Demographic Rates and Energetics of Red Knots Wintering in Florida

MISSING IMAGE

Material Information

Title:
Demographic Rates and Energetics of Red Knots Wintering in Florida
Physical Description:
1 online resource (96 p.)
Language:
english
Creator:
Schwarzer,Amy C
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Interdisciplinary Ecology
Committee Chair:
Percival, Henry F
Committee Co-Chair:
Collazo, Jaime A
Committee Members:
Levey, Douglas J
Brush, Janell

Subjects

Subjects / Keywords:
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre:
Interdisciplinary Ecology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
The rufa subspecies of red knot (Calidris canutus), a long-distance migrant shorebird, has undergone drastic declines in the last 25 years. In this study I examined the Florida wintering population of this subspecies and assessed possible vulnerabilities faced by the population on its wintering grounds. I estimated annual survival and size-adjusted body mass for knots wintering in Florida and compared them to estimates derived from the entire subspecies. Comparisons between Florida and South American birds indicated that annual survival rates between 2005 and 2010 were similar (FL = 0.86-0.94 vs. SA = 0.87-0.92), and that they had similar or higher body mass. My findings do not suggest local factors are the primary explanation for declines of Florida populations, nor do they indicate that adult mortality is the likely mechanism of decline during 2005-2010. Evaluating alternative mechanisms and vital parameters, such as fecundity, during other portions of their annual cycle is necessary. Multiple theories abound for the distribution of non-breeding migrants. Through the comparison of Florida and South American populations, I assessed the likelihood of these theories as they apply to the global red knot subspecies. Although my work was not definitive, I demonstrated how ecological conditions such as food resources were a more likely determinate of winter red knot distribution than evolutionary based theories of fitness and dominance. I also modeled local residency and movement rates among three sites in the Tampa Bay region as a function of prey, body mass index, and distance. Knots remained in traditionally used areas unless conditions (e.g., human recreation) changed markedly. However, mean length of stay (MLS) varied greatly among the three sites during the core winter months (28d at Indian Shores, 44d at Longboat, 69d at Anna Maria). Although I did not measure human disturbance directly, smaller MLS corresponded with higher anthropogenic development and vice versa. Transition probabilities between areas were minimal (<0.08). Finally, I used plasma metabolites to evaluate the energetic state of wintering birds. Plasma metabolites were consistent with maintenance to a slightly positive energy budget and showed no significant differences between years or sites.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Amy C Schwarzer.
Thesis:
Thesis (M.S.)--University of Florida, 2011.
Local:
Adviser: Percival, Henry F.
Local:
Co-adviser: Collazo, Jaime A.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0043434:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 DEMOGRAPHIC RATES AN D ENERGETICS OF RED KNOTS WINTERING IN FLORIDA By AMY C. SCHWARZER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Amy C. Schwarzer

PAGE 3

3 To the red knot and all of its feathered kin and to all the people who inspired me to work with them and for them

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank the Florida Fish and Wildlife Conservation Commission for supporting this work under the State Wildlife Grant program, project 9713 720 6328 Brush, Anne Jackson, and Bambi Clemons) this project would have never gotten off the ground. Dr. Lawrence Niles and Nancy Douglass, w ith their own ground breaking work on red knots in Florida and continued cooperation throughout the length of this project, provided expertise that strengthened e very aspect of my work and allowed me to catch far more birds than I would have done so on my own. My FWC supervisor Janell Brush kept me on task, on budget and out of trouble, for which I am extremely grateful. I also thank Matthew Carroll, Bobbi Carpent er, Ariane Waldstein, Carolyn Enloe, Michael Milleson, Meredith Zahara, Nicole Ranalli, Kimberly Mortimer Branciforte, Mark Peck, Amanda Dey, Joseph Sage, Ned Beecher, David Armitage and the many other agency personnel and volunteers for their assistance i n the field and the lab Ft. De Soto County Park and its two managers, Robert Browning and James Wilson, generously provided camp sites for the field crews throughout the three years of field work as well permission to work on its lands. John Kasbohm and Lower Suwannee National Wildlife Refuge similarly provided living space and research permission in the Cedar Key area. The Florida Cooperative Fish and Wildlife Research Unit donated a trailer for technician housing for the duration of the study. Additi onal thanks to Drs. Chris Guglielmo and Alice Boyle of the University of Western Ontario for technical advice and use of lab space for the analysis of plasma metabolites. Special thanks go to Pat and Doris Leary, the New Jersey Department of Fish, Game an d Wildlife, the Delaware Division of Fish and Wildlife, the Georgia Department of Natural Resources,

PAGE 5

5 Dr. James Fraser from Virginia Polytechnic Institute, Monomoy National Wildlife Refuge, Dr. Alan Baker from the Royal Ontario Museum and www.BandedBirds.or g for sharing resight data. Many people also helped me with data analysis and editing of my thesis. I would especially like to thank my advisor, Dr. Jaime Collazo Without his tireless hours of editing my drafts at very short notice, incredible accessibili ty to his students, positive attitude, and sincere encouragement I would not have finished. I am also very grateful t o my committee, Dr. Franklin Percival Janell Brush, and Dr. Doug Levey for their support and advice on my drafts. Drs. William Kendall, J ames Nichols, James Hines, Richard Barker, and Erin Leone provided extensive statistical advice and assistance with data analyses. This thesis also benefited fro m e arlier reviews from Drs. Conor McGowan and William Kendall.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 2 ANNUAL SURVIVAL AND BODY MASS OF RED KNOTS (CALIDRIS CANUTUS) WINTERING IN F LORIDA ................................ ................................ ... 16 Methods ................................ ................................ ................................ .................. 19 Field Methods ................................ ................................ ................................ ... 19 Data Analysis ................................ ................................ ................................ ... 20 Results ................................ ................................ ................................ .................... 25 Discussion ................................ ................................ ................................ .............. 26 3 LOCAL SURVIVAL, MOVEMENT RATES, AND PLASMA METABOLIT ES OF RED KNOTS WINTERING IN THE TAMPA BAY REGION ................................ .... 38 Methods ................................ ................................ ................................ .................. 40 Study Area ................................ ................................ ................................ ........ 40 Capture Methods ................................ ................................ .............................. 41 Resighting Methods ................................ ................................ .......................... 42 Invertebrate Sampling Methods ................................ ................................ ........ 43 Plasma Metabolite Laboratory Analysis Methods ................................ ............. 43 Data Analysis ................................ ................................ ................................ ... 44 Residency and movement rates ................................ ................................ 44 Prey and plasma metabolites ................................ ................................ ..... 48 Results ................................ ................................ ................................ .................... 49 Discussion ................................ ................................ ................................ .............. 50 4 CONCLUSION ................................ ................................ ................................ ........ 66 Management Recommendations ................................ ................................ ............ 68 Recommendations for Further Research ................................ ................................ 69

PAGE 7

7 APPENDIX A PREY SELECTION IN WINTERING RED KNOTS IN THE TAMPA BAY REGION ................................ ................................ ................................ .................. 71 Methods ................................ ................................ ................................ .................. 71 Study Area ................................ ................................ ................................ ........ 71 Invertebrate Sampling ................................ ................................ ...................... 72 Reference sampling ................................ ................................ ................... 72 Foraging sampling ................................ ................................ ..................... 72 Stomach contents ................................ ................................ ...................... 73 Data Analysis ................................ ................................ ................................ ... 73 Results ................................ ................................ ................................ .................... 74 Reference Sampling ................................ ................................ ......................... 74 Foraging Sampling ................................ ................................ ........................... 74 Stomach Co ntents ................................ ................................ ............................ 74 Discussion ................................ ................................ ................................ .............. 75 B PREY AVAILABILITY AND SELECTION IN MIGRATORY RED KNOTS IN THE CEDAR KEY REGION ................................ ................................ ............................ 80 Methods ................................ ................................ ................................ .................. 80 Study Area ................................ ................................ ................................ ........ 80 Reference Sampling ................................ ................................ ......................... 81 Foraging Sampling ................................ ................................ ........................... 82 Red Knot Surveys ................................ ................................ ............................ 82 Data Analysis ................................ ................................ ................................ ... 82 Results ................................ ................................ ................................ .................... 83 Discussion ................................ ................................ ................................ .............. 84 LIST OF REFERENCES ................................ ................................ ............................... 90 BIOGRAPH ICAL SKETCH ................................ ................................ ............................ 96

PAGE 8

8 LIST OF TABLES Table page 2 1 Model selection for annual survival for adult and juvenile Red Knots banded in Florida from 2005 2010. ................................ ................................ ................. 33 2 2 Parameter estimates from the top model {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY(=0), AHY) PIM} listed in 2 1. ................................ ............. 34 2 3 Comparison of seasonal averages of weights for Florida birds with South American estimates reported in the literature. ................................ .................... 35 3 1 Model selection for daily, local survival of Red Knots in the Tampa Bay region (winters 2008 2009 and 2009 2010). ................................ ....................... 57 3 2 Model selection for survival and movement rates of Red Knots wintering in three sub regions within the Tampa Bay region ................................ ................. 58 3 3 Beta parameter estimates for the multi state model for Red Knots wintering in the Tampa Bay Region ................................ ................................ ....................... 59 3 4 Results of plasma metab olite analysis from the Tampa Bay region .................. 60 A 1 Comparison of reference and foraging samples from the Anna Maria, Indian Shores and Longboat regions in Tampa Bay during the winter and spring. ...... 76 A 2 Number of items and percentage of total items by prey type found in 16 Red Knot stomachs from the Tampa Bay region. ................................ ....................... 77 B 1 Mean horseshoe crab egg/larvae counts by site for 2008 2010. ........................ 86 B 2 Differences in horseshoe crab eggs/larvae buried compared to eggs available. ................................ ................................ ................................ ............ 87

