Bay Mouth (kilometers) variable. This grid was merged with longline stations (also shown) to estimate distance of each station to the mouth of the estuary......................................................................17

Map 1-5: Distribution of VIMS Trawl Survey stations sampled during the monthsof June through September for the years 1995 - 1999..............................................22

Map 1-6: Examples of interpolated variable grids used to display response surfaces for nursery delineation models, a) Distance to Bay Mouth, b)Depth, c) bottom salinity, d) bottom dissolved oxygen. Depth data are from EPA, Chesapeake Bay Program. Salinity and dissolved oxygen grids are interpreted from VIMS Trawl Survey data combined over the period June - September, 1995-1999.................................................................................................23

Map 1-7: Interpolated grid based on CPUE from longline stations....................................28

Map 1-8: Response surface regression tree (CART) modeling for Ecological Model I. Shaded area is portion of Chesapeake Bay with average summer salinity greater than 20.5 psu, depth greater than 5.5 meters, and dissolved oxygen concentration greater than 5.35 ppm. This area is interpreted to represent suitable nursery habitat according to the model......................................... 41

Map 1-9: Response surface regression tree (CART) modeling for Ecological Model II. Shaded area is portion of Chesapeake Bay with average summer salinity greater than 20.5 psu and depth greater than 5.5 meters,. This area is interpreted to represent suitable nursery habitat according to the reduced Ecological Model............................................................................................................42

ix


Map 1-10: Response surface regression tree (CART) modeling for theManagement Model. Shaded area is portion of Chesapeake Bay less than 34.5 km from the mouth of the Bay and depth greater than 5.5 meters,. This area is interpreted to represent suitable nursery habitat according to the reduced Management Model........................................................................................47

Map 1-11: Tag recaptures for juvenile C. plumbeus recaptured in Chesapeake Bay during the same year and subsequent years compared to CART suitable nursery habitat models, a) Recaptures compared with Ecological Model II b) Recapture compared with the Management Model................................ 51

Map 1-12: Telemetry fixes for nine juvenile C. plumbeus manually tracked for 11 to 64 hours (cumulative 350 hours) between 1996 and 1999 compared to CART suitable nursery habitat models. The minimum interval between fixes was six hours to avoid autocorrelation, a) Telemetry fixes compared with Ecological Model II b) Telemetry fixes compared with the Management Model..............................................................................................................................52

Map 1-13: Maps comparing suitable nursery habitat defined by CART modelsand logistic regression models a) Ecological Model b) Management Model.............67

Map 1-14: Comparison of suitable nursery habitat defined by a) CART Ecological Model I and b) CART Ecological Model II for a wet year (low salinity) and a drought year (high salinity)........................................................................................... 74

Map 2-1: Locations of standard stations sampled by the VIMS longline survey from 1973-1999 including stations K (Kiptopeke) and M (Middleground) in the lower Chesapeake Bay........................................................................................... 91

Map 2-2: Distribution of juvenile C. plumbeus tagging by VIMS during summersof 1995 to 2000........................................................................................................... 110

Map 2-3: Long distance tag returns. All recaptures made >200 kilometers fromtagging location, seasons combined.......................................................................... 118

Map 2-4: Short-term tag recaptures. All sharks recaptured the same summer inwhich they were tagged...............................................................................................120

Map 2-5: Evidence of natal homing. Tag recaptures made < 50 kilometers fromtagging location in summers following at least one winter migration...................... 122

Map 3-1: Movements of shark 9631 tracked for 50 consecutive hours (August 26- 28, 1997). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space.............................................................. 161

x


Map 3-2: Movements of shark 9632 tracked for 49 consecutive hours (August 6- 8,1997). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The shark was tracked for 15 additional hours three weeks later (August 27-28). Day and night courses of this second track are shown as light and dark lines respectively. The track is outlined by the Minimum Convex Polygon of activity space.................................... 162

Map 3-3: Movements of shark 9635 tracked for 43 consecutive hours (July 7-9,1998). Light circles are location fixes recorded during the day and darkcircle are location fixes recorded at night. The track is outlined by theMinimum Convex Polygon of activity space............................................................. 163

Map 3-4: Movements of shark 9637 tracked for 50 consecutive hours (July 13-15,1998). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by theMinimum Convex Polygon of activity space.............................................................. 164

Map 3-5: Movements of shark 9639 tracked for 44 consecutive hours (August 12- 14, 1998). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space...............................................................165

Map 3-6: Movements of shark 5285 tracked for 50 consecutive hours (July 26-28,1999). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by theMinimum Convex Polygon of activity space...............................................................166

Map 3-7: Movements of sharks tracked for less than 40 consecutive hours.Shark 9633 tracked for 13 hours (June 20-21, 1997); Shark 9630 tracked for 25 hours (July 8-9, 1997); Shark 9634 tracked for 10 hours (June 9-10,1998); and Shark 5375 tracked for 11 hours (September 2, 1999). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space............................................................................................ 167

Map 3-8: Minimum Convex Polygons for all ten juvenile Carcharhinus plumbeustracked. See Table 3-2 for area calculations associated with polygons................168

Map 3-9: Kernal Activity Space of Shark 9632. 50%, 75%, and 95% kernals areshown. See Table 3-2 for area calculations of kernals for all sharks tracked 169

XI


Map 3-10: Combined tracks for sharks 9631, 9632, 9635, and 5285 showing utilization of deep channel in eastern Chesapeake Bay and increased dispersal during night tracks, a) Daytime fixes only - Minimum Convex Polygon for these fixes combined contained an area of 160.08 km2, b)Nighttime fixes only - Minimum Convex Polygon for these fixes combined contained an area of 181.20 km2............................................................................... 217

Map 3-11: Perimeters of 95% Kernals for eight (9637 excluded) Chesapeake Bay tracks combined over bathymetry grid. Kernals are centered over deep channels throughout the lower estua........................................................................ 218

xii


LIST OF FIGURES

Figure 1-1: Matrix plot of CPUE plus five independent variables (distance tomouth, salinity, minimum set depth, dissolved oxygen, temperature)..................... 31

Figure 1-2: Matrix plot of CPUE plus four independent variables (year, latitude,longitude, maximum set depth).................................................................................... 32

Figure 1-3: Matrix plot of seven independent variables significantly correlatedwith CPUE according Spearman’s Rank Correlation statistics................................. 34

Figure 1-4: Full Ecological Model with CPUE as response and salinity, minimum depth, dissolved oxygen, and temperature as predictors, a) Full tree with predicted CPUE at each terminal node (11 nodes) b) Plot of residual mean deviance versus number of terminal nodes generated by cost-complexity pruning (value of cost-complexity parameter along top of graph).............................38

Figure 1-5: a) Ecological Model regression tree pruned to six terminal nodes with predicted CPUE below each node, b) Results of ten-fold cross-validation of each pruned subtree. Plot of cross-validated residual mean deviance versus number of terminal nodes (value of cost-complexity parameter along top of graph)..............................................................................................................................39

Figure 1-6: Ecological Model final tree pruned to three terminal nodes. Predicted CPUE and histogram shown below each terminal node. Histograms show distribution of CPUE values among the observations in each node.........................40

Figure 1-7: Full Management Model with CPUE as response and distance to mouth of estuary, minimum depth, dissolved oxygen, and temperature as predictors, a) Full tree with predicted CPUE at each terminal node (10 nodes) b) Plot of residual mean deviance versus number of terminal nodes generated by cost-complexity pruning (value of cost-complexity parameter along top of graph)........................................................................................................ 44

Figure 1-8: a) Management Model regression tree pruned to five terminal nodes with predicted CPUE below each node, b) Results of ten-fold cross- validation of each pruned subtree. Plot of cross-validated residual mean deviance versus number of terminal nodes (value of cost-complexity parameter along top of graph)......................................................................................45

Figure 1-9: Management Model final tree pruned to three terminal nodes.Predicted CPUE and histogram shown below each terminal node.Histograms show distribution of CPUE values among the observations ineach node.......................................................................................................................46


Figure 1-10: Plot of univariate selection functions ( w'(z) ). Parameter estimates from univariate logistic regressions for salinity, distance from mouth, depth, depth adjusted for distance, and depth adjusted for salinity (Table 1.12).................66

Figure 2-1: Monthly CPUE (sharks per 100 hooks) for the period 1996-1999 at a) Kiptopeke and b) Middleground stations using both standard hooks and circle hooks.....................................................................................................................98

Figure 2-2: Mean semimonthly C. plumbeus CPUE (sharks per 100 hooks) for Middleground and Kiptopeke stations combined. a)Standard 9/0 hooks - 1990 to 1999; b) 5/0 circle hooks - 1996 to 1999...................99

Figure 2-3: Mean semimonthly C. plumbeus CPUE (sharks per 100 hooks) andsurface temperature (°C) for Middleground and Kiptopeke stations combined for the years 1990-1999 (standard 9/0 hooks only)................................................. 101

Figure 2-4: Fitted line plots from linear regression of mean C. plumbeus CPUE(sharks per 100 hooks) vs. mean surface temperature. Means are for semimonthly intervals, (error bars = SEM) a) immigration period: May 1 - July 31; b) emigration period: July 15 - October 15..................................................103

Figure 2-5: Mean semimonthly C. plumbeus CPUE (sharks per 100 hooks) and day length (hours) for Middleground and Kiptopeke stations combined for the years 1990-1999 (standard 9/0 hooks only).............................................................. 105

Figure 2-6: Fitted line plot from linear regression of mean C. plumbeus CPUE (sharks per 100 hooks) vs. day length (hours) for emigration period (July 15 -October 15) only. Means are for semimonthly intervals, (error bars = SEM ) 107

Figure 2-7: Temporal delineation of C. plumbeus summer nursery using tagrecapture data. Distance between tag and recapture location vs. day of year of recapture...................................................................................................................116

Figure 2-8: Location and temporal delineation of C. plumbeus winter nursery using tag recapture data. Latitude of recapture location vs. day of year of recapture....................................................................................................................... 117

Figure 2-9: Evidence of natal homing from tag recaptures. Distance from tagging location versus days at liberty for all sharks recaptured less than 200 kilometers from the tagging location...........................................................................124

Figure 3-1: Swimming depth and bottom depth recordings for Shark 9630 tracked for 25 consecutive hours. (Bottom depths were interpolated from a bathymetry grid ArcView GIS using corrected fix locations - see text for correction methodology.)............................................................................................177

xiv


Figure 3-2: Swimming depth and bottom depth recordings for Shark 9631 tracked for 50 consecutive hours. (Bottom depths were interpolated from a bathymetry grid ArcView GIS using corrected fix locations - see text for correction methodology.)............................................................................................ 178





Figure 3-7: Shark 9630 swimming depth dynamics, a) Comparison of the distributions of swimming depth during day and night, b) Distribution of the distance from swimming depth to bottom of estuary during day and night 183



Figure 3-10: Shark 9635 swimming depth dynamics, a) Comparison of the distributions of swimming depth during day and night, b) Distribution of the distance from swimming depth to bottom of estuary during day and night............ 186

Figure 3-11: Shark 9637 swimming depth dynamics, a) Comparison of the distributions of swimming depth during day and night, b) Distribution of the distance from swimming depth to bottom of estuary during day and night............ 187

xv



Figure 3-13: Mean swimming depth during Day and Night based on sunrise and sunset for sharks 9630, 9631, 9632, 9635, 9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests. Overali mean daytime swimming depth was significantly deeper than nighttime swimming depth (t-test, T = 2.95, p=0.032). Error bars are standard error of the mean....................................................................................................................... 189

Figure 3-14: Mean swimming depth during Light and Dark based on timing of nautical twilight for sharks 9630, 9631, 9632, 9635, 9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests.Overall mean swimming depth during light was significantly deeper than mean swimming depth during dark (t-test, T = 2.85, p=0.036). Error bars are standard error of the mean..........................................................................................191

Figure 3-15: Mean swimming depth during Lunar High (moon risen) and Lunar Low (moon set) for sharks 9630, 9631, 9632, 9635, 9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests.Overall mean swimming depth was significantly deeper during lunar high than during lunar low (t-test, T = 2.71, p=0.042). Error bars are standard error of the mean..........................................................................................................193

Figure 3-16: Shark 9631 Directional Swimming Data, a) Frequency histogram of deviation of swimming direction from tidal current direction, b) Shark/Currentcorrelation index plotted with tidal current amplitude................................................199

Figure 3-17: Shark 9632 Directional Swimming Data, a) Frequency histogram of deviation of swimming direction from tidal current direction, b) Shark/Currentcorrelation index plotted with tidal current amplitude............................................... 200

Figure 3-18: Shark 9635 Directional Swimming Data, a) Frequency histogram of deviation of swimming direction from tidal current direction, b) Shark/Currentcorrelation index plotted with tidal current amplitude............................................... 201



xvi


Figure 3-21: Shark 5285 Directional Swimming Data, a) Frequency histogram of deiation of swimming direction from tidal current direction, b) Shark/Current correlation index plotted with tidal current amplitude...............................................204

xvii


LIST OF TABLES

Table 1-1: Spearman’s Rank Correlation (rho) for CPUE (sharks per 100 hooks)versus potential predictor variables..............................................................................33

Table 1-2: Spearman’s Rank Correlation (rho) for predictor variables. Onlysignificantly correlated predictor variables are shown................................................35

Table 1-3: Classification and regression tree (CART) pruning for Ecological and Management models. RSM is the residual mean deviance for the overall tree. RSMCV is the residual mean deviance of the ten-fold cross-validated model.............................................................................................................................. 48

Table 1-4: Classification ability of reduced ecological and management tree models. Each model was pruned to three terminal nodes. Classification is measured by the proportion of each estimation parameter is included in the terminal node possessing the highest shark CPUE................................................... 50

Table 1-5: Results of univariate logistic regressions, a) Model significance,classification, Somer’s D measure of model predictive ability; b) Goodness of fit tests for each univariate model (significance indicates lack of f it) ........................58

Table 1-6: Summary of overall logistic regression statistics for full and reducedEcological and Management Models a) overall model significance (-2 log likelihood = G), classification, Somer’s D measure of model predictive ability b) goodness-of-fit tests (p<0.05 indicates lack of fit)..................................................59

Table 1-7: Full Ecological Model logistic regression a) parameter estimates, significance, and odds ratios; b) model goodness-of-fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 27.925, df=4, p<0.0001)..............................................................................................................60

Table 1-8: Reduced Ecological Model (full model - temperature) logisticregression a) parameter estimates, significance, and odds ratios; b) model goodness-of-fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 23.102, df=3, p<0.0001)................................................ 61

Table 1-9: Reduced Ecological Model (salinity and depth ONLY) logisticregression a) parameter estimates, significance, and odds ratios; b) model goodness-of-fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 19.329, df=2, p=0.0001)................................................ 62

Table 1-10: Full Management Model logistic regression a) parameter estimates, significance, and odds ratios; b) model goodness-of-fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 38.876, df=4, p<0.0001)..............................................................................................................63


Table 1-11: Reduced Management Model (distance and depth ONLY) logistic regression a) parameter estimates, significance, and odds ratios; b) model goodness-of-fit statistics; c) model classification. Overall model significance(-2 log likelihood, G = 38.300, df=2, p<0.0001).......................................................... 64

Table 1-12: Parameter estimates, significance, and odds ratios from univariate logistic regressions for salinity, distance, depth, depth adjusted for distance, and depth adjusted for salinity. These parameter estimates were used to plot selection functions in Figure 1.9............................................................................. 65

Table 2-1: Linear regression summary for mean CPUE vs. mean surface temperature a) for the immigration period May 1 - July 31 and b) for the emigration period July 15 - October 15. Means are for semimonthly intervals......................................................................................................................... 102

Table 2-2: Linear regression summary for mean CPUE vs. mean day length a) for the immigration period May 1 - July 31 and b) for the emigration period July 15 - October 15. Means are for semimonthly intervals...................................... 106

Table 2-3: Summary of VIMS juvenile Carcharhinus plumbeus tagging activity1995-2000.................................................................................................................. 111

Table 2-4: Summary of WINTER tag recapture data. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. State refers to the coast nearest the recapture location labeled as NC (North Carolina), SC (South Carolina), NJ (New Jersey).................................................................................................................119

Table 2-5: Summary of SUMMER tag recapture data for returns occurring the same summer as tagging. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. All recaptures occurred in Virginia waters. Area of recapture summarized as Eastern Chesapeake Bay (CB-E), Western Chesapeake Bay (CB-W), and Virginia Eastern Shore (ES)..................................................................121

Table 2-6: Summary of SUMMER tag recapture data for returns occurring following at least one winter season. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. Area of Virginia recapture summarized as Eastern Chesapeake Bay (CB-E), Western Chesapeake Bay (CB-W), and Virginia Eastern Shore (ES), Virginia Beach (VB). Recapture from other states labeled as NC (North Carolina), SC (South Carolina), NJ (New Jersey)................123


Table 3-1: Summary of juvenile Carcharhinus plumbeus tracks.Regions: SECB - Southeastern Chesapeake Bay; SWCB - Southwestern Chesapeake Bay; SCCB - South Central Chesapeake Bay; VAES - Atlantic Side of Virginia’s Eastern Shore.................................................................................158

Table 3-2: Minimum Convex Polygon (MCP) and Kernal activity space, mean swimming speed, and mean swimming rate for all ten juvenile C. plumbeus tracked. Grand means only include tracks of greater than 24 hours in duration. Italicized activity space estimates were not included in calculation of mean activity spaces. (S.D. = standard deviation)............................................. 170

Table 3-3: Results of t-tests for difference between mean swimming depth (meters) during day and night based on sunrise and sunset. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate. (* indicates T statistic is significant at a= 0.05.)........... 190

Table 3-4: Results of t-tests for difference between mean swimming depth(meters) during day and night based on time of nautical twilight. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate. (* indicates T statistic is significant at a= 0.05.)...........192

Table 3-5: Results of t-tests for difference between mean swimming depth (meters) while moon was risen versus while moon was set. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate. (* indicates T statistic is significant at a= 0.05.)...........194

Table 3-6: Results of one-way Analysis of Variance for difference between mean swimming depth (meters) during ebb, slack, and flood tidal current phases.Means are given with standard deviation in parentheses. Results of analyses using raw depth data for each track are followed by ANOVA results using overall mean for each track as a replicate. (* indicates T statistic is significant at a= 0.05.).................................................................................................195

Table 3-7: Calculation of non-statistical correlation index between swimmingdirection and tidal current direction............................................................................ 197


Table 3-8: Carcharhinus plumbeus directional swimming (current correlation) data summary. Concentration index (r), angular dispersion (s), and mean angle are in relation to mean tidal current direction, not the overall distribution of the raw swimming direction data. Significant Rayleigh’s Z indicates the data are not distributed randomly around a circle, but have a mean directionality. Significant V-test results (as determined by the u statistic) indicate that the mean swimming direction is not statistically different from the mean tidal current direction.......................................................... 205

Table 3-9: Activity space (MCP) estimate for C. plumbeus compared with thepublished findings for three species of carcharhiniform sharks.............................. 216

xxi


ABSTRACT

Chesapeake Bay is possibly the largest summer nursery for Carcharhinus plumbeus in the western Atlantic. Longline sampling conducted from 1990-1999 was used to delineate this nursery spatially and temporally. Catch data from 83 longline stations sampled throughout the Virginia Chesapeake Bay were analyzed as a function of nine physical and environmental variables to delineate this nursery spatially. Tree-based models determined which variables best discriminated between stations with high and low catches and indicated that complex distribution patterns could be adequately modeled with few variables. The highest abundance of juvenile sharks was predicted where salinity was greater than 20.5 and depth was greater than 5.5 meters. Longline data from 100 sets made at two standard stations in the lower Bay indicated that immigration occurred in late May and early June and was highly correlated with increasing water temperature. Emigration from the estuary occurred in late September and early October and was highly correlated with decreasing day length. Between 1995 and 2000, 1846 juvenile C. plumbeus were tagged. With two exceptions, recaptures made in summer months were within 50 kilometers of the tagging location. Those recaptured in winter months were caught between 200 and 830 kilometers from the tagging location and indicated that the coastal waters of North Carolina and South Carolina serve as important winter nurseries from late October until May. Tag recaptures made in subsequent summers suggest that most juvenile sandbar sharks return to the same summer nurseries annually. Ultrasonic telemetry was used in investigate the diel activity patterns of juvenile C. plumbeus in Chesapeake Bay. Ten sharks were tracked for 10 to 50 consecutive hours. Swimming direction was correlated with mean direction of tidal currents. Mean activity space was conservatively estimated to be 110 km2, which is two orders of magnitude greater than that reported for other carcharhiniform species. Swimming depth ranged from surface to 40 meters and was significantly deeper during the day (12.8 meters) than during the night (8.5 meters). This diel activity pattern and large activity space is hypothesized to be an adaptation for foraging on patchy prey in a productive, yet dynamic, temperate estuary.

xxii


NURSERY DELINEATION, HABITAT UTILIZATION, MOVEMENTS, AND MIGRATION OF JUVENILE CARCHARHINUS PLUMBEUS IN CHESAPEAKE BAY, VIRGINIA, USA


CHAPTER 1Spatial delineation of summer nursery areas to define essential fish habitat

for juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia

2


3

ABSTRACT

The 1996 amendment to the Magnuson-Stevens Fishery Conservation

and Management Act mandated the delineation of essential fish habitat (EFH) for

all species regulated under federal management plans highlighting the

importance of nursery areas. Carcharhinus plumbeus has been federally

regulated since 1993, though much of its habitat lies in state-controlled waters.

The lower Chesapeake Bay is possibly the largest summer nursery for sandbar

sharks in the western Atlantic. The VIMS Longline Survey was expanded from

1990 to 1999 to include ancillary stations throughout the Virginia Chesapeake

Bay to delineate this nursery spatially. Catch per unit effort (CPUE) data from 83

stations were analyzed as a function of nine physical and environmental

variables. Spearman’s rank correlation was used to determine which variables

would be included in the models. Tree-based regression models determined

threshold values of the variables that were most influential and best discriminated

between stations with high and low CPUE. Minimum cost-complexity pruning

and cross-validation of the pruned tree models were used to develop optimal

trees. The tree models indicated that complex distribution patterns could be

adequately modeled with few variables. The highest abundance of juvenile C.

plumbeus was predicted where salinity was greater than 20.5 and depth was

greater than 5.5 meters. To increase applicability of the models to actual

management practices, distance to the mouth of the estuary was introduced as a

surrogate variable for salinity. The models determined that the highest

abundance of sharks was in areas less than 34.5 km from the mouth of the


4

estuary and in depths greater than 5.5 meters. Several dependent and

independent measures were used to estimate the classification and predictive

ability of these models. Both models performed very well using all measures.

Logistic regression using presence/absence data was used to validate the tree

models. The logistic models agreed very well with the tree models, selecting the

same variable combinations in predicting shark presence and absence. The

threshold variable values were spatially mapped in a Geographic Information

System (GIS). The resulting response surfaces were interpreted to represent

essential nursery habitat for C. plumbeus in Chesapeake Bay.


5

INTRODUCTION

Meek (1916) proposed that the available evidence indicated that females

of viviparous sharks resorted to specific areas of their range for the liberation of

young. He also stated that the carcharhiniform sharks Galeorhinus canis (=

galeus) and Mustelus canis (actually M. mustelus) had been observed to pup in

shallow water with the young remaining after the larger sharks migrated into

deeper water alluding to the use of shallows as protective nursery areas.

Springer (1967) stated that females of many coastal species, with special

reference to carcharhinids, migrate to specific areas in the shallower portions of

their range to pup and the young of many of these species remain in restricted

nurseries that are void of larger sharks that prey on the juveniles. Nurseries for

most species in the western North Atlantic are very discrete geographically and

are usually located in highly productive bays and estuaries (Castro 1987). It is

thought that such areas provide protection from predation as well as abundant

food to promote rapid growth in juveniles (Springer 1967, Branstetter 1990). The

use of such areas is an adaptation that probably evolved with the evolution of the

K-selected reproductive strategy characterized by large parental investment into

few young (Simpfendorfer and Milward 1993). Indeed, Lund (1990) has found

evidence of elasmobranch nursery areas in the fossil record from 320 million

years ago.

Though locations of nurseries are known for many species of coastal

sharks (Clarke 1971, Bass 1978, van der Elst 1979, Medved and Marshall 1981,

Gruber et al. 1988), data characterizing these areas geographically, physically,


6

chemically, and biotically are scant. Castro (1993) described length frequencies

and temporal migrations of nine species of carcharhiniform sharks utilizing Bull’s

Bay, South Carolina, as a nursery, however, environmental characterization of

this bay in relation to shark utilization was not included. Simpfendorfer and

Milward (1993) described temporal utilization and the distribution patterns

(aggregate versus random) for eight species of carcharhinid and sphyrnid sharks

utilizing Australia’s Cleveland Bay as a nursery, though no environmental data

were used to describe the patterns. Morrissey and Gruber (1993a,b) described

spatial utilization and habitat preferences based on depth, bottom type,

temperature, and salinity for juvenile Negaprion brevirostris in the North Sound of

Bimini, Bahamas. This study provided the most comprehensive physical

description of a shark nursery to date.

In 1996, the reauthorization of the Magnuson-Stevens Fishery

Conservation and Management Act through the Sustainable Fisheries Act (SFA)

established a mandate for the National Marine Fisheries Service (NMFS) and

Federal Fishery Management Councils requiring the identification of essential

fish habitat (EFH) for all species regulated by federal fishery management plans.

In short, 39 federal management plans had to be amended to include plans for

identification of EFH for more than 700 fishery stocks (Schmitten 1999). The

SFA defines EFH as “ ...those waters and substrate necessary to fish for

spawning, breeding, feeding, or growth to maturity” (USDOC 1996). Musick

(1999) pointed out, however, that in past EFH investigations, virtually all habitats

where a species is known to occur have been deemed EFH. This is particularly

problematic for highly migratory species such as large coastal sharks. Musick


7

(1999) recommended an EFH designation system based on utilization,

availability, and vulnerability. Using these criteria, estuarine nursery areas in

regions heavily impacted by humans are of particular concern. Packer and Hoff

(1999) proposed identifying 100% of estuarine areas where juvenile Paralichthys

dentatus (summer flounder) were found as EFH because the life stage is

estuarine dependent and the resource is overfished.

Carcharhinus plumbeus, the sandbar shark, is the dominant species in the

directed commercial shark fishery along the east coast of the United States,

constituting more than 2/3 of landings (Anonymous 1996, Castro 1993). It has

been federally regulated since 1993 by the Fishery Management Plan for Sharks

of the Atlantic Ocean (NMFS 1993). The plan covers 39 species of sharks

divided into large coastal, small coastal, and pelagic species groups.

Carcharhinus plumbeus is managed as part of the large coastal species group. It

reaches at least 239 cm in total length (Compagno 1984) and requires at least 15

years to reach sexual maturity (Sminkey and Musick 1995). In addition, this

species has been listed in the World Conservation Union (IUCN) Red List of

Threatened Species as a lower-risk, conservation-dependent species of concern

(IUCN 2000). This species inhabits insular regions to at least 250 m depth

(Garrick 1982) and is highly migratory with several hundred kilometers separating

summer and winter habitats (Bigelow and Schroeder 1948, Springer 1960), a

pattern that begins as neonates (Grubbs and Musick in prep a). These habits

make delineation of EFH problematic at best. Among the most critical EFH data

needs, however, is the delineation of summer pupping and nursery areas (Hoff

and Musick 1990, NMFS 1996).


8

The historical nursery grounds for Carcharhinus plumbeus in the western

North Atlantic are distributed in shallow water habitats on the east coast of the

United States from Long Island, New York (possibly north to Cape Cod,

Massachusetts) to Cape Canaveral, Florida (Springer 1960). Merson (1998)

found that this nursery range has contracted and now only extends from New

Jersey south to South Carolina. A secondary nursery exists in the northwestern

Gulf of Mexico (Springer 1960). The Chesapeake Bay is thought to be the

largest estuarine nursery area for C. plumbeus in the western Atlantic. Mature

females enter the lower Bay as well as the saline lagoons along the Eastern

Shore of Virginia to pup in May and June (Musick and Colvocoresses 1986).

These mature individuals then migrate offshore while the neonates remain in the

estuary until September or October. The young then migrate offshore and south

below Cape Hatteras (Grubbs and Musick in prep. a). The following spring, the

juveniles return to their natal nurseries to take advantage of the high productivity

and the protection provided by the estuary (Grubbs and Musick in prep a). The

young sandbar sharks continue to use the nursery during the warmer months for

the first four to ten years of life (Sminkey 1994, Grubbs and Musick in prep. b).

They then remain coastal year round, presumably only entering Mid-Atlantic

estuaries after reaching maturity to bear young.

The primary objective of this study was to delineate the summer nursery

for juvenile C. plumbeus in Chesapeake Bay spatially to define EFH and provide

a framework for delineation of EFH for other species in a format that can be

applied to management problems. This was accomplished by using the catch-

per-unit-effort (CPUE) from stations sampled by the VIMS longline survey in


9

Chesapeake Bay. These data were analyzed as a function of physical and

environmental variables using classification and regression tree (CART)

modeling and logistic regression to determine which variables were important in

defining suitable nursery habitat. These results were combined in a Geographic

Information System (GIS) framework using ArcView 3.1 and spatial grid

coverages of potential EFH were developed.


10

MATERIALS AND METHODS

Sampling GearSampling was conducted during longline cruises conducted by the Virginia

Institute of Marine Science (VIMS). These cruises are part of a long-term survey

to monitor the abundance and diversity of coastal sharks. Collections were made

with a commercial-style longline consisting of 3/8-inch tarred, hard-laid nylon

main line, which was anchored at each end and marked by a hi-flier marker buoy.

Three-meter gangions were spaced approximately 18 meters apart along the

main line and a large inflatable float was attached to the main line following every

20th gangion. Each gangion was composed of a stainless-steel tuna clip

attached to a 2-meter section of 1/8-inch tarred nylon trawl line, the end of which

was attached to a large barrel swivel. A 1-meter section of 1/16 inch galvanized

aircraft cable was crimped to the swivel and the other end was crimped to a

Mustad 9/0 stainless-steel shark hook. Each longline set consisted of 80-120

gangions. Bait consisted mostly of Brevoortia tyrannus, Atlantic menhaden, and

Scomber scombrus, Atlantic mackerel. Soak times were between three and four

hours. A standard 100-hook set covered about two kilometers. The statistical

unit was catch per unit effort (CPUE) defined as the number of sharks per 100

hooks.

Sampling Design

Since its inception in 1974, the longline survey conducted by VIMS has

routinely included stations in Chesapeake Bay. Two locations in the lower

eastern Bay, Kiptopeke (37°10' N, 76° 00' W) and Middleground (37° 06' N, 76°


11

03’ W), have been standard stations since 1980 (Map 1-1). The historical data

indicated the abundance of juvenile C. plumbeus in this region was very high

from June to September suggesting it is part of a primary summer nursery. The

survey was expanded in 1990 and 1993 to include additional ancillary sampling

in the Bay in an effort to delineate the summer nursery. These preliminary data

indicated the primary nursery was indeed concentrated in the lower eastern part

of the Bay where salinity is highest. Based on these preliminary sets, continued

sampling to delineate the nursery was limited to the Virginia portion of the Bay

south of 37° 50' N, an area of approximately 4000 km2. Logistical constraints

made random sampling of this area impractical, therefore, selection of sampling

sites were chosen haphazardly to maximize coverage. In the ten-year span from

1990 to 1999, 174 longline sets were made in Chesapeake Bay. Grubbs and

Musick (in prep a) indicated that utilization of the primary nursery occurs from

early June through September. Therefore, only sets made in these months were

used for spatial delineation. This reduced the number of sets to 147. Of these,

73 were standard sets made at Kiptopeke and Middleground, and 74 were

ancillary sets sampled to delineate the nursery. Twenty of these ancillary sets

were made between during the years 1990-1994, whereas 54 were made during

the period 1995-1999. To minimize sampling bias toward stations K and M,

mean CPUE was calculated for each of these stations over two-year periods.

This provided four CPUE estimates for station M and five for station K. These

were combined with the ancillary stations giving a total of 83 stations used for the

spatial delineation of the nursery (Map 1-2).


12

Data Processing

Catch-per-unit-effort was recorded for each station. Ail sharks were

measured and a condition factor from “Excellent" to “Poor” was assigned. All

healthy sharks were tagged with Hallprint® nylon dart tags. In addition, surface

temperature, minimum and maximum depth, and time of day were recorded. A

Hydrolab® multiprobe was used to record temperature in degrees Celsius, salinity

in practical salinity units (Lewis and Perkin 1978, JPOTS 1980), and dissolved

oxygen in parts per million at two-meter intervals from surface to bottom for

stations sampled from 1996 to 1999. The Virginia Institute of Marine Science

has conducted a trawl survey for juvenile finfish and blue crabs since 1955. This

survey consists of monthly, random stations sampled throughout Chesapeake

Bay. From 1990 to present these stations included surface and bottom records

of water temperature, dissolved oxygen concentration, and salinity. The number

of stations sampled per month ranged from a mean of 62 in the early 90’s to 119

in 1999. These data were used in ArcView 3.1 with the Spatial Analyst extension

to interpolate grids mapping each of these variables. Maps were made for each

variable for the months June through September 1990-1999. All longline stations

were layered over the appropriate monthly maps. By merging tables,

interpolated environmental variables could be assigned to each station. For

example, Map 1-3a shows the distribution of stations sampled by the trawl survey

for August 1999. From these stations, grids were interpolated for bottom salinity

(Map 1 -3b), temperature (Map 1-3c), and dissolved oxygen (Map 1 -3d) using

ArcView 3.1. The longline stations sampled were merged with these grids and a


13

value for each variable was assigned for the station. This method was used to

assign environmental variables to all stations for which these were not measured

in situ. In addition, a line was drawn from Cape Henry to Fisherman’s Island to

delimit the mouth of Chesapeake Bay. A grid coverage of distance to this line

was interpolated in ArcView 3.1 (Map 1-4).


14

Map 1-1: Locations of standard stations sampled by the VIMS longline survey from 1973-1999 including stations K (Kiptopeke) and M (Middleground) in the lower Chesapeake Bay.


Reproduced

with perm

ission of the

copyright owner.

Further reproduction prohibited

without perm

ission.

76°30'76°p0'76°30’

76°00’ 75°30'76°30‘

76°00* 74°30'

<8*..A

C i)

$

{ ' i

<8>

20/(

40 Kilometers C*0>

75°00‘ 74°30‘

15

Map 1-2: Distribution of longline stations used to delineate the Carcharhinus plumbeus nursery spatially in Chesapeake Bay. Darker circles represent higher CPUE (sharks per 100 hooks).


36°6

0’ 37

°00‘

37°1

0'

37‘?

0‘ 37

^30'

37“4

0*

37^0

*

76°30‘ 76°20' 76°10' 76°00*

Nursery Stations Stations K & M condensed

CPUE (sharks per 100 hooks)O oO 1.0-3.0 • 3.0-6.0

>6.0

10 0 . 1C 20 K i l o m e t e r s

76°30‘ 76°20‘ 76°10‘ 76°00‘

75°50'

75°60'


16

Map 1-3: Stations sampled by the trawl survey in August 1999 and monthly grids of variables, interpolated in ArcView, used to assign values to associated longline stations, a) Stations sampled by the trawl survey, b-d) longline stations sampled in August 1999 with grids of b) bottom salinity (practical salinity units), c) bottom temperature (°C), and d) bottom dissolved-oxygen (parts per million).