PAGE 9

9 LIST OF FIGURES Figure page 2 1 Map showing primary study area (inset map) and secondary study area (main map). ................................ ................................ ................................ ........ 36 2 2 Popu lation trends predicted by heuristic model. ................................ ................. 37 3 1 Map of Tampa Bay region ................................ ................................ .................. 61 3 2 Daily survival probabilities across season for three locations in the Tampa Bay region. ................................ ................................ ................................ ......... 62 3 3 Daily transition probabilities between sub regions in the Tampa Bay region ...... 63 3 4 Plasma metabolites levels with standard error for the Tampa Bay region .......... 64 3 5 Weekly mean densities of preferred prey over the winter and spring at three sub regions within Tampa Bay ................................ ................................ ........... 65 A 1 Weekly mean densities of all preferred prey over the winter and spring at three sub regions within Tampa Bay ................................ ................................ .. 78 A 2 Weekly mean densities of Donax Emerita and Spionidae over the winter and spring at three sub regions within Tampa Bay ................................ .................... 79 B 1 Map of Cedar Key region with major resighting and invertebrate spawning locat ions. ................................ ................................ ................................ ............ 88 B 2 Example invertebrate reference sampling plot. ................................ ................... 89

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requi rements for the Degree of Master of Science DEMOGRAPHIC RATES AN D ENERGETICS OF RED KNOTS WINTERING IN FLORIDA By Amy C. Schwarzer August 2011 Chair: H. Franklin Percival Cochair: Jaime A. Collazo Major: Interdiscipl inary Ecology The rufa subspecies of red knot (Calidris canutus), a long distance migrant shorebird, has undergone drastic declines in the last 25 years. In this study I examined the Florida w intering population of this sub species and assessed possible v ulnerabilities faced by the population on its wintering grounds. I estimated annual survival and size adjusted body mass for knots wintering in Florida and compared them to estimates derived from the entire subspecie s. Comparisons between Florida and Sou th American birds indicated that annual survival rates between 2005 and 2010 were similar (FL = 0.86 0.94 vs. SA = 0.87 0.92), and that they had similar or higher body mass. My findings do not suggest local factors are the primary explanation for declines of Florida populations, nor do they indicate that adult mortality is the likely mechanism of decline during 2005 2010 Evaluating alternative mechanisms and vital parameters, such as fecundity, during other portions of their annual cycle is necessary. Mu ltiple theories abound for the distribution of non breeding migrants. Th rough the comparison of Florida and South American populations I assessed the likelihood of these theories as they apply to the global red knot subspecies. Although my work was

PAGE 11

11 not d efinitive, I demonstrate d how ecological conditions such as food resources were a more likely determinate of winter red knot distribution than evolutionary based theories of fitness and dominance. I also modeled local residency and movement rates among thr ee sites in the Tampa Bay region as a function of prey, body mass index, and distance. Knots remained in traditionally used areas unless conditions (e.g., human recreation) changed markedly. However mean length of stay (MLS) varied gre atly among the thr ee sites du ring the core winter months ( 28d at Indian Shores 44d at Longboat, 69d at Anna Maria) Although I did not measure human disturbance directly smaller MLS corresponded with higher anthropogenic development and vice versa. Transition probabili ties between areas were minimal (<0.08) Finally, I used plasma metabolites to evaluate the energetic state of wintering birds. Plasma metabolites were consistent with maintenance to a slightly positive energy budget and showed no significant differences between years or sites.

PAGE 12

12 CHAPTER 1 INTRODUCTION Red Knots ( Calidris canutus ) consist of at least six subspecies world wide, representing both long distance and short distanc e migrants. In the Americas three subspecies exist ( C.c. roselaari C.c. islandi ca and C.c. rufa ). The C.c. rufa subspecies (hereafter referred to as rufa ) is then subdivided into three distinct over wintering populations found in southern South America (Tierra del Fuego), Brazil, and the southeastern United States. The bulk of the rufa population winters in Tierra del Fuego (~14,000 16,000 individuals). Smaller groups are found in Brazil (~3,000) and the United States (~ 2,000 in Florida) (Niles et al 2008a). Historically Tierra del Fuego hosted an even larger proportion of the wi ntering population (up to ~100,000 birds orth combined) (Niles et al 2008b ). Over the past 25 years the rufa subspecies has declined app roximately 80% (Niles et al. 2008a). This decline h food source, horseshoe crab ( Limulus polyphemus ) eggs (Haramis et al 2007, Karpanty et al. 2006, Tsipoura and Burger 1999) at the final north bound migratory stopover in Delaware Bay. Available d ata also suggest that the rufa population in Florid a is declining (Niles et al. 2008c ). Florida is home to the largest wintering knot population in the southeastern United States ( Niles et al. 2006 ). This prompted the Florida Fi sh and Wildlife Commission in 2005 to support research designed to determine the status of the species within the state and identify factors that could be managed to curb further declines.

PAGE 13

13 An assessment of the status of the sub species wintering in Florida must start with the recogn ition of the possible interplay between habitat quality and survival rates of the throughout its annual cycle. Understanding t his relationship may be complicated by the range of migratory strategies exhibited by segments of the rufa sub species. Post breeding Red Knots migrate south along a gradient of available habitat from the United States to southern S outh America, similar to some other Arctic breeding shorebirds (Myers et al 1986 ) Within this range, h igher concentrations of bird s occur at more southerly latitudes Several hypotheses have been advanced to account for this non uniform distribution pattern. These include greater food resources, less competition for food, and decreased pre dator pressure in southern latitudes (Myers et al. 198 6). Piersma (1997, 2007) suggested that lower disease and parasite loads at non tropical latitudes may also attract more birds. Myers et al. (1986) hinted at other hypotheses, some of which have been documented in other avian taxa. These include sex spe cific resource dominance (e.g. Marra 2000), leap frogging (e.g. Kelly et al. 2002), and the presence of sedentary and migrant segments in a population (e.g. Lauber and Langenberg 1998, Jahn et al. 2004). The presence of short and long d istance migrants w ithin the sub species leads to several possible mechanisms that might account for population declines. For examp le, if one assumes that Tierra del Fuego the migratory terminus of the species, represents the strategy that maximizes annual survival and repr oduction, then lower annual survival rates for Florida Red Knots would be consistent with a marginal population at the extreme northern edge of the species wintering range (e.g., Piersma 1997, 2007). As such the status of knots wintering in Florida could be particularly vulnerable

PAGE 14

14 (sensitive) to variations in habitat quality and anthropogenic pressures. Co nversely, if differences in annual su rvival are negligible along the north south gradient, then population decli nes reported by Niles et al. (2008a) for all rufa populations might be related to factors shared across selected migratory stopovers or the breeding grounds. The objective of this study was to assess the status of Florida Red Knots by examining and contrasting evidence of selected vital paramete rs and energetic status of the population on a global (hemispheric) and local (Florida) scale. Examining evidenc e at multiple scales offers opportunities to ycle and requirements, providing a better foundation to formulate co nservation sc hemes. In Chapter 2, I compare annual survival rates and body condition of knots wintering in Florida with their counterpart s wintering in South America. This comparison also provides a basis for reasonable speculation about migratory strate gies and trade offs. I obtained estimates of annual survival from knots banded in Florida from 2005 to 2010, and from the populations of rufa wintering in South America from capture resight studies conducted at Delaware Bay (McGowa n et al. 2011). I also compare size adjusted estimates of body mass between knots wintering in Florida and South America as a means to assess their relative energetic state or condition. Size adjusted mass estimates have received a great deal of attention in recent years as a t ool to assess trade offs associated with migratory strategies and habi tat quality (Atkinson et al. 2007, Niles et al. 2008b, Baker et al. 2004). Recently, the metric was a focal parameter used to gauge whether knots m igrating through the Dela ware Bay area had met physiological and energy thresholds required for survival and prospective reproducti ve performance (Niles et al. 2008b, Smith et al. 2008 McGowan et al. 2011).

PAGE 15

15 In Chapter 3, I document local dynamics of knots wintering in Flo rida. Specifically, I estimate residency and movement rates for and among three traditionally used wintering areas in the Greater Tampa Bay region -Indian Shores (which includes birds from Indian and Reddington Shores and Indian Rocks Beach) Anna Maria (which includes Anna M aria Island, Ft. D e Soto County Park and Shell Key Preserve) and Longboat (which includes Longboat and Lido Keys). I examine a priori predictions regarding the influence of size adjusted body mass, prey levels and distance on the aforementio ned parameter s. I also examine temporal trends in body condition and levels of plasma metabolites as means to assess the energy state of wintering birds. Plasma metabolites were of particular interest in this assessment because, unlike values of size adjusted mass, t hey provide an indication of the energetic state of birds for a time window of 2 5 days (Williams et al. 1999, Guglielmo et al. 2005, Acevedo Seaman et al. 2006). Seasonal changes in levels of triglyceride glycerol and OH butyrate indicate whether overwintering birds were on a path of energetic maintenance, fat loss (catabolism) or fat production (anabolism). In Chapter 4, I present my concluding remarks and discuss their potential conservation implications. I also suggest several topics for future research that may further our understanding of Red Knots and causes of decline. Finally, I include two appendices summarizing prey types, levels and seasonal fluctuations at the Tampa Bay and Cedar Key regions respective ly. Results represent a broad based attempt to document seasonal patterns of prey with particular interest in documenting the presence and abundance of horseshoe crab eggs and determining prey preferences for Red Knots in Florida.

PAGE 16

16 CHAPTER 2 ANNUAL SURVIV AL AND BODY MASS OF RED KNOTS (CALIDRIS CANUTUS) WINTERING IN FLORIDA The Red Knot ( Calidris canutus ) is a globally distributed sandpiper that breeds in the high Arctic and makes short to long distance migrations to its non breeding sites. There are six rec ognized subspecies: canutus, islandica, piersmai, rogersii, roselaari, and rufa. Three of these subspecies occur in North America and demonstrate alternative migration strategies. Islandica breeds in the eastern Canadian Arctic and migrates along a tra ns Atlantic route to Europe for the non breeding season (Harrington 2001). Rufa breeds in the central eastern Canadian Arctic and migrates south along the Atlantic Flyway ( Niles et al. 2008b, Harrington 2001). The rufa subspecies winters along a gradient with wintering populations in the southeastern United States, Brazil, Argentina and Chile ( Niles et al. 2008b). The majority of the subspecies is found at this migratory extreme. On the other hand, i t has been suggested that the majority of roselaari while migrating south from their Alaskan Arctic breeding grounds do not attempt such long distance migrations, wintering instead in Mexico and Central America ( Niles et al. 2008b). rufa subspecies have been well documented at both wintering and migratory stopover locations ( Morrison et al. 2004, Niles et al. 2008b). Previous hypotheses to account for these declines include: a) direct mortality to underweight individuals during migration (Baker et al. 2004) or b) inability of underweight birds to breed successfully due to physiological constraints ( Morrison et al. 2005 ). Attention has primarily focused on Delaware Bay where knots congregate in large numbers to build up energy reserves for both migration to the breeding grounds

PAGE 17

17 have linked seasonal changes in body mass, as a proxy for energy reserves, with food resources, particularly horseshoe crab ( Limulus pol y phemus ) eg gs (Haramis et al 2007, Karpanty et al. 2006, Tsipoura and Burger 1999 ). Studies examining the relationship between food availability and the species status have shown that as the available supply of hors eshoe crab eggs has declined the percentage of bird s reaching sufficient weight for migration, or at least a theoretical threshold of 180 grams, has decreased annual survival (Baker et al. 2004, Niles et al. 2008b). Th ere was strong support in the data for this hypothesis (McGowan et al. 2011). However, including climatic conditions on the breeding grounds (i.e., snow fall) improved model support substantially, suggesting that other ecological and anthropogenic factors (i.e., climate change) may contribute to declines in knot populations. n limit ed by lack of information on vital parameters. This is partly due to difficulties in obtaining reliable estimates of repr oductive success across the vast and remote geographic extent of the breeding grounds. Survival estimates have only recently become available (McGowan et al. 2011, Baker et al 2004). Most estimates come from Delaware Bay, where the majority of birds from all wintering populations converge prior to migration to the breeding grounds (Harrington 2001). No estimates exist for distinct populations such as those wintering in Florida. Florida supports one of the largest wintering concentrations of Red Knots in the southeastern U.S. (Niles et al. 2008c ). Cumulative data since 2005 are adequate to est imate selected vital parameters, providing an opportunity to explore mechanisms of decline.

PAGE 18

18 I report annual survival for adults and for birds banded as juveniles, a nd body condition estimates for wintering knots in southwestern Florida. Estimates were used to determine if birds wintering in Florida differed from other rufa populations. Lower estimates in Florida than elsewhere would lend support to the possibility that population declines in Florida have a local demographic or physiolog ical component. Specifically I explor ed whether annual survival rates of Florida adults differed from estimates for other populations obtained at Delaware Bay. Although Florida knots pass through Delaware Bay along with South American populations they likely represent less than 10% of tota l migrants in the bay. Therefore I viewed survival estimates from Delaware Bay as being representative of annual survival for the South American wi n tering populations. I also examin ed whether body mass, adjusted for body size (i.e., culmen), was similar to that from birds wintering in South America during the same time period. I focused on this metric because it is commonly used as an indicator of t he energetic status and health of individuals (Baker et al. 2005, Niles et al. 2006) and as a predictor of survival and reproduction ( Baker et al. 2004, McGowan et al. 2011). I use annual survival estimates to inform a discussion about differing migratory strategies and distributions of various knot populations Florida birds are short distance migrants compared to other populations of rufa red knots. One school of migration theory would suggest that there will be a trade off between survival and reproduc tive success (i.e. closer access to the breeding grounds and higher reproductive success is balanced by higher adult mortality) (ex. Belthoff and Gauthereaux 1991, Ketterson and Nolan 1976 ), another suggests that immunological factors such as disease and p arasites determine distribution (i.e. more diseases and parasites are present at more

PAGE 19

19 tropical latitudes) ( Piersma 1997, 2007 ) while a third suggests that ecological factors such as prey availability, predator presence, etc. determines the distribution of birds among populations (Myers et al. 1986). In the first two theories, Florida birds should suffer higher adult mortality than their South American counterparts. In the last theory however, one would expect that birds from all populations would show sim ilar levels of survival. Thus by comparing survival estimates from different populations I explore these theories and discuss possible mechanisms for the distribution and size of rufa knot populations. Finally, I discuss results and conservation implicat ions for the species in Florida and for rufa as a whole. Methods Field Methods Birds wer e captured and banded as part of this project (20 07 2010) and previous project s (2005 2007) along the southwest Gulf coast of Florida at locations including Sanibel Isl and, Cayo Costa Island, and the greater Tampa Bay area including Longboat and Lido Keys (Figure 2 1) Banding took place each winter between 2005 and 2010. Not all locations were trapped each year, but captures took place at two or more locations per win ter. At least 90 Red Knots were captured per winter, exceeding 1,700 captures total for the five year period. Birds were captured using cannon nets Upon capture, each knot was banded with a standard Incoloy U.S. Fish and Wildlife Service band and marked each with a field readable leg flag inscribed with a unique alpha numeric code. Before release each bird was weighed, measured (i.e., wing chord, culmen length, and combined head/bill length), and aged as a hatch year (HY) or after hatch year (AHY). Ban ding protocols were sanctioned under the University of Florida IFAS Animal Research Committee, protocol number 004 08SNR and US Fish and

PAGE 20

20 Wildlife Service Banding Permit 21980. Marked birds were resighted in both a structured and an opportunistic fashion. I conducted resighting efforts following a pre determined schedule during the winters of 2007 2010 in the greater Tampa Bay region. Additionally I obtained data from other organized resighting efforts both within and outside of Florida as well as opportu nistic sightings by the general birding public (Figure 2 1) I obtained these data through the bandedbirds.org database. Data Analysis I modeled age specific survival using the Barker model in Program MARK (Barker 1997, White and Burnham 1999). This mode ling framework uses encounter histories from multiple sources, particularly from sampling units that are greater than the primary sampling unit (e.g., peninsular Florida). In this study, the secondary sampling unit encompassed primarily the Atlantic coast (Figure 2 1) I constructed encounter histories for 1,348 individual Red Knots marked in the primary sampling area and resighted in the primary and/or secondary sampling area. Resightings predominantly came from Delaware, New Jersey and Georgia, althoug h birds were also seen in Virginia, Massachusetts, North Carolina, Ontario (Canada). Use of this data structure ( i.e. multiple data sources with nested primary and secondary sampling units ) with the Barker model permits estimation of true survival, the c omplement of which includes only deaths. This stands in contrast to the traditional mark resight (recaptures only) analyses used in most shorebird studies (e.g., Fernandez et al. 2003, Rice et al. 2007) that produce estimates of apparent survival, the com plement of which includes both deaths and permanent emigration from the sampled areas. Estimates of apparent survival in these traditional analyses are biased low when taken as estimates of true survival because they combine survival and fidelity.

PAGE 21

21 I selec ted this analytical framework, as opposed to a multi state framework, because the Barker model focuses on survival while allowing for the testing of a variety of emigration patterns The use of a multi state model gives state specific survival (e.g., surv ival for each major sampling location, including FL), but also requires estimates of state specific (specific to each location) detection probabilities and transition probabilities to other locations. In addition, because of the temporal asymmetry in samp ling (e.g., Del aware Bay sampled at later time than F lorida ), the multistate approach would need seasonal survival probabilities for the periods that separated these different sampling periods/places. I felt that estimation of these additional parameters were not germane to my main objective, estimation of annual survival, and could probably introduce undue complexity and precision problems. I defined adult and juvenile annual survival as the probability that a bird of either age class banded in Florida be tween October and March of a given winter i was present in either the primary (peninsular Florida) or secondary sampling area (Atlantic seaboard) during year i +1 (Barker 1997). Juvenile survival in this work refers to the probability that a wintering juve nile will survive to the next winter. I used the notation HY FL to distinguish survival during this period from true juvenile survival, estimated from the time of hatching or fledging. Barker models estimate 7 parameters (Barker 1997). Of relevance to t his study were survival (S i ; the probability an animal alive at i is alive at i +1), recapture (or in this study, resight) probability (p i ; the probability an animal at risk of capture at i is captured at i ), and two parameters that relate to movement: F i ( the probability an animal at risk of capture at i is at risk of capture at i

PAGE 22

22 capture at i is at risk of capture at i +1). These parameters concern the probability of being exposed to sampling efforts in t he primary study area (e.g., peninsular Florida) depending on location during winter the previous year. Parameter r i is the probability an animal that dies in i i +1 is found dead and the band reported) and R i an animal that dies in i i +1 without being found dead is resighted alive in i i +1 before it died. Several parameters were fixed (r i = 0, R i HY = 0) because they were not interpretable or necessary in our analysis because data consisted of only live birds (W. Kendall; C olorado State University; R. Barker, University of Otago, New Zealand, pers. HY was fixed because HY FL become AHY and there were no birds marked outside the primary study area to estimate the transition between sampling units. The remaining pa rameter, R i that a marked animal that survives from i to i +1 is resighted alive at some time between i and i +1. I created a candidate set of 1 4 models to evaluate evidence in the data for constant or time specific survival rates, detection rates, and move ment rates for HY FL (hatching year at Florida) and AHY (after hatching year). The standard parameterization in MARK for movement is Markovian (no constraint), that is, the availability of a knot to be captured is conditioned on the state (available or not ) during the previous occasion. However, I included two other parameterizations in the candidate set to assess their support in the data and influence on parameter estimates. These tested for random movements between sampling units (F i i ) and permanent emigration (F i Parameters SE are reported

PAGE 23

23 Careful consideration of assumptions is important to interpret r esults. As indicated before, I assumed that annual survival estimate s reported by McGowan et al. (2011) represented rates of the rufa populations wintering in South America based on their representation at Delaware Bay (i.e., 90 95%; Niles 2008a). We also assumed that survival rates obtained at Florida (wintering grounds) and at Delaware Bay (stopover) had similar biological interpretation as they represent estimates based on a complete annual cycle. Finally, the Barker model makes three assumptions. It assumes that that every color marked bird had the same probability o f being resighted in sampling period i given that it was present in the population at the time the survey was conducted and that every marked bird present at the primary sampling area in year i had the same probability of being present on the primary sampl ing area in sampling period i +1. Second, marks (i.e., color bands) were not lost and that color bands were correctly recorded. Third, the reporting rate of marks of dead animals depends only on the stratum in which the animal was during the immediately p receding live observation. T he second assumption was met as banding techniques and color schemes have long history of use and dependability. Meeting the third assumption was not necessary because data in this study did not include dead recoveries. Assum ptions about emigration were tested using three movement parameterizations (i.e., permanent, random, Markov; Barker 1997). There was no evidence for permanent emigration, but not all birds had the same probability of returning to the primary sampling area (Florida) annually (assumption 1). Moreover, I cannot claim that every bird in peninsular Florida and elsewhere had equal probability of being captured (resighted) Thus, I used the