37^0*_________ 37*30*_________ 37 TO*_________3710-_________ 3 7 W 37^0*_______ 37*30* 37TO* 37*10*

c£> ►_ OD ’Ti O •- <~1 -7 ir> to--r-^Nc<rufNr<isoj•-iniSf'CDO'Q'-Mrl’Ti/lN » - -r-i-r-NfirtNNN A

r*> n in id ^o» • • . ■ • r*.V N fl ^ tf) (fl A

6661 isnBnw joj!nsd) ou0 Dc*n.*!0 djs3:ui

5 6 6 1 i^ n B n y (oidd) p u s u a B ^ o paAjossiQ

E g

■07. IS OCOC AZ*tC J3U1C A k t t JtKtt Mat atut jouts oaur

■OKif O t . i t A i t O U lt JM Li£


17

Map 1-4: Interpolated grid of Distance to Bay Mouth (kilometers) variable. This grid was merged with longline stations (also shown) to estimate distance of each station to the mouth of the estuary.


37°10‘

.OCoZC ,0Zol£ .0 UXC .OOblS


18

Data Analysis

The null hypothesis states:

Ho: The distribution of juvenile Carcharhinus plumbeus in Chesapeake

Bay is completely random.

The working alternative hypothesis states:

Hi: Juvenile Carcharhinus plumbeus distribution in Chesapeake Bay is

not random, but is a function of some combination of nine measured

environmental and physical variables.

All stations were mapped in a Geographic Information System (GIS) using

ArcView 3.1 software. Using the ArcView extension, Spatial Analyst, a grid was

interpolated using the Inverse Distance Weighted method using the three nearest

neighbors and a power of three. This grid interpolated areas where high

(CPUE>6.0), moderate (CPUE between 3.0 and 6.0), and low (CPUE between

1.0 and 3.0) shark abundance would be expected based on the data. These

areas were interpreted to represent the spatial delineation of primary, secondary,

and tertiary nursery areas.

Spearman’s rank correlation (rho) was calculated between CPUE and nine

measured variables using SPSS statistical software. The non-parametric

Spearman’s Rho was chosen due to the non-normality of the CPUE data, even

after multiple transformation attempts and due to its low sensitivity to outliers.

These results were used to determine which variables best explained the

observed abundance distribution. Catch per unit of effort was modeled as a


19

function of these habitat parameters with Classification and Regression Trees

(CART) using S-Plus statistical software. CART models use the method of

binary-recursive partitioning to split data according to the predictor variables in a

way that minimizes response variance within resulting nodes. The best splits are

those that separate high and low values of the response variable (Breiman et al.

1984). Partitioning continues until nodes are homogeneous or there are too few

observations to continue (S-Plus 2000). In this study, the minimum node size

was set at one and five observations to grow initial trees. The initial regression

trees tend to overfit the data and are more complex and difficult to interpret than

needed (S-Plus 2000, Norcross et al. 1997). A three-step process was used to

simplify these trees. In the first step minimal cost-complexity pruning was

applied to the full tree using S-Plus statistical software. Pruning reduces the

number of terminal nodes by successively clipping the least important nodes

from the full tree (S-Plus 2000). For each subtree constructed the cost-

complexity measure, Ra(T) , is calculated wherein:

Ra(T) = R(T) + a \f

The variable a is a cost complexity parameter, R(T) is the total deviance of

subtree T, and f is the tree complexity defined as the number of terminal nodes

(Breiman et al. 1984). For each value of or, a subtree is constructed that

minimizes the cost-complexity, thereby producing a nested sequence of subtrees

of progressively decreasing complexity and increasing deviance. Steps two and

three involved selecting the best subtree. Ten-fold cross-validation was

performed on the pruned tree sequence using S-Plus. This procedure uses 90%


20

of the data to construct a tree, then tests this tree using the 10% of the data held

out. This is repeated ten times for the initial tree and each pruned subtree. The

deviance, R"(T) , is calculated for each subtree. This deviance is typically

lowest at intermediate tree sizes (Breiman et al. 1984). The optimum tree size

was chosen as the one wherein RCV(T) is at a minimum.

The ability of a given tree to predict areas of high abundance of C.

plumbeus was assessed by three post-hoc estimates of classification. The first

was the proportion of total sharks caught in the survey that resided in the

terminal node predicting highest shark abundance. The second was the

proportion of sets with CPUE>3.0 that resided in the terminal node predicting

highest shark abundance. The threshold of 3.0 was selected arbitrarily to

represent primary nursery areas. The third was the proportion of sets with

CPUE>1.0 that resided in the terminal node predicting highest shark abundance.

The threshold of 1.0 was selected arbitrarily to represent primary and secondary

nursery areas combined.

Spatial grid coverages were developed for all predictor variables using

ArcView 3.1 GIS. A grid coverage of distance to Bay mouth was also interpolated

as described in the previous section (Map 1-6 a). Greater than 2000 trawl-survey

stations were sampled during the months June through September in the years

1995-1999. The distribution of these stations is shown in Map 1-5. These

stations were used to interpolate spatial grids of average summer temperature,

dissolved oxygen, and salinity (Map 1-6, c, d). A bathymetry grid was

interpolated from TIN (triangular irregular network) data from the EPA,


21

Chesapeake Bay Program (Map 1-6 b). Crucial values of those variables that

were most influential in explaining increased C. plumbeus CPUE distribution

according to the CART models were combined spatially and mapped. The

resulting spatial-grid coverages were interpreted to represent suitable nursery

habitat and thus define EFH according to the model.


22

Map 1-5: Distribution of stations sampled by the trawl survey during the months of June through September for the years 1995 -1999.


37°40' 37°30‘ 37°20* 37°10‘ 37aQ0'

!0 I

I

i

SI*bl

S i

o2

.ofc^e .0 SoLt .0 Zolt .Oiolt .oobie


76°3

0’ 76

°201

76°1

0‘ 76

°00‘

76°6

0‘ 76

°40*

76

°30*

23

Map 1-6: Examples of interpolated variable grids used to display response surfaces for nursery-delineation models, a) Distance to Bay Mouth, b) Depth, c) Bottom salinity, d) Bottom dissolved-oxygen. Depth data are from EPA, Chesapeake Bay Program. Salinity and dissolved-oxygen grids are interpreted from data collected by the trawl survey combined over the period June • September, 1995-1999


37*®*________J 7 *3 0 *_________ 37*20* 37-10*

233~*siasssM M H(MBui)mdaajiaM

37-i®* 37*30* J7T0* 37*10*

V CM rt *t /I ID A

6 6 5 1 *5 6 6 L J^> q u j3 idvS -o un r irudd) p u s u o fi/< o paAjos^iQ

J9WS MUt OLtf 4Xk£C 4€jc mu c ou u otkzr

OUiX

37-40*_________ 37*30* 37*20* 37-10*

666 1*9 6 6 : io q u jo id o s - o u n r 'n*»d) piiQ /iiu in s oc*u*iodjo;ui

AK4X J ttU t J X U t O U tt


24

Testing the Model

Logistic regression was used to validate the CART nursery models.

CPUE was coded as a binary-response variable for all stations used in the CART

models. Stations with CPUE equal to zero were coded as 0 and those with

CPUE greater than zero were coded as 1. The logistic regression models used

the same predictor variables used in the CART models. All continuous predictors

were coded as categorical variables to improve model stability and goodness-of -

fit (Hosmer and Lemeshowe 1989). Highly correlated variables were not

analyzed in the same model to avoid multicollinearity. Logistic regression is a

predictive analysis that employs the principle of maximum likelihood rather than

least squares for parameter estimation. Therefore, it is free of many of the

restrictive assumptions of ordinary least squares regression. Logistic regression

does not assume the dependent variable is normally distributed and

homoscedastic, that error terms are normally distributed, or that a linear

relationship exists between dependent and independent variables (Menard,

1995). Multiple logistic regression has been used in similar habitat-selection

studies (Manly et al. 1993, Norcross et al. 1999). The resulting probability

function, \v(z), which predicts resource selection (Manly et al. 1993), takes the

form:

a+fix.X\+...+Ppxp


25

wherein a represents a constant, and Pp is the coefficient of the independent

variable xp. For this study, this was the probability of catching one juvenile C.

plumbeus given the model variables. Values of w'{z) <0-5 predicted shark

absence and those >0.5 predicted shark presence.

Univariate logistic regressions were conducted on each of the

independent variables. Multivariate regressions were performed using the same

variable combinations used in the CART models. All variables with coefficient

estimates not significantly different from zero were removed from the models.

The overall significance of each model was determined by the -2 log-likelihood

statistic, G, which approximates a chi-square distribution (Agresti 1990). Model

fit was assessed using three goodness-of-fit tests: Hosmer-Lemeshow, Deviance

chi-square, and Pearson chi-square. Predictor variables from significant

univariate and multivariate models were compared to the variables selected by

the CART models. The probability function estimating resource selection was

plotted for the most significant univariate models. The value of the predictor

variable wherein w’ (^)=0.5 was taken as a reference and compared to the

critical values dividing stations of high and low shark abundance in the CART

models. Areas of Chesapeake Bay that the predictor variables predicted the

probability of catching a shark to be greater than 50% were mapped spatially and

the resulting grid coverages were added to the maps from the CART modeling

for comparison.

The CART models were further validated by two independent data

sources. During the summers of 1997, 1998, and 1999 a total of nine juvenile C.


26

plumbeus were tracked manually in Chesapeake Bay using ultrasonic telemetry

for periods ranging from 10 to 64 hours to examine habitat utilization and diel

activity patterns (Grubbs and Musick in prep. c.). Position fixes were recorded at

10-minute intervals. The position fixes were overlaid on the model nursery grid

as a post-hoc means of validating the model. To avoid autocorrelation of

samples, only fixes separated by at least six hours were used. The percentage

of fixes within the model nursery grid was used as an estimate of model

validation.

From 1995 to 1999, 1615 juvenile C. plumbeus were tagged with Hallprint

nylon dart tags to examine population structure and long-term movements

(Grubbs and Musick, in prep. a,c). To date, 45 recaptures have been reported.

Information is complete for 40 of these. Of the 31 recaptures occurring in

summer months all except one were recaptured inside the Bay, just outside the

mouth of the Bay, or in seaside lagoons along the Virginia Eastern Shore. Of the

22 that were recaptured inside the Bay, 13 were recaptured the same summer

they were tagged and 9 were recaptured in subsequent years. These recapture

locations were overlaid to examine the fit of the nursery model. The percentage

of these long-term and short-term recaptures that occurred within the model

nursery was used as an additional estimate of model validation.


27

RESULTS

The highest abundance of juvenile C. plumbeus was found in the southern

portion of the Bay. The distribution of longline stations sampled in Chesapeake

Bay between 1990 and 1999 and associated CPUE, after accounting for clumped

sampling at Kiptopeke and Middleground, are shown in Map 1-2. A grid

interpolated from these data showing areas of high (CPUE > 3.0) and low (CPUE

>1.0, <3.0) estimated abundance is shown in Map 1-7. These highlighted areas

represent the primary and secondary nursery areas according to the data. This

nursery grid is relatively patchy, which is not surprising given the relatively low

sample size and the extremely high variance associated with CPUE data

collected using longline sampling. The grid does, however, suggest that the

primary nursery is concentrated south of 37°30' N latitude and is relatively evenly

distributed longitudinally. Interestingly, the primary and secondary nursery areas

include the outer mouth of the York River but very little of the James River

located closer to the mouth. This may be explained by the higher volume of

freshwater discharge and increased industrial and agricultural runoff typical of the

lower James. The standard longline stations, Kiptopeke and Middleground, are

well within the primary nursery area according to these data. Those areas

outside the grid coverage are not in the described nursery and are interpreted to

have very low densities of juvenile C. plumbeus.


28

Map 1-7: Interpolated grid based on CPUE from longline stations.


op.00TO

<o|

76°30' 76°15' 76°00'

Nursery GridInverse Distance Weighted Interpolation Neighbors = 3, Power = 3 H 1.0 - 3 . 0

3 . 0 - 6 . 0 > 6.0

oS?h-TO

•Or .h-TO

r* A \/-v

OPKTO

14 K i l o m e t e r s

76°30' 76°15' 76°00'


75°45‘

75°45‘

.Stto9

C io

SIe

.91-

olt

.O&L

t .9f

c, It

.00.

89

29

/: Spearman’s Rank Correlation

Nine independent variables were used in the analysis and all were

significantly correlated to CPUE except maximum set-depth and year, therefore

these two were not included in the nursery models. The results of the

Spearman’s Rank Correlation are shown in Table 1-1. Matrix plots of the

independent variables and the dependent variable, CPUE, are shown in Figures

1-1 and 1-2. Distance to bay mouth, bottom temperature, and latitude had highly

significant, negative correlation coefficients with CPUE (p<0.001, 2-tailed).

Bottom salinity had a highly significant, positive correlation coefficient with CPUE

(p<0.001, 2-tailed). Significant positive correlation coefficients (p<0.05, 2-tailed)

were found between CPUE and bottom dissolved-oxygen, minimum set-depth,

and longitude also.

Figure 1-3 is a matrix plot of the seven predictor variables that had

significant rank-correlation coefficients. Many of these variables were correlated

with each other Table 1-2. Latitude was highly correlated with all six other

variables, and was therefore eliminated from consideration for the CART models.

Longitude was highly correlated to three of the other variables. In addition, the

significant correlation of longitude to CPUE was largely driven by one point (Fig.

1-3), therefore it was also eliminated from model consideration. The remaining

five variables (distance to Bay mouth, bottom salinity, bottom dissolved-oxygen

concentration, bottom temperature, and minimum set depth) were retained for

the CART models though several of them were correlated (Fig. 1-3). The highest

correlation was between distance to Bay mouth and salinity (Table 1-2). This is


30

intuitive as distance to Bay mouth was initially selected as a potential predictor

variable only to act as a surrogate to salinity to increase applicability of the

models to management problems. Due to this high correlation these two

variables were not included together in any model.


31

Figure 1-1: Matrix plot of CPUE plus five independent variables (distance to mouth, salinity, minimum set-depth, dissolved oxygen, temperature)


80 “60

40120 ~

o l

2015_

1 (f5o-

2826242220

20T51050

3025201510

86420

i i i i i G) 70 40 60 80 i i i i (a 5 10 1fi 20 1 i i i 20 22 24 26 28

CPUE

0 -PO - oo o § % °

°oo0

° v K Lo a a jP * '

o-*<^8oQOt]8O0BB o

% o °

® °<9&°‘b

0 8 O

° °%% 8

Distance to Bay Mouth

0 D

%

c

c

5 Q O jxg^gcr r> ^r>

3

O

BottomSalinity

—

° 00 oo

oo o uo

o

° O i

co

" F - 88 o 8

— a---------

o

3

O

O

O

QJ ^ h° 8 o

MinimumDepth

0

o

a 3

0

c

8 o

>

1 f t° §

o

. _°

8 °

O o °

° ° o

°o °

o ° o°°

BottomDissolvedOxygen

o o o ooooo

n oo

a> o c<8 < # > {r °00 O 000

o1 \ \ \ \

03D O O O O

ooooo ocgotaert) (D m x d ra o o o o

oo o o

1 1 1 1 I

o oo o o o

^«QOOOOO 00 oSosqqoo oo

o o o

1 T 1 1

coo o oo

0 oooooo OOQQOQQOOOQ O

oo o o

1 I 1 I

O COO ODD

0 OOOGDO O OOTnii» iB in |iiiiii

o ooupJJ iw —o o o q ^ o oo o

000o

\ T I I

BottomTemperature

—1-------1-------1-------1-------1—

0 5 10 15 20 10 15 20 25 30 0 2 4 6 8

32

Figure 1-2: Matrix plot of CPUE plus four independent variables (year, latitude, longitude, maximum set-depth).


OOGD

me %

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission

33

Table 1-1: Spearman’s Rank Correlation (rho) for CPUE (sharks per 100 hooks) versus potential predictor variables.

SPEARMAN’S NONPARAMETRIC CORRELATION (rho)VARIABLE Correlation Coefficient Significance NDistance to Bay Mouth -0.535** <0.001 83Bottom Salinity 0.447** <0.001 83Bottom Temperature -0.351** <0.001 83Bottom Dissolved Oxygen 0.232* 0.034 83Minimum Set-Depth 0.265* 0.016 83Maximum Set-Depth 0.177 0.109 83Latitude -0.401** <0.001 83Longitude 0.248* 0.024 83Year -0.047 0.676 83•‘ Correlation is significant at the .01 level (2-tailed)

•Correlation is significant at the .05 level (2-tailed)


34

Figure 1-3: Matrix plot of seven independent variables significantly correlated with CPUE according Spearman’s Rank Correlation.


Reproduced

with perm

ission of the

copyright ow

ner. Further

reproduction prohibited

without

permission.

J I I I L -766 l -76.2 l i i i i i 5 1iO 1i5 20 i i i I UlQ li5 20 25 20,I I I I I

Latitude

.37.8

.37.637.437.2370

-76.0-762-764-76.6

f io'OP

Longitude COf t C? %Oo

Distance to Bay Mouth

cP

»

80604020!o

2015’10‘5'0 -

00

MinimumDepth

8 O

O GO (DODO

QD O

O OOO O_o mmm o000 o 00ODOOOOt

CD mtM I®000

S «CDo 00o

000 o 00O 000000

00W &00 o o

BottomTemperature

o 00 00 o 1000C po 00

0000 CDD

ooaso o 00

0(o o o v o cxy” o

aoD o

28262422

’2030252015

10H

°8o o

BottomSalinity O V&OQOO ^

o9> o

Oo00

< r—i— r

e o

o00

cP—]— i— r

BottomDissolvedOxygen

“ i— i— i— r 0 2 4 6 8

i i i i r370 374 37.8

i i i i i r0 20 40 60 80

1 I I I 20 22 24 26 28

35

Table 1-2: Spearman’s Rank Correlation (rho) for predictor variables. Only significantly correlated predictor variables are shown.

SPEARMAN’S NONPARAMETRIC CORRELATION (rho)___________________FACTOR COMBINATION Correlation

Coefficient(rho)

Sign.(2-tailed)

N

Latitude vs. Bottom Temperature 0.483“ <0.001 83Latitude vs. Bottom Salinity -0.613“ <0.001 83Latitude vs. Bottom Dissolved-Oxygen -0.685“ <0.001 83Latitude vs. Distance to Bay Mouth 0.900“ <0.001 83Latitude vs. Minimum Set-Depth 0.314“ 0.004 83Latitude vs. Maximum Set-Depth 0.296“ 0.007 83Longitude vs. Bottom Salinity 0.347“ 0.001 83Longitude vs. Minimum Set-Depth 0.389“ <0.001 83Longitude vs. Maximum Set-Depth 0.512“ <0.001 83Bottom Salinity vs. Bottom Temperature -0550“ <0.001 83Bottom Salinity vs. Bottom Dissolved-Oxygen 0.302“ 0.006 83Distance to Bay Mouth vs. Bottom Temperature 0.563“ <0.001 83Distance to Bay Mouth vs. Bottom Salinity -0.737“ <0.001 83Distance to Bay Mouth vs. Bottom D. Oxygen -0.652“ <0.001 83Bottom Dissolved-Oxygen vs. Minimum Set-Depth -0.223* 0.043 83Minimum Set-Depth vs. Maximum Set-Depth 0.721“ <0.001 83‘ ‘ Correlation is significant at the .01 level (2-tailed)

‘ Correlation is significant at the .05 level (2-tailed)


36

II: Classification and Regression Trees (CART)

Two primary models were used in this analysis. Both used CPUE

as the dependent variable. The first model included bottom salinity, bottom

dissolved oxygen, bottom temperature, and minimum set depth as the predictor

(independent) variables and is referred to as the Ecological Model. In the second

model, distance to Bay mouth replaced bottom salinity whereas all other

variables remained unchanged. The introduction of the distance variable was

intended to increase applicability of the associated models to resource-

management problems. This second model is, therefore, referred to as the

Management Model. The initial trees for both models were allowed to grow to

completion with a minimum node size of one observation. The resulting trees

overfit the data in both models. Each full tree had 26 terminal nodes. A rule

limiting the minimum node size to five observations was applied. Given 83 total

observations in the analysis, this limited the number of terminal nodes to a

maximum of 16 for each full tree.

A. The Ecological Model

The full ecological tree model (min. node = 5) again was overly complex,

consisting of 11 terminal nodes (Fig. 1-4a) and had a residual mean deviance of

13.51 (Table 1-3). All four predictor variables were used in constructing this tree.

Figure 1-4b was generated using minimal cost-complexity pruning, plotting


37

deviance versus number of terminal nodes. The numbers along the top of the

graph are the values of the cost-complexity parameter. Most of the reduction in

deviance was explained by the first six terminal nodes, therefore the initial tree

was pruned accordingly (Figure 1-5a). The resulting tree was much easier to

interpret but came at the cost of increased residual mean deviance of 15.78.

Only salinity, depth, and dissolved oxygen were used to build the tree. It

suggests that the primary nursery areas (CPUE>3.0) are located where bottom

salinity is greater than 20.5, depth is greater than 5.5 meters, and dissolved

oxygen is greater than 5.35 parts per million (ppm). The response surface of

areas in the Bay that meet these criteria is shown in Map 1-8 termed Ecological

Model I.

Cross-validation suggested that the tree model pruned to six nodes also

overfit the data. A plot of cross-validation deviance versus tree size (Figure 1 -5b)

indicated that deviance was minimized at only three terminal nodes (Table 1-3),

therefore, the tree was pruned again (Figure 1-6). The residual mean deviance

increased to 21.62, but the resulting tree was simple and easily interpreted. It

suggested the primary nursery is located in areas where bottom salinity is greater

than 20.5 and depth is greater than 5.5 meters. The geographic response

surface for these criteria is shown in Map 1-9 termed Ecological Model II.


38

Figure 1-4: Ecological Model with CPUE as response and salinity, minimum depth, dissolved oxygen, and temperature as predictors, a) Full tree with predicted CPUE at each terminal node (11 nodes) b) Plot of residual mean deviance versus number of terminal nodes generated by cost-complexity pruning (value of cost-complexity parameter along top of graph).


B.SAUN|7Y<20.!

B.SAUN TY<19.5 MIN.IX5.5

0.2632 2.7610 0.6364B.SAUN TY<22.5

MIN.IX7.5

5.1530 14.0700

B.DQ<5.35

B.TEMP<24 B.SAUNirY<24.75

5.1040 1.4480B.TEMP<23.3 B.TEMP<23.1

12.6800 8.1620 5.0400 6.9450

400 340 170 110 51 41 32 11 -Inf

4 6 8

Number of Teririnal Nodes10


39

Figure 1-5: Ecological Model, a) Tree pruned to six terminal nodes with predicted CPUE below each node, b) Results of ten-fold cross-validation of each pruned subtree. Plot of cross-validated residual mean deviance versus number of terminal nodes (value of cost-complexity parameter along top of graph).


aBottom Salipity<20.5psu

Minimum Depth<5.5m0.9358

Bottom Salinity<22.5psu0.6364

Minimum 0epth<7.5m B. Dias. Oxygen<5.35ppm

5.1530 14.0700 Z8540 8.0050

400 340

oos

I sc CN(B>

4 6 8

Number of Terminal Nodes

10


40

Figure 1-6: Final tree of the Ecological Model pruned to three terminal nodes. Predicted CPUE and histogram shown below each terminal node. Histograms show distribution of CPUE values among the observations in each node.


<0n

to

IOin a

inin

Qin

tototoCM

CM

00IOtO

o


41

Map 1-8: Response surface for Ecological Model I. Shaded area is portion of Chesapeake Bay with average summer salinity greater than 20.5, depth greater than 5.5 meters, and dissolved oxygen concentration greater than 5.35 ppm. This area is interpreted to represent suitable nursery habitat according to the model.


76°30‘ 76°15' 76°00' 75045,

oP .COco

ECOLOGICAL M O D E L I Salinity > 20.5 psu Depth > 5.5 mDissolved Oxygen > 5.35 ppm

uC D

1

102*h-n

u

£

os?.n

<*>

IOr.h-cn

u-.2<71

OP-CO 3

IO2* .<oCO

<dO)*

14 K i l o m e te t S '

76°30‘ 76°151 76°00' 75°45'


42

Map 1-9: Response surface for Ecological Model II. Shaded area is portion of Chesapeake Bay with average summer salinity greater than 20.5 and depth greater than 5.5 meters. This area is interpreted to represent suitable nursery habitat according to the reduced Ecological Model.


76°30‘ 76°15’ 76°00' 75°45'

oP .GOm

ECOLOGICAL M O D E L II Salinity > 20.5 psu Depth > 5 . 5 m

NWl

oP.h-n

i0r.h-n

oP-kn

inr .ton

14 K i l o m e t e r s Ki l o m e t e r s ’

76°30' 76°15* 76°001 75°45’


,9to9

£ .OO

oZE

.Sio

l£ ,0£

oL2

.9fc>

Z£

,0Q,

8E

43

B. The Management Model

The full management tree (min. node = 5) again was overly complex,

consisting of 10 terminal nodes (Fig. 1-7a) and had a residual mean deviance of

15.33 (Table 1-3). Note that even in this full tree, only distance to Bay mouth and

minimum depth were used to build the model. Figure 1 -7b was generated using

minimal cost-complexity pruning, plotting deviance versus number of terminal

nodes. The first five terminal nodes explained most of the reduction in deviance.

The initial tree was therefore pruned accordingly (Fig. 1-8a), resulting in a tree

that was much easier to interpret but came at the cost of a slight increase in

residual mean deviance to 16.18 (Table 1-3). All of the high CPUE stations were

actually grouped by the first two splits in this tree, indicating that the third and

fourth splits, which divided the one terminal node into three, was unnecessary.

Cross-validation agreed that the tree model pruned to five nodes overfit the

data. A plot of cross-validation deviance versus tree size (Fig. 1-8b, Table 1-3)

indicated that deviance was minimized at only three terminal nodes, therefore,

the tree was again pruned accordingly (Fig. 1-9). The residual mean deviance

increased to 18.97, but the resulting tree was simple and easily interpreted. It

suggested the primary nursery is located in areas less than 34.6 kilometers from

the Bay mouth where depth is greater than 5.5 meters. The geographic

response surface for these criteria is shown in Map 1-10 termed the

Management Model.


44

Figure 1-7: Management Model with CPUE as response and distance to mouth of estuary, minimum depth, dissolved oxygen, and temperature as predictors, a) Full tree with predicted CPUE at each terminal node (10 nodes), b) Plot of residual mean deviance versus number of terminal nodes generated by cost-complexity pruning (value of cost-complexity parameter along top of graph).


DISTf:34.6

MIN.IX5.5

0.7692DIST =31.4

DIST<6.4

10.9000

DIST< 39.55

3.1120 0.4304

12.6700

DIST429.95

DIST< 14.05

MINI <7.15 DIST:20.3

8.2980

5.7820 8.4180 3.0250 5.9290

480 470 140 110 35 31 29 17 -Inf

eo

4 6 8

Number of Terrrinal Nodes10


45

Figure 1-8: Management Model, a) Regression tree pruned to five terminal nodes with predicted CPUE below each node, b) Results of ten-fold cross- validation of each pruned subtree. Plot of cross-validated residual mean deviance versus number of terminal nodes (value of cost-complexity parameter along top of graph).


Distanccj<34.6km

Minimum DepttK5.5m0.9667

DistancB<31.4km0.7692

Distance<6.4km12.6700

10.9000 S.1400

480 470

8'5

o8

4 6

Number of Terminal Nodes

10


46

Figure 1-9: Final tree of the Management Model pruned to three terminal nodes. Predicted CPUE and histogram shown below each terminal node. Histograms show distribution of CPUE values among the observations in each node.


CO

oo>.to

<0

1010

CO JC

1010

C40>U>


47

Map 1-10: Response surface for the Management Model. Shaded area is portion of Chesapeake Bay less than 34.5 km from the mouth of the Bay and depth greater than 5.5 meters. This area is interpreted to represent suitable nursery habitat according to the reduced Management Model.


76°30‘ 76°15’ 76°00’ 75°45‘

o

PGOr>

M A N A G EM EN T M O DEL Distance to M outh < 34.5 Depth > 5.5 m

IO

CO

o9MCO

■or .<o

o9-h-co

IO

IOl*»

76°30* 76°15' 76°00* 75°45'


.9fc>9e

.O

Ooi£

.Sio

L£

,0£o

i£

,9Po

L£

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-3: Pruning of Classification and Regression Trees (CART) for Ecological and Management models. RSM is the residual mean deviance for the overall tree. RSMCV is the residual mean deviance of the ten-fold crossvalidated model.

MODELc = > Ecological (Salinity, Depth, Temp, DO) Management (Distance, Depth, Temp, DO)

# Nodes VariablesUsed

RSM RSMCV # Nodes VariablesUsed

RSM RSMCV

FULL TREE 11 ALL 13.51(973/72)

29.32(2111/72)

10 Distance,Depth

15.33(1119/73)

31.41(2293/73)

1st PRUNED TREE 6 Salinity,Depth

15.78(1215/77)

28.03(2158/77)

5 Distance,Depth

16.18(1262/78)

29.24(2281/78)

2nd PRUNED TREE 3 Salinity,Depth

21.62(1729/80)

23.63(1890/80)

3 Distance,Depth

18.97(1517/80)

24.32(1946/80)

400

49

III: Testing the Tree Models

A. Tree Model Classification and Validation

Both of these very simple spatial models performed very well in all post-hoc

measures of classification (Table 1-4). The terminal node predicting highest

shark abundance in Ecological Model II contained 74.1% of all sets with CPUE

>1.0 and 91.2% of all sets with CPUE >3.0. That of the Management Model

performed even better, with 81.5% of all sets with CPUE >1.0 and 97.1% of all

sets with CPUE >3.0. Of the total number of sharks caught, 90.1% were in the

high abundance terminal node of Ecological Model II and 88.7% were in that of

the Management Model.

The ability of these models to classify and delineate the nursery correctly

was assessed by two independent data sources, tag-recapture data and

telemetry data. Nineteen of 22 (86.4%) tag recaptures from the Bay were within

the response surface for Ecological Model II whereas 20 of these 22 (81.8%)

were within the response surface of the Management Model (Maps 1-11 a, b). A

total of nine sharks manually tracked for a cumulative 350 hours generated 67

location fixes temporally separated by at least six hours. All 67 (100%) of these

location fixes were within the response surfaces for both models (Maps 1-12 a,b).


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-4: Classification ability of reduced ecological and management tree models. Each model was pruned to three terminal nodes. Classification is measured by the proportion of each estimation parameter is included in the terminal node possessing the highest shark CPUE.

Estimation Parameter (% encompassed by model)

Ecological (S.Z.T.DO) Salinity/Min.Depth (3)

Management (D.Z.T.DO) Distance/Min.Depth (3)

Sets where CPUE>1.0 74.1% (40/54) 81.5% (44/54)

Sets where CPUE>3.0 91.2% (31/34) 97.1% (33/34)

Total Sharks Caught 90.1% 88.7%

Tag Recaptures 86.4% (19/22) 81.8% (18/22)

Telemetry Fixes 100% (67/67) 100% (67/67)

U1o

51

Map 1-11: Tag recaptures for juvenile C. plumbeus recaptured in Chesapeake Bay during the same year and subsequent years compared to CART models of nursery habitat, a) Recaptures compared with Ecological Model II, b) Recapture compared with the Management Model


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

a.76*?0' 76“ 10'

a_ Recapture Location w (>100 days at liberty „ Recapture Locations

(<100 days at liberty■ Ecological Model II

76*00'

76*20' 76*10' 76*00 '

b.76*?0' 76*10'

★ Recapture Locatio (>100 days at liberty]

_ Recapture Locatio ™ (<100 days at liberty]

I I Management Model

76*00'

K

76*20 76*10' 76*00'

oobic

.o

u/e

.ozU

e.Q

CoiE

!»»

Ie

52

Map 1-12: Telemetry fixes for nine juvenile C. plumbeus manually tracked for 11 to 64 hours (cumulative 350 hours) from 1996 through 1999 compared to CART models of nursery habitat. The minimum interval between fixes was six hours to avoid autocorrelation, a) Telemetry fixes compared with Ecological Model II, b) Telemetry fixes compared with the Management Model.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

a.76*20' 76° 10' 76°00‘

Telemetry Fixes w 6-hour minimum interval

(# denotes individual sharks tracked) |■ I Ecological Model II

b. 76*20' 76°1 O' 76*oo'

^Telemetry Fixes w 6-hour minimum interval

(# denotes individual sharks tracked)|I I Management Model

76-20' 76-10' 76-00' 76“20 76“10 76“00'

lOOk/E

.o

U/e

.o

t>LZ

.o

eU

e,

or

„/

E

53

B. Logistic Regression

Univariate logistic regressions were performed using each independent

variable and CPUE as the dependent or response variable. All five continuous

predictor variables used in the CART models were interval coded to increase

model stability (Table 1-5a). The -2 log-likelihood statistic (G) was significant

(p<0.05) for all univariate models except that for minimum set depth (p=0.059).

The percent correct classification for the univariate regressions ranged from 69

and 77% (Table 1-5a). In all cases, the univariate models were more successful

in predicting presence than absence. Overall model significance was greatest for

distance to Bay mouth (G=27.7, p<0.0001) and bottom salinity (G=16.4,

p<0.0001), comparable to the results of the tree models. Pearson, Deviance,

and Hosmer-Lemeshow goodness-of-fit tests were insignificant for all univariate

models except bottom temperature (Table 1-5b). The tests were all highly

significant (p<0.001) for temperature indicating this univariate model fit the data

very poorly. Insignificant test statistics for all other variables indicated these

models adequately fit the data.

Multivariate regressions were performed using the same combinations of

variables as in the Ecological and Management CART models. The most

significant multivariate logistic models are summarized in Table 1-6. The logistic

regression of Ecological Model I including salinity, depth, and dissolved oxygen

(dropping temperature) predicted presence of juvenile C. plumbeus best (Table

1-8), as in the tree models. The overall model was highly significant (G=23.102,

df=3, p<0.0001) and all goodness-of-fit tests suggested the model adequately fit


54

the data. The model had 78.31% correct classification, but predicted presence

(88.89%) much better than absence (58.62%). This model had a Somer’s D

estimate of model predictive ability of 0.61. The scale of Somer’s D ranges from

zero for a model with no predictive ability to one for a perfect predictive model.

The model including temperature was also highly significant (G=27.925, df=4,

p<0.0001) but the Pearson goodness-of-fit chi-square (p=0.08) suggested this

model did not fit the data as well as the model without temperature (Table 1-7).