PAGE 24

24 median procedure in MARK to estimate a variance inflation factor to adjust for possible assumption violations (Burnham and Anderson 2002). Due to the use of the median procedure m odels were ranked by Q A ICc instead of AICc where the model with the minimum Q AICc was the model with the most support in the data. The difference in Q AICc units between the best supported model and any other model ( Q AICc) was used to calculate model weights ( Q AICc w i ), which indicate the relative likelihood of the model given the data (Burnham and Anderson 2002). Models with Q AICc 2 were considered models with highest support. Finally, I compared average body mass (gr) of knots captured in Florida with reported values for knots captured in South America during December January (core of winter) and late February early March (spring). Birds in Florida were captured in 2005, 2006, 2008 and 2009; those from South America were captured between 2000 2005 ( Niles et al. 2006, Bak er et al. 2005 ). Averages from South America were adjusted for sex, molt pattern (complete or not), culmen (35 mm) and season (wint er and spring). While I was unable to sex Florida individuals due to lack of molecular data, tests were restricted to birds that had completed molt, had a culmen of 35 35.9 mm, and were capture d either in winter or spring. Comparisons consisted of one sample t tests because standard errors for South American birds were not reported in the literature Accordingly, I treated the average body mass from South America as the hypothesized value () and asked whether the body mass from the sample of birds captured in Florida came from a bird population with the hypothesized mean value. I used a grand mean

PAGE 25

25 SE (all years) for the Florida estimates. This was done to incorporate annual variation given that hypothesized means came from data collected in different years. Tests were 2 tailed, favoring a more conservative test (non directional). Test significance was set 0.05. Results Model selection was adjusted for overdispersion; median 29 Survival of adults and juveniles banded in Florida (HY FL ) was best explained by a model that featured constant survival for juveniles banded in Florida (HY FL ), time specific survival for adults (AHY), time specific detection probability, and age specific movement probability (Tabl e 2 1). There was additional support for two other models. One included permanent emigration instead of an age specific movement prob ability (Q AICc w i = 0.19) and the other featured the same model terms as the top model except that it supported the inclusion of time specific survival rates for both adults and juveniles marked in Florida (Q AICc w i = 0.15). I will focus on the model with highest support ( Q AICc w i = 0.62 ) to summarize parameter estimates and subsequent discussion of results. Under this model, annual adult survival rates varied between 0.85 0.02 and 0.94 0.03 (Table 2 2), for an average survival rate of 0.91 0.01 over 5 years. Survival rate for juveniles marked in Florida was 0.94 0.06. Capture probability varied between 0.50 0.05 and 0.99 0.07. The probability of birds banded as juveniles in Florida to return as adults was 0.70 0.08. Similarly, the probabi lity for adults captured in peninsular Florida to return the following year was 0.81 0.05. The probability of adults moving into peninsular Florida, given that they were outside peninsular Florida the year before was 0.22 0.08 (Table 2 2).

PAGE 26

26 Aver age, size adjusted body mass (g ) for Florida and South American birds are summarized in Table 2 3. Florid a birds averaged 128.8 0.79 g d uring winter and 120.3 1.26 g during spring. Average, size adjusted body mass of Florida birds was significantly higher than average values reported for South America (P < 0.001) with the exception of birds captured in Chile. Body mass from Florida birds could have come from a population with a mean of 128.5 g (P > 0.05, Table 2 3). Discussion The 5 year average survival estimate for knots wintering in Florida was 0.91 0.01. The 95% confidence intervals of this estimate encompassed the survival rates reported for two body mass groups for Delaware Bay from 1997 to 2007 ( McGowan et al 2011 ). S 80 g) and light (<180 g ) birds were 0.92 and 0.91 respectively Similarities between rates also existed for the period of overlap between studies. Between 2005 and 2007, survival for heavy birds in Delaware Bay ranged between 0.91 and 0.92, and between 0.87 and 0.90 for light birds. Florida estimates varied between 0.86 and 0.94 during those years. Admittedly, the presence of knots from Florida in Delaware Bay may have contributed to similarities. This possibility leads to two possible interpretations albeit both coincident with our fundamental inference. Either annual survival rates of Florida knots came from a distribution of rates represented by knots at Delaware Bay (predominantly from South America) or they did not, in which case they still expe rience d similar survival ra tes to birds wintering elsewhere The application of the Barker framework was particularly insightful in that it provided estimates of fidelity, which were quite high ( 0.70 0.81 ). Conversely, there is a portion of rufa that does not return to Florida in any given year. Accounting for these

PAGE 27

27 patterns of emigration and fidelity is one reason why my estimates were, on average, higher than the 0.84 reported previously for the overall rufa population (Baker et al 2004). I cannot clai m, however, that my estimates are completely unbiased because some birds might have still occurred in areas the network of observers or I did not sample. The annual survival rate of juveniles banded in Florida (HY FL ) was 0.94. The high rates of survival could be accounted by the fact that my estimate does not include the portion of the juvenile annual cycle associated with highest levels of mortality (Martin 1995 Oppel and Powell 2010). Banding takes place in winter after this period of high mortality has already passed. High rates could also be explained by the possibility that knots often do not breed until their second adult year, many skipping migration to and from the breeding grounds during their first adult year or exhibiting only partial migrat ion. Thus, juveniles banded in Florida could also exhibit high survival rates by avoiding the risks of migration during their first year of adult life. Florida birds were equal to and often significantly heavier than their South American counterparts samp led at various locations across their wintering grounds. Florida birds on average are morphometrically larger than Chilean or Argentine birds and overlap in size with the Brazilian population ( Niles et al. 2008b). However I accounted for this by comparin g birds of similar size based on culmen measurements (~35 mm), molt and season (Baker et al 2005, Niles et al 2006). Comparisons of adult survival estimates and wintering weight estimates among Florida and South American birds point towards two important conclusions. First, Florida knots did not have a lower annual survival than birds wintering in South

PAGE 28

28 America. If there is a tradeoff(s) between staying in Florida and migrating to South America, it does not seem to be at the annual survival parameter le vel. Second, given that body mass can exert an influence on survival and reflect the physiologic state of the individual (sensu McGowan et al. 2011), then Florida birds were not at a disadvantage relative to birds wintering in South America. All to gether my results suggest that decline in population numbers in Florida is not a uniquely local phenomenon, although some variation in annual survival has to have a local component. Adult survival rates and those of juveniles banded in Florida (HY FL ) are reasona bly high, making it unlikely that direct mortality during migration or on the breeding grounds are the main causes of declines (Baker et al 2004). The question then becomes what are plausible mechanisms that could explain population declines in Florida an d elsewhere for rufa over the last twenty years? One possibility is that declines are linked to a process(es) or ca use(s) that may impinge the sub species across portions where their populations overlap. Recently, increasing attention has been given to tw o other vital parameters: breeding productivity and recruitment rates. McGowan et al. (2011) found that adult survival is positively influenced by annual snow cover on the breeding grounds. These researchers suggested that higher levels of snowfall provi de the moisture needed in this water limited system for sufficient production of insects (Noy grounds. Depressed food levels might reduce fecundity either by precluding birds from breeding or produci ng low fledging rates (McGowan et al. 2011). Unfortunately, researchers have been unable to obtain reliable estimates of fecundity and recruitment for rufa largely because breeding grounds occur in remote locations and over a vast

PAGE 29

29 landscape. Thus resear chers have resorted to using fecundity estimates from the islandica subspecies for current models. Insights about the importance of parameters that comprise fecundity might be gained if the ratio of juveniles (HY) to adults (AHY) in wintering grounds was a suitable index of breeding productivity. Although timing of migration for juveniles and adults differs, it appears that juveniles and adults share the same wintering grounds in Florida. The value of the index would also depend on meeting the assumption that juvenile and adult capture probability is similar (Bart et al. 1999). Here I assumed that the ratio was a suitable index and used it to explore the importance of fecundity with a heuristic, deterministic model. I did so by calculating the predicted population change (positive or negative) of a population of 10,000 individuals (or the estimated Red Knot population in Florida during the mid 1980s) based on the percentage of juveniles in the population of 1% to 13%, at 0.5% increments. This range was d erived from empirical estimates of the percent of juveniles in catches over the past 5 years, which ranged from 1.65 to 13.86, for an average of 7.34. Population response was modeled with the model presented below. I used my estimates of adult and juveni le (HY FL ) survival. The use of HY FL assumed that the observed age ratio reflected mortality rates from the time of fledging to arrival in Florida. Population in year t+1 = ((pop in year t*% adults*adult survival) + (pop in year t*% juveniles* HY FL survi val ))/% adults A population containing 2 3% juveniles (HY/AHY) closely matched the actual declines observed in the Florida population. Even the observed average of ~ 7% produced a decline, albeit of less severe proportions. The population became stable a t

PAGE 30

30 ~8.75% and increased at the highest observed level of 13% ( Figure 2 2 ). This exercise, while heuristic, was useful in demonstrating that observed declines could be reproduced with low juvenile percentages in light of my estimates of survival for adults and HY FL I also ran simulations under another scenario. On a second set of simulations, I used the 0.84 adult survival rates previously reported by Baker et al (2004). Results strongly suggested that at that rate of survival, and even using plausible but biologically high fecundity and juvenile recruitment rates, population declines would have been much sharper and would not conform to the magnitude of observed decline. Thus, these simulations hint at the possibility that the demographic basis of obse rved declines may indeed include a strong fecundity component. I also used annual estimates of survival to gain insights about the migration strategies exhibited by subpopulations of rufa staying in Florida or going further to South America. Myers et al. (1986) advanced several hypotheses aimed at assessing why sub populations of Sanderlings ( Calidris alba ) exhibited a similar situation, including higher numbers wintering in South America. These were: a) milder climatological conditions further south resu lting in lower physiological stress, b) lower predation rates on far south wintering grounds, c) resource availability is higher on southern wintering grounds, and d) resources are more stable further south providing greater predictability. Of these, Myer s et al. suggested that availability and/or the predictability of food resources were stronger determinants of the distribution of wintering Sanderlings. Put in another way, fewer Sanderlings occurred on northern shores (e.g., California) because the food resources did not exist to support a larger population. Admittedly,

PAGE 31

31 Myers et al. were not able to completely discount the possibility that the threat of predation also played a role in influencing the observed distribution of Sanderlings. Red Knots resem ble Sanderlings in very important ways and my estimates of annual survival provided a basis for reasonable speculation. First, similar survival estimates between birds wintering in Florida and the population converging in Delaware Bay lends support to one annual survival constraints, were stronger determinants of the differing migratory strategies of knots. Moreover, as they contended for Sanderlings, the availability of food resources likely pla yed a strong role in influencing numbers of knots across its wintering range. While direct comparisons of food availability among wintering locations are not currently possible, reports from South America (Harrington 2001 ) and Florida ( Appendix A Chapter 3 ) suggest that prey items sought by knots, particularly the clam Donax are more abundant in southern South America during the core of winter (December February). Certainly, the core winter months have the lowest availability of prey items for rufa in F lorida ( Appendix A ). Anecdotal evidence suggests that predation by avian predators may be higher in South America than in Florida ( L. Niles, pers. comm.), but research is needed before discarding this competing hypothesis. It is interesting to note that Florida and Brazilian birds are morphometrically larger than the Tierra del Fuego birds. This may allow these populations to stay further north for one of two reasons. One is that larger birds may be more adept at competing for scarce food resources and therefore able to dominate northern beaches, where presumably food resources are scarcer (Ketterson and Nolan 1976). Another is that a larger body size may allow birds to persevere under physiologically demanding conditions, either due to

PAGE 32

32 lower food resou rces or more variable weather patterns (Ketterson and Nolan 1976, Belthoff and Gauthreaux 1991). Sex ratio differences between populations are likely not an influence on average size for a given population since previous studies have found that the Florid a and South American populations have a sex ratio of roughly 50:50 (Niles et al. 2006). I assessed the status of knots in Florida using a broad geographic framework and two metrics, annual survival rates and body condition. Results suggested that knots wi ntering in Florida were not at a disadvantage when compared to the rest of the hemispheric population. Rather my results suggested that the underlying factors behind observed population declines are affecting the whole population, likely in stages of thei r annual cycle where all segments of the population could be similarly influenced. This possibility advocates for obtaining, for example, reliable estimates of fecundity and the early trajectory of juvenile survival (i.e., fledging to wintering grounds). Efforts to complete demographic and vulnerability assessment of the species. Such data gain added importance in the advent of climate change and exploration of the kn adapt to this change

PAGE 33

33 Table 2 1. Model selection for annual survival for adult and juvenile Red Knots banded in Florida from 2005 2010. Model QAICc QAICc QAICc Weights Model Likelihood Num. Par QDeviance {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY ( =0 ) AHY) PIM} 5695.16 0.00 0.62 1.00 13.00 650.73 {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0), F, F'=0 (permanent emigration} 5697.49 2.33 0.19 0.31 12.00 655.08 {S(age*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY ( =0 ) AHY) PIM} 5697.95 2.79 0.15 0.25 16.00 647.46 {S(HY(.), AHY(t) p(t) r(=0) R(.) R'(=0) F(HY(.), AHY(t)) F'(HY(=0), AHY(t))} 5701.14 5.98 0.03 0.05 18.00 646.61 {S(age) p(t) r(=0) R(.) R'(=0) F(age) F'(HY(=0), AHY (.)} 5719.51 24.35 0 0 10.00 681.13 {S(age) p(age*t) r(=0) R(.) R'(=0) F(age) F'(HY (=0), AHY (.))} 5725.07 29.91 0 0 13.00 680.64 {S(HY(.), AHY*t) p(.) r(=0) R(.) R'(=0) F(age) F'(HY (=0), AHY (.)} 5769.43 74.27 0 0 10.00 731.05 {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0) F = F') (random movement)} 5781.10 85.94 0 0 12.00 738.69 {S(HY*t, A=(.)) p(.) r(=0) R(.) R'(=0) F(age) F'(HY (=0), AHY (.)} 5803.37 108.21 0 0 9.00 767.00 {S(HY (.), AHY (.)), p(.), r (=0), R (.), R' (=0), F (HY (.), AHY (.)), F' (HY= 0, AHY (.)} 580 3.86 108.70 0 0 9.00 767.50 {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY=0, AHY) PIM} 7337.17 1642.01 0 0 12.00 650.73 {S(age*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY=0, AHY) PIM} 7339.00 1643.84 0 0 15.00 647.46 {S(HY(.), AHY(t)) p(t ) r(=0) R(.) R'(=0) F(HY(.), AHY(t)) F'(HY(=0), AHY(t))} 7341.94 1646.78 0 0 17.00 646.61 {S(age) p(age*t) r(=0) R(.) R'(=0) F(age) F'(HY (=0), AHY (.))} 7377.49 1682.33 0 0 13.00 680.43 Model parameters were: S = survival probability from peri od i to i +1, p = capture probability, r = probability of dead encounters (fixed to zero in these models since data only relied on live encounters), R = probability of live encounters tudy area (also fixed to zero), F = probability that primary study will return to the study area, t = variation over time. Lower QAIC c weight indicates better support for a model.

PAGE 34

34 Table 2 2. Parameter estimates from the top model {S(HY(.), AHY*t) p(t) r(=0) R(.) R'(=0) F(age) F'(HY ( =0 ) AHY) PIM} listed in Table 2 1 Parameter Estimate SE Lower 95% CI Upper 95% CI Juvenile survival for 2005 2010 0.9462 0.067674 0.564754 0.995829 Adult survival (2005 2006) 0.9368 0.031946 0.837347 0.977115 Adult survival (2006 2007) 0.8597 0.023143 0.80803 0.899328 Adult survival (2007 2008) 0.9166 0.024765 0.853536 0.954077 Adult survival (2008 2 009) 0.8796 0.024143 0.823766 0.91953 p (2005 2006) 0.4981 0.051233 0.399121 0.597298 p (2006 2007) 0.6881 0.039864 0.605291 0.760568 p (2007 2008) 0.6962 0.039726 0.61336 0.768141 p (2008 2009) 0.9980 0.076422 8 0.000001 1 1.0000000 R 0 0 0 0 R 0.3181 0.01105 0.296863 0.340158 R' 0 0 0 0 Juvenile F 0.6996 0.082881 0.518136 0.834607 Adult F 0.8083 0.02381 0.757283 0.85071 Juvenile F' 0 0 0 0 Adult F' 0.2218 0.059403 0.126782 0.358785 Model parameters were: S = survival from time t to t+1, p = capt ure probability, r = probability of dead encounters (fixed to zero in these models since data only relied on probability of dead encounters outside the study area (also fixed to zero), F = probability that a bird banded in the study area will remain in the area in the next period (fidelity) = probability that a bird seen outside the primary study will return to the study area S for 2009 veniles was fixed (=0) because juveniles become adults in year t+1 and none were not banded in the secondary sampling area. All birds were banded while wintering in Florida.

PAGE 35

35 Table 2 3. Comparison of seasonal averages of weights for Florida birds with S outh American estimates reported in the literature. Winter Florida Average: 128.8g +/ 0.795 SE (n = 121) Location Year Season Mean (g) Sample Size Complete molt p value Source Argentina 2000 winter 124.5 441 Na <.0001* Baker et al 2005 Argentina 20 01 winter 124 447 Na <.0001* Baker et al 2005 Argentina 2002 winter 121.5 250 Na <.0001* Baker et al 2005 Argentina 2003 winter 124 137 Na <.0001* Baker et al 2005 Argentina 2004 winter 120 94 No <.0001* Baker et al 2005 Chile 2005 2006 winter 128.5 102 Yes 0.698 Niles et al 2006 Spring Florida Average: 120.3 g +/ 1.265 SE (n = 36) Brazil 2004 2005 spring 114 38 Yes <.0001* Baker et al 2005 Florida averages derived from birds with complete molt and culmen length of 35.0 35.9 mm. Win ter (Dec Jan) Florida averages include data from 2005 2006, 2006 2007, 2008 2009, and 2009 2010. Early spring averages include data from 2008 2009 and 2009 2010. Birds from winter 2007 2008 were excluded because molt data were unavailable. All South Ame rican estimates are model predictions for birds with a culmen length of 35.0 35.1 mm. Contrasts were made with one sample t tests.

PAGE 36

36 Figure 2 1. Map showing main Florida banding and resighting locations within the primary study area (inset map), as well as the major resighting locations in the secondary study area (main map).