The model including only salinity and depth was also highly significant

(G=19.329, df=2, p=0.0001) and the goodness-of-fit statistics suggested this

model adequately fit the data as well (Table 1-9). Model classification (74.70%)

and Somer’s D (0.56) were slightly lower than the model that included dissolved

oxygen. Significance of individual variable coefficients was determined by the

Wald statistic evaluated at 0.05 (Tables 1-7,1-8, 1-9). Salinity was the most

significant factor in all three of these models and the odds ratio ranged from1.32

to 1.35 indicating an increased likelihood of shark presence with increased

salinity. Minimum set depth was also a significant factor in the models that

included dissolved oxygen and/or temperature but was insignificant in the

reduced model that included only salinity and depth. Its odds ratio ranged from

1.19 to 1.28 indicating an increased likelihood of shark presence with increased

depth.

The logistic regression of the Management Model using all four variables

used in the CART model was highly significant (G=38.876, df=4, p<0.0001) and

all goodness-of-fit tests suggested the model adequately fit the data (Table 1-10).


55

The model had 84.34% correct classification (88.89% for presence; 75.86% for

absence) and a Somer’s D of 0.76. The coefficients for dissolved oxygen and

temperature were insignificant, therefore they were dropped from the model. The

reduced Management Model including only distance to mouth and minimum set

depth was also highly significant (G=38.300, df=2, p<0.0001), and the

percentage of correct classification and Somer’s D were virtually unchanged from

the full model (Table 1-11). Distance to mouth was the most significant variable

(p<0.001) and its odds ratio was 0.91 in both models indicating a decrease in

likelihood of shark presence with increased distance from the mouth of the Bay.

Minimum set depth was also highly significant in both models (p<0.01) and its

odds ratio was 1.48 in the full model and 1.50 in the reduced model indicating

and increased likelihood of shark presence with increased depth.

These results were in close agreement with the CART models. In the

Ecological Models, salinity and depth were the most important variables

influencing the distribution of juvenile C. plumbeus using CART and logistic

regression. Both methods suggested dissolved oxygen might also be influential.

For the Management Models, distance to Bay mouth and depth were the most

important variables influencing distribution using both techniques.

Using both regression tree modeling and multivariate logistic regression, it

was found that complex habitat selection patterns could be adequately modeled

with only two variables. To examine potential differences in the response

surfaces delineated using each technique, the selection functions, w '( j) , were

plotted for each influential variables using the coefficients from their univariate


56

models (Table 1-12, Fig. 1-9). Values of the selection function, w'{x), greater

than 0.5 predict shark presence and values less than 0.5 predict shark absence.

The value of the independent variable as it crosses this threshold can be

compared with the splitting-rule value of the variable in the CART models.

Differences in actual variable reference points were probably simply a function of

the use of a continuous response variable, CPUE, in the CART models and the

use of a binary response variable, presence/absence, in the logistic models. The

selection function for distance to bay mouth started at a value greater than 0.9 at

zero kilometers and gradually declines. The curve predicted shark presence

when distance to mouth is less than 44.5 km compared to a value of 34.5 km

selected by the CART models (Fig. 1-9a). The selection function for salinity

indicated that shark presence was predicted when salinity is greater than 19.7

(Fig. 1 -9b). This corresponded very closely with the value 20.5 from the CART

models. Three selection functions were plotted for the minimum set depth

variable. The univariate regression model (G=3.551, p=0.59) for this variable

and the corresponding coefficient (Wald=3.22, p=0.0729) were not significant.

This model, however, included all stations whereas the CART models first

divided the stations based on salinity or distance to bay mouth. It then used only

those stations with salinity greater than 20.5 in the Ecological Model and those

stations less than 34.5 km from the mouth in the Management Model to make a

second split based on the depth parameter. Based on this fact, two additional

univariate logistic regressions were performed. The first used only stations where

distance to mouth was less than 44.5 km, the reference from the univariate


57

selection function using distance to mouth as the predictor. This model was

highly significant as was the model coefficient for depth (Wald chi-square = 8.35,

p=0.0039). The second used only stations where salinity was greater than 19.7,

the reference from the univariate selection function using salinity as the predictor.

This model was also highly significant as was the model coefficient for depth

(Wald chi-square = 6.63, p = 0.01). When plotted, the selection functions for

these two models predicted presence of C. plumbeus ( w’( j ) > 0.5) in water

deeper than 3.9 and 3.65 meters, respectively (Figure 1-9c), compared with 5.5

meters selected by the CART models.

Though the models appeared to be in close agreement, small differences

in salinity, distance to mouth, and depth correspond to large differences in area

defined (Map 1-13). The area delineated by the CART Ecological Model (salinity

>20.5, depth >5.5 m) was layered over that delineated by the logistic regression

(salinity >19.7, depth >3.65 m) in Map 1-13a. In this case, the area defined by

the logistic model was 39% greater than that defined by the CART model. The

area delineated by the CART Management Model (distance to mouth <34.5 km,

depth >5.5 m) was layered over that delineated by the logistic regression

(distance to mouth <44.5 km, depth >3.9 m) in Map 1-13b. The area defined by

the logistic model was 51% greater than that defined by the CART model.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-5: Results of univariate logistic regressions, a) Model significance, classification, Somer’s D measure of model predictive ability; b) Goodness-of-Fit tests for each univariate model (significance indicates lack of fit)

a.Variable G df Sign. Classification

% CorrectSomer’s D

Pres Abs OverallBottom Salinity(<16;17-19;20-22;23-25; >26psu)

16.398 1 <0.0001 90.74 48.28 75.90 0.49

Distance from Bay Mouth(<15km; 15-30km; 30-45km; 45-60km; >60km)

27.721 1 <0.0001 90.74 51.72 77.11 0.64

Minimum Set-Depth (1m intervals; 2-20m)

3.551 1 0.059 98.15 13.79 68.67 0.20

Bottom Dissolved-Oxygen (<2;3;4;5;>6 ppm)

4.823 1 0.028 98.15 13.79 68.67 0.26

Bottom Temperature 10.540 1 0.001 87.04 55.17 75.90 0.48(1°C intervals; 20-28°C)

b.Variable Goodness-of-Fit Tests (probabilities)

Bottom SalinityPearson

0.722Deviance

0.716Hosmer-Lemeshow

0.545

Distance from Bay Mouth 0.593 0.536 0.593

Minimum Set-Depth 0.395 0.285 0.110

Bottom Dissolved-Oxygen 0.658 0.637 0.779

Bottom Temperature 0.001 0.005 0.009

cnoo

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-6: Summary of overall logistic-regression statistics for full and reduced Ecological and Management Models a) overall model significance (-2 log likelihood = G), classification, Somer’s D measure of model predictive ability b) Goodness-of-Fit tests (p<0.05 indicates lack of fit)

a.Model G df Sign. Classification

% CorrectSomer’s D

Pres Abs Overall

Ecological Model 1 (S, Z, DO, T) 27.925 4 <0.0001 83.33 62.07 75.90 0.66Ecological Model 2 (S, Z, DO) 23.102 3 <0.0001 88.89 58.62 78.31 0.61Ecological Model 3 (S, Z) 19.329 2 0.0001 85.19 55.17 74.70 0.56Management Model 1 (Dist, Z, DO, T) 38.876 4 <0.0001 88.89 75.86 84.34 0.76Management Model 2 (Dist, Z) 38.300 2 <0.0001 90.74 72.41 84.34 0.75

b.Model____________________________Goodness-of-Fit Tests (probabilities)

Pearson Deviance Hosmer-LemeshowEcological Model 1 (S, Z, DO, T) 0.088 0.244 0.345Ecological Model 2 (S, Z, DO) 0.291 0.116 0.158Ecological Model 3 (S, Z) 0.143 0.102 0.305Management Model 1 (Dist, Z, DO, T) 0.046 0.683 0.767Management Model 2 (Dist, Z) 0.962 0.937 0.476

OlCD

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-7: Logistic regression for the full Ecological Model a) parameter estimates, significance, and odds ratios; b) model Goodness-of-Fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 27.925, df=4, p<0.0001)

a.Variable Coefficient Standard

ErrorWald

Chi-squaredDF Probability Odds Ratio

Constant 2.2494 5.5338 0.1652 1 0.6844Bottom Salinity 0.2262 0.0911 6.1648 1 0.0130* 1.35Minimum Set-Depth 0.2973 0.1308 5.1648 1 0.0231* 0.65Bottom Temperature -0.4338 0.2047 4.4913 1 0.0341* 1.25Bottom Dissolved-Oxygen 0.4694 0.2405 3.8081 1 0.051 1.60

b.Goodness-of-Fit Tests

Method Chi-squared DF ProbabilityPearson 85.343 69 0.088Deviance 76.743 69 0.244

Hosmer-Lemeshow 8.973 8 0.345

c.Classification Table

Predicted Percent CorrectObserved Present AbsentPresent 45 9 83.33%Absent 11 18 62.07%

Overall 75.90%

05O

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1*8: Logistic regression for the Ecological Model excluding temperature a) parameter estimates, significance, and odds ratios; b) model Goodness-of-Fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 23.102, df=3, p<0.0001)


ErrorWald


Constant -9.0980 2.4716 13.5493 1 0.0002Bottom Salinity 0.2766 0.0877 9.9535 1 0.0016** 1.32Minimum Set-Depth 0.2462 0.1178 4.3661 1 0.0367* 1.28Bottom Dissolved-Oxygen 0.4267 0.2270 3.5341 1 0.0601 1.53

b.Goodness-of-Fit Tests

Method Chi-squared DF ProbabilityPearson 64.486 59 0.291Deviance 72.199 59 0.116Hosmer-Lemeshow 11.855 8 0.158



Overall 78.31%

CD

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-9: Logistic regression for the Ecological Model using only salinity and depth a) parameter estimates, significance, and odds ratios; b) model Goodness-of-Fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 19.329, df=2, p=0.0001)


ErrorWald


Constant -7.2477 2.0868 12.0622 1 0.0004Bottom Salinity 0.3037 0.0856 12.5821 1 0.0005** 1.35Minimum Set-Depth 0.1726 0.1047 2.7205 1 0.0991 1.19

b.Goodness-of-Fit TestsMethod Chi-squared DF ProbabilityPearson 46.172 37 0.143Deviance 48.267 37 0.102Hosmer-Lemeshow 9.459 8 0.305



Overall 74.70%05ro

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-10: Logistic regression for the full Management Model a) parameter estimates, significance, and odds ratios; b) model Goodness-of-Fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 38.876, df=4, p<0.0001)


ErrorWald


Constant 5.6059 5.7920 0.9368 1 0.3331Distance from Bay Mouth -0.0896 0.0264 11.4793 1 0.0007** 0.91Minimum Set-Depth 0.3916 0.1508 6.7440 1 0.0094** 1.48Bottom Temperature -0.1820 0.2491 0.5339 1 0.4650 0.83Bottom Dissolved-Oxygen 0.0095 0.2975 0.0010 1 0.9745 0.56




Overall 84.34%

o>CO

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-11: Logistic regression for the Management Model using only distance and depth a) parameter estimates, significance, and odds ratios; b) model Goodness-of-Fit statistics; c) model classification. Overall model significance (-2 log likelihood, G = 38.300, df=2, p<0.0001)


ErrorWald


Constant 1.3305 0.9317 2.0394 1 0.1533Distance from Bay Mouth -0.0984 0.0230 18.2326 1 <0.0001** 0.91Minimum Set-Depth 0.4046 0.1423 8.0848 1 0.0045** 1.50




Overall 84.34%o>-p.

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 1-12: Parameter estimates, significance, and odds ratios from univariate logistic regressions for salinity, distance, depth, depth adjusted for distance, and depth adjusted for salinity. These parameter estimates were used to plot selection functions in Figure 1.9.

Variable Coefficient StandardError

WaldChi-square

df Sign. OddsRatio

Bottom Salinity 0.3013 0.0833 13.0714 1 0.0003 1.35Constant -5.9388 1.8079 10.7911 1 0.0010

Distance from Bay Mouth -0.0728 0.0170 18.4001 1 <0.0001 0.93Constant 3.2405 0.6998 21.4444 1 <0.0001

Minimum Set Depth 0.1586 0.0892 3.2159 1 0.0729 1.17Constant -0.5493 0.6822 0.6720 1 0.4124

Minimum Set Depth 0.5067 0.1753 8.3498 1 0.0039 1.66(stations < 44km from Mouth of Bay)Constant -1.9755 1.0817 3.3348 1 0.0678

Minimum Set Depth 0.3673 0.1427 6.6276 1 0.0100 1.44(Stations with Salinity >20) Constant -1.3420 0.9633 1.9411 1 0.1635

ocn

66

Figure 1-10: Plot of univariate selection functions ( w’* ( /) ) . Parameter estimates from univariate logistic regressions for salinity, distance from mouth, depth, depth adjusted for distance, and depth adjusted for salinity (Table 1.12).


Sele

ctio

n Fu

nctio

n Se

lect

ion

Func

tion

Sele

ctio

n Fu

nctio

nSelection Function - Distance From Bay Mouth

— 0(3 2405 - 0 0728(Dtslance)

■j + 0(3.2405 - 0 0728(D*stance)

0.5

00 10 20 30 40 50 60 70 80 90

Distance From Bay Mouth (KM)

Selection Function - Salinity

W(X) = e('5 94 * 0 30135<Salinjty)1 + p(-5 94 <■ 0.30135(Salioitv)

0.5

0 5 10 15 20 25 30 35 40Salinity (psu)

Selection Function - Minimum Depth

W (x) = e<a + b (Salinity)1 + g(a + b (Salinity)

All Stations (a = -0.5493; b = 0.15857)<44km from Bay Mouth (a = -1.976; b = 0.5067) Salinity >19.5psu (a = -1.3421; b = 0.3673)

0 5 10 15 20 25 30 35 40 45Bottom Depth (m)


Map 1-13: Maps comparing suitable nursery habitat defined by CART models and logistic regression models: a) Ecological Model, b) Management Model.


37*16*

•00.8C .9U/C

38*50' 37*46' 37*16* 37*00* 36*46*

mCD£ = CO^ in aO i/) £ cn a —£ 5 i< E QQ. c, . . S Q ~o - =L^ ir, l-~ myj id —

9 ° AO CNs A_ j |E E

3§»<=■ J l - rO cc =/<-K< Cm l ) j

CO .00,80 .9fcie .O C g /S .Sl*/B AUt .97,90


.97.44 .0

04

/ .9

14

/ .0

04

/ .97.9/

40

4/

.91

4/

.06

4/

68

DISCUSSION

Delineation of important shark nurseries and defining essential fish habitat

(EFH) have been identified as crucial data needs for proper management of

stocks (Hoff and Musick 1990, Musick eta l. 1993, NMFS 1996, Benaka 1999).

Shark nurseries have been superficially identified for several carcharhinid

species (Castro 1993, Simpfendorfer and Milward 1993), including Carcharhinus

plumbeus (Medved and Marshall 1981, Musick and Colvocoresses 1986, Merson

1998), but none of these were quantified or delineated. Morrissey and Gruber

(1993 b) were the first to investigate and quantify habitat selection for an

elasmobranch, concluding that juvenile Negaprion brevirostris selected

shallower, warmer areas with sandy or rocky substrate. Merson (1998)

investigated habitat selection in C. plumbeus in Delaware Bay and found no

selection based on temperature, salinity, bottom depth, and tidal cycle, however

these variables were analyzed using only t-tests. This study represents the first

attempt to quantify habitat selection spatially and delineate the corresponding

nursery to define essential fish habitat for an elasmobranch. The statistical

procedures used in this study have been used on a number of teleost species.

Regression tree modeling was used to determine important habitat variables in

nursery areas for several species of flatfishes in Alaskan waters (Norcross et al.

1995, 1997). Multivariate logistic regression has been used to investigate habitat

preference by juvenile flatfishes around Kodiak Island, Alaska (Norcross et al.

1999) and by juvenile Paralichthys dentatus in Chesapeake Bay (Kraus and


69

Musick in prep). This is the first time these techniques have been used for

elasmobranchs.

Although estuarine habitats are extremely complex and dynamic, this

study suggests that much of the variability in distribution of juvenile sandbar

sharks in Chesapeake Bay may be explained by very few variables. Abundance

of juvenile C. plumbeus, based on CPUE data from longline sampling, was

positively correlated with bottom salinity, depth, and dissolved oxygen, and

negatively correlated with distance to the mouth of the Bay and bottom

temperature. Regression tree modeling indicated salinity was the most important

environmental variable influencing the distribution of juvenile sharks in the

estuary and suggested a preference for areas where salinity is greater than 20.5.

The model selected depths greater than 5.5 meters as the second most

important variable defining nursery habitat and also suggested a preference,

perhaps tolerance-based, for areas with dissolved oxygen concentrations greater

than 5.35 ppm. Cross-validation of the regression tree model indicated

truncation of the model to three terminal nodes, which corresponded to station

splits based on salinity and depth only, was sufficient to explain the data. This

model performed very well in all measures of classification. The corresponding

response surface map (Map 1-9) indicated that suitable nursery habitat

encompasses most of the lower Bay south of 37° 20’ N latitude and extends as

far as 37° 40’ N on the eastern side of the Bay. This reflects the haloclinal tilting

typical of the estuary due to freshwater riverine influx from the western side of the

Bay coupled with tidal influx of high-salinity oceanic water from the south. One

major criticism with this model, however, is that whereas depth is a relatively


70

stable variable, salinity is extremely dynamic seasonally and annually. This

model attempted to model this variable as fixed in space and time. To illustrate

the effect of annual variability on the response surface, salinity and dissolved

oxygen grid coverages were interpolated for data from July 1996, representing a

very wet year, and July 1999 representing a drought year. Map 1-14 shows the

response surfaces for Ecological Models I and II applied to these two months.

According to Ecological Model I, the amount of suitable habitat in 1999 (794 km2)

was 50% greater than that of 1996 (524 km2). According to Ecological Model II,

only using salinity and depth, the discrepancy was even greater. The amount of

suitable habitat in July 1999 (2134 km2) was more than 180% greater than that in

July 1996 (741 km2). This is supported by anecdotal evidence of sharks being

caught by recreational fishers as far up the York River as VIMS in 1999.

Schwartz (1960) reported juvenile C. plumbeus as far north as Flag Pond in

Calvert County (four specimens, 1958) and the West River in Anne Arundel

County (one specimen, 1959) in the Maryland portion of the Bay. Perhaps these

were rare forays into marginal habitats or perhaps salinity was anomalously high

during this period. This was prior to the development of any directed shark

fisheries along the East Coast, however. Therefore abundance may have been

much higher and competition may have forced the utilization of these areas.

The first model delineated nursery EFH for juvenile C. plumbeus in

Chesapeake Bay according to the environmental parameters sampled. The

model was very simple to understand ecologically and was based primarily on

selection for high salinity regions. It was called the Ecological Model due to its

insight into the ecology of the organism. A second goal of this study, however,


71

was to develop an EFH model that would be of use for management of the

population through the establishment of regulations limiting exploitation within the

nursery. Protecting or regulating geographic areas based on dynamic variables

such as salinity are difficult at best, even during periods of stability. Annual and

seasonal variability render management impossible based on salinity Salinity is

highly correlated with distance to the mouth in Chesapeake Bay; therefore

distance was introduced as a surrogate variable for salinity in a second model.

The regression trees indicated distance to mouth was the most important

variable influencing shark distribution in the Bay, and predicted presence at

stations less than 34.5 km from the Bay mouth. As in the Ecological Model, this

model indicated higher abundance of sharks at depths greater than 5.5 meters.

This model also performed extremely well in all measures of prediction and

classification. Because both variables used in this model are stagnant, the

response area from this model (Map 1-10) is stable. Though distance to mouth

may have no direct ecological significance or influence, the resulting response

surface encompasses most of the suitable habitat from the Ecological EFH

Model but does not fluctuate due to dynamic influential variables. This model

provides a much more functional management tool for regulating the nursery and

was therefore called the Management Model.

As a means of validating these models, logistic regression modeling was

performed on the data. Presence/absence was substituted for CPUE as the

response variable. The results from univariate and multivariate models agreed

closely with the CART models. Both methods determined that distribution of

juvenile sandbar sharks in Chesapeake Bay was influenced by salinity and water


72

depth and that distance to Bay mouth serves well as a surrogate to salinity for

management applications. Both methods also suggested that bottom

temperature and dissolved oxygen concentration may also influence this

distribution, though less than salinity and depth. Dissolved oxygen concentration

in Chesapeake Bay is extremely dynamic. Large hypoxic areas are established

in late summer. Most of these areas are in deeper water north of the EFH area

delineated in this study. It is hypothesized, however, that in certain summers this

hypoxic zone may indeed cause constriction of the summer nursery just as low

salinity (wet years) did in 1996.

The results of this study provide a framework for delineating EFH for shark

nursery grounds. Nursery EFH for most shark species occurs in state-controlled

waters. The spatial models resulting from this methodology can be used by state

and regional regulatory agencies to limit harvest in areas determined to

constitute essential nursery habitat. In the summer of 1996, a directed

commercial fishery developed in Chesapeake Bay, targeting juvenile

Carcharhinus plumbeus. About 20,000 kg of juvenile sharks were landed, mostly

in three-week period from mid-June to the beginning of July. All of the sharks

landed came from the area delineated as EFH in this study. As discussed, the

amount of suitable habitat according to these models was severely constricted in

1996 due to low overall estuarine salinity. This may have concentrated the

juvenile sharks making them particularly vulnerable to the gillnet fishery. VIMS

CPUE data indicated this may have equated to the harvest of as much as 75% of

the Chesapeake Bay nursery population and resulted in severe juvenescence of

that portion of the Atlantic stock (Grubbs and Musick in prep b). Minimum size


73

limit regulations have recently been established, however the development of a

temporary no-take zones during the summer months, as determined by EFH

modeling, may be a better method for preventing such destructive fishing

practices from resurfacing.

Future research directions with these EFH models will involve including

water quality and other potentially influential anthropogenic variables.

Interestingly, all of the models in this study identified the lower western portion of

Chesapeake Bay as suitable nursery habitat, yet the CPUE data indicated very

low abundance of juvenile C. plumbeus in this region. This is the most urbanized

region of the lower estuary and is subject to intense urban and agricultural run-off

through the James River. It is hypothesized that these factors have severely

degraded otherwise suitable nursery habitat in the region.


74

Map 1-14: Comparison of suitable nursery habitat defined by a) CART Ecological Model I and b) CART Ecological Model II for a wet year (low salinity) and a drought year (high salinity).


37^16' 36*45'

■OObSC .OfcZS .9UIC

37*16' 37*00'

.OObSC ,9 U C


76=3

0' 76

*16'

76*0

0' 76

*46'

76*3

0' 76

*16'

76*0

0' 76

*46'

75

LITERATURE CITED

Agresti, A. 1990. Categorical data analysis. John Wiley & Sons, Inc. New York. 558pp.

Anonymous. 1996. Final Report. MARFIN Award NA47FF0008. Gulf and South Atlantic Fisheries Development Foundation. March 1996.

Bass, A. J. 1978. Problems in studies of sharks in the southwest Indian Ocean, pp. 545-594 In: Sensory biology of sharks, skates, and rays (E. S.Hodgson and R. F. Mathewson, eds.). U.S. Gov't. Printing Office, Washington, DC.

Benaka, L. R. 1999. Summary of panel discussions and steps toward an agenda for habitat policy and science, pp. 455-459 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

Bigelow, H. B. and W. C. Schroeder. 1948. Sharks, pp. 59-546 In: Fishes of the western north Atlantic. Part 1. Vol. 1. (J. Tee-Van, C. M. Breder, S. F. Hildebrand, A. E. Parr, and W. C. Schroeder, eds.). Mem. Sears Foundation for Marine Research, Yale Univ. New Haven, CT.

Branstetter, S. 1990. Early life-history strategies of carcharhinoid and lamnoid sharks of the northwest Atlantic, pp. 17-28 In: Elasmobranchs as living resources: advances in the biology, ecology, systematics and the status of the fisheries (H. L. Pratt, Jr., S. H. Gruber, and T. Taniuchi, eds.),. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 90.

Breiman. L., J. H. Friedman, R. A. Olshen, and C. J. Stone. 1984. Classification and regression trees. Wadsworth International Group. Belmont, CA.358 pp.

Castro, J. I. 1987. The position of sharks in marine biological communities, pp. 11-17 In: Sharks, an inquiry into biology, behavior, fisheries, and use. Oregon State University Extension Service, Corvallis. (S. Cook, ed.).

Castro, J. I. 1993. The shark nursery of Bulls Bay, South Carolina, with a review of the shark nurseries of the southeastern coast of the United States. Environ. Biol. Fishes 38: 37-48.

Clarke, T. A. 1971. The ecology of the scalloped hammerhead, Sphyrna lewini, Hawaii. Pac. Sci. 25: 133-144.


76

Compagno, L. J. V. 1984. FAO species catalogue. Vol. 4 Sharks of the world. Part 2 - Carcharhiniformes. FAO Fish. Synop. 4(2): 250-655.

Garrick, J. A. F. 1982. Sharks of the genus Carcharhinus. NOAA Tech. Rep. NMFS Circ. 445.

Grubbs, R. D. and J. A. Musick. Long-term movements, migration, and temporal delineation of summer nurseries for juvenile Carcharhinus plumbeus in the Mid-Atlantic Bight, in prep a.

Grubbs, R. D. and J. A. Musick. Population Trends, Juvenescence, and mortality estimation for juvenile Carcharhinus plumbeus in Chesapeake Bay and Virginia coastal waters, in prep b.

Grubbs, R. D. and J. A. Musick. Movements of juvenile Carcharhinus plumbeus in Chesapeake Bay. in prep c.

Gruber, S. H., D. R. Nelson, and J. F. Morrissey. 1988. Patterns of activity and space utilization of lemon sharks, Negaprion brevirostris, in a shallow Bahamian lagoon. Bull. Mar. Sci. 43 (1): 61-76.

Hoff, T. B., and J. A. Musick. 1990. Western North Atlantic shark-fisherymanagement problems and informational requirements, pp. 455-472 In Elasmobranchs as living resources: advances in the biology, ecology, systematics and the status of the fisheries (H. L. Pratt, Jr., S. H. Gruber, and T. Taniuchi, eds.), U.S. Dep. Commer., NOAA Tech. Rep. NMFS 90.

Hosmer, D. W. and S. Lemeshow. 1989. Applied logistic regression. John Wiley & Sons, Inc. New York. 307pp.

IUCN (International Union for Conservation of Nature and Natural Resources). 2000. 2000 IUCN Red List of Threatened Animals. IUCN, Gland, Switzerland.

JPOTS. 1980. The Practical Salinity Scale 1978 and the International Equation of State of Seawater 1980. Tenth report of the Joint Panel on Oceanographic Tables and Standards (JPOTS), Sidney, B.C. Canada, 1-5 September 1980, sponsored by UNESCO, ICES, SCOR, IAPSO. UNESCO technical papers in marine science 36.

Kraus. R. T. and J. A. Musick. Modeling habitat occurrence of young-of-year summer flounder, Paralichthys dentatus, in the Chesapeake Bay. in prep.

Lewis, E. L. and R. G. Perkin. 1978. Salinity: its definition and calculation. J. Geophys. Res. 83: 466-478.


77

Lund, R. 1990. Chondrichthyan life history styles as revealed by the 320 million years old Mississippian of Montana. Environ. Biol. Fish. 27: 1-19.

Manly, F. F. J., L. L. McDonald, and D. L. Thomas. 1993. Resource selection by animals: Statistical design and analysis of field studies. New York: Chapman and Hall, 177 pp.

Medved, R. J., and J. A. Marshall. 1981. Feeding behavior and biology of young sandbar sharks, Carcharhinus plumbeus (Pisces, Carcharhinidae), in Chincoteague Bay, Virginia. Fish. Bull. 79(3): 441-448.

Meek, A. 1916. The migrations offish. Edward Arnold, London. 427 pp.

Menard, S. W. 1995. Applied logistic regression analysis. Sage Publications, Thousand Oaks, CA. 98pp.

Merson, R. R. 1998. Nurseries and maturation of the sandbar shark. Ph.D. Dissertation. University of Rhode Island. 150 pp.

Morrissey, J. F., and S. H. Gruber. 1993a. Home range of juvenile lemon sharks, Negaprion brevirostris. Copeia 1993 (2): 425-434.

Morrissey, J. F., and S. H. Gruber. 1993b. Habitat selection of juvenile lemon sharks, Negaprion brevirostris. Environ. Biol. Fishes 38: 311-319.

Musick, J. A. 1999. Introduction to Part 2: Essential fish habitat identification, p. 41 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

Musick, J. A., and J. A. Colvocoresses. 1986. Seasonal recruitment ofsubtropical sharks in Chesapeake Bight, U.S.A. pp. 301-311 In: Workshop on recruitment in tropical coastal demersal communities (A. Yanez y Arancibia and D. Pauley, eds.), FAO/UNESCO, Campeche, Mexico, 21-25 April 1986. I.O.C. Workshop Rep. 44.

Musick, J. A., S. Branstetter, and J. A. Colvocoresses. 1993. Trends in shark abundance from 1974 to 1991 for the Chesapeake Bight region of the U.S. Mid-Atlantic Coast, pp. 1-18 In: Conservation Biology of Elasmobranchs (S. Branstetter, ed.), U.S. Dep. Commer., NOAATech. Rep. NMFS 115.

NMFS. 1993. Fishery management plan for sharks of the Atlantic Ocean. National Oceanic and Atmospheric Administration, National Marine Fisheries Sen/ice, U.S. Department of Commerce.


78

NMFS. 1996. 1996 Report of the shark evaluation workshop. June, 1996. NOAA, National Marine Fisheries Service, Southeast Fisheries Science Center, Miami, Florida.

Norcross, B. L., A. Blanchard, and B. A. Holladay. 1999. Comparison of models for defining nearshore flatfish nursery areas in Alaskan waters. Fish. Oceanogr. 8(1): 50-67.

Norcross, B. L., B. A. Holladay, F. J. Muter. 1995. Nursery area characteristics of Pleuronectids in coastal Alaska, USA. Neth. J. Sea Res. 34:161-175.

Norcross, B. L., F. J. Muter, B. A. Holladay. 1997. Habitat models for juvenile Pleuronectids around Kodiak Island, Alaska, USA. Fish. Bull. 95: 504- 520.

Packer, D. B. and T. Hoff. 1999. Life history, habitat parameters, and essential habitat of mid-Atlantic summer flounder, pp. 76-92 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

S-Plus. 2000. Guide to Statistics, Volume 1. Data Analysis Products Division, MathSoft, Seattle, WA.

Schmitten, R. A. 1999. Essential fish habitat: opportunities and challenges for the next millennium, pp. 3-10 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

Schwartz, F. J. 1960. Measurements and the occurrence of young sandbar shark, Carcharhinus milberti, in Chesapeake Bay, Maryland. Ches. Sci. 1 (3): 204-206.

Simpfendorfer, C. A., and N. E. Milward. 1993. Utilization of a tropical bay as a nursery area by sharks of the families Carcharhinidae and Sphyrnidae. Environ. Biol. Fishes 37: 337-345.

Sminkey, T. R. 1994. Age, growth, and population dynamics of the sandbar shark, Carcharhinus plumbeus, at different population levels. Ph.D. Dissertation, Va. Inst. Mar. Sci., College of William and Mary. 99 pp.

Sminkey, T. R. and J. A. Musick. 1995. Age and growth of the sandbar shark, Carcharhinus plumbeus, before and after population depletion. Copeia 1995(4):871-883.

Springer, S. 1960. Natural history of the sandbar shark, Eulamia milberti. U.S. Fish. Wildl. Serv., Fish. Bull. 61: 1-38.


79

Springer, S. 1967. Social organization of shark populations, pp. 149-174. In Sharks, Skates, and Rays (Gilbert, P. W., R. F. Mathewson, and D. P. Rail, eds.) pp. 111-140.

Van der Elst. 1979. A proliferation of small sharks in the shore-based Natal spots fishery. Environ. Biol. Fish. 4: 349-362.

USDOC (U.S. Department of Commerce). 1996. Magnuson-Stevens Fishery Conservation and Management Act as amended through October 11, 1996. National Oceanic and Atmospheric Administration Technical Memorandum NMFS-F/SPO-23.


CHAPTER 2Migratory movements, philopatry, and temporal delineation of summer

nurseries for juvenile Carcharhinus plumbeus in the Chesapeake Bay region.

80


81

ABSTRACT

The purpose of this study was to delineate temporally the migration

patterns of juvenile Carcharhinus plumbeus, sandbar sharks, in Chesapeake

Bay, to determine the location of wintering areas, and to determine if philopatry

or homing to natal summer nurseries in subsequent years occurs. Between 1990

and 1999, 100 longline sets were made at standard stations in the lower

Chesapeake Bay to delineate the migration patterns temporally. These data

indicated that immigration to the Bay occurred in late May and early June and

was highly correlated with increasing water temperature. Emigration from the

estuary occurred in late September and early October and was highly correlated

with decreasing day length. It is hypothesized that day length is the

environmental trigger to begin fall and spring migrations, whereas temperature

may elicit the response to move into the estuaries that serve as summer

nurseries. Between 1995 and 2000, 1846 juvenile C. plumbeus were tagged.

Only 2.4% were recaptured and reported. This low recapture rate is believed to

be due to under-reporting by the commercial-fishing sector in both summer and

winter nurseries. With two exceptions, recaptures made in summer months were

within 50 kilometers of the tagging location. Those recaptured in winter months

were caught between 200 and 830 kilometers from the tagging location. These

recaptures indicate that wintering areas are concentrated off the coast of North

Carolina between 33°30’N and 34°30’N latitude in water less than 20 meters

deep. However, recaptures were made as far south as Hilton Head, South


82

Carolina and as far as 200 kilometers from shore. Return data indicated that

these winter nurseries were utilized from late October until late May. Ninety-

three percent of tag recaptures made in subsequent summers were in the same

summer nursery area in which they were tagged. These data suggest that most

juvenile sandbar sharks return to the same summer nurseries every summer.

Additional data are needed to examine the strength and longevity of this homing

response and to determine if females actually return to their natal nursery to

deliver their own young years later.


83

INTRODUCTION

A majority of species of marine fishes is migratory to some degree (Meek

1916). Migration enables animals to utilize resources or habitats that are only

temporarily, usually seasonally, suitable or tolerable (Aidley 1981). Migratory

movements may be broadly divided, based on primary purpose, into alimentary,

climatic, and gametic migrations (Heape 1931). The majority of studies of fish

migration have been concerned with the timing of the gametic migrations and

philopatry (natal homing) of diadromous species of commercial importance.

These primarily have involved salmonids (Hallock 1970, Smith 1973, Mundy

1984, Tarbox 1988), clupeids (Talbot and Sykes 1958, Melvin et al. 1986,

Friedland and Haas 1988), anguilliformes, and a few perciforms such as Morone

saxatilis (Chapoton and Sykes 1961, Boreman and Lewis 1987). A large body of

research also has been dedicated to the oceanic migrations of scombrids and

other commercially important pelagic teleost species (Mather 1962, Seckel

1972). Research involving the migratory movements of sharks has been sparse

due to their historically low economic value.

In 1996, the reauthorization of the Magnuson-Stevens Fishery

Conservation and Management Act through the Sustainable Fisheries Act (SFA)

established a mandate for the National Marine Fisheries Service (NMFS) and

Federal Fishery Management Councils requiring the identification of essential

fish habitat (EFH) for all species regulated by federal fishery management plans.

In short, 39 federal management plans had to be amended to include plans for

identification of EFH for more than 700 fishery stocks (Schmitten 1999). The


84

SFA defines EFH as “...those waters and substrate necessary to fish for

spawning, breeding, feeding, or growth to maturity" (USDOC 1996). Musick

(1999) pointed out, however, that in past EFH investigations, virtually all habitats

where a species is known to occur have been deemed EFH. This is particularly

problematic for highly migratory species. Musick (1999) recommended an EFH

designation system based on utilization, availability, and vulnerability. Using

these criteria, estuarine nursery areas in regions heavily impacted by humans

are of particular concern.

Most estuaries are seasonally dynamic, therefore estuarine species

assemblages tend to be highly seasonal as well. The diversity and abundance of

fishes in estuaries tend to be highest in spring and summer seasons and lowest

in winter (McErlean 1973, Merriner et al. 1976, Cowan and Birdsong 1985). Most

species migrate to the highly productive estuaries to utilize abundant food

sources (alimentary migration) or to bear young (gametic migration) which in turn

benefit from increased food availability and potentially higher growth rates

(Harden Jones 1968). Physiological limitations later force migrants into climatic

emigration from the estuary to avoid intolerable conditions. Chesapeake Bay

represents one of the most seasonally dynamic marine environments in the world

with temperature extremes ranging as much as 30°C between summer and

winter. The demersal fish fauna of the region is dominated by a few boreal

species in winter and many highly migratory sub-tropical and temperate species

in summer (Musick et al. 1986). Many of these species rely on the Bay as crucial

seasonal nursery habitat. These patterns of species diversity also include the


85

elasmobranchs. The winter batoid fauna consists of only two species of rajids

whereas the summer fauna includes one rajid, four dasyatids, two gymnurids,

and two myliobatids, all of which are known to undergo parturition in the Bay

(Grubbs, unpublished). The shark fauna is dominated by Squalus acanthias in

winter, but consists of several species of carcharhiniform and lamniform sharks in

summer (Musick and Colvocoresses 1986). Musick and Colvocoresses (1986)

stated that these migratory movements are probably driven largely by seasonal

temperature changes.

The sandbar shark, Carcharhinus plumbeus, is the most common shark in

Chesapeake Bay and is the dominant species in the directed commercial shark

fishery along the east coast of the United States, constituting more than 65% of

landings (Anonymous 1996). It has been federally regulated since 1993 (NMFS

1993) and is currently regulated by the Fishery Management Plan for Atlantic

Tunas, Swordfish, and Sharks (NMFS 1999). In addition, this species has been

listed in the World Conservation Union (IUCN) Red List of Threatened Species

as a lower-risk, conservation-dependent species of concern (IUCN 2000).

Carcharhinus plumbeus is a large, coastal species reaching at least 239 cm in

total length (Compagno 1984) and requiring at least 15 years to reach sexual

maturity (Sminkey and Musick 1995). It inhabits insular regions to at least 250 m

depth (Garrick 1982) in the western Atlantic from Massachusetts, USA, to Brazil

and is highly migratory with several hundred kilometers separating summer and

winter habitats (Bigelow and Schroeder 1948, Springer 1960). These habits

make them particularly vulnerable to fishery exploitation and render delineation of


86

EFH problematic at best. The spatial and temporal delineation of summer

pupping and nursery areas was defined as a crucial data need in defining EFH

for this species (Hoff and Musick 1990, NMFS 1996). The historical nursery

grounds for Carcharhinus plumbeus in the western North Atlantic are distributed

in shallow, principally estuarine habitats on the east coast of the United States

from Long Island, New York (possibly north to Cape Cod, Massachusetts) to

Cape Canaveral, Florida (Springer 1960). Merson (1998) found that this nursery

range has contracted and now only extends from New Jersey south to South

Carolina. A secondary nursery exists in the northwestern Gulf of Mexico

(Carlson 1999). Chesapeake Bay may be the largest nursery area for C.

plumbeus. Mature females enter the lower Bay as well as the saline lagoons

along the Eastern Shore of Virginia in May and June to pup (Musick and

Colvocoresses 1986). These mature individuals then migrate offshore while the

neonates remain in the highly productive estuarine waters until fall (Musick and

Colvocoresses 1986) feeding on the abundant blue crabs (Callinectes sapidus),

Atlantic menhaden (Brevoortia tyrannus), summer flounder (Paralichthys

dentatus), and miscellaneous sciaenid fishes (Medved and Marshall 1981,

Cowan and Birdsong 1985). Young sandbar sharks continue to use Chesapeake

Bay as a nursery during the warmer months for the first four to ten years of life

(Sminkey 1994, Grubbs and Musick in prep. d). They then remain coastal year

round, presumably only entering Mid-Atlantic estuaries after reaching maturity to

bear young.


87

Grubbs and Musick (in prep a) found that the primary nursery in

Chesapeake Bay is limited to southern portion of the Bay, particularly the

southeastern portion, where salinity is greater than 20.5 and depth is greater

than 5.5 meters. The temporal pattern of utilization and timing of migratory

movements have not been formally investigated, which is one of the objectives of

this study. Young sandbar sharks are particularly susceptible to local overfishing

when aggregated in the estuary. In addition, it has often been hypothesized that

many carcharhinid sharks are philopatric to their natal nurseries. Though this

has not been tested rigorously, if true this would render populations even more at

risk due to local overharvesting and habitat degradation. Grubbs and Musick (in

prep a) reported that more than 20,000 kilograms of juvenile sandbar sharks

were harvested in the principal nursery area of Chesapeake Bay in 1996. The

constriction of suitable nursery habitat due to overall reduced salinity and

increased hypoxia in Chesapeake Bay in 1996 may have concentrated the

sharks making them especially vulnerable. State regulations have been

implemented to protect juvenile sharks in these crucial habitats (Pruitt 1997).

However, as Camhi (1998) pointed out, protection of juveniles in summer

nurseries will have little effect if protection is not also afforded in wintering areas.

Commercial longline and drop-net fishers have exploited aggregations of juvenile

sandbar sharks heavily on near-shore wintering grounds off North Carolina

(Anonymous 1996). Before protection is possible, these winter nurseries must

first be delineated spatially and temporally.


88

The primary objective of this study was to delineate temporal utilization of

Chesapeake Bay by juvenile C. plumbeus and to determine what environmental

factors act as catalysts to begin the migratory movements. This was

accomplished using CPUE data collected by a longline survey conducted by the

Virginia Institute of Marine Science to monitor shark abundance and from

recapture data obtained as a result of the implementation of a multi-year tagging

program targeting juvenile sandbar sharks. A second objective was to determine

the location of the winter nursery areas and investigate the temporal utilization of

these regions using the tag-recapture data. A final objective was to test the

hypothesis of natal nursery philopatry, or homing, by juvenile sharks in

subsequent summers.


89


Sampling Gear

Sampling was conducted during VIMS coastal shark monitoring longline

cruises. Collections were made with commercial-style longlines consisting of

3/8-inch tarred, hard-laid nylon main line, which was anchored at each end and

marked by a hi-flier marker buoy. Three-meter gangions were spaced

approximately 18 meters apart along the main line and a large inflatable buoy

was attached to the main line following every 20th gangion. Each gangion was

composed of a stainless-steel tuna clip attached to a 2-meter section of 1/8-inch

tarred nylon trawl line, the end of which was attached to a large barrel swivel. A

1-meter section of 1/16-inch galvanized aircraft cable was crimped to the swivel

and the other end was crimped to a Mustad-9/0, stainless-steel shark hook.

Each longline set consisted of 80 - 120 gangions. Bait consisted mostly of

Brevoortia tyrannus, Atlantic menhaden, and Scomber scombrus, Atlantic

mackerel. Soak time for each set was between three and four hours. A standard

100-hook set covered about two kilometers. Beginning in 1996, sampling was

expanded to include sets using 5/0 circle hooks in addition to the standard hook

sets. The circle hooks are smaller, therefore they are more efficient at capturing

juvenile, especially neonate, sandbar sharks. In addition, due to the tendency of

circle hooks to hook the sharks in the corner of the mouth, it was believed that

mortality would be lowered using these hooks. The statistical unit was catch per

unit effort (CPUE) defined as the number of sharks per 100 hooks for each set.


90

Sampling Design

Since its inception in 1974, the VIMS longline survey has routinely

included stations in Chesapeake Bay. Two locations in the lower eastern Bay,

Kiptopeke (37°10' N, 76° 00' W) and Middleground (37° 06' N, 76° 03* W), have

been standard stations since 1980 (Map 2-1). Monthly data collected from May

to October at these two stations using the standard gear during the ten-year

period from 1990 to 1999 were used to delineate migration patterns and nursery

usage temporally in Chesapeake Bay. Data collected using the alternate gear

with circle hooks during the four-year period from 1996 to 1999 were analyzed

separately.

Data Processing

CPUE, defined as the number of juvenile Carcharhinus plumbeus per 100

hooks fished, was recorded for each station. Pre-caudal length (PCL) and

stretched total length (TL) were measured for each shark. PCL was defined as

the distance from the tip of the snout to the caudal peduncle, and TL was defined

as the distance from the tip of the snout to the tip of the caudal fin when

stretched in line with the body axis. In addition, surface temperature, minimum

and maximum depth, and time of day were recorded. A Hydrolab® multiprobe

was introduced to the survey in 1996. It was used to record temperature, salinity,

and dissolved oxygen at two-meter intervals from surface to bottom for stations

sampled from 1996 to 1999.


91

Map 2-1: Locations of standard stations sampled by the VIMS longline survey from 1973-1999 including stations K (Kiptopeke) and M (Middieground) in the lower Chesapeake Bay.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

76°,3 O'

Standard Longline Stations fished by the VIMS Shark Ecology Program 1973-2000

m il k t78o30• 76°00 ' 75<>3 0 ’

CO

CD

40 Kilometers

75“0 0 ‘ 7 4°30 ‘

92

Survival Factors and Tagging

A survival factor was assigned to all juvenile C. plumbeus captured on

each set. The factor consisted of five rather subjective levels: (1) Excellent, (2)

Good, (3) Fair, (4) Poor, and (5) Dead. Sharks that struggled vigorously while

being measured and swam away rapidly following release were assigned survival

factor 1. Those that struggled moderately but swam strongly upon release were

assigned survival factor 2. Sharks that struggled very little while being measured

and swam slowly upon release were assigned survival factor 3. Those that did

not struggle at all but showed nictitating membrane and jaw response and

attempted to swim when released but with apparent equilibrium disruption were

assigned survival factor 4. Sharks that showed no nictitating membrane or jaw

responses were assumed dead and assigned survival factor 5.

All sharks determined to be in excellent, good, or fair condition based on

pre-release criteria were tagged using Hallprint ® nylon-tipped dart tags. A

stainless-steel applicator was used to apply the tags. The tags were inserted into

the musculature just below the first dorsal fin with an angle of attack of 30-40°

relative to the sagittal plane of the shark’s body. The tag was pushed through

the basal cartilages of the first dorsal fin with the barb directed posteriorly

ensuring the tag was locked behind this plate. Data from tag returns were used

to investigate long-term movements and migration patterns. They also provided

an additional means to validate the temporal nursery-delineation patterns

determined using the CPUE data.


93

Data Analysis

The sampling period was May 1 to October 15 of each year. This period

was divided into eleven semimonthly intervals. Mean CPUE was calculated for

each interval using the ten-year, standard-hook data set and the four-year, circle-

hook data set. The results of the two data sets were plotted separately. These

plots were used to determine the timing of the summer immigration to

Chesapeake Bay and the fall emigration from the Bay. Temporal trends in CPUE

were compared to surface temperature, surface salinity, day length, and lunar

phase to investigate potential stimuli for migration. The influence of these

environmental factors (independent variables) on CPUE (dependent variable)

was investigated using linear regression over immigration and emigration periods

independently.

Tag-return data were used also to investigate the timing of summer and

fall migrations for juvenile sandbar sharks. All recaptures were mapped using

ArcView 3.1 GIS. Distance from tagging location and recapture location was

measured as the shortest distance between the two points without crossing land.

Data from tag recaptures made in Chesapeake Bay were used to estimate when

sharks first arrive to the estuary in the summer and when they leave the estuary

in the fall. Recaptures made during the winter and spring were used to

determine the general location of the primary wintering grounds for the juvenile

sharks. In addition, these data were used to determine the timing of their arrival

to and departure from the wintering grounds. Finally, recaptures made in


94

subsequent summers, those having gone through at least one winter prior to

recapture, were used to determine whether or not these juvenile sharks return to

their natal estuary as a summer nursery (i.e. evidence of philopatry) or move to

new areas in subsequent years.


95

RESULTS

I: Temporal Nursery Delineation

Chesapeake Bay is utilized only as a summer nursery for juvenile C.

plumbeus, therefore abundance is strongly seasonal. The monthly CPUE (catch

per unit effort) values over the period from 1996-1999 using both standard 9/0

hooks and the smaller 5/0 alternative hooks are shown in Figure 2-1. The

Kiptopeke and Middle ground stations are plotted separately. Though this

seasonal trend is apparent in Figure 2-1, these data also illuminate the high

degree of inter-annual and intra-annual variability typical of the system. To

delineate the overall temporal-utilization pattern for the nursery, standard hook

data for the ten-year period 1990-1999 were combined. In addition, circle-hook

data for 1996-1999 were combined and the results compared to the standard

survey data.

A total of 100 standard longline sets (9/0 hooks) were made at the

Kiptopeke (59 sets) and Middleground (41 sets) stations from 1990 to 1999. No

C. plumbeus were caught during the early May time interval. The first sandbar

sharks appeared in the survey during the late May interval but the CPUE was

very low. CPUE continued to increase through June, peaking in July. These

data indicate that immigration to Chesapeake Bay occurs from late May through

June (Fig. 2-2a). Mean CPUE began to decline in August, particularly later in the

month, and continued to decline through October. The most dramatic decrease

occurred from early September through early October. The mean CPUE for the


96

early-October time interval was driven by two anomalously high sets in 1990 and

1991. The data indicate that in most years nearly all sharks have vacated the

Bay by mid-October.

In addition, 59 sets were made between 1996 and 1999 using the

alternative gear with the smaller 5/0 circle hooks. Variance was much higher for

this data set, especially for the early July period due to low sample size.

Nevertheless, these data show the same seasonal trend in CPUE as the

standard data (Fig. 2-2b). No sharks were caught during the May 1-15 time

period and very few were caught from May 16-31. Most immigration occurred

from early June to early July and emigration occurred from early September to

early October. Interestingly, CPUE was lower in the late July and early August

time periods than in early July or late August. This dip may represent the

dispersal of smaller juveniles to more suitable regions of the nursery farther from

the mouth of the estuary. Perhaps this is an adaptation to avoid predators such

as larger juvenile C. plumbeus and large Carcharias taurus (sandtiger sharks)

that are found in the lower Bay and are known to feed on neonate and small

juvenile sandbar sharks.

Sea-surface temperature, day length (photoperiod), salinity, and lunar

phase were investigated as potential environmental catalysts for these migratory

movements. Only the ten-year data set using standard gear was used for this

analysis due to lower variance and larger sample sizes. The immigration period

was defined as the six sampling intervals from May 1 to July 31 and the

emigration phase was defined as the six sampling units from Juiy 16 to October


97

15. Highest CPUE was observed during the July 16-31 interval, therefore it was

included in the immigration and emigration periods as an end point. Surface

salinity at the two standard stations was typically lowest in May and highest in

August or September. Salinity was highly variable between years sampled,

although it was not significantly correlated with CPUE. Lunar phase also was not

significantly correlated with CPUE. Salinity and lunar phase, therefore, will not

be discussed. Alternatively, regression analyses indicated that surface

temperature and day length are good predictors of shark migration. The results

for these two variables will be discussed individually.


98

Figure 2-1: Monthly CPUE (C. plumbeus per 100 hooks) for the period 1996- 1999 at a) Kiptopeke and b) Middleground stations using both standard hooks and circle hooks.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Monthly Longline CPUE 1996-1999 a) Kiptopeke Station

50 j ♦ ■■Circle Hooks4 Q Standard Hooks

30Jg 20 8 10 A A

J M M J S N J M M J S N J M M J S N J M M J S N fc 1996 | 1997 | 1998 | 1999(/)£ b) Middleground Station

— 40UJ? 30Q-° 20

10

J M M J S N J M M J S N J M M J S N J M M J S N 1996 I 1997 I 1998 I 1999

= No sampling

99

Figure 2-2: Mean semi-monthly CPUE (C. plumbeus per 100 hooks) for Middleground and Kiptopeke stations combined, a) Standard 9/0 hooks 1990 to 1999; b) 5/0 circle hooks -1996 to 1999.


CPUE

(s

hark

s pe

r 10

0 ho

oks)

CP

UE

(sha

rks

per

100

hook

s)a) Temporal CPUE of C. plumbeus in Chesapeake Bay

Standard Hooks (1990-1999)

V v<0 V . « r V .< o V

^ ^ ^ ^

) Temporal CPUE of C. plumbeus in Chesapeake Bay Circle Hooks (1996-1999)


100

Sea-Surface Temperature

Mean surface temperature from data collected in situ mirrored the

temporal trend in CPUE very closely, particularly during the immigration period

(Fig. 2-3). No sharks were caught when temperatures were below 18°C and sets

with CPUE>2.0 were observed only when temperature was greater than 21°C.

Peak shark CPUE was observed when temperature was approximately 26°C.

Linear regression using mean surface temperature as the independent variable

and mean CPUE as the dependent variable for the immigration period only was

highly significant (p<0.001, ^=0.98) indicating temperature may act as a factor

triggering immigration to Chesapeake Bay (Fig. 2-4a, Table 2-1 a). During the

emigration period, mean surface temperature also mirrored CPUE, though there

appeared to be a temperature lag (Fig. 2-3). CPUE began to decline a full month

prior to significant declines in temperature, which never cooled below 21 °C

during the emigration phase. Linear regression indicated that CPUE is

significantly correlated with surface temperature during the emigration period

(p=0.02, ^=0.77), and the intercept of the fitted line suggested all sharks leave

the Bay prior to the water cooling to 20°C (Fig. 2-4b, Table 2-1 b). The

relationship, however, was not as strong as during the immigration period and

the observed time lag suggested that temperature may not be the environmental

trigger for these sharks to leave the estuary and begin the fall migration.


101

Figure 2-3: Mean semi-monthly CPUE (C. plumbeus per 100 hooks) and surface temperature (°C) for Middleground and Kiptopeke stations combined for the years 1990-1999 (standard 9/0 hooks only).


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Temporal CPUE of C. plumbeus and Surface TemperatureChesapeake Bay (1990-1999)

18

to 1 5

oo£o 12 o

<D9

(0jc

(06

LU3Q.O 3

/ / ' / ' / /' / *I CPUE —♦ —Temp.

Tem

pera

ture

(d

egre

es

C)

102

Table 2-1a: Linear-regression summary for mean CPUE vs. mean surface temperature for the immigration period May 1 - July 31. Means are for semimonthly intervals.

Dependent Variable: Mean CPUE Independent Variable: Mean Surface Temperature

df SS MS F PRegression 1 199.07 199.07 195.61 0.0002Residual 4 4.07 1.02Total 5 203.14

Multiple R2 0.9899Adjusted R2 0.9800Standard Error 1.0088Observations 6

Table 2-1 b: Linear-regression summary for mean CPUE vs. mean surface temperature for the emigration period July 15 - October 15. Means are for semimonthly intervals.

Dependent Variable: Mean CPUE Independent Variable: Mean Surface Temperature

df SS MS F PRegression 1 58.65 58.65 13.40 0.0216Residual 4 17.51 4.38Total 5 76.16



103

Figure 2-4: Fitted line plots from linear regression of mean CPUE (C. plumbeus per 100 hooks) vs. mean surface temperature. Means are for semimonthly intervals, (error bars = SEM) a) Immigration period: May 1 July 31; b) Emigration period: July 15 - October 15.


CPUE

CP

UE

C. plumbeus CPUE vs. Temperature (Means) May 1 - July 31

18 -rR2 = 0.980015 -

12 -

9 --

6 -

3 -

15 17 19 21 23 25 27Temperature (degrees C)

C. plumbeus CPUE vs. Temperature (Means) July 15 - October 15

20 TR = 0.7701

16 -

8 -

4 -

27 26 25 24 23 22 21 20 19Temperature (degrees C)


104

Day Length (Photoperiod)

Day length was defined as the time between sunrise and sunset. These

data were calculated for each set using data supplied by the United States Naval

Observatory Astronomical Applications Department. This variable is annually

conservative, therefore means for the time intervals were calculated for one year

only. Day length changed little during May 1 to July 31 period, varying by only 42

minutes (Fig. 2-5). Linear regression analysis of CPUE as a function of day

length for the immigration period was insignificant (Fig. 2-6a, Table 2-2a). Day

length declined continuously during the emigration period, dropping by

approximately 2.6 hours, and mirrored CPUE remarkably well during this period

(Fig. 2-5). The linear-regression analysis was highly significant (p<0.001,

(^=0.96), suggesting day length may be a significant cue triggering emigration

from the estuary (Fig. 2-6b, Table 2-2b).


105

Figure 2-5: Mean semi-monthly CPUE (C. plumbeus per 100 hooks) and day length (hours) for Middleground and Kiptopeke stations combined for the years 1990-1999 (standard 9/0 hooks only).


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Temporal CPUE of C. plumbeus and Day Length Chesapeake Bay (1990-1999)

15

14

13

12

11

10

l CPUE • Day Length

Day

Leng

th

(hou

rs)

106

Table 2-2a: Linear-regression summary for mean CPUE vs. mean day length for the immigration period May 1 - July 31. Means are for semimonthly intervals.

Dependent Variable: Mean CPUE Independent Variable: Mean Day Length

df SS MS F pRegression 1 30.56 30.56 0.71 0.4474Residual 4 172.59 43.15Total 5 203.14


Table 2-2b: Linear-regression summary for mean CPUE vs. mean day length for the emigration period July 15 - October 15. Means are for semimonthly intervals.

Dependent Variable: Mean CPUE Independent Variable: Mean Day Length

df SS MS F pRegression 1 73.22 73.22 99.45 0.0006Residual 4 2.95 0.74Total 5 76.16



107

Figure 2-6: Fitted line plot from linear regression of mean CPUE (C. plumbeus per 100 hooks) vs. day length (hours) for emigration period (July 15 - October 15) only. Means are for semimonthly intervals, (error bars = SEM)


a)

1113Q_

C. plumbeus CPUE vs. Day Length (Means) July 15 - October 15

20

16

12

8

4

0

15 14.8 14.6 14.4 14.2 14Day Length (hours)

b) C. plumbeus CPUE vs. Day Length (Means) July 15 - October 15

16 --

4 --

15 14 13 12 11 10Day Length (hours)


108

Mark and Recapture

A total of 1,846 juvenile C. plumbeus were tagged in Virginia waters from

1995 through 2000. The distribution of this tagging effort is shown in Map 2-2

and a summary is given in Table 2-3. More sharks were tagged in 1999 than any

other year (n=439, 23.4% of total), followed closely by 1998 (n=416, 22.5% of

total). Sampling and tagging only occurred during the months of May through

October (Table 2-3a). The fewest sharks were tagged in May (n=51, 2.8%) and

the most were tagged in July (n=550, 29.8%). Approximately 61.3% (n=1131) of

the sharks were tagged inside Chesapeake Bay whereas 14.3% (n=264) were

tagged in seaside lagoons and tidal creeks along Virginia's Eastern Shore, and

24.4% (n=451) were tagged in Virginia coastal waters (Table 2-3b). To date, 45

shark recaptures have been reported giving an overall recapture rate of only

2.4%, which is less than half of that reported for juvenile C. plumbeus in

Delaware Bay (Merson 1998). It is believed that the low recovery rate is due to

severe under-reporting by commercial gillnet fishers in the summer nursery and

winter longline and drop-net fishers in North Carolina waters. Four of the

reported recaptures were discarded due to incomplete data, leaving 41 that were

used in the analysis.

Recreational fishers returned more tags than any other group (n=25).

Commercial vessels accounted for ten of the reported recaptures. Six of the

commercial recaptures were reported by independent fisheries observers on

commercial vessels whereas only four were reported by the commercial fishers

themselves. These results coupled with the overall low coverage of commercial


vessels by observers indicate that under-reporting in the commercial sector was

extreme. In fact, four of six winter recaptures by commercial vessels in North

Carolina waters were reported by a single observer, Mr. Chris Jensen, formerly

of the Gulf & South Atlantic Fisheries Development Foundation. In addition to

fishery recaptures, five of the tag returns were recaptured by the VIMS longline

survey and researchers from the North Carolina Aquarium returned one tag.

Recapture data also were obtained for two sharks tagged by the VIMS shark-

ecology program using tags supplied by the National Marine Fisheries Service

prior to the inception of the VIMS tagging program bringing the total returns used

in the analysis to 43.

Sharks tagged with VIMS dart tags were recaptured after a mean of 267

days and ranged between 4 and 1,109 days at liberty. Distance between tag and

recapture locations was calculated as the shortest distance between the points

using land as a bounding graphic in ArcView 3.2 GIS. In other words, sharks

were not allowed to cross over land to reach the recapture location giving a

conservative but realistic point-to-point distance measure. The mean distance

between tag and recapture locations was 104 kilometers and ranged from 0 to

830 kilometers. In addition, the two recaptured sharks tagged by VIMS using

NMFS tags were recaptured after 2,049 and 561 days at liberty and were 300

and 560 kilometers from the tagging location, respectively.


110

Map 2-2: Distribution of juvenile C. plumbeus tagging by VIMS during summers of 1995 to 2000.


77“2

0' 77

°10'

77*0

0*

76*5

0' 78

*40'

76*3

0*

76*2

0' 76

*10'

76*0

0' 75

*50'

75*4

0' 75

*30'

75°2

0' 75

*10'

75°0

0' 74

*50'

74*4

0' 74

*30’

74*2

0’ 74

*10'

74*00' 37*50' 37*40' 37*30' 37*20' 37*10* 37*00' 36*50' 36*40' 36*30' 36*20'

■ooO)a

JC

CD%

in «

in1A

• •

*fs,s

«or«*

U)rv

s

' 5 ,,r - ^______________ L ^ A iW

y«jr I" > /

r

oo<0

(0r*.s

♦ /

.OSoZC X»U.Z .QColZ JCZaLZ .Ot^C .OOo C .0So9C .0* <3Z J0ZJ9Z OZ,9Z


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 2-3: Tagging summary for juvenile Carcharhinus plumbeus tagged by VIMS (1995 - 2000).

a lYear Total

# (% total)May

# (% year)June

# (% year)July

# (% year)August

# (% year)September # (% year)

October # (% year)

1995 169 (9.2%) 0 0 36 34 99 01996 219(11.9%) 10 37 54 61 30 271997 357 (19.3%) 2 31 132 55 96 411998 416 (22.5%) 9 44 149 137 62 151999 439 (23.8%) 30 107 138 121 21 222000 246 (13.3%) 0 52 41 49 64 40T o t# (%) 1846 51 (2.8%) 271 (14.7%) 550 (29.8%) 457 (24.8%) 372 (20.2%) 143 (7.7%)

b lYear Total # Chesapeake Bay

# (%)Eastern Shore Lagoons

# (%)Virginia Coastal

#(%)1995 169 125 (74.0%) 8 (4.7%) 36 (21.3%)1996 219 117(53.4%) 33(15.1%) 69 (31.5%)1997 357 233 (65.3%) 38(10.6%) 86 (24.1%)1998 416 272 (65.4%) 53 (12.7%) 91 (21.9%)1999 439 301 (68.6%) 69(15.7%) 69(15.7%)2000 246 83 (33.7%) 63 (25.6%) 100 (40.7%)Total 1846 1131 (61.3%) 264 (14.3%) 451 (24.4%)

112

II: Migration PatternsOf the 43 tag returns, 31 (72%) were recaptured less than 50 km from the

tagging location (Fig. 2-7). Twenty-five of these were recaptured in the

Chesapeake Bay nursery and six were in nurseries on Virginia’s Eastern Shore.

The earliest of these recaptures occurred on May 28 and the latest return was on

October 15 (Fig. 2-7) suggesting that these summer nursery areas are utilized

from late May to mid-October. These results are in perfect agreement with the

temporal delineation pattern of the summer nursery interpreted using the longline

CPUE data (Fig. 2-2).

All of the remaining 12 recaptures were more than 200 kilometers

(mean=384 km) from the tagging location following a mean of 550 days at liberty.

Eleven of these were recaptured south of the tagging location whereas only one

was recaptured north of its tagging origin (Fig. 2-8, Map 2-3). With only one

exception, all southern recaptures occurred during the winter and spring when

they are suspected to be in winter nursery areas (Table 2-4). These data

indicate that the primary winter nurseries are located in near shore areas along

the Outer Banks of North Carolina between 34° 30' N and 35° 30’ N latitude (Fig.

2-8). The shark tagged with NMFS tag R9670 was approximately age two when

tagged and was recaptured in this region more than 5.5 years later. This

suggests that this region is used as a wintering area for at least the first seven

years of life. These wintering areas may extend much farther south, however.

One shark tagged as a neonate was recaptured the following May in the Inter

coastal Waterway in Hilton Head, South Carolina (32° 9’ N, 80° 50' W), 830 km


from the tagging location. A shark tagged with NMFS tag 211102 was

recaptured 1.5 years later in of January and was nearly 200 km off the coast of

Charleston, South Carolina (32° 50’ N, 77° 50’ W). This shark was at least seven

years old when tagged, suggesting older juveniles may utilize deeper, offshore

southern regions as wintering areas.

These data also provide information concerning the timing of these

migratory movements. Sharks were recaptured in the wintering areas as early in

the fall as October 25 and as late in the spring as May 23 (Fig. 2-8). This

corresponds remarkably well with the timing of the immigration and emigration

from the summer nursery in Chesapeake Bay. These data indicate that

Chesapeake Bay and lagoons along Virginia’s Eastern Shore act as summer

nursery and refuge habitat from late May to mid-October and coastal areas of

North Carolina and South Carolina provide important winter habitat from late

October to late May.

Ill: PhilopatryThirty-three of the 43 recaptured juvenile sharks were caught between

May 28 and October 15. Of these, 17 were recaptured the same year they were

tagged (Map 2-4, Table 2-5) after a mean of 30 days at liberty (range = 4-82

days). The mean distance between tag and recapture locations was 15 km

(range = 0-37 km). These recaptures indicate that the sharks do not leave the

protective nursery during this period, but actively move throughout the estuary.

For instance, one shark was recaptured approximately 32 kilometers from its


114

tagging location only four days later whereas another was recaptured within one

kilometer of the tagging location after 44 days. These findings agree with those

obtained using ultrasonic telemetry (Grubbs and Musick, in prep. b). Ten juvenile

sandbar sharks were continuously tracked for periods of 10 to 50 hours.

Although none of the tracked sharks moved out of the estuary, their daily

straight-line movements averaged 34 kilometers per day and activity spaces

ranged from 35 to 250 square kilometers.

Fourteen recaptured sharks were caught in the Chesapeake region but in

subsequent summers (Map 2-5, Table 2-6) after a mean of 461 days at liberty

(range = 225-1,109). The mean tag-recapture distance for this group was 17

kilometers (range 0-48 km). Ten were recaptured after approximately one year

at liberty, three after two years at liberty, and one after three years at liberty (Fig.

2-9). The age at recapture based on age at tagging estimated from length-at-age

data from Sminkey (1994) ranged from one to four years.

Two sharks recaptured during the summer months were not recaptured in

Virginia waters, shown as yellow circles in Figures 2-7and 2-8. One of these was

tagged in Virginia coastal waters along Virginia Beach in October 1999. These

near-shore waters are part of the migration route for sharks from nurseries north

of Chesapeake Bay. This shark was recaptured the following August in Little

Egg Harbor, part of Barnegat Bay, in New Jersey (Map 2-3, Figures 2-7,8). This

shark was probably tagged during its fall migration to southern wintering grounds,

then returned the following summer to its natal summer nursery in New Jersey.

Therefore, this does not indicate a departure from natal homing. The same


115

cannot be said for the second shark, which was recaptured in the Cape Fear

River in North Carolina in July 1998, one year after being tagged in Chesapeake

Bay (Map 2-3, Fig. 2-7,8). This animal may have migrated to this region the

previous fall as a wintering area but migrated inshore to the river the following

summer rather than returning to its natal nursery to the north. Therefore, in this

data set, 14 of 15 recaptures (93%) returned to their natal summer nursery in

subsequent years. These data provide the first strong evidence for philopatry or

natal homing in this species.


116

Figure 2-7: Temporal delineation of summer nursery of C. plumbeus using tag recapture data. Distance between tag and recapture location vs. day of year of recapture.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Mark/Recapture Distance vs. Day of Year

<DOCJS(/>b

1000

800

600

400

200

J F M A M Month

May 28

O

O

s o

Summer Nursery

N D

October 15

117

Figure 2-8: Location and temporal delineation of winter nursery of C. plumbeus using tag recapture data. Latitude of recapture location vs. day of year of recapture.


Latit

ude

of

Rec

aptu

re

vs.

Dat

e

Q

Z

o( 0

<

(Uc/}

oc

CO

E

□<DCO13zh_0)*«■*C

CD

E

O A O O S i O I O ^ f O N r^ c o « r t M « r t M n n

(saajBap) apn)pe~|


118

Map 2-3: Long-distance tag returns. All recaptures made >200 kilometers from tagging location, seasons combined.


Long-Distance Tag Returns - >200 km (shortest distance with land boundary)

• Tagging Locations (Summer)

▲ Recapture Locations’ (Summer) MDRecapture Locations (Winter)

Virginia

North Carolina

South Carolina

hoo

cohoo

COLooo

CO—<1o

COhoo

cohoo

CO-Ao

COhcoo

oT"co

100 100 200 KilometersT

N+coisoo

coo

81' 80° 79e 78° 77° 76° 75° 74°


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 2-4: Summary of tag recapture data from WINTER months. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. State refers to the coast nearest the recapture location labeled as NC (North Carolina), SC (South Carolina).

Tag# Date Date Tag Tag Recap Recap Distance Days at State GroupTagged Recap. Latitude Longitude Latitude Longitude km large

NMFS-R9670 8/18/92 3/29/98 37.2000 -76.0333 34.7533 -76.0217 300 2049 NC CommercialNMFS-211102 7/4/95 1/15/97 36.9167 -75.7000 32.8333 -77.8333 560 561 SC CommercialVIMS-0146 9/18/95 3/24/98 36.7500 -75.8667 34.5267 -75.9617 270 918 NC CommercialVIMS-0517 7/17/97 11/15/97 37.2667 -76.1000 35.2000 -75.6833 260 121 NC CommercialVIMS-0978 8/17/98 11/17/99 37.2596 -76.8989 35.5167 -75.4633 200 457 NC RecreationalVIMS-1026 8/18/98 5/23/00 37.2000 -76.0400 32.1500 -80.8333 830 644 SC RecreationalVIMS-1073 8/19/98 10/25/00 37.1333 -75.9833 35.1917 -75.6000 330 787 NC CommercialVIMS-1043 8/19/98 2/10/99 37.1450 -75.9900 34.6333 -76.6333 390 176 NC CommercialVIMS-1162 9/30/98 2/10/99 37.2883 -75.7833 35.0333 -75.9667 300 133 NC CommercialVIMS-1968 8/21/00 10/30/00 37.1000 -76.0667 35.0833 -75.8500 280 70 NC Commercial

120

Map 2-4: Short-term tag recaptures. All sharks recaptured the same summer in which they were tagged.