PAGE 37

37 Figure 2 2. Red knot p opulation trends predicted by heuristic model based on percentage of hatch year birds seen within wintering population and differential survival of hatch year (0.94) and adult birds (0.91). Starting population of 10,000 was based on estimates from surveys conducted in the mid 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 Hypothesized Population Size 1% juv 2% juv 3.5% juv 6% juv 8.75% juv 11% juv 13% juv

PAGE 38

38 CHAPTER 3 LOCAL SURVIVAL, MOVE MENT RATES, AND PLAS MA METABOLITES OF RE D KNOTS WINTERING IN THE TAMPA BAY REG ION Over the past ten years the Red Knot ( Calidris canutus ) has been the focus of intense interest by researchers across the globe. This shorebird species is a long distance migrant and in Europe has been frequently used for studies regarding migratory strategies and migration related physiology (e.g., Piersma 2007, van Gils and Piersma 2004, van Gils et al. 2005a and 2005b). Many studies within the Americas have been more conservation focused, developed to explain the decline in numbers of the rufa subspecies seen a t winter and migratory stopover sites over the past 30 years (e.g., Morrison et al. 2004, Baker et al. 2004, Karpanty et al. 2006 Atkinson et al. 2007 Niles et al. 2009 ) However, much of this research has ignored the rufa population that winters in the southeastern United States, instead placing emphasis on the population wintering in South America (particularly in Argentina and Chile). Florida is home to the largest concentration of wintering rufa knots in the United States. The main concentration o f knots on the southwest Florida coast is found in the greater Tampa Bay region. Aerial surveys conducted durin g 1980 1982 estimated approximately 6,500 10,000 wintering knots from Cape Romano in the south to Anclote Key in the north (Harrington et al. 19 88, Morrison and Harrington 1992) Aerial surveys conducted over the past five years suggested that this population had dwindled to approximately 1,000 2,500 knots (Niles et al 2008c). This decline has emphasized the importance of information that may h elp highlight threats and pressures facing these birds during their stay in Florida. I combined trapping and resighting data with sampling of the intertidal invertebrate community to augment our understanding of the winter dynamics of knots in the Tampa

PAGE 39

39 Ba y region. Specifically, I estimated local daily survival probability for the Tampa Bay region. I also estimated local daily survival probability and movement probabilities among three areas (states) known to harbor wintering knots in the greater Tampa Ba y region. I used residency rates to gauge the strength and duration (mean length of stay) with which birds used areas over winter, and assessed evidence in the data in support of hypothesized influence of a body mass index and prey levels on residency rat es. High residency rates are associated with areas where shorebirds meet energetic requirements (Alerstam and Lindstrom 1990, Lyons and Haig 1995, Rice et al. 2 007). I also determined the extent of exchange or movement between the three traditionally use d areas in the region by wintering knots. Beach habitats are linear and do not present obvious barriers that might curtail connectivity. However, wintering knots in Florida encounter higher levels of coastal developmen t and anthropogenic disturbance. A greater understanding of the extent of movement and how distance might impinge on it is of value to conservation planners. The ability or willingness of individuals to move among areas can provide wintering knots the means to access alternative foraging a reas, ameliorate predation risks, and avoid or minimize anthropogenic disturbance (Farmer and Parent 1997, Skagen 1997, Belisle 2005). I also assessed plasma metabolites to quantify the physiological state of knots as a means to explore further the condi tion of knots wintering in Florida. Body mass has been used in many studies of Red Knots as to assess body condition and energetic state in knots (e.g., Baker et al. 2004, Niles et al. 2006, Atkinson et al. 2007 ) and, more recently, to examine the implica tions of different body mass levels on survival (McGowan et al. 2011). By this standard I demonstrated that knots wintering in Florida

PAGE 40

40 are in similar or better condition than their South American counterparts (Chapter 2). However, body mass estimates onl y represent a point estimate, often derived from capturing an individual once in a season (Ferna ndez et al. 2003, Dinsmore and Collazo 2003, Rice et al. 2007) and the energetic trajectory of the individual is not clear. It is possible, for example, that h eavier birds may be physiologically stressed and burning energy, while lighter birds may be rapidly assimilating energy sources. Plasma metabolites provide insights into the short term (2 5 days) energetic path of knots post capture ( Cersale and Guglielmo 2006, Guglielmo et al. 2005, William s et al. 1999). Here I report levels of three metabolites collected from birds during three periods over the wintering season and asked if there was any evidence that birds were on a path to a positive or negative ener gy balance Samples were collected in late October/early November, early January, and late February/early March, spanning the core of winter in I also report metabol ite levels at Delaware Bay where birds are building up fat reserves to complete the last leg of their migration to the breeding grounds and assist with the onset of reproduction (McGowan et al. 2011 ) Finally, I discuss the conservation implications of my findings for knots wintering in Florida. Methods Study Area The Tampa Bay region (Figure 3 1) along the Gulf Coast of Florida encompasses ~74 linear kilometers of ocean side beaches. These beaches are predominantly composed of fine grained sand, face the Gulf of Mexico, and are generally subject to low energy wave action and small tidal amplitude (~0.25 0.75 m). Only two locations, Gulfport and Shell Key, contain broad tidally exposed sand or mud flats that were also

PAGE 41

41 regularly surveyed. The locations re present a wide range of human activity, including the highly developed, tourist rich Clearwater Beach to the less developed, private beaches of Longboat Key to the multi use publicly owned preserves of Ft. De Soto County Park and Shell Key Preserve. In thi s study I grouped all data for analyses into three sub regions of roughly equal extent: 1) Indian Shores which included all territory from Clearwater Beach to Madeira Beach and Gulfport, 2) Anna Maria which included Shell Key Preserve, Ft De Soto County Pa rk and Anna Maria Island, 3) and Longboat which included Longboat and Lido Keys (Figure 3 1). I used these three areas as state variables in the multi state models. Capture Methods I trapped knots during four periods during winter 2008 2009 and 2009 2010 The periods were late October/early November, early January, late February/early March and mid April. I spaced trapping efforts to establish weight and plasma metabolite profiles for the Tampa Bay population while minimizing trap stress on the birds. Birds were trapped using a cannon net array consisting of two cannons and an 8 x 25 m net specifically for shorebird trapping. I trapped opportunistically on beaches where birds were present and in catchable conditions. Catchable conditions were defined as the time when beach topography allowed sufficient access and area to set up equipment without forcing birds from the beach. I distribute d capture locations throughout the region each period to capture variation among beaches. Upon capture, each knot wa s banded with a standard Incoloy U.S. Fish and Wildlife Service band and marked each with a field readable leg flag inscribed with a unique alpha numeric code. Before release each bird was weighed, measured (i.e., wing chord, culmen length, and combined h ead/bill length), and aged as a hatch year (HY) or after hatch year (AHY).

PAGE 42

42 I collected blood from a subset of birds at each capture. Samples were collected from up to 30 individuals. All birds were processed within 30 minutes of the time of initial captu re to minimize the possibility that stress altered plasma metabolite levels ( Guglielmo et al. 2002, Dietz et al. 2009 ). I bled birds following the standard method of piercing the brachial vein and collecting blood in one or two 70 l hep a rinized capillary tubes. I used cotton swabs and applied pressure to stop bleeding and held birds for approximately two minutes to assure bleeding had stopped. I kept the blood chilled on ice. Each night I spun the blood to separate the plasma using a centrifuge set at 3 000 rpm for ten minutes. I then transferred the plasma into micro centrifuge tubes and stored it at 80 F until lab analysis. All b irds were banded under a US Fish and Wildlife Service Bird Banding permit held by the Florida Fish and Wildlife Conservati on Commission and the capture, handling, and blood sampling protocols were approved by the University of Florida IFAS Animal Research Committee permit 004 08SNR. Resighting Methods I surveyed all beaches for marked birds at the following locations: Clearw ater Beach to Madeira Beach (including Indian Shores), Gulfport, Shell Key Preserve, Fort De Soto County Park, Anna Maria Island, Longboat Key and Lido Key (Figure 3.1) Surveys occurred during October to mid May of winter 2008 2009 and 2009 2010. There were 17 total survey periods and time intervals between surveys were 69 14 14 10 7 7 7 7 7 7 7 7 8 6 10 and 3 days, respectively. During each survey I would cover all locations within the study area within a consecutive 3 day period. I sca nned all beaches for the presence of Red Knots. If knots were located, a second observer (if present) and I recorded general data including start time, weather conditions, number of

PAGE 43

43 knots present, number of knots banded, knot activity (e.g., foraging, roo sting), tide stage, and estimated numbers of other bird species present. After recor ding general information, I scanned flocks using a spotting scope (Vortex Skyline 80mm with 20 60x zoom or Kowa TSN 660 with 20 60x zoom). I recorded all band and flag co de information on individuals within the flock. I scanned the flock multiple times and if possible from several angles to maximize the leg flags detected. When two complete scans produced no new flags for either observer, we ended the scanning session an d recorded the end time. Invertebrate Sampling Methods I sampled for invertebrates at Ft. De Soto, Shell Key, Lido Key, Longboat Key and Indian Shores in 2010. I selected these locations based on the consistent presence of foraging Red Knots in previous y ears. I established two reference transects at each of the five locations Each transect consisted of a fixed starting point, with four points total spaced ten meters apart. Transects were placed randomly along selected beaches and were at least 200 m a part. I sampled both transects weekly at each location starting the third week of January. I sampled in the surf zone, parallel to the waterline. At each point I collected sediment cores 5 cm deep and 10 cm in diameter and stored them in a lock seal pla stic bag for future analysis. After sample collection, I sieved each sample with a 1 mm wide mesh sieve. I collected, identified to family (or genus if possible) and counted each invertebrate in the sample. I measured common invertebrates in order to cre ate an index of biomass. Plasma Metabolite Laboratory Analysis Methods I analyzed plasma samples at the facilities of Dr. Chris Guglielmo at the University of Western Ontario. I assayed lipid metabolites in a microplate spectrophotometer. I

PAGE 44

44 measured t ota l triglyceride and free glycerol sequentially by endpoint assay ( Sigma Aldrich glycerol reagents A and B; Williams et al. 1999 ). True triglyceride concentration was calculated by subtracting glycerol from total triglyceride. OH butyrate was measured by kinetic assay (R Biopharm OH butyrate kit and Stanbio standards ; Williams et al. 1999 ). I chose these metabolites because they are easily measured and have a strong relationship with the rate of change in fat levels ( Guglie lmo et al. 2005 Williams et al. 1999). Triglyceride is a product of fat production and OH butyrate and glycerol are products of fat metabolization. Comparison of the three indicates whether a bird is actively laying down fat, maintaining current condi tions, or burning fat (Guglielmo et al. 2005 ). Data Analysis Residency and movement rates I used two types of modeling frameworks in program MARK (White and Burnham 1999) to estimate re sidency and movement rates. I pro gram MARK (White and Burnham 1999) to estimate re sidency rates for the Tampa Bay Region Residency rates ( ) were defined as the probability that a knot, banded in the Tampa Bay Region on day i, remained in the area until day i +1. I examined 18 models of residency and recapture rates, which included the four basic models (i.e., constant and time specific residency and recapture rates) Daily residency probabilities were calculated to account for unequal time interval length between c apture periods. The daily probability is the n number of days in the interval. Shorebird mean length of stay (MLS), expressed in

PAGE 45

45 days, was estimated using the mean life expectancy equation deri ved by Brownie et al. (1985): MLS = 1/ln (daily residency probability). In addition, I used a multi state analytical framework to estimate daily survival and transition probabilities within the Tampa Bay Region (Kendall and Nichols 2004). Residency ra tes ( ) were defined as the probability that a knot, banded at state j on day i, remained at state j to day i +1. Movement rates (Psi, ) were defined as the probability that a knot, banded in state j on day i mov ed to another state on day i +1. States in the model were: Indian Shores (I), Anna Maria (A), and Longboat (L). The extent of each state is provided in the study area section above. I examined 10 models of residency and recapture rates, which included ba sic models (i.e., constant and time specific residency, recapture and movement rates). Both model sets included year effects (2009, 2010; code=g), seasonal terms (i.e., linear or quadratic) and a body mass index as an individual covariate. Multi state mo dels included two additional covariates (i.e., distance, prey density) and interaction terms. Below I describe all model terms and covariates, and when appropriate, I provide the hypothesized influence of the covariate on residency or movement rates. Unl ess indicated, covariates were modeled as additive factors. Body Mass Index This index is a size adjusted value (BMI) calculated as b ody mass divided by culmen length of the individual used in the two model sets. In the knot literature body mass is ofte n used as a proxy of body condition or health of the individual (e.g., Baker et al. 2004, Niles et al. 2008b ). Adjusting for size allowed us to use all birds across sites. I hypothesized that residency rates should be positively influenced by BMI as bird s in good body condition would chose to remain to exploit

PAGE 46

46 resources and minimize unnecessary risks accrued with movements (Alerstam and Lindstrom 1990, Taft and Haig 2006 a, b ). Seasonal Variation I evaluated evidence for seasonal patterns in daily surviv al probability usi ng two terms, a linear (T) and quadratic trend (TT). P revious studies have shown that a variety of factors such as timing of arrival or departure from Florida, may c ause variation in daily survival pr o bability (e.g., Dinsmore and Co ll a z o 2003 ) Inter State Distance This covariate was defined as the nearest neighbor distance from the perimeter of each sampling or beach state as defined in the study area section to the closest point of a neighboring sampling state. I used Google Earth t o calculate these distances. The dista nce between Anna Maria and Longboat was 13 km, between Anna Maria and Indian Shores was 38 km, and between Longb oat and Indian Shores wa s 53 km The hypothesized influence of this covariate on movement rates was nega tive. Because there are no shorebird studies assessing connectivity in linear, coastal habitats, I used two salient findings from Farmer and Parent (1997) and Taft and Haig (2006a, b) to help me interpret results. These were that connectivity should occu and that movement rates 0.08 were associated with states that were effectively iso lated. Accordingly, I hypothesized that all states would be isolated, with the greatest degree of connectivity occurring between Anna Maria and Longb oat. Prey Density I calculated prey levels as the average number of prey per sample per season for a given sub region (392.5 cm 2 per sample). I included all possible prey items. Preferred prey sources for the knots in Tampa Bay included Donax clams, mole crabs (genus Emerita ) and marine worms. Although other items were included in the

PAGE 47

47 estimated means, non preferred items account for only 0.31% of the overall prey found in samples. The sampling period was divided into four seasons: fall, winter, early spring and late spring. Increased residency in suitable habitat not only is beneficial energetically, but reduces the risks of mortality incurred as birds move among wetlands (Farmer and Parent 1997). I hy pothesized that prey density would have a positive influence on residency rate and negative on movement rates (Weber and Haig 1996). Interaction Terms I evaluated a linear and a quadratic term to account for the fact that prey undergoes marked s easonal fluctuations (Appendix A for prey). I wanted to sort the effects of this factor given that migration (arrival or departure), at the beginning and end points of the winter season, also influence daily residency rates. I used l (Burnham and Anderson 2002). Models were ranked by AICc, where the model with the minimum AICc was the model with the most support in the data. The difference in AICc units between the best suppor ted model and any other model ( AICc ) was used to calcul ate model weights (AICc w i ), which indicate the relative likelihood of the model given the data (Burnham and Anderson 2002). Models with AICc 2 were considered models with highest support. I used the median procedure in MARK to estimate the variance inflation factor using recaptures only data (Burnham and Anderson 2002). There is no procedure in MARK to obtain an estimate c hat to adjust models for overdis persion in multi state models. T hus careful consideration of model assumptions was important to interpret multi state results. First, I assumed that every color marked bird had the same probability of being

PAGE 48

48 resighted in sampling period i and that every m arked bird had the same probability of surviving from sampling period i to i+1, assuming that it was alive and present in the population at the time the survey was conducted. Second, I assumed that emigration (i.e., departure) was permanent. I believe th at this assumption was met because there was support in the data used to estimate annual survival for a model testing for permanent emig ration (Q AICc w i = 0.19; Chapter 2 ). Third, I assumed that marks (i.e., color bands) were not lost and all correctly re corded. Parameter estimates were reported as estimate standard error. A covariate was considered significant within a given model if the 95% confidence intervals of the beta value for the covariate did not overlap zero. Prey and plasma metabolites I co mpared estimates of prey density using a linear model (PROC MIXED, SAS 2002). Model terms were region (Anna Maria, Long Boat, Indian Shores), week (16 weeks, week 8 was excluded due to lack of samples), and interaction terms (i.e., week*region, week*week, week*week*region, week*week*week). This model treated site [region] and transect [region*site] as random effects. The response variable, prey density, was log transformed prior to running the model. I also compared plasma metabolites levels across the winter (seasons = late October/early November, early January, late February/early March). For triglyceride I used a linear model using PROC MIXED (SAS 2002). Model terms were season year, and BMI (adj usting for ddfm) Due to problems with normality in the data for OH butyrate and glycerol, I used a non parametric model ( PROC npar1way SAS 2002 ) Differences in means between groups were tested using the Wilcoxon two 0.05. Response variables were each of three metabolit es measured in this study.

PAGE 49

49 Results Model selection for the residency models was adjusted for overdispersion; median 18 Residency rates were best described by a model that accounted for variation in survival rates by year ( Q AICc w i = 0. 61; Table 3 1 ). Encounter probab ilities were influenced by year, varying by time within year (g*t) A plausible alternative was a model that featured the same terms but included body mass ( QAICc 2; QAICc w i = 0.38). Body mass had a positive e ffect on daily survival, as predicted, but its influence was weak (Beta = 0.208 0.178 ; 95%CIs overlapped zero). Daily survival and movement rates were best described by a model that accounted for variation in survival rates with a linear trend term (T) f or every site, prey abundance and its interaction with time of season, and distance between sites (AICc w i = 0. 62; Table 3 2 ). Encounter probabilities were influenced by year (yr 1 = 0.21, yr 2 = 0.36), but within year detection probability were similar a cross states. A plausible alternative was a model that featured the same terms influencing daily survival but included body mass (AICc w i = 0.38 ). Body mass had a negative effect on daily survival, but its influence was weak (Beta = 0.113 0.177; 95%CI s overlapped zero). Daily survival rates among sites were similar (95% CIs overlapped, F igure 3 2 ). Rates declined linearly as the season progressed, but were positively and strongly influenced by seasonal prey abundance ( T*Prey Beta = 0.026 0.002 ; 95 %CIs did not overlap zero; Table 3 3 ). While daily survival at Anna Maria and Longboat Key remained relatively high and roughly equal throughout the season, daily survival probabilities at Indian Shores dipped substantially during weeks 10 14 ( Mar 22 Apr 23 ). The probability of moving in or out of each site was negatively, but weakly influenced by

PAGE 50

50 distance among sites (Beta 95%CIs overlapped zero). Daily transition probabilities were low be tween all three sites (Figure 3 3). Only one value, the transiti on probability of birds moving from Longboat Key to Anna Maria, was higher than 0.08 (Figu re 3 3). There were no significant differences for metabolites when examined among a reas and years in Florida. Therefore results from both years and multiple areas w ere lumped together for the final analysis. Plasma metabolite levels decreased across the season for triglyceride, glyce rol and OH butyrate (Figure 3 4). Concentrations of triglyceride, OH butyrate, and glycerol were significantly higher in October/November than in January or March (Table 3 4). There were no significant differences between January and March for any metabolite The elevated concentrations of all three metabolites in October /November may be a residual effect from migration, with birds arriving with both higher energy stores (triglycerides) and signs of higher energy expenditure ( OH butyrate glycerol ). While triglyceride levels in Florida ranged from 1.08 to 2.07, concentrations in Delaware Bay averaged 9.59 ( SE 4.82) far more typical for a refueling site. Similarly concentrations of glycerol averaged 2.54 ( SE 2.24) and OH butyrate averaged 0.58 ( SE 0.31) compared to 0.03 0.05 and 0.70 0.83 respectively. Elevated glycerol is also expected at high levels of refueling since it is a by product of high triglyceride production. Discussion Knots overwintering in Florida are a sedentary population that conserves energy and maintains body condition. Residency rates in the greater Tampa Bay region translated to a mean length of stay of 52 d in 2008 2009 and 99 d in 2009 2010. Annual differences likely reflected local variations on the internal state of k nots (e.g., body condition) or habitat quality, which could encompass prey fluctuations, weather

PAGE 51

51 and anthropogenic factors (e.g., human disturbance). To explore how body condition, prey levels and distance might operate, I modeled daily survival and movem ent rates for three locations known to harbor wintering knots. I found that estimates of daily survival were high for all three locations, albeit exhibiting a seasonal negative trend. The negative trend is consistent with a resident population that event ually transitions to a migratory phase. Residency was highest in the Anna Maria region, followed by Longboat and Indian Shores, respectively. During the core of winter (January March), mean length of stay was 69 d for Anna Maria, 44 d for Longboat and 26 d for Indian Shores. Despite the sedentary nature of birds at all three locations, estimates of daily survival were consistently lower at Indian Shores, falling to their lowest value (0.75) during April 5 April 14. At this point mean length of stay was only 3.47 days. As hypothesized, residency rates were positively and strongly influenced by average prey density at each location. Prey levels, however, were lowest at Indian Shores and may explain why the pattern of daily survival at Indian Shores was lo wer than the other two locations (Figure 3 5). Previous studies have linked body condition to residency at stopover and wintering sites ( Skag en and Knopft 1994, Lyons and Haig 1995 ). As body mass is an expression of energetic state, it is reasonable to a ssume that it would affect whether birds chose to remain at or vacate from a location. Yet body mass exerted only a weak influence on residency. Similar findings were reported for wintering Semipalmated Sandpipers in the eastern Caribbean (Rice et al. 20 07). Body mass is a valuable proxy for bird condition, but because it is almost invariably based on a single capture estimate, it is hard to discern if the bird is on a positive, sustaining or negative energetic path. In this sense, my assays of plasma