76°5

0’ 76

°40’

76°3

0’ 76

°20'

76°1

0*

76°0

0’ 75

°50’

75°4

0' 75

°30'

37° 20 ' ________________37° 10' 37° 00 ' Z B °5 0 '

L r , - m £ r

i

.ozoze .OkZC .OOoZC .0So9€


76°5

0*

76°4

0' 76

°30’

76°2

0’ 76

°10*

76

°00'

75°5

0' 75

°40’

75°3

0’

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 2-5: Summary of tag recapture data from SUMMER months for returns occurring the same summer as tagging. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. All recaptures occurred in Virginia waters. Area of recapture summarized as Eastern Chesapeake Bay (CB-E), Western Chesapeake Bay (CB-W), and Virginia Eastern Shore (ES).

Tag# DateTagged

DateRecap.

TagLatitude

TagLongitude

RecapLatitude

RecapLongitude

Distance Days at Area km large

Group

VIMS-0186 5/29/96 8/19/96 36.7500 -75.8667 37.0833 -76.0167 37 82 CB-E RecreationalVIMS-0196 6/24/96 7/30/96 37.1833 -76.0333 37.4000 -76.0450 24 30 CB-E RecreationalVI MS-0252 7/24/96 8/24/96 37.1833 -76.0333 37.0917 -75.9917 11 31 CB-E RecreationalVI MS-0297 8/21/96 8/31/96 37.1667 -76.0833 37.1383 -76.2333 14 10 CB-W RecreationalVIMS-0489 7/17/97 8/2/97 37.1000 -76.0667 37.2000 -76.2500 20 16 CB-W RecreationalVIMS-0533 7/17/97 8/9/97 37.4167 -76.1667 37.4000 -76.2000 3.5 23 CB-W RecreationalVIMS-0480 7/17/97 9/14/97 37.1000 -76.0667 37.3583 -76.2717 34 59 CB-W RecreationalVIMS-0578 8/12/97 8/21/97 37.2667 -75.8967 37.2667 -75.8967 0 9 ES RecreationalVIMS-0652 9/17/97 10/15/97 37.1000 -76.0667 37.1550 -76.2500 17.5 28 CB-W CommercialVIMS-0927 7/29/98 8/30/98 37.1833 -76.0333 37.1667 -76.9967 4 32 CB-E RecreationalVIMS-0948 7/30/98 8/15/98 37.2067 -76.1200 37.2167 -76.2667 13 15 CB-W RecreationalVIMS-0988 8/17/98 9/30/98 37.2517 -75.8987 37.2600 -75.8983 0.5 44 ES VIMSVIMS-1210 5/25/99 5/29/99 37.2600 -75.8983 37.0450 -76.0667 32 4 CB-E RecreationalVIMS-1409 7/8/99 8/15/99 37.1820 -76.2130 37.2833 -76.3067 14 38 CB-W RecreationalVIMS-1389 7/8/99 9/10/99 37.1830 -76.0190 37.1500 -76.9800 5 64 CB-E RecreationalVIMS-1576 8/18/99 8/29/99 37.2200 -76.3200 37.1383 -76.0833 23 11 CB-E CommercialVIMS-1962 8/21/00 8/25/00 37.2000 -76.0333 37.2000 -76.0283 0.5 4 CB-E Recreational

122

Map 2-5: Evidence of natal homing. Tag recaptures made < 50 kilometers from tagging location in summers following at least one winter migration.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

76°40' 76°30' 76°20* 76°10’ 76°00' 75 50

aVirginia Beach

10 0 10 20 Kilometers

75°40‘ 75°30'

N

+

76°40' 76°30' 76°20' 76°1O' 76°00'

Long-term Tag Returns# Recapture Location O Tagging Location #= DAL (days at large)

75°50' 75°40* 75°30’

•05o9

C ,00

oZC

.OUZ

G M

olt

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 2-6: Summary of tag recapture data from SUMMER months for returns occurring following at least one winter season. Distance was measured as the most direct route from tagging location to the recapture location using land as an impenetrable boundary. Area of Virginia recapture summarized as Eastern Chesapeake Bay (CB-E), Western Chesapeake Bay (CB-W), and Virginia Eastern Shore (ES), Virginia Beach (VB). Recapture from other states labeled as NC (North Carolina), SC (South Carolina), NJ (New Jersey).

Tag# Date Date Tag Tag Recap Recap Distance Days at Area or GroupTagged Recap. Latitude Longitude Latitude Longitude km large State

VIMS-0064 8/19/95 7/11/96 37.1833 -76.0333 37.2667 -75.8967 30 326 ES VIMSVIMS-0135 9/14/95 5/28/96 37.1167 -76.1000 37.1833 -76.0333 9.5 225 CB-E VIMSVIMS-0353 8/21/96 9/12/98 37.1667 -76.0833 36.7500 -75.9333 48 752 VB RecreationalVIMS-0439 7/15/97 7/6/99 37.2611 -75.8980 37.2583 -75.8983 0.75 721 ES VIMSVIMS-0509 7/17/97 7/31/98 37.2000 -76.0333 33.9500 -77.9667 540 379 NC NC AquariumVIMS-0507 7/17/97 8/1/00 37.1833 -76.0167 37.3000 -75.8000 45 1109 ES CommercialVIMS-0650 9/17/97 7/15/98 37.1000 -76.0667 37.0333 -76.0667 7 301 CB-E RecreationalVIMS-0648 9/17/97 8/28/99 37.0930 -76.0670 37.0450 -76.0667 4.5 710 CB-E RecreationalVIMS-0678 9/18/97 8/15/98 37.2167 -76.0500 37.2017 -76.0317 2.5 331 CB-E RecreationalVIMS-0718 10/2/97 6/30/98 37.1833 -76.0167 37.0917 -75.9917 10.75 271 CB-E RecreationalVIMS-0877 7/8/98 7/8/99 37.2105 -76.0620 37.2000 -76.2183 14 365 CB-W VIMSVIMS-0921 7/29/98 6/11/99 37.1867 -76.0317 37.1383 -76.2333 19 317 CB-W CommercialVIMS-0908 7/29/98 6/28/99 37.0667 -76.0967 37.1667 -76.4000 32 334 CB-W RecreationalVIMS-0973 8/12/98 6/25/99 37.1345 -76.0878 37.0770 -76.2230 13.75 317 CB-W RecreationalVIMS-0998 8/17/98 8/24/99 37.2925 -75.7900 37.3217 -75.8417 5.5 372 ES RecreationalVIMS-1618 10/6/99 8/5/00 36.7500 -75.8700 39.5900 -74.2857 350 304 NJ Recreational

N>CO

124

Figure 2-9: Evidence of natal homing from tag recaptures. Distance from tagging location versus days at liberty for all sharks recaptured less than 200 kilometers from the tagging location.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

60o>cEb ^O) r£ I 40E c 2 .2§ § 20S• MMa

Days at Liberty vs. Tag/Recapture Distance Distance < 200 km Only

3651

730 1,095

• Age 0 ■ Age 1 ♦Age 2 A age 4

Days At Liberty

125

DISCUSSION

The results of this study indicated that Chesapeake Bay is utilized as a

summer nursery for C. plumbeus primarily from mid-May to mid-October. The

CPUE data indicated that juvenile sharks began immigrating to the Bay after May

15 with the majority of sharks entering the estuary after June 15. The CPUE

peaked in late July indicating full recruitment to the estuary by this time.

Immigration to the Bay was not significantly correlated with salinity, day length, or

lunar phase and was highly correlated with increasing water temperature as had

been reported previously by Musick and Colvocoresses (1986). Sharks were

absent from longline catches when surface temperature was below 18°C and

CPUE greater than 2.0 sharks per 100 hooks was observed only when

temperature was greater than 21 °C. Peak CPUE occurred when temperature

was approximately 26°C. Merson (1998) reported sharks in Delaware Bay when

water temperature was as low as 15.4°C.

Mean CPUE began to decline in August, particularly later in the month. I

hypothesize that this CPUE decline in August represents dispersal throughout

the nursery rather than the beginning of the emigration period. Similar dispersal

trends have been observed in the movement patterns of other fishes such as

Morone saxatilis in Chesapeake Bay (Moore and Burton 1975). The CPUE data

suggested emigration from the estuary in preparation for the fall migration to

wintering grounds occurred in late September and early October. The timing of

emigration also was correlated significantly with surface temperature, though the

relationship was not as strong as during the immigration period as emigration


126

began prior to significant declines in temperature. Interpolation of the regression

line suggested all sharks vacated the estuary prior to temperature falling to 20°C.

The timing of the emigration movement was highly correlated with day length.

Day length has been shown to initiate migration in birds (Berthold 1975) and

Aidley (1981) suggested it might be a possible trigger for fishes. The data

presented in this study indicate that temperature serves as a migratory catalyst

for juvenile C. plumbeus to enter Chesapeake Bay and day length serves as the

stimulus to emigrate from the Bay in fall. It is possible that day length also

serves as the stimulus to begin spring migrations from wintering grounds. In

other words, day length may signal juvenile sharks to begin migrating north, yet

they remain in coastal waters until temperature stimulates a movement into the

estuarine nursery. Additional data from the wintering grounds are needed to test

this hypothesis.

Tag-recapture data indicate that the migrations of juvenile C. plumbeus

are spatially extensive. The wintering grounds are concentrated south of the

summer nurseries as expected. They appear to be concentrated in near shore

waters off the coast of North Carolina and extend to southern South Carolina and

to at least 200 kilometers from shore. The earliest recapture in Chesapeake Bay

occurred on May 28 whereas the latest recapture in the Bay occurred on October

15, supporting the conclusion that the estuary serves as a nursery ground from

late May to the middle of October. The earliest fall recapture on the southern

wintering grounds occurred on October 25 and the latest spring recapture in this

area was on May 23, indicating these areas serve as important wintering areas


127

from late October to the middle of May. These data support the timing of the

migratory movements based on CPUE data.

Gerking (1959) defined homing as “going to a place formerly occupied

instead of equally probable places." Tag-recapture data indicate that most

juvenile C. plumbeus return to their natal summer nurseries for at least their first

four years of life. Ninety-three percent of sharks recaptured in subsequent

summers (n = 15) were recaptured within 50 kilometers (mean = 17 km) of the

tagging location. This is considered strong evidence of philopatry, considering

that these sharks have been shown to travel extensively within the nursery with

activity spaces averaging 100 km2 (Grubbs and Musick in prep b) and that

wintering areas are more than 200 km from the summer nurseries. Additional

data are needed to examine the duration of this philopatric behavior and

determine if females maintain this bond to adulthood, returning when mature at

about 15 years old (Sminkey 1995) to deliver their own pups. In addition, future

research directions should focus on determining what environmental cues the

juvenile sharks use to discern their natal nursery. Olfaction has been well

documented as the principal stimulus for homing in diadromous fishes (Hasler

and Scholz 1980). Harden Jones (1968) suggested marine fishes also might

use olfactory cues from groundwater seepage to locate home habitats.

These data elucidate the importance of defining seasonal essential fish

habitat for the various life stages of highly migratory species. Grubbs and Musick

(in prep a) defined EFH for juvenile C. plumbeus in Chesapeake Bay. These

estuarine areas are highly susceptible to anthropogenic degradation and must be


128

afforded some level of protection. Fishing regulations have been implemented to

protect these juvenile sharks in these summer nursery habitats. No regulations

exist, however, on the wintering grounds, which are just beginning to be

investigated. Sharks appear to be densely aggregated in these regions and are

therefore highly susceptible to commercial fishing gear and can easily be over-

exploited. Protection of juvenile sandbar sharks while in crucial summer habitats

may prove fruitless unless protection in winter and migratory habitats is

implemented as well.


129

LITERATURE CITED

Aidley, D. J. 1981. Animal Migration. Society of Experimental Biology - Seminar Series 13. 264 p. Cambridge University Press.

Anonymous. 1996. Final Report. MARFIN Award NA47FF0008. Gulf and South Atlantic Fisheries Development Foundation. March 1996.

Bigelow, H. B. and W. C. Schroeder. 1948. Sharks, pp. 59-546 In: Fishes of the western north Atlantic. Part 1. Vol. 1. (J. Tee-Van, C. M. Breder, S. F. Hildebrand, A. E. Parr, and W. C. Schroeder, eds.). Mem. Sears Foundation for Marine Research, Yale Univ. New Haven, CT.

Berthold, P. 1975. Migration: control and metabolic physiology, pp. 77-128 In: Avian Biology: Volume 5 (D. S. Farner and J. R. King eds.). Academic Press. New York.

Boreman, J. and R. R. Lewis. 1987. Atlantic coastal migration of striped bass, pp. 331-339 In: Common strategies of anadromous and catadromous fishes (M. J. Dadswell, R. J. Klauda, C. M. Moffitt, R. L. Saunders,R. A. Rulifson, and J. E. Cooper, eds.), American Fisheries Society Symposium 1.

Camhi, M. 1998. Sharks on the line: A state-by-state analysis of sharks and their fisheries. National Audubon Society. New York.

Carlson, J. K. 1999. Occurrence of neonate and juvenile sandbar sharks,Carcharhinus plumbeus, from the northeastern Gulf of Mexico. Fishery Bulletin 97(2): 387-391.

Chopoton R. B. and J. E. Sykes. 1961. Atlantic coast migrations of large striped bass as evidenced by fisheries and tagging. Trans. Am. Fish. Soc. 90 (1): 13-20.

Compagno, L. J. V. 1984. FAO species catalogue. Vol. 4. Sharks of the world. Part 2 - Carcharhiniformes. FAO Fish. Synop. 4(2): 250-655.

Cowan, J. H. Jr., and R. S. Birdsong. 1985. Seasonal occurrence of larval and juvenile fishes in a Virginia Atlantic coast estuary with emphasis on drums (family Sciaenidae). Estuaries 8(1): 48-59.

Friedland, K. D. and L. W. Haas. 1988. Emigrations of juvenile Atlanticmenhaden Brevoortia tyrannus from York River estuary. Estuaries 11 (1): 45-50.


130

Garrick, J. A. F. 1982. Sharks of the genus Carcharhinus. NOAATech. Rep. NMFS Circ. 445.

Gerking, S. D. 1959. The restricted movements of fish populations. Biological Reviews 34: 221-242.

Grubbs, R. D. and J. A. Musick. Spatial delineation of summer nursery areas to define essential fish habitat for juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia, in prep a.

Grubbs, R. D. and J. A. Musick. Movements of juvenile Carcharhinus plumbeus in Chesapeake Bay. in prep b.

Grubbs, R. D. and J. A. Musick. Population trends, juvenescence, and mortality estimation for juvenile Carcharhinus plumbeus in Chesapeake Bay and Virginia coastal waters, in prep c.

Hallock, R. J., R. F. Elwell. and D. H. Fry, Jr. 1970. Migrations of adult king salmon, Oncorhynchus tshawytscha, in the San Joaquin Delta, as demonstrated by the use of sonic tags. California Department of Fish and Game, Fish Bulletin: 151. 92 p.

Harden Jones, F. R. 1968. Fish Migration. St. Martin’s Press. New York.

Hasler, A. D. and A. T. Scholz. 1980. Artificial imprinting: a procedure forconserving salmon stocks, pp. 179-199 In: Fish behavior and its use in the capture and culture of fishes (J. E. Bardack, J. J. Magnuson, R. C. May, and J. M. Reinhart, eds.). Manila: International Center for Living Aquatic Resources Management.

Heape, W. 1931. Emigration, migration and nomadism. Heffer, Cambridge.369 pp.

Hoff, T. B., and J. A. Musick. 1990. Western North Atlantic shark-fisherymanagement problems and informational requirements, pp. 455-472 In: Elasmobranchs as living resources: advances in the biology, ecology, systematics and the status of the fisheries (H. L. Pratt, Jr., S. H. Gruber, and T. Taniuchi, eds.), U.S. Dep. Commer., NOAATech. Rep. NMFS 90.

IUCN (International Union for Conservation of Nature and Natural Resources). 2000. 2000 IUCN Red List of Threatened Animals. IUCN, Gland, Switzerland.

Mather, F. J. 1962. Transatlantic migration of two large bluefin tuna. J. du Conseil 27: 325-327.


131

McErlean, A. J. 1973. Abundance, diversity and seasonal patterns of estuarine fish populations. Estuarine and Coastal Marine Science 1(1): 19-36.

Medved, R. J., and J. A. Marshall. 1981. Feeding behavior and biology of young sandbar sharks, Carcharhinus plumbeus (Pisces, Carcharhinidae), in Chincoteague Bay, Virginia. Fish. Bull. 79(3): 441-448.

Meek, A. 1916. The migrations offish. Edward Arnold, London. 427 pp.

Melvin, G. D., M. J. Dadswell, and J. D. Martin. 1986. Fidelity of American shad, Alosa sapidissima (Clupeidae), to its river of previous spawning. Can. J. Fish. Aq. Sci. 43(3): 640-646.

Merriner, J. V., W. H. Kriete and G. C. Grant. 1976. Seasonality, abundance, and diversity of fishes in the Piankatank River, Virginia (1970-1971).Ches. Sci. Vol. 17, No.4:238-245

Merson, R. R. 1998. Nurseries and maturation of the sandbar shark. Ph.D. Dissertation. University of Rhode Island. 150 pp.

Moore, C. J. and D. T. Burton. 1975. Movements of striped bass, Moronesaxatilis, tagged in Maryland waters of Chesapeake Bay. Transactions of the American Fisheries Society 104 (4): 703-709.

Mundy, P. R. 1984. Migratory timing of salmon in Alaska: with an annotated bibliography on migratory behavior and relevance to fisheries research. Alaska Department of Fish and Game: Informational leaflet No. 234.

Musick, J. A. 1999. Introduction to Part 2: Essential fish habitat identification, p. 41 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

Musick, J. A., and J. A. Colvocoresses. 1986. Seasonal recruitment ofsubtropical sharks in Chesapeake Bight, U.S.A. pp. 301-311 In: Workshop on recruitment in tropical coastal demersal communities (A. Yanez y Arancibia and D. Pauley, eds.), FAO/UNESCO, Campeche, Mexico, 21-25 April 1986. I.O.C. Workshop Rep. 44.

Musick, J. A., J. A. Colvocoresses and E. J. Foell. 1986. Seasonality and the distribution, availability and composition offish assemblages in Chesapeake Bight, pp. 451-474 In: A. Yanez-Arancibia (ed.) Fish Community Ecology in Estuaries and Coastal Lagoons: Towards and Ecosystem Integration.


132

NMFS. 1999. Final fishery management plan for Atlantic tunas, swordfish, and sharks. April 1999. National Oceanic and Atmospheric Administration, National Marine Fisheries Service, U.S. Department of Commerce.

NMFS. 1996. 1996 Report of the shark evaluation workshop. June, 1996. NOAA, National Marine Fisheries Service, Southeast Fisheries Science Center, Miami, Florida.

NMFS. 1993. Fishery management plan for sharks of the Atlantic Ocean. National Oceanic and Atmospheric Administration, National Marine Fisheries Service, U.S. Department of Commerce.

Pruitt, W. A. 1997. Pertaining to sharks. Virginia Marine Resources Commission. 4 VAC 20-490-10 et seq.

Schmitten, R. A. 1999. Essential fish habitat: opportunities and challenges for the next millennium, pp. 3-10 In: Fish habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society, Symposium 22, (L. R. Benaka, editor). Bethesda, Maryland. 459 pp.

Seckel, G. R. 1972. Hawaiian-caught skipjack tuna and their physical environment. Fishery Bulletin 70:763-787.

Sminkey, T. R. 1994. Age growth, and population dynamics of the sandbar shark, Carcharhinus plumbeus, at different population levels. Ph.D. Dissertation, Va. Inst. Mar. Sci., College of William and Mary. 99 pp.

Sminkey, T. R. and J. A. Musick. 1995. Age and growth of the sandbar shark, Carcharhinus plumbeus, before and after population depletion. Copeia 1995(4):871-883.

Smith, H. D. 1973. Timing of Babine Lake sockeye salmon stocks in the north- coast commercial fishery as shown by several taggings at the Babine counting fence and rates of travel through the Skeena and Babine Rivers. Fisheries Research Board of Canada. Technical report; no. 418. 31 p.

Springer, S. 1960. Natural history of the sandbar shark, Eulamia milberti. U.S. Fish. Wildl. Serv., Fish. Bull. 61:1-38.

Talbot, G. B. and J. E. Sykes. 1958. Atlantic coastal migrations of American shad. Bull. U.S. Bureau Fish. 58: 473-490.

Tarbox, K. E. 1988. Migratory rate and behavior of salmon in upper Cook Inlet, Alaska, 1983-1984. Alaska Department of Fish and Game, Fishery Research Bulletin 88-05.


133

USDOC (U.S. Department of Commerce). 1996. Magnuson-Stevens Fishery Conservation and Management Act as amended through October 11,1996. National Oceanic and Atmospheric Administration Technical Memorandum NMFS-F/SPO-23.


Chapter 3Short-term Movements and Swimming Depth of Juvenile Carcharhinus

plumbeus in Chesapeake Bay

134


ABSTRACT

135

Manual telemetry was used in investigate the diel activity patterns of

juvenile Carcharhinus plumbeus (sandbar sharks) in Chesapeake Bay.

Ultrasonic transmitters equipped with depth sensors were attached to ten sharks

externally and manually tracked for 10 to 50 consecutive hours. Mean activity

space was conservatively estimated to be 110 km2. This estimate is two orders

of magnitude greater than that reported for other carcharhiniform species.

Swimming depth ranged from surface to 40 meters. Depths of more than 40

meters were common during daylight hours whereas depths greater than 20

meters were rare during the night. Mean daytime swimming depth was 12.8

meters whereas mean nighttime swimming depth was 8.5 meters. This

difference was statistically significant. The activity spaces of nearly every shark

were centered over one of the three deep channels in the lower Chesapeake

Bay. Most sharks tracked remained in the deep channels during the day but

ventured out of the channels to shallower waters during the night. This diel

activity pattern and large activity space is hypothesized to be an adaptation for

foraging on patchy prey in a highly productive, but highly seasonal, temperate

estuary. I challenge the demersal classification of the species because the

sharks were at least three meters from the bottom more than 50% of the tracking

duration and at least six meters from the bottom more than 35% of the duration.

Swimming direction was correlated with mean direction of tidal current. A mean


136of 75.8% of all fixes was in the general direction of the mean current. Mean

swimming direction and mean current direction were statistically similar. Angular

dispersion was estimated at 56.8° relative to the tidal-current direction.


137

INTRODUCTION

TelemetryBiotelemetry methods are valuable tools that enable biologists to monitor

the movements, behavior, and physiological parameters of free-ranging animals.

They enable the collection of data from animals that are not readily visible with

little interference of their natural behaviors and provide more data than are

normally collected with techniques such as mark and recapture (Winter 1992).

The development of telemetry equipment with underwater applications began in

the 1950s (Trefethen 1956). A publication by Johnson (1960) marked the first

actual application of these underwater biotelemetry techniques. He fitted adult

Oncorhynchus tshawytscha with external ultrasonic transmitters and tracked

them on their migration up the Columbia River in Washington. Following this

pioneering work, many researchers recognized the tremendous advantages

underwater telemetry could provide for understanding behavioral patterns of

subjects that lived in such a concealing medium. This is reflected in the increase

to 147 publications and reports utilizing these techniques to study 40 species of

fishes, 10 species of mammals, 4 species of reptiles, and 5 species of

invertebrates by 1975 (Stasko and Pincock 1977).

Many of the early studies involving fishes concentrated on location-only

tracking of diadromous fishes such as Oncorhynchus gorbuscha (Stasko et al.

1973), Anguilla anguilla (Tesch 1972), and Alosa sapadissima (Dodson et al.

1972) during their annual migrations and near obstructions such as dams. Other


138

authors used location-only telemetry to study the daily activity patterns of pelagic

oceanic fishes such as Euthynnus pelamis (Yuen 1970) and freshwater fishes

such as Esoxlucius in lakes and impoundments (Malinin 1971). Other early

applications of underwater biotelemetry involved sensored transmitters to detect

parameters such as body temperature (Carey and Lawson 1973), ambient

temperature (Scarriota 1974, Coutant 1975), swimming depth (Standora et al.

1972, Stasko and Rommel 1974), swimming speed (Standora et al. 1972), and

heart rate (Nomura and Ibaraki 1969).

Elasmobranch TelemetryBass and Rascovich (1965) developed the first ultrasonic transmitters for

tracking large, fast fishes. They field tested the equipment on one adult Sphyrna

lewini off Palm Beach, Florida, and one adult Carcharhinus plumbeus off Jupiter

Inlet, Florida, in 1963. The animals were tracked for 2 hours and 4 hours,

respectively. Since this time telemetric techniques have been used with at least

35 species of elasmobranchs. Many of these were tracked for very short periods

or as preliminary reports. In fact, as of 1990, only six species (Prionace glauca,

Carcharhinus amblyrhynchos, Negaprion brevirostris, Sphyrna lewini, Squatina

californica, and Heterodontus francisci) were represented by more than 10

trackings (Nelson 1990). This number has increased dramatically over the past

decade.


139

Activity SpaceMany volumes of literature have been published on the subjects of activity

spaces and home ranges. An understanding of spatial and temporal activity

patterns is essential to studying the ecology of a species. Most of the pioneering

studies in this area as well as the analytical development were from studies of

mammals and birds. Very few studies have been conducted on the activity

patterns of juvenile elasmobranchs, especially in nursery areas. The North

Sound of Bimini, Bahamas, represents a semi-protected nursery for juvenile

Negaprion brevirostris that has been studied extensively. Morrissey and Gruber

(1993) used ultrasonic telemetry techniques to track 38 juvenile N. brevirostris in

this nursery for periods ranging between 1 and 153 days. They found that these

sharks had very well-defined activity spaces that ranged between 0.23 and

1.26 km2 with a high degree of site fidelity. Size of the activity space was

positively correlated with size of the shark. This relationship has been thoroughly

demonstrated in mammals (McNab 1963), but rarely applied to fishes.

Holland et al. (1993) tracked six juvenile Sphyrna lewini in Kaneohe Bay,

Oahu, Hawaii, a known nursery for this species. Sharks were tracked for up to

12 days and had activity spaces between 0.46 and 3.52 km2. The activity space

was greatly expanded during evening hours, presumably due to foraging, and

restricted during the day, presumably due to refuging.

Two published studies have detailed attempts to track the movements of

juvenile C. plumbeus while they were in their nurseries. Huish and Benedict

(1977) tracked 10 animals reported to be Carcharhinus obscurus in Cape Fear,


140

North Carolina. Due to the small size of these animals (65 - 95.5 cm TL) it is

believed that they were actually C. plumbeus (C. obscurus are 90-110 cm TL at

birth). This potential misidentification is common with juveniles of these two

species. Transmitters were implanted into the peritoneal cavity in five subjects

and attached externally to the other five. The sharks were tracked for only 0.7 -

13.0 hours and traveled with the tidal currents at rates of 0.1 to 1.3 kilometers per

hour. Medved and Marshall (1983) tracked three C. plumbeus using ultrasonic

telemetry and 20 C. plumbeus using tethered balloons in the area of

Chincoteague, VA. Animals were tracked for periods ranging from 1.0 to 10.7

hours and traveled between 1.87 and 14.68 kilometers. Twenty of the twenty-

three subjects spent the majority of the time traveling with the tidal currents.


141


Study Area

The Chesapeake Bight is characterized by one of the most seasonally

dynamic environmental regimes in the world. Sea-surface temperature may vary

seasonally by as much as 30°C. The demersal fish fauna in the region is largely

migratory, dominated by boreal species in winter and sub-tropical species in

summer (Musick and Colvocoresses 1986). Chesapeake Bay and the seaside

lagoons of Virginia's Eastern Shore are heavily utilized as pupping and nursery

areas for Carcharhinus plumbeus. Pregnant females enter the lower Bay and

seaside lagoons in late May through June to liberate young and then retreat to

deeper offshore waters. The young remain in the nursery through the summer,

emigrating in fall to over-wintering areas south of Cape Hatteras, North Carolina

(Grubbs and Musick, in prep b).

Chesapeake Bay is a coastal-plain estuary characterized by extremely

variable environmental parameters. The hydrographic regime is dominated by

two-layer estuarine circulation in which low-salinity water from numerous rivers

flows over the dense, high-salinity ocean water. The freshwater flow is greatest

in the spring and leads to a stratified water column. This stratification is re

enforced by the warming of surface waters during this period and is maintained

throughout the summer. This stratification causes the development of a

thermocline, pycnocline, and obvious halocline, as weli as local oxyclines as


142

waters below the mixing layers get depleted of oxygen and are not replenished.

The hydrography of the Bay is also influenced by the effect of Coriolis, which

deflects incoming oceanic water masses to the right as they enter the lower

estuary. Combined with the fact that most low salinity riverine input is from the

western side of the Bay, these factors lead to a halal tilting of higher salinities

along the eastern portion of the estuary. This hydrography is hypothesized to

have profound effects on the distribution of primary and secondary nursery areas

for juvenile C. plumbeus in this estuary. The stratification in Chesapeake Bay

breaks down in the fall as surface waters cool and sink causing mixing of surface

and bottom layers. The timing of this destratification and its effect on salinity and

temperature may affect the temporal utilization of the nursery.

The hydrographic regime of the seaside creeks and lagoons along the

eastern shore are much less variable than the main estuary in terms of salinity

and dissolved oxygen. The constriction of tidal creeks, however, produces much

higher current velocities on average. Environmental clines rarely form because

the water column remains mixed and very saline as the dynamic tidal currents

constantly flush it. In addition, these creeks and lagoons have little riverine

influence. These hydrographic differences may affect the habitat utilization

patterns of young C. plumbeus in the two habitats.


143

Telemetry Gear

The telemetry system consisted of ultrasonic transmitters, a directional

hydrophone, an ultrasonic receiver, and a pulse-interval decoder or stopwatch.

The ultrasonic transmitters used were model DT-96 manufactured by

Sonotronics in Tucson, Arizona. These transmitters produced pulses between

32 and 40 kHz, which is well above the auditory range of elasmobranchs.

Transmitters of the same frequency were differentiated by unique aural codes.

For example, a transmitter with code 258 produced 2 pulses followed by a pause,

5 pulses followed by a pause, then 8 pulses followed by a pause, and repeated.

Each transmitter was equipped with a pressure sensor that increased the pulse

interval with increasing pressure enabling swimming depth to be calculated. The

transmitters were 18 mm in diameter and 95 mm long with an average battery life

of 15 months. The weight of the transmitters was 14 grams in water, which

corresponded to approximately 1.4% of the body weight of the smallest C.

plumbeus in the nursery. This is well below the 2.0% maximum limit that is the

standard for telemetric studies with fishes (Mellas and Haynes 1985).

Transmissions were received by a Dukane Model N30A5B Underwater

Acoustic Locator System that consisted of a directional hydrophone and an

ultrasonic receiver that was sensitive to acoustic signals between 25 and 40 kHz.

Pulse interval was measured by a Pulse Interval Docoder developed by

Ultrasonic Telemetry Systems or simply by using a stopwatch to time 30 pulses.

Dividing this time by 30 gave the pulse interval. Due to interference from intense


144

background noise the stopwatch method was most often used. A simple

regression equation provided the conversion of pulse-interval to depth.

Transmitter Attachment

Four considerations must be made when determining the proper method

of transmitter attachment for a given study; handling time and stress, post

attachment trauma on the subject, desired duration of transmitter attachment and

acoustic contact with subject, and likelihood of recovery of the transmitter at the

conclusion of tracking.

Ingestion of the transmitter offers the least handling time and stress as

well as the shortest post-release trauma. This method was successfully used on

juvenile Sphyrna lewini in Hawaii (Holland et al. 1993). It also offers the

possibility of recovery if the transmitter is made positively buoyant and the

subject is tracked until the transmitter is voided from the body. There are several

negative aspects of this method. The transmitter may be voided through

regurgitation and therefore provide only brief tracking durations. It may block the

digestive track of subjects causing long-term trauma or death. The presence of

the transmitter in the stomach may also alter natural behaviors such as foraging

if the transmitter is large enough to occupy a significant proportion of the

stomach volume.

Intraperitoneal surgical implantation offers tracking durations limited only

by the life of the transmitter and is assumed to cause no alteration of natural


145

behavior following an appropriate post-surgery recovery period. This method

requires longer handling times, more severe stress, and the recovery period due

to post-surgical trauma may be long. This method was successfully used with

juvenile Negaprion brevirostris in Bimini, Bahamas (Morrissey and Gruber 1993).

In addition, surgically implanted transmitters were found to have no significant

effects on growth, hematocrit analysis, leukocyte counts, and condition of

juvenile C. plumbeus held in captivity on the Eastern Shore of Virginia (Leigh

1997). This is probably the most desirable method for long-term tracking

experiments.

External attachment methods are characterized by relatively short

handling times and recovery times. The amount of attachment trauma depends

on the type of external attachment but also may be very low. The duration of the

attachment varies between methods from a few days to a few months. This may

be maximized by carefully selecting the proper attachment method. External

attachment methods also offer an increased probability of transmitter recovery

through the use of positively buoyant transmitters and degradable attachment

links with known degradation rates. In choosing an external method of

attachment, researchers must also consider long-term physical damage such as

tissue necrosis that may occur due to the attachment as well as possible

inhibitory effects the external attachments may have on normal behaviors such

as swimming if the appropriate method is not chosen. One last consideration is

the degree to which an externally attached transmitter may foul with epibiota in a


146

given environment or become entangled with macroalgae. Huish and Benedict

(1977) compared external attachment to intraperitoneal implantation in the C.

plumbeus (published as C. obscurus) they tracked briefly in Cape Fear, North

Carolina. They found no differences in behavior or rates of movement between

the two methods. Blaylock (1990) found that externally attached transmitters did

not affect swimming speed significantly in juvenile Rhinoptera bonasus if the

transmitter weighed less than 3% of the body weight of the ray. The transmitter

significantly slowed swimming speed if it weighed more than 7% of the ray’s body

weight.

External attachment was selected in this study because the primary

considerations were short handling times, low trauma due to attachment, short

recovery periods, and unaltered natural behaviors. A small 3-mm hole was

punched in the lower central portion of the first dorsal fin and a small piece of

surgical tubing was inserted through the hole. A section of 50-pound

monofilament was threaded through the tubing down both sides of the dorsal fin

and attached to the transmitter through two perpendicular holes through the

transmitter end-cap. The attachment resulted in the transmitter trailing just

behind the dorsal fin, slightly nose-down, while the subject was swimming. This

method required relatively short handling time (approximately two minutes) and

relatively little stress. It was assumed that subjects would resume normal

behavior within one or two hours after release. There should have been little

inhibitory drag caused by these small-diameter transmitters with such an


147

attachment and fouling was minimal over the short duration of the study.

Tracking commenced immediately following release. Due to the physically

dynamic environment in this study and the relatively small size of the subjects, it

was assumed that transmitters would not be recoverable though this attachment

method allowed transmitters to be readily identified and returned should the

subject be recaptured by VIMS, recreational anglers, or commercial fishermen.

Tracking Methodology and Experimental Design

Tracking was conducted primarily from a 25-foot Wellcraft Nova V-bottom

research vessel. Other vessels also were used. Sharks were caught by hook

and line. The first healthy C. plumbeus that was less than 65 centimeters in pre-

caudal length was immediately equipped with a transmitter and released.

Manual tracking began upon release to ensure contact was maintained. Fixes

were recorded at ten-minute intervals. Location of the tracking vessel was

determined using a Northstar differential global positioning system (GPS).

Water depth was recorded using a boat-mounted depth sounder. Position of the

shark was estimated from a compass heading and estimate of the distance to

transmitter based on signal strength. A series of range tests were conducted

prior to each tracking cruise to enhance the crew’s experience and accuracy of

estimation of distance from signal strength. Ideally, sharks were tracked from a

distance of 100 to 200 meters. Pulse interval also was measured during each

location fix and decoded into swimming depth. Salinity, dissolved oxygen, and


148

temperature data for the entire water column were recorded in two-meter

intervals with a Hydrolab® Reporter 2 Multiprobe and Surveyor 3 Data Logger

once per hour of tracking. Other environmental information such as tidal cycle,

cloud cover, wind speed and direction, and sea state also were recorded at one-

hour intervals. Sharks were tracked as long as conditions allowed for a

maximum of 50 cumulative hours.

Data Analysis

I Activity Space

Latitude and longitude data were converted to decimal degrees format

using Microsoft Excel 97. The tracks were mapped spatially using ArcView GIS

3.1. These data were used to test the null hypothesis that juvenile C. plumbeus

in Chesapeake Bay are nomadic, showing no site fidelity or home range.

Numerous statistical methods exist for the analysis of activity space or

home-range data. The concepts of “home range" and “activity space" have been

abused in the telemetry literature (White and Garrott 1990). This appears

particularly true in studies utilizing underwater telemetry. Many models exist for

estimating home range and activity space and estimates of core areas of activity

may vary significantly for the same data set depending on the method of analysis

(Wray et al. 1992b). Many of the estimators also make assumptions that are

rarely met in practice. For example, Kemal (Worten 1989) and Jennrich-Tumer

(Jennrich and Turner 1969) estimators are dependent on a bivariate normal


149

distribution of locations. Implicit in the parametric models is the assumption of

independence of observations (Cresswell and Smith 1992). Temporally

autocorrelated data are often unavoidable in underwater telemetry studies due to

logistical constraints on tracking. This is true of the data collected in this study.

The degree of autocorrelation has profound effects on activity-space estimates

(Cresswell and Smith 1992). Time to independence of location fixes can be

calculated using several methods such as those of Swihart and Slade (1985a)

and independence then can be obtained by eliminating fixes until all are

separated by sufficient time. This often leads to a data set that is so reduced and

fragmented that it is no longer ecologically meaningful. Non-parametric home-

range estimators such as the Harmonic Mean Method (Dixon and Chapman

1980) and Dirichlet Tesselation (Wray et al. 1992a) are less affected by

autocorrelated data, yet they are grid-based models and quite cumbersome in

practice. Many of these grid-based non-parametric methods also suffer from

severe calculation problems due to dependence on grid origin and spacing

(White and Garrott 1990). The Minimum Convex Polygon (Mohr 1943, Mohr

1947, Winter 1977) is the oldest, most widely used method of activity-space

estimation due to its simplicity and flexibility (White and Garrott 1990). It is

created by connecting all outer location fixes to create a single convex polygon.

The area within the polygon is calculated as an estimate of activity space. This

method is non-statistical and is not affected by temporally autocorrelated data

(Swihart and Slade 1985b), yet it has a tendency to over-estimate activity space


150

due to the assumption of concavity (no allowance for convexity) and it has no

associated confidence interval (White and Garrott 1990).

Despite their shortcomings, Minimum Convex Polygon (MCP) and Kernal

activity spaces were calculated for all tracks in this study to enable comparison

with other published studies. These activity spaces were calculated and spatially

modeled using the Animal Movement Analysis Extension for Arc-View GIS 3.1

(Hooge and Eichenlaub 1997). Site attachment was evaluated for each track

using the Site Fidelity Test in the ArcView Animal Movements Extension.

Swimming speed and swimming rates also were calculated for the six

tracks of longest duration. Cumulative distances between fix locations were

calculated using the ArcView GIS Spatial Analyst Extension. These cumulative

distances were divided by time to give a swimming speed, which was converted

to kilometers per hour. This speed was converted to a swimming rate based on

the total length of the shark and was given in body lengths per second.

II Depth Telemetry

Swimming-depth data collected during the tracks were used to test the null

hypotheses that swimming-depth patterns are temporally and spatially random

and that these juvenile sharks utilize the entire water column. Only tracks of 24

hours or more in duration were used in the depth-telemetry analysis. The

pressure sensor on transmitter 5285 failed, therefore this fifty-hour track was also

eliminated from this analysis. Swimming depth and bottom depth were plotted


151

over cumulative time for each of the six remaining tracks to investigate temporal

trends in these two factors as well as the relationship between the two factors.

Initial plots using raw water-depth data, which was measured at the tracking

vessel, revealed a problem inherent with working in an estuary that is so dynamic

in three-dimensional space. The shark’s swimming-depth estimate was deeper

than the water depth in as many as 10% of all fixes. This is obviously

impossible; the depth of the estuary in the areas tracked is highly variable and

the distance between the tracking vessel and the shark was enough to cause the

discrepancy. Fixes were, therefore, corrected to estimate the location of the

shark rather than that of the tracking vessel. This was accomplished by

converting all latitude and longitude values to universal transverse mercator

(UTM) coordinates. Then these coordinates were corrected using the bearings

(converted to radians) and distance to shark estimates recorded during tracking

using the following equations:

UTMxb = UTMxa + (Distance x (SIN(bearing)))

UTMyb = UTMya + (Distance x (COS(bearing)))

The new UTM coordinates were plotted and water depths for these locations

were estimated by merging these data with a bathymetry grid interpolated from

TIN (triangular irregular network) data from the EPA, Chesapeake Bay Program.

This assigned a water-depth value to each new fix location. This correction

reduced the number of nonsensical data points to less than one percent of all

fixes. The mean difference between raw and corrected water-depth estimates


152

ranged from 1.1 to 4.0 meters. These differences were greatest for those tracks

made during the first summer of tracking, 1996 (range 3.1 - 4.0 m), and

decreased dramatically in 1997 and 1998 (range 1.1 -1 .7 m). This can probably

be attributed to a learning curve for tracking efficiency and an increase in return-

volunteer experience.

Mean swimming depth was calculated for each shark relative the

following potentially influential variables: day/night as defined by sunrise and

sunset, day/night as defined by nautical twilight, moon position (risen/set), and

tidal current (flood/slack/ebb). The difference between mean depth for each

factor was evaluated using a t-test (two-tailed, a=0.05) for the day/night and

moon phase variables and a one-way ANOVA (a=0.05) to examine the influence

of tidal-current phase. These tests were conducted for each track of 24 hours or

more, then on the overall means of all of these tracks combined. In an effort to

visualize the results, histograms of the depth distribution were plotted as a

function of significant factors. The difference between swimming depth and

corresponding water depth also was calculated for all fixes. The distribution of

these differences was used as a preliminary investigation of the degree of

association between the sharks and the bottom, and the proportion of the water

column actually utilized.


153

III Correlation with Tidal Current

Movement direction and direction of tidal current were used to test the null

hypothesis that the directional movements of juvenile C. plumbeus are random

and are not influenced by or correlated with tidal currents. Data manipulation

was required to estimate the mean direction of travel for a shark at a given time.

Five-point moving averages of fix locations were calculated for the six longest

tracks to filter out the effects of tidal and wind-driven currents on the tracking

vessel between fixes,. These averaged tracks were converted to UTM

coordinates and then spatially plotted in ArcView GIS 3.1. The Animal

Movements Extension for ArcView (Hooge and Eichenlaub 1998) has a function

that automatically calculates successive distance between geographic points.

The distance between successive fixes was calculated for each track. Given the

distance (D) between location A and location B and the UTM coordinates of the

two locations, the bearing {0) required to reach B from A could be calculated

using the following equation:

0= ArcSIN ((UTMxa - UTM xJ/D) X (180P/J]) (Eq. 3.1)

Current was not measure in situ, therefore tidal-current bearing and

magnitude was estimated using data downloaded from the software program

Tides and Currents 2.0 (Nautical Software Inc.). This program has tidal current

calculations for more than 150 location in Chesapeake Bay for any date needed.


154

Current data were downloaded for each fix using the closest station and depth

possible. The current regime of the Bay is very complex and dynamic due to

many hydrographic features such as riverine input from the north and west,

oceanic tidal flux from the southeast, and the two-layer estuarine circulation that

develops as a function of the varying salinity between these sources. Due to

these factors, the accuracy of the current data from this software program may

be questioned. Nevertheless, this hypothetical inaccuracy only made the test of

the null hypothesis more conservative, and therefore was accepted.

Correlation of movements with tidal currents was evaluated using several

indices. The estimated tidal current bearing was subtracted from the shark

movement bearing for each fix. Negative values were corrected by adding

360 degrees to them. These values were converted from degrees to radians. If

the null hypothesis was true, then the data would be highly dispersed. If the null

hypothesis was incorrect suggesting that shark movements were correlated with

tidal directions, then the resulting data set would have a unimodal distribution

about zero degrees. This was tested in several ways. An angular index of

concentration (r) was calculated for each track using the following equations,

where is the corrected difference between shark and current bearings at each

of n fixes (Zar 1996):

n£ cos3

* = ^ -------- (Eq. 3.2)n


155

I sin3Y = -i=LY = (Eq. 3.3)

n

r = V F 7 7 r (Eq. 3.4)

This index of concentration has no units and ranges from zero for data that is

very dispersed to one for data that are all concentrated in exactly the same

direction as the tidal currents. The angular deviation (Zar 1996) was then

calculated as an index of dispersion (s) using the following equation:

This index is analogous to standard deviation in linear analyses and the units are

in degrees.

Two simple tests were utilized to examine the degree of dispersal in the

data statistically. Both tests examine the null hypothesis that the data are

distributed randomly or uniformly around a circle. Rayleigh’s test for circular

uniformity was applied initially. The statistic (z) was calculated from the index of

concentration (r, Eq. 3.3) and the sample size (n) using the following equation:

If z was significant (a=0.05), the null hypothesis was rejected, indicating the

existence of a mean direction of travel. This test, however, gives no indication

that the mean direction is distributed around zero degrees and correlated with

tidal currents. The V-test (Zar 1996), a modification of Rayleigh’s z, was used to

(Eq. 3.5)

Rn

R = n r z = n r '

(Eq. 3.5)


156

examine this potential correlation. This test also examines the alternative

hypothesis that the data are not distributed randomly or uniformly about a circle,

but more specifically states that the distribution has a specific mean angle

specified by the researcher. If the sharks’ movements were correlated with tidal

currents, the data would be distributed about a mean of zero degrees as stated

above. The mean angle ( a ) was first calculated from the values of X(Eq. 3.2), Y

(Eq. 3.3), and r(Eq. 3.4) using the following relationships:

- ysina = — (Eq. 3.6)

r

cosa = — (Eq. 3.7)r

The statistic V then was calculated using the following equation wherein // is the

predicted mean angle:

V = R cos (a - / j ) (Eq, 3.8)

The significance of V is evaluated using critical values of u wherein:

u = V (Eq. 3.9)

Rejection of the null hypothesis suggests the data are indeed distributed with a

mean angle statistically similar to the predicted angle.


157

RESULTS

Tracking of sandbar sharks began in June of 1997. Four sharks were

tracked during the first summer. Five additional sharks were tracked in the

summer of 1998, and one shark was tracked in 1999. In all, ten juvenile C.

plumbeus, six males and four females, were tracked for 10 to 50 consecutive

hours (Table 3-1). One shark was relocated two weeks after the initial track and

monitored for 15 additional hours, giving a total of 64 hours for that animal. The

size of the sharks, measured as pre-caudal length (PCL), ranged from 44 to 62

centimeters. The target duration was 50 consecutive hours for each track. This

goal was achieved for only four of the subjects, and two others were tracked for

more than 40 hours. Weather and sea conditions are extremely unpredictable in

Chesapeake Bay, especially for small vessels at night. This was the usual cause

for early termination tracks.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-1: Juvenile Carcharhinus plumbeus tracking summary.Regions: SECB - Southeastern Chesapeake Bay; SWCB - Southwestern Chesapeake Bay; SCCB - South Central Chesapeake Bay; VAES - Atlantic Side of Virginia’s Eastern Shore

Transmitter (# / kHz)

Date Sex PCL(cm)

Duration(hours)

Vessel Reason for Termination StartRegion

9633 / 37.0 June 20-21, 1997 9 56 13 RA/ Kingfisher Contact Lost / Weather SECB

9630 / 38.4 July 8-9, 1997 9 44 25 R/V Shearwater Weather SECB

9632 / 38.4 August 6-8, 1997 d 56 4 9 + 1 5 RA/ Shearwater Engine Failure SECB

9631 / 37.0 August 26-28, 1997 9 49 50 R/V Shearwater Goa! Reached SECB

9634 / 38.0 June 9-10, 1998 d 47 10 R/V Shearwater Weather / Water too Shallow VAES

9635 / 38.0 July 7-9. 1998 d 44 43 R/V Shearwater Weather SECB

9637 / 39.0 July 13-15, 1998 d 45 50 R/V Shearwater Goal Reached SECB

9639 / 40.0 August 12-14, 1998 d 52 44 R/V Shearwater Weather / Vessel Interference SWCB

5375 / 38.0 September 2, 1998 d 50 11 R/V Shearwater Weather SCCB

5285 / 38.0 July 26-28, 1999 9 62 50 R/V Tern Goal Reached SECB

CJl00

159

I Activity Space

Maps 3-1 to 3-7 spatially display the movement patterns for each of the

ten tracks. Each track is outlined by its minimum convex polygon (MCP)

representing the activity space for that individual. All sharks tracked were highly

active, covering large portions of the lower Bay, yet none of them exited the Bay

at any time during a track. Estimated MCP and Kernal activity spaces for all ten

sharks are shown in Table 3-2. MCP activity spaces for all ten sharks are shown

in Map 3-8. MCP estimates ranged from 4.20 to 275.84 km2. Only sharks

tracked more than 24 hours (n=7) were used in calculating mean activity spaces.

The mean MCP activity space for these sharks was 110.26 km2 (S.D. = 77.60

km2). A single example of Kernal activity spaces is shown in Map 3-9. The 50%,

75%, and 95% kernals are shown for shark 9632. Mean kernal activity spaces

were calculated using estimates from the seven longest tracks. The mean 50-

percent kernal activity space (core area) was 18.76 km2 (range 7.95 - 64.01 km2).

The mean 75-percent kernal was 75.53 km2 (range 25.79 - 290.63 km2) and the

mean 95-precent kernal was 140.10 km2 (range 62.28 - 382.66 km2). Site-fidelity

tests, conducted using the Animal Movement Analysis ArcView Extension

(Hooge and Eichenlaub 1998), were insignificant. This suggests these animals

were highly nomadic while in the summer nursery and did not establish home

ranges or, alternatively, home ranges were established but they were very large

and the data were insufficient (duration too brief) to elucidate them. In either

case, the estimates presented here underestimate the true activity space. There


160

was no correlation between shark size and activity space in this study.

Mean swim speed (km/hr) and swim rate (body lengths/sec) were

calculated for the six sharks tracked for the longest duration. The mean swim

speed was 1.44 km/hr (S.D., 0.24). This is comparable to the estimate of 1.21

km/hr reported by Medved and Marshall (1983), but more than twice the estimate

of 0.66 km/hr reported by Huish and Benedict (1977). The mean swimming rate

was estimated to 0.594 (S.D., 0.161) body lengths per second. Gruber et al.

(1988) demonstrated, that such straight-line estimations of swimming speed or

rate may underestimate the true value by as much as a factor of two.


161

Map 3-1: Movements of shark 9631 tracked for 50 consecutive hours (August 26-28,1997). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space.



75"48' 3r15' 3r10' 37"'5'


162

Map 3-2: Movements of shark 9632 tracked for 49 consecutive hours (August 6-8,1997). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The shark was tracked for 15 additional hours three weeks later (August 27-28). Day and night courses of this second track are shown as light and dark lines respectively. The track is outlined by the Minimum Convex Polygon of activity space.


78*46' 3ri5 ' 37"10'

f t

w 0110 4>< ® * i XS 2S 3 v X O .35 ° jo a o i £ 2 z s u at

«

/+

f t

.0ZolZ ,9 UiC .0U£


163

Map 3-3: Movements of shark 9635 tracked for 43 consecutive hours (July 7-9,1998). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space.


75”46' 37°15' 37”10l

.SU IE .OUiC


164



75*46’ 37*15' 37*^0*

n •

S ^ H" “

s s * s &f i s ° 2 s «2 S n « s► - ^ o • L j

IVV

9

f

/4 -

tr o -r co r-j cr oC-J ^ r j r-j r j o r > ^> • " r j iri o t co r j iri o

o t co — cj r j cj o r> *t

.OZoIC .suie .Ok, iz


165

Map 3-5: Movements of shark 9639 tracked for 44 consecutive hours (August 12-14,1998). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space.


37°15* 3 n o '

n ®« s «

“ ^ 2 a. > o -c a 2 s s23VV e i - . n *

2*»<fc

+suie Ok£C ft>Z£


166



75*46' 37*15' 37*10'

,9UiC .ouze


167

Map 3-7: Movements of sharks tracked for less than 40 consecutive hours. Shark 9633 tracked for 13 hours (June 20-21,1997); Shark 9630 tracked for 25 hours (July 8-9,1997); Shark 9634 tracked for 10 hours (June 9-10, 1998); and Shark 5375 tracked for 11 hours (September 2,1999). Light circles are location fixes recorded during the day and dark circle are location fixes recorded at night. The track is outlined by the Minimum Convex Polygon of activity space.


7 6 *2 5 ’ 76 *20 ’ 76 *15 ’

A Track 9633 1 (13 hours)^ t> Day Fixes

» Night Fixes I 1MCP 9633

Track 9630 (25 hours)

b DayFixes ■ Night Fixes

I 1 MCP 9630


a DayFixes a Night Fixes

I 1 MCP 9634


o DayFixes • Night Fixes

MCP 5375I I

*

.6»4C

.OV

oie

,Sl.2

£

168

Map 3-8: Minimum Convex Polygons for all ten juvenile Carcharhinus plumbeus tracked. The area calculations associated with polygons can be seen in Table 3-2.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

76°10'

Mnknum Convex Polygons

76°1076°20 76°15

76°00' 75°66‘ 75 50* 75 45 “

Track 9630

Track 9633 Track 9632 Track 9631 Track 9634

Track 9639 Track 9635 Track 9637 Track 5375 Track 5285

13 hours

25 hours49 hours50 hours10 hours

44 hours 43 hours 50 hours11 hours 50 hours

76°5‘I " T"...

76°00''" T " "

75°50'i

75°45‘

V&Ie

.ou

>z

e,9

^e

.v

Llz

169

Map 3-9: Kernai Activity Space of Shark 9632. 50-percent, 75-percent, and 95-percent kernals are shown. The area calculations associated with polygons can be seen in Table 3-2.


76°15' 76°10‘

76°16' 76°10‘ 76°6' 76°00'

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3*2: Minimum Convex Polygon (MCP) and Kernal activity space, mean swimming speed, and mean swimming rate for all ten juvenile C. plumbeus tracked. Grand means only include tracks of greater than 24 hours in duration. Italicized activity-space estimates were not included in calculation of mean activity spaces. (S.D. = standard deviation)

Shark # Duration(hours)

MCP(km2)

50% Kernal (km2)

75% Kernal (km2)

95% Kernal (km2)

Mean Speed km / hr

Swimming Rate body lengths / sec

9634 10 4.20 1.58 3.66 7.23 not calculated not calculated




9635 43 96.38 10.12 47.76 116.10 1.39 0.553

9639 44 97.94 16.60 47.64 102.03 1.53 0.719

9631 50 65.96 10.39 25.79 78.35 1.48 0.535

9637 50 275.84 64.01 290.63 382.66 1.71 0.779

5285 50 39.59 12.36 33.77 62.28 1.00 0.327

9632 4 9 + 1 5 121.64 7.95 52.21 135.74 1.55 0.652

MEAN(tracks>24hrs)

46.57 110.26 18.86 75.53 140.10 1.44 0.594

S.D. 11.71 77.60 7.59 95.37 109.61 0.24 0.161

~n IO

171

II Depth Telemetry

Only tracks with at least 24 cumulative contact hours (n=7) were used in

this analysis. The pressure sensor failed on one of these transmitters (#5285)

leaving six depth tracks for the analysis. Swimming depth and “corrected”

bottom depth are plotted versus cumulative track time in Figures 3-1 to 3-6. Five

of the six sharks tracked utilized depth ranges from less than five meters to more

than 30 meters. The depth plots for these five sharks suggested that these deep

waters were used primarily during daylight hours whereas the shallowest portions

of the tracks were used during the night. Shark 9631 demonstrated this pattern

most conclusively (Fig. 3-2). This pattern was also obvious, though not as clean,

for sharks 9530, 9632, 9635, and 9637. Shark 9639 showed no obvious diel

pattern in depth use. The bottom depth during this track only varied between

seven and fifteen meters (Fig. 3-6).

Swimming-depth data were separated into day and night fixes based on

the time of sunset and sunrise for that day. These data sets were binned

separately into three-meter depth intervals and plotted on vertical histograms

arranged such that the proportion of daytime fixes within a given depth stratum

was plotted on a vertical axis opposite the proportion of nighttime fixes for that

same stratum (Fig. 3-7a to 3-12a). This species generally has been thought of

as highly demersal. The difference between the swimming depth and the

corrected bottom depth was calculated to investigate this claim. These data also

were separated into day and night fixes plotted on histograms similar to the raw


172

swimming depth data to investigate potential diel patterns in deviation from the

bottom of the estuary (Fig. 3-7b to 3-12b).

Three general patterns emerged from the histograms. Swimming depth

was obviously deeper during the day for sharks 9630, 9631, and 9632. The

majority of depth fixes recorded during the night were less than 12 meters deep,

whereas the majority of fixes taken during the day were greater than 12 meters

deep. This pattern was most obvious for shark 9630, which spent greater than

50% of night fixes in less than 9 meters of water whereas more than 50% of day

fixes were greater than 21 meters deep (Fig. 3-7a). Sharks 9631 and 9632 had

very few depth fixes below 21 meters during the night whereas many daytime

fixes were greater than 30 meters (Fig. 3-8a and 3-9a). Sharks 9635 and 9637

also utilized deeper water during the day but the pattern was somewhat different.

The shallower mean nighttime depth was due to the absence of occasional deep

excursions that were made during the day. The majority of day and night fixes

were less than 9 meters deep, yet there were no night fixes deeper than 15

meters whereas day fixes extended to more than 21 meters for shark 9635 (Fig.

3-10a) and more than 30 meters for shark 9637 (Fig. 3-11a). The histogram for

shark 9639 was nearly symmetrical (Fig. 3-12a) as all day and night fixes for this

shark were less than 15 meters deep.

The histograms plotting the difference between swimming depth and

bottom depth (Fig. 3-7b to 3-12b) were nearly symmetrical in all cases indicating

that day and night did not influence the distance from the bottom the sharks


173

swam. The overall pattern of these histograms does suggest that the demersal

characterization of juvenile C. plumbeus may be incorrect. In all six cases, more

than 50% of the swimming-depth fixes were more than three meters from the

bottom (range 51.8 - 80.6%). A mean of 37.6% of fixes were more than six

meters from the bottom (range 21.6 - 53.9%) and a mean of 19.1% of fixes were

more than nine meters from bottom (range 6.3 - 33.6%). If C. plumbeus were

truly demersal, the majority of fixes would be less than three meters from the

bottom. These data also indicate that these juvenile sharks make occasional

vertical movements to the surface even in waters deeper than 30 meters.

Anecdotal evidence supports these findings, as free-swimming juvenile C.

plumbeus were observed swimming on the surface on many occasions during

the tracking periods. These observations occurred in water depths greater than

40 meters on three separate occasions.

The statistical significance of the diel pattern in swimming depth was

evaluated using a series of t-tests. The swimming-depth data departed

significantly from the normal distribution for all individual shark tracks. The t-test

was utilized despite the violation of this assumption, as it is robust to such

departures for large sample sizes (Brown and Rothery 1993). T-tests also were

used to investigate significant differences in the overall mean (averaged

individual means) swimming depth for each factor. These data did meet the

assumption of normality.

Two-sample t-tests (a=0.05), assuming unequal variances, initially were


174

conducted to compare mean swimming depth during night and day, based on

sunrise and sunset using all fixes. Each shark was analyzed individually. Mean

daytime swimming depths ranged from 7.55 to 18.08 meters and mean nighttime

swimming depths ranged from 5.60 to 12.23 meters. Swimming depths were

significantly deeper during the day for five of the six sharks analyzed (Table 3-3).

Swimming depth was not significantly different between day and night for shark

9639 only. Mean swimming depths for all six sharks during day and night are

shown in Figure 3-13. The overall mean (averaging individual means) daytime

swimming depth was 12.80 (S.D. 5.01) meters and the grand mean swimming

depth during nighttime was 8.46 (S.D. 2.30) meters. To test the significance of

this difference, a paired t-test was conducted using only the day and night means

for each shark. This overall t-test was significant (p=0.032) indicating that

daytime swimming depths were significantly deeper than nighttime swimming

depths.

Swimming-depth differences also were analyzed using alternative

definitions of day and night. Rather than defining day as the period between

sunrise and sunset and night as the period between sunset and sunrise, they

were defined based on nautical twilight. According to this definition, day begins

when the sun rises to 12° below the horizon and ends when the sun sets to 12°

below the horizon. Day and night are commonly equated with light and dark, yet

day and night defined by sunrise and sunset relegate most of the crepuscular

periods, which contain significant light, to part of night. Nautical twilight relegates


175

this period to day, and is therefore a better measure of light and dark. If the diel

pattern in depth utilization observed above was a function of light level, then the

difference should be even more significant based on this definition. The mean

depths for each shark based on nautical twilight are shown in Figure 3-14. The

results of the t-tests using this definition are shown in Table 3-4. Whereas the

mean difference between day and night swimming-depth was greater using

nautical twilight, 4.99 meters versus 4.34 meters using sunrise/sunset, the results

of the t-tests were nearly identical to those obtained using definitions based on

sunrise and sunset. Daytime swimming depth was significantly deeper for all

sharks except 9639, and the overall t-test was also significant (p=0.036).

Lunar position, based on moonrise and moonset, also was evaluated as a

potential influential variable on swimming depth. For lack of a better term, in this

discussion the period between moonrise and moonset will be called lunar high

and the period between moonset and moonrise will be referred to as lunar low.

The mean depths for each track during lunar low (set) and lunar high (risen) are

shown in Figure 3-15. Swimming depth was significantly deeper during lunar

high for four of the six sharks analyzed (Table 3-5). The overall mean swimming

depth was 12.47 (S.D. 3.86) meters during lunar high and 9.73 (S.D. 3.95)

meters during lunar low. This difference was significant (p=0.042) indicating that

lunar high swimming depths were significantly deeper than those during lunar low

were.

Tidal currents, estimated from the software program Tides and Currents


176

2.0 (Nautical Software Inc.), were used to investigate the potential influence of

tidal phase on shark swimming depth. Tidal currents were divided into three

categories, Ebb, Slack, and Flood. Slack was defined as any period when the

current was less than 0.4 knots. This value was chosen primarily to achieve near

equal sample sizes. A one-way Analysis of Variance was conducted using

swimming depth as the dependent variable and current phase as an independent

variable with three levels. The results, shown in Table 3-6, were significant

(p<0.01) for four of the six sharks but no true pattern existed between the sharks.

Swimming depth was deepest during Ebb tide for two of the sharks, during Slack

tide for one shark, and during Flood tide for one shark. An overall ANOVA was

conducted on the means for these three phases, and the results were

insignificant (p=0.815), suggesting that tidal-current phase had little influence on

swimming depth.


177

Figure 3-1: Swimming depth and bottom depth recordings for Shark 9630 tracked for 25 consecutive hours. (Bottom depths were interpolated from a bathymetry grid using corrected fix locations in ArcView GIS - see text for correction methodology.)


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Shark #9630 Cumulative Time (hours)

0:00:00 12:00:00 24:00:00 36:00:00 48:00:00

0

5

10

1 15 £| 20

25

30

3 5 d a y n ig h t d a y n ig h t d a y

-^S w im m ing Depth — Bottom Depth (corrected)

178



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


0:00:00 12:00:00 24:00:00 36:00:00 48:00:00

0

5

10

15

£ 20

& 25 O

30

35

40

45

-•-Swimming Depth — Bottom Depth (corrected)

NIGHTNIGHT

179

Figure 3*3: Swimming depth and bottom depth recordings for Shark 9632 tracked for 49 consecutive hours. (Bottom depths were interpolated from a bathymetry grid using corrected fix locations in ArcView GIS • see text for correction methodology.)


Shark #9632

00:00

Cumulative Time (hours)

12:00:00 24:00:00 36:00:00 48:00:00

j

y M jk ’J& fi

DAY NIGHT DAY NIGHT DAY

Swimming Depth — Bottom Depth (corrected)

180



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

12:00:00


24:00:00 36:00:00 48:00:000:00:00

DAY NIGHT DAY NIGHT


DAY

181



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

12:00:00


24:00:00 36:00:00 48:00:000:00:00

D 25

DAY NIGHT DAY NIGHT


DAY

182



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

0:00:00 12:00:00


24:00:00 36:00:00 48:00:00

O 12

DAY NIGHT DAY NIGHT


DAY

183

Figure 3-7: Shark 9630 swimming-depth dynamics, a) Comparison of the distributions of swimming depth during day and night, b) Distribution of the distance from swimming depth to bottom of estuary during day and night.


a) 50

0-3

3-6

_ 6-912Q>| 9-12

£ 12-15 &&CT)C

15-18

I 18-21

I 21-24

24-27

27-30

>30

Swimming Depth Distribution - Shark #9630% of Day Fixes % of Night Fixes

30 20 10 0 10 20 30 40 50

□ Day ■ Night

b)

Day Fixes: N=83 Night Fixes: N=35

Distance from Swimming Depth to Bottom - Shark #9630

27-30 □ Day ■ Night

- 24-27

21-24

« 18-21

5 15-18

Q 12-15

« 9-12

30 20 10 0% o f Day Fixes

10 20 30 40% of Night Fixes


184



a)Swimming Depth Distribution - Shark #9631

50 40 30% o f Day Fixes

20 10

0-3

3-6

— 6-9(0w4>"3 9-12E£ 12-15CL0™ 15-18O)c1 18-21

SCO 21-24

24-27

27-30

>30

% of Night Fixes10 20 30 40 50

□ Day ■ Night

b)>30

27-30

24-27

| 21-24I

18-21

| 15-18

12-15

9-12

6-9

3-6

0-3



D"□ Day ■ Night

60 50 40 30 20 10% of Day Fixes

10 20 30% of Night Fixes

40 50 60


185




% o f D ay F ixes 50 40 30 20 10

0-3

3-6

— 6-9(Ak.0)% 9 -1 2E5 12 -15Q.0)° 15 -18mc| 18-21

ito 2 1 -2 4

2 4 -2 7

2 7 -3 0

> 30

% o f N ig h t F ixes 10 20 30

n

50

□ D ay ■ N ig ht

b)>30

2 7 -3 0

2 4 -2 7"3TI 2 1 -2 4

8 18‘21 I 15 -18

12 -15

9 -1 2

6 -9

3-6

0 -3



□ D ay ■ N ight

Da

60 50 40 30 20 10

% of Day Fixes

10 20 30

% of Night Fixes

40 50 60


186



Swimming Depth Distribution - Shark #96353 ) % of Day Fixes % of Night Fixes

' 50 40 30 20 W _____ 0 10 20 30

0-3

3-6

_ 6-9#r . . .1 9-12E,£ 12-15 a a>™ 15-18O)cI 18-215OT 21-24

24-27

27-30

>30

50

□

□ Day ■ Night

b)>30

27-30

24-27«T$ 21-24

S18-21

15-18


■e 12-15

9-12

6-9

3-6

0-3

Distance from Swimming Depth to Bottom - Shark #9635.

□ Day ■ Night

D□

r w

60 50 30 20 10% of Day Fixes


50 60


187




50

0-3

3-6

6-9

0)3 9-12££ 12-15 a.«° 15-18O)cI 18-21 1M 21-24

24-27

27-30

>30

b)>30

27-30

24-27

I 21-24

18-21

15-18

12-15

9-12

6-9

3-6

0-3

40% of Day Fixes


50

□□

Q

n

□□

□ Day ■ Night


Distance from Swtnming Depth to Bottom - Shark #9637

□ Day ■ Night

0

□tm

60 50 40 30 20 10% of Day Fixes


40 50 60


188



Swimming Depth Distribution - Shark #9639

a) 50

0-3

3-6

_ 6-9m4>■S 9-12

£ 12-15 a<u» 15-18 c| 18-21

ico 21-24

24-27

27-30

>30

40% of Day Fixes

30 20 10 0% of Night Fixes 10 20 30 40 50

r " i

□ Day ■ Night

b)>30

27-30

24-27



a Day ■ Night

* 21-24

18-21

a 15-18

12-15

30 20 10% of Day Fixes

10 20 30% o f Night Fixes

60


189

Figure 3-13: Mean swimming depth during Day and Night based on sunrise and sunset for sharks 9630, 9631, 9632, 9635, 9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests. Overall mean daytime swimming depth was significantly deeper than nighttime swimming depth (t-test, T = 2.95, p=0.032). Error bars are standard error of the mean.


(iu) iftdap ueeiu


□ Su

n Ri

sen

■ Su

n S

et

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-3: Results of t-tests for difference between mean swimming depth (meters) during day and night based on sunrise and sunset. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate.(* indicates T statistic is significant at a= 0.05.)

Shark # Sun Risen N Mean (S.D.)

Sun Set N Mean (S.D.)

T statisticAi * M i

P (sign.)

9630 82 16.88 (7.18) m 35 9.72 (6.29) m 5.40* <0.00019631 156 18.08 (8.26) m 121 12.23 (4.86) m 7.37* <0.00019632 152 17.09 (9.50) m 105 8.10 (6.66) m 8.92* <0.00019635 148 8.57 (6.85) m 99 7.00 (2.48) m 2.55* 0.0129637 172 8.65 (8.12) m 103 5.60 (2.81 )m 4.51* <0.00019639 118 7.55 (2.62) m 104 8.11 (2.88) m -1.29 0.13

ALL means 6 12.80 (5.01) m 6 8.46 (2.30) m 2.95* 0.032

COo

191

Figure 3-14: Mean swimming depth during Light and Dark based on timing of nautical twilight for sharks 9630, 9631, 9632, 9635, 9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests. Overall mean swimming depth during light was significantly deeper than

mean swimming depth during dark (t-test, T = 2.85, p=0.036). Error bars are standard error of the mean.


Mean

De

pth

with

Naut

ical

C N ^ - C O C O O C N ^ - C O O O O T" r" r* r" r* M

(ui) iftdap ueeui


□ Li

ght

■ Da

rk

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-4: Results of t-tests for difference between mean swimming depth (meters) during day and night based on time of nautical twilight. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate.(* indicates T statistic is significant at a= 0.05.)

Shark # Twilight Day N Mean (S.D.)

Twilight Night N Mean (S.D.)

T statisticM \ * / * 2

P (sign.)

9630 89 16.93 (7.08) m 28 7.74 (4.58) m 8.03* <0.00019631 180 17.84 (7.89) m 97 11.23 (4.39) m 8.97* <0.00019632 171 16.73 (9.38) m 86 6.84 (5.67) m 10.49* <0.00019635 168 8.31 (6.53) m 79 7.16 (2.39) m 2.03* 0.0449637 198 8.55 (7.65) m 77 4.83 (2.26) m 6.18* <0.00019639 136 7.57 (2.60) m 86 8.18 (2.95) m -1.57 0.12

ALL means 6 12.66 (4.97) m 6 7.66 (2.10)m 2.85* 0.036

CDro

193

Figure 3-15: Mean swimming depth during Lunar High (moon risen) and Lunar Low (moon set) for sharks 9630, 9631, 9632, 9635,9637, and 9639. An asterisk (*) indicates a significant difference according to individual t-tests. Overall mean swimming depth was significantly deeper during lunar high than during lunar low (t-test, T = 2.71, p=0.042). Error bars are standard error of the mean.


Mea

n De

pth

with

M

oon

Rise

n vs

Set

(ui) iftdap ueaui


! □ Ri

sen

■ S

et

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-5: Results of t-tests for difference between mean swimming depth (meters) while moon was risen versus while moon was set. Means are given with standard deviation in parentheses. Results of t-tests using raw depth data for each track are followed by t-test results using overall mean for each track as a replicate.(* indicates T statistic is significant at a= 0.05.)

Shark # Lunar High (risen) N Mean (S.D.)

Lunar Low (set)N Mean (S.D.)

T statisticA ^

P (sign.)

9630 60 14.73 (7.51) m 57 14.74 (7.85) m -0.01 0.999631 163 17.19 (7.80) m 114 13.15 (6.51) m 4.67* <0.00019632 120 15.83 (8.74) m 137 11.31 (9.73) m 3.92* <0.00019635 138 9.17 (6.73) m 109 6.39 (2.39) m 4.33* <0.00019637 137 10.29 (8.36) m 138 4.75 (2.81) m 7.37* <0.00019639 115 7.58 (2.54) m 107 8.05 (2.95) m -1.26 0.21

ALL means 6 12.47 (3.96) m 6 9.73 (3.95) m 2.71* 0.042

CD

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-6: Results of one-way Analysis of Variance for difference between mean swimming depth (meters) during ebb, slack, and flood tidal current phases. Means are given with standard deviation in parentheses. Results of analyses using raw depth data for each track are followed by ANOVA results using overall mean for each track as a replicate. (* indicates T statistic is significant at a= 0.05.)

Shark # EBB Current N Mean (S.D.)

SLACK Current N Mean (S.D.)

FLOOD Current N Mean (S.D.)

F statistic T statisticM x * M i

9630 34 14.55 (9.49) m 39 16.76 (5.84) m 44 13.08 (7.20) m 2.47 0.0899631 80 17.76 (8.64) m 97 14.42 (7.08) m 10 14.81 (6.71) m 5.14* 0.0069632 76 16.91 (9.88) m 94 13.55 (9.74) m 87 10.23 (7.87) m 10.73* <0.0019635 77 7.74 (4.53) m 81 10.14 (6.63) m 89 6.12 (4.62) m 12.11* <0.0019637 106 5.00 (2.99) m 74 8.44 (5.81) m 95 9.59 (9.31 )m 13.48* <0.0019639 65 7.84 (3.27) m 63 7.22 (2.64) m 94 8.19 (2.36) m 2.38 0.095

ALL means 6 11.63 (5.43) m 6 11.76 (3.73) m 6 10.34 (3.18) m 0.21 0.815

COcn

196

III Tidal Current Correlation

Only the six sharks tracked for more than 40 consecutive hours were used

in this analysis. The degree of correlation between swimming direction and tidal-

current direction was first assessed by subtracting the current bearing from the

swimming direction bearing for each fix. Histograms were created for each shark

using 30-degree bins. In the broadest of terms, the proportion of fixes from 0-90°

departure were interpreted as traveling with (not against) the tidal current

whereas those from 90-180° were interpreted as traveling against (not with) the

tidal current. These histograms are shown in Figures 3-16a to 3-21 a. The

proportion of fixes “with” the tidal current ranged from 63.8 %to 85.2% (mean

75.8%). A non-statistical correlation index was developed to aid in visualizing the

relationship between current direction and shark swimming direction. The

bearing calculated for each shark fix was subtracted from the mean current

bearing. The correlation index was assigned to each fix based on this departure

using Table 3-7. The mean correlation index was calculated for each hour of

tracking. This mean correlation index was plotted on a y-axis versus cumulative

track time on the x-axis. Tidal current speed then was plotted on a second y-axis

using positive values for flood currents and negative values for ebb currents.

These plots indicate a high degree of correlation between tidal currents and

swimming direction (Fig. 3-16b to 3-21 b). This was especially true of sharks

9635 and 5285. More than 70% of ail fixes deviated less than 60° from the

current bearing for these two tracks (Fig. 3-18a and 3-21 a).


197

Table 3-7: Calculation of non-statistical correlation index between swimming direction and tidal-current direction.

Swimming direction - Current direction Tidal-Current Phase CorrelationIndex

0-30 or 330-360 FLOOD 3

30-60 or 300-330 FLOOD 2

60-90 or 270-300 FLOOD 1

90-120 or 240-270 FLOOD -1

120-150 or 210-240 FLOOD -2

150-210 FLOOD -3

0-30 or 330-360 EBB -3

30-60 or 300-330 EBB -2

60-90 or 270-300 EBB -1

90-120 or 240-270 EBB 1

120-150 or 210-240 EBB 2

150-210 EBB 3


198

Summary data for the current correlation analyses are compiled in Table

3-8. The Concentration Index (r) ranged from 0.25 to 0.61 (mean = 0.50) and the

mean Angular Dispersion (s) was 56.8° (S.D. 7.1°). These indices were

calculated relative to the tidal-current direction indicating a relatively high degree

of correlation between current and swimming directions. The mean angle of

deviation from the tidal-current direction was 32.7° (range 13°-51°). Rayleigh’s Z

tests were highly significant (p<0.001) for all six sharks indicating that swimming-

direction fixes were not randomly dispersed around a circle, but had a significant

mean directionality. The V-test examines the difference between this mean

direction and that of the tidal current. It tests the same null hypothesis of random

dispersal as Rayleigh’s Z but with the more stringent alternative hypothesis of a

mean directionality that does not differ significantly from that of the tidal current.

These results were also highly significant (p<0.001) for all sharks indicating that

the mean swimming direction did not differ significantly from the mean tidal-

current direction.


199

Figure 3-16: Shark 9631 Directional-Swimming Data, a) Frequency histogram of deviation of swimming direction from tidal-current direction, b) Shark/Current correlation index plotted with tidal-current amplitude.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

b)

co•4=«£Oo

0-30

Shark #9631 (Current Bearing - Track Bearing)

18.T%"with current against current

x 20

30-60 60-90 90-120

Directional Difference (degrees)

120-150

Shark #9631 - Correlation with Current

150-180

1.503

1.002

0.501

0.000

-0.501

1.002

-1.50312 15 18 21 24 27 30 33 36 39 42 45 486 90 3

•Shark Correlation •CurrentCumulative Hours

Curr

ent

Spee

d (k

nots

)

200



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


33.1%66.9%with current against current

x 20

b)

co•■Cro

eoO

0-30 30-60 60-90 90-120


120-150

Shark #9632 - Correlation with Current

150-180

3

2

1

0

1

2

-312 15 18 21 24 27 30 33 36 39 42 45 4890 3 6

1.5

1

0.5

0

-0.5

-1

-1.5


Curr

ent

Spee

d (k

nots

)

201



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

a) Shark #9635 (Current Bearing - Track Bearing)50

40

</> __ a> 30 x

20

10

b)

co5£oo

14.8%85.2%against currentwith current

0-30 30-60 60-90 90-120


120-150

Shark #9635 - Correlation with Current3

2

10

1

2

-39 12 15 18 21 24 27 30 33 36 39 4230 6

150-180

1.5

1

0.5

0

-0.5

-1

-1.545 48


Curr

ent

Spee

d (k

nots

)

202



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

a)

M0)X

b)

c01£oO


27.0%73.0%with current igainst current

0-30 30-00 60-90 90-120


120-150 150-180


2

1

o

1

2

312 15 18 21 24 27 30 33 36 39 42 45 483 6 90

1.5

1

0.5

0

-0.5

-1

-1.5

•Shark Correlation ■CurrentCumulative Hours

Curr

ent

Spee

d (k

nots

)

203



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

0-30


36.2%63.8%against currentwith current

30-60 60-90 90-120


120-150 150-180

b)

coy=J2£oo


2

1

0

1

-2

-39 12 15 18 21 24 27 30 33 36 39 42 45630

1.50

1.00

0.50

0.00

-0.50

- 1.00

-1.5048

■Shark Correlation ■CurrentCumulative Hours

Curr

ent

Spee

d (k

nots

)

204



Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Shark #L5285 (Current Bearing - Track Bearing)

84.3% 15.7%with current against current

° 20

b)

c0*3101oo

0-30 30-60 60-90 90-120 120-150


Shark #L5285 - Correlation with Current

150-180

3 1.5

2

1 0.5

0

1 -0.5

2

3 -1.50 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48

■Shark Correlation ■Current Cumulative Hours

Curr

ent

Spee

d (k

nots

)

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-8: Summary of Carcharhinus plumbeus directional swimming analysis. Concentration index (r), angular dispersion (s), and mean angle are in relation to mean tidal-current direction, not the overall distribution of the raw data. Significant Rayleigh’s Z indicates the data are not distributed randomly around a circle, but have a mean directionality. Significant V-test results (as determined by the u statistic) indicate that the mean swimming direction is not statistically different from the mean tidal current direction.

Shark # 9631 9632 9635 9637 9639 5285 MEAN (SD)

Duration (hours) 50 49 43 50 44 50 47.67 (3.27)

n (samples size) 277 257 242 282 229 287 262.33 (23.51)

r (concentration index) 0.59 0.47 0.61 0.46 0.25 0.60 0.497 (0.138)

s (angular dispersion) 52° 59° 51° 59° 69° 51° 56.8° (7.1°)

Mean Angle 37° 51° 16° 41° 38° 13° 32.7° (14.9°)

Rayleigh’s Z 96.53* 56.59* 89.86* 60.90* 14.77* 102.22* 70.15* (33.01)

p (signif.) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

V statistic 130.93 75.71 141.96 98.12 46.21 167.06 110.00 (44.95)

u (from V-test) 11.13* 6.68* 12.91* 8.26* 4.08* 13.95* 9.50* (3.82)

p (signif.) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

% fixes with current 81.6% 66.9% 85.2% 73.0% 63.8% 84.3% 75.8% (9.2%)

toocn

206

DISCUSSION

I Activity Space

The determination of three-dimensional activity space and examination of

short-term movement patterns are essential for defining crucial habitats for a

species or life stage of a species. These data are particularly crucial for

exploited species and those that utilize near-shore habitats, which are prone to

degradation, as nurseries or breeding grounds. Chesapeake Bay is utilized as

an important summer nursery area for juvenile C. plumbeus (Grubbs and Musick

in prep a,b). This temperate estuary is particularly susceptible to habitat

degradation due to agricultural pollution, industrial pollution, and habitat

destruction for industrial and residential development.

Tracking experiments reported in this study indicated that juvenile C.

plumbeus are extremely active. Mean activity space was 110.26 km2 over

tracking periods from 25 to 50 hours. Site fidelity tests suggested these sharks

are nomadic and do not establish true activity spaces or they establish activity

spaces much too large to be surveyed during the time frame of these

experiments. In fact, on three occasions, trips were made to search for previously

telemetered sharks. The range of the transmitters was approximately one

kilometer, therefore the entire lower Bay (south of 37°20’ N) was gridded at an

interval of 750 meters. This grid was completely searched systematically on

each occasion. No sharks were ever relocated in this fashion, supporting this

hypothesis. One shark, #9632, was tracked for 49 hours, then relocated during a


207

subsequent track three weeks later and tracked for 15 additional hours. The

initial MCP for this shark was 121 km2. The additional data did not increase the

MCP for this animal (Map 3-2).

Although tracking experiments have been published previously for juvenile

C. plumbeus in western Atlantic estuaries (Huish and Benedict 1977 - published

as C. obscurus, Medved and Marshall 1983), the duration of tracks was very

short ( 1 -13 hours), so activity-space was not reported. Several authors have

published activity-space estimates for other carcharhiniform species inhabiting

protective nurseries (Morrissey and Gruber 1993, Holland et al. 1993), but these

studies were conducted in tropical lagoons. These studies showed that these

species establish well-defined core areas of activity that were more confined

during daylight hours. This is the first such report for a highly migratory

elasmobranch in a temperate nursery. The mean MCP activity space from this

study is compared with the results of three studies conducted in tropical

ecosystems in Table 3-9. The subjects of these studies were juvenile Negaprion

brevirostris in the Bahamas, juvenile Sphyma lewini in Hawaii, and adult

Carcharhinus amblyrhynchos in the Marshall Islands. The mean activity space

reported here for juvenile C. plumbeus in Chesapeake Bay (110.26 km2) was 162

times greater than that reported for juvenile N. brevirostris (0.68 km2, Morrissey

and Gruber 1993), 87 times greater than that for juvenile S. lewini (1.26 km2,

Holland et al 1993), and 26 times greater than that reported for adult C.

amblyrhynchos (4.20 km2, McKibben and Nelson 1986). In fact, the


208

underestimated mean activity space reported here for juvenile C. plumbeus was

22 times greater than that reported for adult Carcharodon carcharias, white

sharks in the Farallon Islands off the California coast (4.94 km2, Goldman and

Anderson 1999). This mean activity space was comparable, however, to the

results of one large (230 cm TL) N. brevirostris tracked for 113 hours in the

Bahamas by Gruber et al. (1988) that exhibited an activity space of 93 km2.

The data presented herein were insufficient to examine differences in the

size of activity spaces during day and night statistically, yet examination of Map

3-10 suggests a similar pattern as described for the tropical species. The

daytime tracks of four sharks tracked in the same area are layered on one map

and the nighttime tracks are layered on the other. The nighttime tracks are

slightly more dispersed than the daytime tracks. The minimum convex polygon

for the combined day tracks was 160.1 km2 whereas the MCP for the night tracks

was 181.2 km2 supporting this hypothesized pattern.

This tremendous difference in activity space between this study and

similar studies on other species may simply be due to behavioral differences

between the species. It is more likely, however, that these differences are due to

the ecosystem differences rather than the species. Productivity in tropical

ecosystems is generally low, compared to those in temperate regions. Increased

productivity in certain tropical communities, such as the mangrove-fringed lagoon

in the Bahamas where Morrissey and Gruber (1993) tracked the juvenile N.

brevirostris gain most production through the detrital pathway from production of


209

leaf matter rather than from phytoplankton. This productivity is fairly uniform

throughout the edge of the lagoon and most of the inhabitants are year-round

residents. Prey organisms also are distributed uniformly along the mangrove

fringe. Morrissey and Gruber (1993) found that juvenile lemon sharks in this area

establish small, very stable home ranges that stretch along this mangrove fringe.

The home ranges of adjacent sharks overlapped considerably. This is an

efficient strategy for utilizing the uniformly distributed abundance of prey while

minimizing interspecific competition, a potential evolutionary advantage because

many are siblings and most are related.

Production in temperate estuaries such as Chesapeake Bay is very high.

This production is driven largely by spring and fall phytoplankton blooms. This

productivity supports a tremendous biomass of consumers. This seasonal

production combined with a climate that is only seasonally habitable, drives most

of the estuaries inhabitants to be highly migratory and highly seasonal as well.

Medved et al. (1985) found that the primary food for juvenile C. plumbeus in

Chincoteague Bay, Virginia, primarily consisted of Callinectes sapidus (blue

crabs) followed by Brevoortia tyrranus (Atlantic menhaden) and Paralichthys

dentatus (summer flounder). All three of these species are extremely abundant

in Chesapeake Bay during the warmer months. Nevertheless, they are all

migratory, however, and extremely patchy in distribution. The aggregations of

Callinectes migrate throughout the Bay and Brevoortia travel in immense, highly

mobile schools. The young sharks utilize this area as a summer nursery to


210

maximize food intake for growth while being afforded some degree of protection

from larger predators. It would be highly inefficient to establish small, stable

home ranges and wait for passing schools of fishes or aggregations of

crustaceans. Therefore, they are more nomadic which enables them to seek

patchy prey actively.

Shark 9632 provides speculative evidence of this hypothesis. This shark

initially was tracked for 49 consecutive hours (August 6-8,1997). During the first

day, the shark remained in deep waters (20-40 meters) off Cape Charles. The

first evening the shark moved inshore toward the East to shallow water (1-3

meters) near the mouth of a tidal creek. The following day this pattern repeated

in its entirety, though a different creek was utilized. Large schools of Brevoortia

were observed in the shallow areas on both evenings. Shark 9632 was tracked

for 15 additional hours three weeks later (August 27). Again the shark spent

daylight hours in the deeper waters of the eastern channel; yet at night, it moved

westward only to the edge of the channel in 15-20-meter depth. Large

aggregations of Micropogonias undulatus (Atlantic croaker), another prey

species, were observed in the area during this time. Perhaps C. plumbeus

movements change daily due to changes in abundance of and proximity to

dynamic prey aggregations. Their true activity space would therefore be defined

by the combined movement patterns of these prey species.

The validity of this hypothesis can be examined in two different ways.

The most obvious test would be to track juvenile Carcharhinus plumbeus in a


211

tropical nursery. Hawai’i supports a resident tropical population of C. plumbeus.

Little is known about the ecology of this population, but It is known that juveniles

occupy areas of deep reef around 80 meters deep. This is a completely different

habitat than the Chesapeake Bay nursery. A tracking and tagging study of this

population could be compared to the Chesapeake Bay data. If the pattern

observed in the current study were a species characteristic, then Hawaiian C.

plumbeus also would have very large activity spaces. If it were a function of the

temperate ecosystem, then the Hawaiian sharks would exhibit the tropical pattern

of a small, highly stable home range. Secondly, the activity spaces of additional

temperate carcharhinids need to be examined to see if the pattern is similar to

that presented here. Juvenile Sphyrna lewini utilize seaside lagoons throughout

the Mid-Atlantic Bight, from Maryland to South Carolina. The activity spaces of

these animals should be examined. If the activity spaces are similar to those

presented in the current study, this would be interpreted as evidence that the

pattern is ecosystem specific. If it is species specific, however, the young S.

lewini should have small, well-defined activity spaces similar to those exhibited in

Kaneohe Bay, Hawaii (Holland etal. 1993).

II Swimming Depth

The distribution of swimming-depth data ranged from zero to more than

forty meters, which corresponds to the deepest portions of Chesapeake Bay.

Swimming depth was significantly deeper during daylight hours. The overall


212

mean swimming depth during the day, as defined by sunrise and sunset, was

12.8 meters whereas the mean depth at night was only 8.5 meters. Most of the

sharks tracked stayed in deep channels during the day, then ventured into

shallower water during the night. This pattern is hypothesized to be a nocturnal

foraging strategy. Other tracking studies have reported similar results, though

these were performed on highly pelagic adult elasmobranchs. Scarriota and

Nelson (1977) reported that adult Prionace glauca were epipelagic during the day

and moved inshore to shallower areas to feed on squid at night. Carey and

Scharold (1990) reported an increase in deep vertical movements during daylight

hours by adult P. glauca. They suggested this pattern had foraging and

thermoregulatory functions. The deepest waters in the lower Chesapeake Bay

(-45 meters) are found in a slough that begins just offshore of Kiptopeke State

Park (37°10’ N, 76°00’ W) and extends northwest for nearly twenty kilometers.

Seven of the ten sharks tracked in this study utilized this deep channel. Four of

the ten spent the majority of daylight hours in this channel. Those sharks that did

not utilize this area used other similar channels such as the Chesapeake

Channel or the York River Entrance Channel. Outlines of the 95-percent Kernal

activity spaces for eight of the nine tracks conducted in Chesapeake Bay are

shown in Map 3-11. Shark 9637 was not included in this map as it had the

largest activity area of any shark tracked and never crossed its own path.

Kernals of activity are therefore meaningless. The kernals for the remaining eight

tracks are layered over the bathymetry of the estuary. At least one deep channel


213

is incorporated in each of these Kernals. In fact many of the Kernals are actually

centered by a deep channel. This indicates that these deep sloughs are

important nursery habitat for these juvenile sharks.

Swimming depth was also significantly deeper during lunar high (period

between moonrise and moonset) than lunar low (period between moonset and

moonrise). Additional data are needed to investigate this influence. It is

hypothesized that the significance of this factor during this study may have been

due to a significant interaction effect with sunrise and moonrise being highly

correlated during four of the tracks. This interaction has not been quantified, and

it is possible that this lunar influence on swimming depth is valid and may have a

function in nocturnal foraging strategies.

This study also provides evidence contrary to the belief that Carcharhinus

plumbeus is a demersal species. The majority of all swimming depth fixes were

estimated to be more than three meters from the bottom. In fact, more than one-

third of the fixes were more than six meters from the bottom. This indicates that

this species spends a significant portion of its time in mid-water or at the surface.

Interestingly, the first and third most abundant prey items identified by Medved

et al. (1985) were highly demersal species (Callinectes sapidus, Paralichthys

dentatus) whereas the second most abundant prey item was a mid-water

planktivore (Brevoortia tyrranus). Stillwell and Kohler (1993) found that the

majority of prey items consumed by larger C. plumbeus were demersal

elasmobranchs (Raja sp.) and demersal teleosts (Lophius americanus, Gadidae,


214

Bothidae, etc.). Yet they also reported that as much as 10% of the prey items

were mid-water teleosts (Pomatomus saltatrix and Scomber scombrus) and

various pelagic cephalopods.

Ill Correlation Tidal Currents

Swimming direction was significantly correlated with mean tidal direction.

A mean of 75.8% of all fixes deviated less than 90° from the current direction,

meaning they were not against the current. The mean angular dispersion of the

swimming direction was 56.8° relative to the tidal current direction. These data

suggest juvenile C. plumbeus utilize the momentum of tidal currents to conserve

energy. Chesapeake Bay is characterized by strong two-layer estuarine

circulation driven by a stiff halocline. Currents above and below the halocline

may travel in very different directions. Perhaps the vertical movements observed

within the water column function to exploit this two-layer circulation to facilitate

movement.

Future Directions

This study has provided interesting insights into the behavioral ecology

and distribution of crucial habitats of juvenile Carcharhinus plumbeus in

Chesapeake Bay. The proper determination of activity space for these animals

would require much more data than collected here, however. The daunting task

of collecting these data could be accomplished much more easily using newer


215

“chat” tags, archival transmitters from which stored data can be downloaded

during tracking. In addition to the data presented here, environmental data

(salinity, dissolved oxygen, and temperature) were collected at one to two hour

intervals during all tracking periods. These data will be analyzed in the future

and may provide important information concerning crucial nursery habitats and

the environmental factors that influence activity space.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Table 3-9: Activity-space (MCP) estimate for C. plumbeus compared with the published findings for three species of carcharhiniform sharks.

Species Negaprionbrevirostris

Sphyrna lewini Carcharhinusambiyrhynchos

Carcharhinusplumbeus

ReferenceMorrissey and Gruber 1993

Holland et al. 1993 McKibben and Nelson 1986

Current Study

N 17 6 26 7

Duration (range) 1-135 days 9-72 hrs 2-6 days 25-64 hrs

Size Range (PCL) 47-65 cm 35-37 cm 137-167 cm 44-62 cm

Mature? NO NO YES NO

Ecosystemtropicalmangrove lagoon

tropical sheltered bay

tropicalcoral atoll lagoon

temperateestuary

Mean MCP 0.68 km2 1.26 km2 4.20 km2 110.26 km2

MCP Range 0.23-1.26 km2 0.46-3.52 km2 0.19-53.00 km2 39.59-275.84 km2

to

Os

217

Map 3-10: Combined tracks for sharks 9631, 9632, 9635, and 5285 showing utilization of a deep channel in eastern Chesapeake Bay and increased dispersal during night tracks, a) Daytime fixes only - Minimum Convex Polygon for these fixes combined contained an area of 160.08 km2, b) Nighttime fixes only - Minimum Convex Polygon for these fixes combined contained an area of 181.20 km2.


Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

218

Map 3-11: Perimeters of 95-percent Kernals for eight (9637 excluded) Chesapeake Bay tracks combined over bathymetry grid. Kernals are centered over deep channels throughout the lower estuary.


37*10'

.0Z.Z6 .SUZC ,0U>£C .s .ze


219

LITERATURE CITED

Bass, G. A., and M. Rascovich. 1965. A device for the sonic tracking of large fishes. Zoologica 50 (8): 75-83.

Blaylock, R. A. 1990. Effects of external biotelemetry transmitters on thebehavior of the cownose ray Rhinoptera bonasus. J. Exp. Mar. Biol. Ecol. 141: 213-220.

Brown, D. and P. Rothery. 1993. Models in Biology: Mathematics, Statistics, and Computing. Wiley & Sons, West Sussex, England.

Carey, F. G., and K. D. Lawson. 1973. Temperature regulation in free- swimming bluefin tuna. Comp. Biochem. Physiol. 44A: 375-392.

Carey, F. G., and J. V. Scharold. 1990. Movements of blue sharks (Prionace glauca) in depth and course. Marine Bio. 106: 329-342.

Coutant, C. C. 1975. Temperature selection by fish - a factor in power-plant impact assessments, pp. 575-595. In: Environmental effects of cooling systems at nuclear power plants. Int. Atomic Energy, Vienna.

Cresswell, W. J. and G. C. Smith. 1992. The effects of temporallyautocorrelated data on methods of home range analysis, pp. 272-284 In: (Priede, G. and S. M. Swift, eds). Wildlife Telemetry: remote monitoring and tracking of animals. Ellis Horwood Limited, West Sussex, England.

Dixon, K. R. and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040-1044.

Dodson, J. J., W. C. Leggett, and R. A. Jones. 1972. The behavior of adult American shad (Alosa sapidissima) during migration from salt to fresh water as observed by ultrasonic tracking techniques. J. Fish. Res. Board Can. 29: 1445-1449.

Goldman, K. J. and S. D. Anderson. 1999. Space utilization and swimming depth of white sharks, Carcharodon carchanas, at the South Farallon Islands, central California. Environ. Biol. Fishes 56: 351-364.

Grubbs, R. D. and J. A. Musick. Spatial delineation of summer nursery areas to define essential fish habitat for juvenile Carcharhinus plumbeus in Chesapeake Bay, Virginia, in prep a.


220

Grubbs, R. D. and J. A. Musick. Migratory movements, philopatry, and temporal delineation of summer nurseries for juvenile Carcharhinus plumbeus in the Chesapeake Bay region, in prep b.

Gruber, S. H., D. R. Nelson, and J. F. Morrissey. 1988. Patterns of activity and space utilization of lemon sharks, Negaprion brevirostris, in a shallow Bahamian lagoon. Bull. Mar. Sci. 43 (1): 61-76.

Holland, K. N., B. M. Wetherbee, J. D. Peterson, and C. G. Lowe. 1993.Movements and distribution of hammerhead shark pups on their natal grounds. Copeia 1993 (2): 495-502.

Hooge, P. N. and W. M. Eichenlaub. 1998. Animal movement extension toArcView. ver. 1.1. Alaska Biological Science Center, USGS, Anchorage, AK.

Huish, M. T., and C. Benedict. 1977. Sonic tracking of dusky sharks in the Cape River, North Carolina. J. Elisha Mitchell Scien. Soc. 93 (1): 21-26.

Jennrich, R. I. and F. B. Turner. 1969. Measurement of non-circular home range. J. Theor. Biol. 22: 227-237.

Johnson, J. H. 1960. Sonic tracking of adult salmon at Bonneville Dam, 1957. U.S. Fish Wildl. Serv. Fish. Bull. 60: 471-485.

Leigh, T. H. 1997. Effects of surgically implanted, dummy transmitters on captive, juvenile sandbar sharks, Carcharhinus plumbeus. Master’s thesis, Hofstra University, Hempstead, NY. 57 pp.

Malinin, L. K. 1971. Home range and actual paths offish in the river pool of the Rybinsk reservoir. Biol. Vnutr. Vod. Inform. Byull. 22 (25): 158-166. (Transl. from Russian by Fish. Res. Board Can. Transl. Ser. 2282, 26p.)

McKibben, J. M., and D. R. Nelson. 1986. Patterns of movement and grouping of gray reef sharks, Carcharhinus amblyrhynchos, at Enewetak, Marshall Islands. Bull. Mar. Sci. 38 (1): 89-110.

McNab, B. K. 1963. Bioenergetics and the determination of home range size. Amer. Nat. 97: 133-140.

Medved, R. J., and J. A. Marshall. 1983. Short-term movements of young sandbar sharks, Carcharhinus plumbeus. Bull. Mar. Sci. 33 (1): 87-93.


221

Medved, R. J., C. E. Stillwell, and J. J. Casey. 1985. Stomach contents of young sandbar sharks, Carcharhinus plumbeus, in Chincoteague Bay, Virginia. Fish. Bull. 83 (3): 395-402.

Mellas, E. J., and J. M. Haynes. 1985. Swimming performance and behavior of rainbow trout (Salmo gairdneri) and white perch (Morone americana): effects of attaching telemetry transmitters. Can. J. Fish. Aquat. Sci. 42: 488-493.

Mohr, C. 0 . 1947. Table of equivalent populations of North American small mammals. Am. Midi. Nat. 37: 223-249.

Mohr, C. 0 . 1943. A summary of North American small mammal censuses.Am. Midi. Nat. 29: 545-587.

Morrissey, J. F., and S. H. Gruber. 1993. Home range of juvenile lemon sharks, Negaprion brevirostris. Copeia 1993 (2): 425-434.

Musick, J. A., and J. A. Colvocoresses. 1986. Seasonal recruitment of subtropical sharks in Chesapeake Bight, U.S.A. pp. 301-311. In:Workshop on recruitment in tropical coastal demersal communities (A. Yanez y Arancibia and D. Pauley, eds.), FAO/UNESCO, Campeche,

Mexico, 21-25 April 1986. I.O.C. Workshop Rep. 44.

Nelson, D. R. 1990. Telemetry studies of sharks: a review, with applications in resource management, pp. 239-256. In: Elasmobranchs as living resources: advances in the biology, ecology, systematics and the status of the fisheries (H. L. Pratt, Jr., S. H. Gruber, and T. Taniuchi, eds.), U.S. Dep. Commer., NOAATech. Rep. NMFS 90.

Nomura, S., and T. Ibaraki. 1969. Electrocardiogram of the rainbow trout and its radio transmission. Jpn. J. Vet. Sci. 31: 135-147.

Sciarrotta, T. C. 1974. A telemetric study of the behavior blue shark, Prionace glauca, near Santa Catalina, California. Master’s Thesis. California State Univ., Long Beach, Calif.

Sciarrota, T. C. and D. R. Nelson. 1977. Die! behavior of the blue shark(Prionace glauca) near Santa Catalina Island, California. Fish. Bull. 75 (3): 519-528.

Standora, E. A. Jr., T. C. Sciarrota, D. W. Ferrel, H. C. Carter, and D. R. Nelson.1972. Development of a multichannel, ultrasonic telemetry system for the


222

study of shark behavior at sea. Calif. State Univ. Long Beach Found. Tech. Rep. 5: 69 pp.

Stasko, A. B., R. M. Horral, A. D. Hasler, and D. Stasko. 1973. Coastal movements of mature Fraser River pink salmon (Oncorhynchus gorbuscha) as revealed by ultrasonic tracking. J. Fish. Res. Board Canada. 30: 1309-1316.

Stasko, A. B., and D. G. Pincock. 1977. Review of underwater biotelemetry, with emphasis on ultrasonic techniques. J. Fish. Res. Board Can. 34 (9): 1261-1285.

Stasko, A. B., and S. A. Rommel Jr. 1974. Swimming depth of adult American eels (Anguilla rostrata) in a saltwater bay as determined by ultrasonic tracking. J. Fish. Res. Board Can. 31:1148-1150.

Stillwell, C. E. and N. E. Kohler. Food habits of the sandbar shark Carcharhinus plumbeus off the U.S. northeast coast, with estimates of daily ration. Fish. Bull. 91 (1): 138-150.

Swihart, R. K. and N. A. Slade. 1985a. Testing the independence of observations in animal movements. Ecology 66:1176-1184.

Swihart, R. K. and N. A. Slade. 1985b. Influence of sampling interval on estimates of home range size. J. Wildl. Manage. 49:1019-1025.

Tesch, F. W. 1972. Experiments on telemetric tracking of spawning migrations of eels, Anguilla anguilla, in the North Sea Helgol. Wiss. Meeresunters. 23:165-183. (Transl. from German by Fish Res. Board Can. Transl. Ser. 2724, 29 p.)

Trefethen, P. S. 1956. Sonic equipment for tracking individual fish. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 179:11 p.

White, G. C. and R. A. Garrott. 1990. Analysis of Wildlife Radio-tracking Data. Academic Press, New York.

Winter, J. D. 1992. Underwater Biotelemetry, pp. 371-395 In: Fisheries Techniques (Nielsen, L. A. and D. L. Johnson, eds.) American Fisheries Society, Bethesda, Maryland.

Winter, J. D. 1977. Summer home range movements and habitat used by four largemouth bass in Mary Lake, Minnesota. Trans. Am. Fish. Soc. 106: 323-330.


223

Worton, B. J. 1989. Kernal methods for estimating the utilisation distribution in home range studies. Ecology 70:164-168.

Wray, S., W. J. Cresswell, and D. Rogers. 1992a. Dirichlet tesselations: a new, non-parametric approach to home range analysis, pp. 247-255 In: Wildlife Telemetry: remote monitoring and tracking of animals. (Priede, G. and S. M. Swift, eds.). Ellis Horwood Limited, West Sussex, England.

Wray, S., W. J. Cresswell, P. C. L. White, and S. Harris. 1992b. What,if anything, is a core area? An analysis of the problems of describing internal range configurations, pp. 256-251 In: Wildlife Telemetry: remote monitoring and tracking of animals. (Priede, G. and S. M. Swift. Ellis, eds.) Horwood Limited, West Sussex, England.

Yuen, H. S. H. 1970. Behavior of skipjack tuna, Katsuwonus pelamis, asdetermined by tracking with ultrasonic devices. J. Fish. Res. Board Can. 27: 2071-2079.

Zar, J. H. 1996. Biostatistical Analysis. Third Edition. Prentice Hall. New Jersey.


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

APPENDIX 1 Station Data for Spatial Delineation

Set Yr date Stat’n Latitude Longitude Reg- TempCPUE Surf.

1 90 6/5 K 37.1667 -76.0000 2.002 90 7/6 Anc 37.2830 -76.2830 0.003 90 7/13 Anc 37.2670 -76.3670 0.004 90 7/16 K 37.1667 -76.0000 3.005 90 7/20 K 37.1667 -76.0000 24.006 90 7/20 M 37.0830 -76.1330 13.987 90 8/20 M 37.0830 -76.1330 12.008 90 8/20 K 37.1667 -76.0000 8.579 90 8/21 K 37.1667 -76.0000 1.00

10 90 8/24 Anc 37.2000 -76.3000 14.0011 90 9/4 K 37.1667 -76.0000 19.0012 90 9/28 M 37.0830 -76.1330 4.001391 6/4 K 37.1667 -76.0000 9.0014 91 7/8 Anc 37.2330 -76.2830 1.001591 7/19 K 37.1667 -76.0000 29.0016 91 8/12 M 37.0830 -76.1330 15.0017 91 8/12 K 37.1667 -76.0000 37.0018 91 9/9 Anc 37.2330 -76.2830 11.2519 91 9/20 K 37.1667 -76.0000 6.0020 91 9/20 K 37.1667 -76.0000 10.0021 92 6/15 Anc 37.4000 -76.6500 0.0022 92 7/2 K 37.1667 -76.0000 30.0023 92 7/20 K 37.1667 -76.0000 21.0024 92 8/11 Anc 37.2330 -76.2830 3.0025 92 8/17 Anc 37.3830 -76.1670 20.0026 92 8/18 Anc 37.4000 -76.0170 18.33

Salinity Salinity D.O. D.O. Dist. to min- max- setSurface Bottom Surf. Bottom mouth Depth Depth time

25 7.9 6.3 4 8 320 6.3 34.0 3 8 319 6.5 40.0 5 7 324 6.9 6.3 7 24 524 6.9 6.3 12 21 224 5.7 13.8 6 8 225 5.3 13.8 6 7 426 6.3 6.3 7 24 526 6.3 6.3 18 24 420 6.3 31.7 6 8 324 6.1 6.3 18 21 424 5.1 13.8 11 14 225 7.9 6.3 7 24 420 5.1 31.1 5 7 324 6.9 6.3 13 26 225 5.3 13.8 6 9 526 6.3 6.3 7 24 521 5.0 31.1 6 11 324 6.1 6.3 15 22 324 6.1 6.3 7 7 412 5.2 62.9 3 4 524 6.9 6.3 10 20 224 6.9 6.3 7 14 521 6.3 31.1 5 10 321 4.0 34.5 9 16 522 4.1 31.8 9 14 2

TempBottom

1725252424242323232524241725242323252424212424252525

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


27 92 8/18 K 37.1667 -76.0000 16.0028 92 9/14 K 37.1667 -76.0000 23.0029 93 7/21 K 37.1667 -76.0000 26.0030 93 7/21 Anc 37.4500 -76.1000 12.0031 93 7/21 Anc 37.4830 -76.0330 1.0032 93 7/22 Anc 37.0830 -76.1830 1.0033 93 7/22 Anc 37.4000 -76.0500 10.0034 93 8/16 M 37.0830 -76.1330 8.0035 93 8/16 K 37.1667 -76.0000 14.0036 93 8/16 Anc 37.3330 -76.0500 1.2537 93 8/26 Anc 37.2330 -76.2830 0.0038 93 8/17 Anc 37.6830 -76.0000 1.0039 93 8/17 Anc 37.6670 -76.1170 0.0040 93 8/17 Anc 37.6670 -76.2330 0.0041 93 8/17 Anc 37.1000 -76.1670 1.0042 93 9/21 K 37.1667 -76.0000 18.0043 94 6/16 K 37.1667 -76.0000 23.7544 94 6/16 M 37.0830 -76.1330 13.7545 94 7/6 K 37.1667 -76.0000 7.5046 94 8/11 K 37.1667 -76.0000 2.0047 94 8/11 Anc 37.2330 -76.2830 0.0048 95 7/5 M 37.0830 -76.1330 17.0049 95 7/5 K 37.1667 -76.0000 14.0050 95 7/5 Anc 37.8330 -76.2330 0.0051 95 7/5 Anc 37.7670 -76.0170 1.0052 95 7/6 Anc 37.6330 -76.1000 0.0053 95 7/6 Anc 37.4580 -76.1140 2.0054 95 7/26 M 37.0830 -76.1330 24.0055 95 7/27 K 37.1667 -76.0000 12.94


26 6.3 6.3 7 24 524 6.1 6.3 11 13 524 6.9 6.3 9 15 221 3.3 39.1 9 13 521 3.9 41.6 7 13 423 5.4 18.4 7 15 322 4.8 32.3 7 7 225 5.3 13.8 11 11 426 6.3 6.3 20 22 423 4.1 25.4 9 9 521 6.3 31.1 7 11 319 3.0 62.9 8 11 219 3.0 62.4 11 13 318 1.0 65.8 9 11 425 4.9 17.0 6 11 524 6.1 6.3 18 22 525 7.9 6.3 12 20 324 6.8 13.8 8 8 424 6.9 6.3 10 17 326 6.3 6.3 9 15 321 6.3 31.1 5 7 425 5.7 13.8 9 12 127 6.0 6.3 6 8 315 3.0 83.2 11 15 517 5.4 72.6 9 15 119 3.0 58.8 11 13 222 4.2 40.5 7 11 325 5.7 13.8 9 9 327 6.0 6.3 15 19 3

>iro

TempBottom

2324242424242423232525262626232417172423252526262526262526

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


56 95 7/26 Anc 37.4070 -76.1000 0.0057 95 7/26 Anc 37.3800 -76.0690 0.0058 95 7/26 Anc 37.5020 -76.1950 0.0059 95 7/27 Anc 37.2340 -76.2380 5.0060 95 8/15 Anc 36.9410 -76.0890 0.0061 95 8/19 M 37.1170 -76.0740 5.0062 95 8/19 K 37.1920 -76.0350 10.0063 95 9/12 M 37.1350 -76.0725 7.0064 95 9/12 K 37.1770 -76.0210 2.3865 95 9/19 Anc 37.2010 -76.0570 3.0066 95 9/20 Anc 37.1030 -76.0860 8.0867 96 6/24 M 37.0950 -76.0820 0.00 2668 96 6/24 K 37.1830 -76.0340 2.5069 96 6/25 K 37.1830 -76.0340 9.00 2570 96 6/28 Anc 36.9500 -76.1950 1.00 2671 96 6/28 Anc 37.0770 -76.2230 1.00 2672 96 7/4 Anc 37.3420 -76.2100 1.00 2473 96 7/24 M 37.0940 -76.0590 1.25 2274 96 7/24 Anc 37.2290 -76.0400 0.00 2575 96 7/24 K 37.1880 -76.0350 6.2576 96 7/24 Anc 37.8330 -75.9330 0.00 2677 96 7/24 Anc 37.7300 -75.9470 0.00 2678 96 7/25 Anc 37.1670 -76.0000 1.00 2679 96 7/25 Anc 37.1330 -76.0000 1.00 2380 96 7/25 Anc 37.2670 -76.0670 1.67 2781 96 7/25 Anc 37.1670 -76.0830 11.67 2682 96 8/13 M 37.0940 -76.0660 5.00 2583 96 8/13 Anc 37.1900 -76.2150 5.00 2584 96 8/20 K 37.1830 -76.0340 11.25 25


24 4.8 34.7 9 16 524 4.8 30.9 9 11 518 4.1 47.8 9 9 123 3.0 27.9 8 8 329 5.3 6.5 7 9 528 5.4 9.5 8 8 424 5.7 10.6 19 38 428 6.0 10.0 9 19 228 6.0 8.6 7 11 527 6.0 12.6 15 27 528 6.0 10.1 8 8 2

19 25 9 6.3 9.9 6 9 427 6.2 9.9 5 27 5

19 27 7 6.2 9.9 5 27 317 20 8 5.4 15.7 5 8 615 16 8 7.8 21.4 5 8 617 19 7 4.4 33.2 6 11 324 27 7 6.5 7.5 8 11 320 26 7 5.8 13.8 2 16 4

21 6.6 10.3 13 19 413 15 8 4.6 79.8 9 17 614 18 9 2.8 68.2 4 13 119 26 7 6.0 6.3 2 13 425 25 7 6.6 3.5 4 7 516 26 8 5.5 19.5 8 13 518 23 8 6.3 12.3 7 9 320 26 7 6.2 8.4 6 9 317 25 7 5.9 23.9 6 10 421 24 9 6.8 9.9 13 19 1

>IGO

TempBottom

2626262522242725252525222525242624212124252523232120252326

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Set Yr date Stat’n Latitude Longitude Reg- Temp Temp Salinity Salinity D.O. D.O. Dist. to min- max- setCPUE Surf. Bottom Surface Bottom Surf. Bottom mouth Depth Depth time

85 96 8/21 Anc 37.1670 -6.0830 10.00 26 25 15 23 9 5.5 12.3 7 9 186 96 8/21 Anc 37.5580 -76.2420 0.00 26 26 13 15 8 8.2 54.9 8 10 687 96 8/21 Anc 37.5000 -76.2570 0.00 26 25 14 17 9 3.7 50.3 4 6 688 96 9/3 Anc 37.2330 -76.2830 1.00 26 25 17 24 8 4.9 31.1 4 689 96 9/17 Anc 37.1670 -76.0830 1.00 24 24 18 19 7 6.1 12.3 7 9 390 96 9/17 K 37.1830 -76.0320 3.75 24 24 19 20 7 6.5 9.6 6 19 491 97 6/9 M 37.1210 -76.0710 1.00 18 16 19 27 6 5.2 9.6 6 9 492 97 6/9 K 37.1860 -76.0240 0.00 18 16 19 29 5 5.0 9.7 7 24 493 97 7/16 M 37.0950 -76.0700 11.25 26 25 22 25 5 5.0 8.3 6 9 294 97 7/16 K 37.1700 -76.0180 4.35 26 24 22 25 5 4.1 7.9 7 24 295 97 7/16 Anc 37.2130 -76.3450 0.00 28 26 19 21 5 4.6 35.8 5 6 696 97 7/16 Anc 37.3370 -76.3660 0.00 28 28 18 18 5 4.9 43.3 4 7 197 97 7/17 M 37.0970 -76.0700 16.67 25 22 24 26 6 5.4 8.4 6 9 298 97 7/17 K 37.1970 -76.0270 13.64 26 23 22 26 5 4.0 10.5 7 24 299 97 7/17 Anc 37.2790 -76.1010 4.00 27 22 20 27 6 4.4 21.6 20 27 4

100 97 7/17 Anc 37.4220 -76.1710 6.67 27 25 16 21 6 3.6 39.1 7 10 5101 97 7/17 Anc 37.3100 -76.1950 11.67 28 25 17 22 6 3.2 30.1 8 11 5102 97 8/13 M 37.0920 -76.0660 9.00 26 25 21 24 7 6.4 8.3 6 9 1103 97 8/13 K 37.2010 -76.0330 4.00 26 25 22 25 7 4.7 11.1 7 24 1104 97 8/13 Anc 37.5660 -76.0300 1.00 27 25 16 24 6 3.8 50.6 13 15 4105 97 8/13 Anc 37.6360 -76.0420 1.00 27 25 16 24 6 3.1 58.4 12 23 5106 97 8/13 Anc 37.7350 -75.9180 0.00 27 26 17 19 6 3.7 68.8 12 20 6107 97 8/13 Anc 37.7440 -76.0240 0.00 27 26 15 18 7 5.4 69.8 10 17 6108 97 8/14 Anc 37.7620 -76.1650 0.00 27 26 15 20 6 3.1 74.1 8 26 2109 97 8/14 Anc 37.6380 -76.1980 0.00 27 25 16 23 6 1.5 61.6 10 11 3110 97 8/14 Anc 37.5630 -76.1410 3.33 27 25 15 24 6 2.1 52.3 11 12 4111 97 8/14 Anc 37.4910 -76.1920 0.00 27 25 16 22 6 2.7 46.3 5 9 4112 97 8/14 Anc 37.1950 -76.2990 0.00 27 27 20 20 6 6.3 30.9 2 6 511397 9/17 M 37.0930 -76.0660 6.00 24 24 27 26 7 7.1 8.3 7 12 5

h - 00 00 CM N- CO CO in M" O 03 00 CO CO 03 03 CO CO 03 CO f CM ■M" 03 00 CO COT— t—

T— T—

00 CO CO 03 O O M" CO in ■t CO 00 co Is- CO M" o O CO Is- CO CO q o T—co COCO 03 ■t- 03 03 o CO 00 •r- r f 03 CO ■*—' rL ''f r 00 yL o 00 03 00 00 oo 03 d r- oT- T— T— T— T—CM T— CM T— CM CM T— r— CM T“

o c o c o c o c o i ' - f ' - i o c o i o i ' - ' f r c o c o r ' - C M o r - . c o i n i n i n c o i n o o i ' - c o c o c oOOCMCMCMCMCMCMCMCMCMCMCMCMCMCNCMCMCMCMCMCMCMCMCMCMCNCMCMCM

( j j f O n - l f i T r T f C O C O r f l f i C D r - T - C N j T - T - C O

03 E1/3

r ^ O T - f v j T t C O C S K O N Q O C O O O ^ - C O W l N £ ~_CM t— t— t — CN CM r r r r M r rCD mE Q

c CL .b 0)E O2 S*-• 3CO OQ E

g c n c n o o M - o o i c o i n o N c o c o o M - i - c D c o . o d d d d d d d d d d d d d d t ^ d d

q * N□ CD

N N C 0 S ( 0 0 ) C 0 C M C N O 0 ) 0 ) f l 0 C 0 N S S

Q %Q CO

£ E c o = *sCD Oco m

£ 8 = -cCD 3 CO CO

E c 0 o

H CDQ.E t0 3 I - CO

LU0 3 ?0 CLa: o 0■O 303 Co

O O C O C O T f ' t C O C O ' t ' t C O C M C O A-5

CMCOeOCOWOS'tCOCOOasCMi - CM CM t- CM CM t- CM t- CM

O 3 C O a ) O 3 N N 0 O O T j - 0 Md d i n r - . ' i ^ d d i ^ d ' t d d

h- h - C30 CO 03 0 S S S N

0 N M - 0 S O O 0 O M S 0 M , 0 0 C M t- M - 0 0 0CN C M C M i - i - C N C M t- i - t- i - t-CM CM CM CM CM CN CM CM CM

i* - co cm co coCM CM CM CM CM

M " O - M - C O C O ^ T t c O C M C O C M C 3 S S N S 0 l O U 3 O T - T - O O i n 0 C O N NCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCM

■M- ' t i n •M" i n Is- Is- 03 00 00 00 03 r - r - Is- c o c o i n i n CM CM CO CO fo- Is- |o- OOCM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM

O O CM r - CO i n Is- CO o O CO Is- o i n i n CO CO o o i n m m o m m m O oO O 00 c o CO i - CO CO o O CO CO o CM Is- CO CO o i n o CM lo o CM h - 00 O inCM CO CM c d 00 c d d cd T - o d d d r — d d c d i n CM d c d q co d d d d o i ^

T_ T“ T_

O O o o o o o o o o o o Is- m CO 03 CO CO CO 00 CO 00 CM T f lo- r - COM 1CM 03 o o CM CO i n i n 00 o o o i n o CO CO 00 CO 03 00 CM CO CO 03 CO 03 00 ti n r* - T— 00 CM CO CO o CM f CM CO 03 CO 00 CO CO CO CM CO o CO CO i n COO o o o O o CM t - CM T - T - 03 o o o o o o d o o O o CM o od c d c d c d c d cd d d d d CO CO d d d d d d d d d c - d d d d d d d

r - 1'- . r* - Is- Is- r - . h - r-. h - r - N - r - Is- f - Co to- c - Is- lo- i (o- lo- fo- |o- Is- lo- lo-

0■o3

CD

CS o o o o o o o o O o o CO r - — oo CO i n O r— r - . i n N- CO 03 O ■M" CO i nh - CM i n |o- |o- i n o T—co o f CO CO h - N- m r-- N- T— o h - CO i n 00 COT“ CM CO T— 00 00 03 T—CM CO o CO 03 03 h - s 03 03 00 03 00 03 oo 03 00 o 00 COCM f- T— o o O T~ CM CO T— O t— T— CM O T—T— o T— O T— O T— CM ofol 1^ r-^ r-J 1^ r-^ 1^ is: 1^ r-J 1 1^ 1^ r - i 1^ 1^ 1^ 1^ r-‘ d 1^ 1^ 1^ i ^ N-CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO co co

m o O a o o o o o o avw c c c c c c c c c cV) X S < < < < < < < < < 2 * < X.

CO 00 00 03 03 03 03 03 03 03 o o 03 oo 00 03 03 CM CM ' t CO CO CO CO CO 00 CO0) T~ T— ■*— CM CM CM CM CM CM CM CO CO T— T— T— T— T— CM CM CO CM CM r-.ro"O 03 03 03 d d n : r*- h - n : 00 CO 00 00 00 03 03 CO CO d CO CO cok_ h - r - r-. 00 00 oo oo 00 00 00 00 00 00 CO 00 oo 00 00 00 03 03 03 03 03 03 03 03 03 03> - 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03

t in CO h - 00 03 O T— CM CO t in CO r - 00 03 O r— CM CO ■t- in CO r-. CO 03 O T— CM0 T— T— T~ T- T - T~ CM CM CM CM CM CM CM CM CM CM CO CO CO CO CO CO CO co CO CO f ■ f M"CO ■*“ t” '— T“ T“ T— T- T— T~ T- ■»” T- T—T—


Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Set Yr date Stat’n Latitude Longitude Reg- Temp Temp Salinity Salinity D.O. D.O. Dist. to min- max- setCPUE Surf. Bottom Surface Bottom Surf. Bottom mouth Depth Depth time

143 99 8/18 Anc 37.2535 -76.2596 18.57 28 28 23 24 9 5.6 30.6 6 11 5144 99 8/18 Anc 37.2856 -76.316 0.00 29 28 23 24 8 6.0 36.5 6 8 5145 99 8/18 Anc 37.2226 -76.3325 0.00 28 28 22 23 7 6.4 34.7 5 9 6146 99 9/14 M 37.0925 -76.0698 2.50 24 24 30 30 8 7.5 8.3 6 9 4147 99 9/14 K 37.1820 -76.0166 0.00 24 24 26 29 8 7.2 8.7 6 22 4

>iO)

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

APPENDIX 2-A Temporal Delineation Data

Date Sunrise Sunset DL Stat standard Standardhrs CPUE hooks

05/29/90 5:48 20:20 14.53 K 2.00 10006/05/90 5:46 20:24 14.63 K 2.00 10007/16/90 5:58 20:27 14.48 K 3.00 10007/20/90 6:01 20:25 14.40 K 24.00 10007/20/90 6:01 20:25 14.40 M 13.98 9308/20/90 6:26 19:49 13.38 M 12.00 10008/20/90 6:26 19:49 13.38 K 8.57 7008/21/90 6:27 19:48 13.35 K 1.00 10009/04/90 6:39 19:28 12.82 K 19.00 10009/28/90 6:59 18:52 11.88 M 4.00 10010/10/90 7:09 18:34 11.42 K 15.00 10010/11/90 7:10 18:33 11.38 M 9.00 10006/04/91 5:46 20:24 14.63 K 9.00 10007/19/91 6:00 20:25 14.42 K 29.00 10008/12/91 6:20 20:03 13.72 M 15.00 10008/12/91 6:20 20:03 13.72 K 37.00 10009/20/91 6:52 19:04 12.20 K 6.00 10009/20/91 6:52 19:04 12.20 K 10.00 6007/02/92 5:50 20:31 14.68 K 30.00 10007/20/92 6:02 20:24 14.37 K 21.00 10008/18/92 6:25 19:51 13.43 K 17.00 10009/14/92 6:48 19:12 12.40 K 23.00 10007/21/93 6:02 20:24 14.37 K 26.00 10008/16/93 6:24 19:54 13.50 M 8.00 10008/16/93 6:24 19:54 13.50 K 14.00 10009/21/93 6:53 19:02 12.15 K 18.00 100

circle Circle C.plum. Temp- Temp- Sal- Sal-CPUE hooks Circle surf, bottom surf, bottom

18 1820 2026 25

25 2527 2727 272726 2621 2223 2320 20

27 2726 26

25 2525 2523 2326 23

24 2425 2527 2523 2324 23

C.plum.Stan.

223

24131261

194

1599

291537

66

3021172326

81418

>i-vi

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

Date Sunrise Sunset DL Stat standard Standard C.plum. circle Circle C.plum. Temp- Temp- Sal- Sal-hrs CPUE hooks Stan. CPUE hooks Circle surf. bottom surf. bottom

06/16/94 5:45 20:30 14.75 K 23.75 80 1906/16/94 5:45 20:30 14.75 M 13.75 80 11 20 2007/06/94 5:52 20:31 14.65 K 7.50 80 6 28 2308/11/94 6:19 20:05 13.77 K 2.00 100 2 26 2207/05/95 5:52 20:31 14.65 M 17.00 100 17 26.4 2607/05/95 5:52 20:31 14.65 K 14.00 100 14 26.1 26.107/26/95 6:06 20:21 14.25 M 24.00 100 24 28.607/26/95 6:06 20:21 14.25 K 12.94 85 11 29.1 28.608/19/95 6:26 19:51 13.42 K 5.00 100 508/19/95 6:26 19:51 13.42 M 10.00 100 1009/12/95 6:45 19:17 12.53 M 7.00 100 709/12/95 6:45 19:17 12.53 K 2.38 84 205/09/96 6:02 20:05 14.05 M 0.00 50 0 0.00 50 0 15.6 13.7 18.3 25.305/09/96 6:02 20:05 14.05 K 0.00 50 0 0.00 50 005/28/96 5:49 20:20 14.52 K 0.00 80 0 7.50 40 3 17.95 15.3 18.9 27.205/28/96 5:49 20:20 14.52 M 1.25 80 1 18.3 17.1 17 25.806/24/96 5:47 20:32 14.75 M 3.75 80 3 0.00 40 0 25.9 22.3 18.7 24.806/24/96 5:47 20:32 14.75 K 2.50 80 2 12.50 40 506/25/96 5:48 20:32 14.73 K 9.00 100 9 25.4 24.8 19.3 26.207/24/96 6:05 20:22 14.28 M 0.00 80 0 2.50 40 1 22.3 20.5 24.4 27.307/24/96 6:05 20:22 14.28 K 6.25 80 5 17.50 40 707/25/96 6:06 20:21 14.25 M 11.67 60 7 27.50 40 11 25.8 20.4 17.6 27.908/13/96 6:21 20:02 13.68 M 5.00 80 4 12.50 40 5 24.8 24.7 20 26.108/20/96 6:27 19:49 13.37 K 22.50 80 18 22.50 40 9 24.95 25.9 16.5 23.508/21/96 6:27 19:49 13.37 M' 10.00 60 6 20.00 40 8 25.95 24.69 15.4 2309/17/96 6:50 19:08 12.30 M’ 0.00 80 0 2.50 40 1 24.1 24 18.1 19.309/17/96 6:50 19:08 12.30 K 3.75 80 3 2.50 40 1 24.4 23.9 19.1 20.310/02/96 7:03 18:45 11.70 M 0.00 60 0 0.00 40 0 21.6 21.1 20.2 2410/02/96 7:03 18:45 11.70 K 0.00 60 0 0.00 40 0 21.6 20.9 22.3 26.3

>I00

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


05/12/97 6:00 20:07 14.12 M 0.00 80 0 0.00 30 0 15.49 14.77 20.3 23.405/12/97 6:00 20:07 14.12 K 0.00 80 0 0.00 30 0 15.44 14.62 19.4 22.705/27/97 5:49 20:19 14.50 K 1.00 100 1 18 1606/09/97 5:46 20:27 14.68 M 1.00 100 1 0.00 40 0 17.64 15.9 19.2 27.406/09/97 5:46 20:27 14.68 K 0.00 100 0 0.00 40 0 17.61 15.89 18.5 29.207/16/97 5:59 20:27 14.47 M 11.25 80 9 12.50 40 5 25.82 24.84 22.4 24.907/16/97 5:59 20:27 14.47 K 4.35 69 3 15.00 40 6 26.2 23.48 21.7 25.407/17/97 6:00 20:27 14.45 M 16.67 60 10 28.00 50 14 24.63 22.38 23.9 26.407/17/97 6:00 20:27 14.45 K 13.64 88 12 10.26 39 4 25.6 23.39 22.1 25.508/13/97 6:21 20:02 13.68 M 9.00 100 9 16.00 50 8 25.71 24.84 21.3 24.108/13/97 6:21 20:02 13.68 K 4.00 100 4 12.00 50 6 26.07 24.82 21.8 25.109/17/97 6:50 19:08 12.30 M 6.00 100 6 19.00 100 19 24.41 24.27 26.7 26.809/18/97 6:51 19:07 12.27 K 2.00 100 2 21.28 94 20 24.03 24.19 24.6 29.609/18/97 6:51 19:07 12.27 M 6.00 50 3 11.58 95 11 24.06 23.85 27.1 28.210/02/97 7:03 18:46 11.72 M 2.00 100 2 4.00 25 1 20.65 20.58 25.9 27.610/02/97 7:03 18:46 11.72 K 1.00 100 1 4.00 25 1 20.63 21.04 23.9 25.305/07/98 6:05 20:03 13.97 M 0.00 80 0 0.00 40 0 18.2 16.35 18.2 23.205/07/98 6:05 20:03 13.97 K 0.00 80 0 0.00 40 0 18.03 15.46 19.4 22.605/27/98 5:50 20:19 14.48 M 0.00 80 0 0.00 40 0 19.91 18.85 18.8 23.205/27/98 5:50 20:19 14.48 K 0.00 80 0 0.00 40 0 19.46 19.23 19.3 24.706/29/98 5:49 20:32 14.72 M 16.67 60 10 22.50 40 9 23.96 23.24 19.4 22.806/29/98 5:49 20:32 14.72 K 8.33 60 5 20.00 40 8 25.22 22.46 16.9 2607/29/98 6:09 20:18 14.15 M 12.50 80 10 27.50 40 11 26.58 23.59 19.5 26.807/29/98 6:09 20:18 14.15 K 13.75 80 11 22.50 40 9 26.79 23.82 20.2 26.708/18/98 6:25 19:52 13.45 M 11.25 80 9 22.50 40 9 27.15 27.28 25.4 2608/18/98 6:25 19:52 13.45 K 13.75 80 11 37.50 40 15 27.34 27.42 25 2709/22/98 6:54 19:01 12.12 M 5.00 80 4 10.00 40 4 24.96 25.02 23.5 27.309/22/98 6:54 19:01 12.12 K 2.50 80 2 17.50 40 7 24.93 25.37 24.8 28.110/01/98 7:02 18:48 11.77 M 6.25 80 5 12.50 40 5 24.08 24.04 25.7 26.7

>ICO

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.


10/01/98 7:02 18:48 11.77 K 5.00 80 4 5.00 40 2 24.32 24.32 24.7 26.605/19/99 5:55 20:13 14.30 M 0.00 80 0 0.00 40 0 18.73 17.57 24 26.505/19/99 5:55 20:13 14.30 K 0.00 80 0 0.00 40 0 17.26 16.87 25.5 28.706/04/99 5:47 20:24 14.62 K 6.43 140 9 29.00 100 29 21.79 20.3 22.8 24.5

6/14/99 5:46 20:29 14.72 M 3.75 80 3 17.50 40 7 22.28 20.93 24.76/14/99 5:46 20:29 14.72 K 0.00 80 0 10.00 40 4 22.84 20.96/23/99 5:47 20:32 14.75 M 3.75 80 3 22.50 40 96/23/99 5:47 20:32 14.75 K 5.00 80 4 12.50 40 5

7/8/99 5:53 20:31 14.63 M 6.25 80 5 2.50 40 1 26.27 24.72 27.2 27.57/8/99 5:53 20:31 14.63 K 3.75 80 3 50.00 40 20 27.29 25.57 26.1 27.3

8/18/99 6:25 19:53 13.47 M 0.00 80 0 5.00 40 2 27.31 27.31 25.9 25.98/18/99 6:25 19:53 13.47 K 17.50 80 14 32.50 40 13 27.7 27.06 25.9 27.99/14/99 6:47 19:14 12.45 M 3.75 80 3 32.50 40 13 23.76 23.76 30.4 30.29/14/99 6:47 19:14 12.45 K 0.00 80 0 15.00 40 6 23.96 24.03 26 29.210/6/99 7:06 18:41 11.58 M 1.25 80 1 20.00 40 8 20.53 20.6 28.7 29.510/6/99 7:06 18:41 11.58 K 0.00 80 0 5.00 40 2 20.41 20.5 28.9 29.7

>Io

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

APPENDIX 2-B Tag-Return Data

Tag# Date Date Tag Tag Recp Recp Dist days T TT L R R GroupTagged Recap. Latitude Longitude Latitude Longitude km PCL PCL TL

R9670 8/18/92 3/29/98 37.2000 -76.0333 34.7533 -76.0217 300 2049 66 91 73 100 Commercial211102 7/4/95 1/15/97 36.9167 -75.7000 32.8333 -77.8333 560 561 120 167 Commercial

64 8/19/95 7/11/96 37.1833 -76.0333 37.2667 -75.8967 30 326 75 102 79 108 VIMS135 9/14/95 5/28/96 37.1167 -76.1000 37.1833 -76.0333 9.5 225 49 66 51 70 VIMS146 9/18/95 3/24/98 36.7500 -75.8667 34.5267 -75.9617 270 918 78 78 69 94 Commercial186 5/29/96 8/19/96 36.7500 -75.8667 37.0833 -76.0167 37 82 58 78 Recreational196 6/24/96 7/30/96 37.1833 -76.0333 37.4000 -76.0450 24 30 42 57 Recreational252 7/24/96 8/24/96 37.1833 -76.0333 37.0917 -75.9917 11 31 47 64 Recreational297 8/21/96 8/31/96 37.1667 -76.0833 37.1383 -76.2333 14 10 46 64 Recreational353 8/21/96 9/12/98 37.1667 -76.0833 36.7500 -75.9333 48 752 47 64 67 91 Recreational439 7/15/97 7/6/99 37.2611 -75.8980 37.2583 -75.8983 0.75 721 69 94 82 112 VIMS489 7/17/97 8/2/97 37.1000 -76.0667 37.2000 -76.2500 20 16 47 65 Recreational533 7/17/97 8/9/97 37.4167 -76.1667 37.4000 -76.2000 3.5 23 42 57 Recreational480 7/17/97 9/14/97 37.1000 -76.0667 37.3583 -76.2717 34 59 42 57 Recreational517 7/17/97 11/15/97 37.2667 -76.1000 35.2000 -75.6833 260 121 66 66 70 95 Commercial509 7/17/97 7/31/98 37.2000 -76.0333 33.9500 -77.9667 540 379 45 61 63 86 NC Aquarium507 7/17/97 8/1/00 37.1833 -76.0167 37.3000 -75.8000 45 1109 63 87 Commercial578 8/12/97 8/21/97 37.2667 -75.8967 37.2667 -75.8967 0 9 47 65 Recreational648 9/17/97 8/28/99 37.0930 -76.0670 37.0450 -76.0667 4.5 710 49 66 69 94 Recreational652 9/17/97 10/15/97 37.1000 -76.0667 37.1550 -76.2500 17.5 28 49 66 Commercial650 9/17/97 7/15/98 37.1000 -76.0667 37.0333 -76.0667 7 301 46 64 Recreational678 9/18/97 8/15/98 37.2167 -76.0500 37.2017 -76.0317 2.5 331 48 65 Recreational718 10/2/97 6/30/98 37.1833 -76.0167 37.0917 -75.9917 10.75 271 53 72 Recreational877 7/8/98 7/8/99 37.2105 -76.0620 37.2000 -76.2183 14 365 57 78 65 88 VIMS927 7/29/98 8/30/98 37.1833 -76.0333 37.1667 -76.9967 4 32 49 65 Recreational921 7/29/98 6/11/99 37.1867 -76.0317 37.1383 -76.2333 19 317 41 56 61 84 Commercial

>I

Reproduced

with perm

ission of the

copyright owner.


without perm

ission.

908 7/29/98 6/28/99 37.0667 -76.0967 37.1667948 7/30/98 8/15/98 37.2067 -76.1200 37.2167973 8/12/98 6/25/99 37.1345 -76.0878 37.0770998 8/17/98 8/24/99 37.2925 -75.7900 37.3217988 8/17/98 9/30/98 37.2517 -75.8987 37.2600978 8/17/98 11/17/99 37.2596 -76.8989 35.5167

1026 8/18/98 5/23/00 37.2000 -76.0400 32.15001043 8/19/98 2/10/99 37.1450 -75.9900 34.63331073 8/19/98 10/25/00 37.1333 -75.9833 35.19171162 9/30/98 2/10/99 37.2883 -75.7833 35.03331210 5/25/99 5/29/99 37.2600 -75.8983 37.04501409 7/8/99 8/15/99 37.1820 -76.2130 37.28331389 7/8/99 9/10/99 37.1830 -76.0190 37.15001576 8/18/99 8/29/99 37.2200 -76.3200 37.13831618 10/6/99 8/5/00 36.7500 -75.8700 39.59001962 8/21/00 8/25/00 37.2000 -76.0333 37.20001968 8/21/00 10/30/00 37.1000 -76.0667 35.0833

-76.4000 32 334 41 56 59 81 Recreational-76.2667 13 15 58 79 Recreational-76.2230 13.75 317 48 65 Recreational-75.8417 5.5 372 77 103 78 107 Recreational-75.8983 0.5 44 49 67 51 70 VIMS-75.4633 200 457 46 63 63 86 Recreational-80.8333 830 644 46 64 78 107 Recreational-76.6333 390 176 63 85 Commercial-75.6000 330 787 44 59 67 91 Commercial-75.9667 300 133 55 75 Commercial-76.0667 32 4 55 74 Recreational-76.3067 14 38 43 59 Recreational-76.9800 5 64 47 63 Recreational-76.0833 23 11 47 64 Commercial-74.2857 350 304 69 92 Recreational-76.0283 0.5 4 46 64 Recreational-75.8500 280 70 48 67 Commercial

nursery delineation, habitat utilization, movements, and

Documents