PAGE 52

52 me tabolites addressed an important void in our understanding of the condition of knots during winter in Florida. Analyses revealed two salient results: a) there were no significant differences between locations during any of the sample periods, and b) despi te a general trend of dropping triglyceride levels across the season s catabolic metabolites remained low and stable during the core of winter This indicated that knots at all three locations while consuming less energy during the core winter months, we re not undergoing a severe or even moderate fast despite di fferences in prey availability at each location Winter birds, unlike migratory birds, are likely concerned with maintaining body condition and minimizing predation risks rather than improving it through building fat and muscle ( Warnock and Bishop 1998, Van Der Veen and Sivars 2000, Gentle and Gosler 2001 ). Body mass may not always be the best indicator of energetic state, and thus, may not illuminate the relationship between residency and energet ic state. This work suggested that a lighter bird would have no more or less incentive to move than a heavier bird. Residency patterns hinted at the role that human disturbance may play within the Tampa Bay region. Among the three areas, Indian Shores is the most highly developed area with condos and motels just behind the dune area, high levels of human foot traffic, the winter, birds at Indian Shores had lower daily s urvival probabilities than knots at other locations. Longboat, which has significantly less land development and human foot traffic, had the second best survival and Anna Maria, which encompasses areas closed to human access, had the highest daily surviva l rates. In addition, birds at Indian Shores underwent a period of greatly reduced daily survival between mid March and

PAGE 53

53 mid April, while the other locations did not. This period coincided with a marked increase in human activity (e.g., recreational), cul minating in a surge of presence and activities along the coast during spring break (pers. obs). I surmised that the drop in daily survival was caused by either knots temporarily or permanently vacating the area in response to this surge. Although this in ference is speculative, it is noteworthy that estimates of food availability actually increased during the period in question. Therefore it weakens the possibility that prey played a major role based on the strong and positive relationship between residen cy and prey levels reported in this study. Results thus far suggested that movements out of an area can be triggered by changes in habitat quality (e.g., prey levels, human activity). This process highlights the importance of alternative areas where knot s can continue to meet overwintering needs. Proximity of good quality sites reduces exposure to predation risks and energetic costs ( Skagen 1997, Farmer and Weins 1999 ). In this study I assessed this process and found that distance among locations exert ed a negative, but weak influence on movement rates. Still, average transition probabilities between locations within Tam pa Bay were quite low (Figure 3 3). Previous studies on shorebirds suggested that t distances greater than 2 km are perceived by shorebirds as effectively isolated (Farmer and Parent 1997, Taft and Haig 2006b). It follows that locations within Tampa Bay area might be perceived by knots in a similar fashion. The highest transition prob ability was from Longboat to Anna Maria (0.14), and this movement was uni directional (transition probability in the opposite speculate that knots from Longboat may take advanta ge of the protected areas within

PAGE 54

54 Anna Maria as reflected by high movement rates and daily survival at Anna Maria. This possibility can be viewed as an example of functional connectivity (Belisle 2005). The implications of my results were particularly rel evant to areas like Indian Shores. It suggested that the Anna Maria and Long Boat areas provided little functional value (e.g., place to reside) when birds at Indian Shores were displaced. My results indicate that birds remained in the area at high rates and seemed reticent to move long distances to other traditionally used areas. As such, displaced birds could incur unnecessary energetic costs and may undermine their overwinter survival (Tarr et al. 2010 ) The magnitude of these costs will largely depe nd on whether displaced knots engage on short distance movements and continue to meet their needs. I did not address this possibility in this study. The topic of movement rates, that is, how far and how frequently an individual moves is of conservation in terest for other reasons. Knots use linear habitats (beaches), which in the context of other landscape studies, present few barriers to movement. Yet knots seem to rely on traditional areas with low exchange rates among them, at least in Tampa Bay. Obvi ous benefits from returning to these areas include familiarity with resources and threats, particularly if favorable in years past. Indeed evidence at a larger scale indicated that adults and juveniles (0.81 and 0.70, respectively) return annually to Flor ida (Chapter 2) and knots in at least one other wintering population show high inter annual site fidelity (Leyer et al. 2006). Traditionally used areas, therefore, serve a demographic function, providing overwinter requirements to multiple segments of the rufa population wintering in Florida. These areas in Florida could be viewed as the more discrete units (e.g., wetlands) used to formulate the

PAGE 55

55 conservation underpinnings of the Hemispheric Shorebird Reserve Network (Myers et al. 1987). Generally, it is believed that persistence of a species is a function of its numeric strength provided the fate of individuals is independent. With shorebirds, however, the numeric strength of the population is undermined because they aggregate in few, traditionally used areas both during migration and the non breeding season. Thus persistence effectively becomes a function of the availability of such areas (Myers et al. 1987). Beach areas in Florida will likely become more discretized in the advent of climate change (e .g., sea level rise causing fragmentation and erosion of coastal habitat) and continued human encroachment. Even though coastal beaches facilitate structural connectivity, I have presented evidence that functional connectivity is dependent on inter site d istance and habitat quality. Additional attention should be given to elucidating the scales at which these processes operate in coastal beach habitats so that increased knowledge could inform conservation design and planning. In this study I documented im portant aspects of the local dynamics of wintering knots. I found that most birds reside for extended periods of time in traditional winter areas, seemingly moving only if necessary. I also documented that birds are maintaining themselves energetically. Of the locations under consideration, knots at Indian Shores were most vulnerable to high human disturbance and low prey densities. Exchange among these traditional areas was minimal, emphasizing the importance in maintaining the integrity of traditional ly used areas. Thus, attention should also be given to the geographic context within which traditional areas are imbedded to determine if some measures of protection could be implemented (e.g., alternative

PAGE 56

56 habitats). As suggested above, unnecessary movem ents could expose these birds to increased predation, loss of foraging opportunities and perhaps higher mortality ( Parent and Farmer 1997 ). This work showed how variation in site quality and other factors could account for the differences in mean length o f stay. But even during a year of low mean length of stay (i.e., 2009), presumably caused by birds that permanently emigrated from the region, annual survival for that winter was not the lowest estimate available over a five year period (Chapter 2). This lends support to the belief that local dynamics contribute to annual survival variation, but cannot fully account for recent population declines.

PAGE 57

57 Table 3 1. Model selection for daily, local survival of Red Knots in the Tampa Bay region (winters 2008 200 9 and 2009 2010). Model QAICc QAICc QAICc Weights Model Likelihood Num. Par QDeviance {Phi(g) p(g*t) PIM} 3614.286 0 0.6159 1 33 3546.710 {Phi(g+BC) p(g*t) PIM}} 3615.230 0.9444 0.3841 0.6236 34 3545.558 {Phi(g*t) p(t) PIM} 3668.002 53.7162 0 0 37 35 92.022 {Phi(g*TT+BC) p(t) PIM} 3741.341 127.0552 0 0 21 3698.698 {Phi(g*T) p(t) PIM} 3754.162 139.8761 0 0 19 3715.633 {Phi(g*t+TT) p(g) PIM} 3773.527 159.2409 0 0 8 3757.427 {Phi(g*t+TT+BC) p(g) PIM} 3775.549 161.2627 0 0 9 3757.424 {Phi(g*TT) p(g*TT ) PIM} 3790.720 176.4346 0 0 8 3774.621 {Phi(g*TT+BC) p(g*TT) PIM} 3792.720 178.4340 0 0 9 3774.596 {Phi(g*t+T) p(g) PIM} 3799.422 185.1361 0 0 6 3787.364 {Phi(g*T) p(g*T) PIM} 3815.113 200.8274 0 0 6 3803.055 {Phi(g) p(g) PIM} 3865.412 251.1259 0 0 4 3857.384 {Phi(g) p(g*T) PIM} 3866.445 252.1593 0 0 5 3856.404 {Phi(.) p(.) PIM} 3894.693 280.4073 0 0 2 3890.685 {Phi(g) p(t) PIM} 3896.658 282.3720 0 0 17 3862.233 Parameters were: Phi = daily survival, g = year, p = detection probability, t = time, B MI = body mass index, T = linear trend in time, TT = quadratic trend in time. Lower QAICc weight indicates better support for a model. = 1.18

PAGE 58

58 Table 3 2. Model selection for survival and movement rates of Red Knots wintering in th ree traditionally used sub regions within the Tampa Bay region (winters 2008 2009 and 2009 2010). Model AICc AICc AICc Weights Model Likelihood Num. Par Deviance {S(T/site+prey+T*p rey), p (By YR but = site ), Psi (Dist)} 5277.45 0 0.675 1 14 5249.16 {S(T/site +prey+T*Prey+BMI), p (By YR but = sites ), Psi (Dist)} 5279.09 1.634 0.298 0.441 15 5248.76 {S(TT/site +prey+TT*Prey), p (By YR but = sites ), Psi (Dist)} 5285.55 8.098 0.012 0.017 15 5255.22 {S(TT/site +prey+TT*Prey+BMI), p (By YR but = sites ), Psi (Dist)} 5287.23 9.773 0.005 0.007 16 5254.85 {S(T/site +prey), p (By YR but = sites ), Psi (Dist)} 5287.35 9.89 0 0.004 0.007 13 5261.09 {S(TT/site ), p (By YR but = sites ), Psi (Dist)} 5288.09 10.636 0.003 0.004 13 5261.84 {S(TT/site +prey), p (By YR but = sites ), Psi (Dist)} 5289.96 12.503 0 0.002 14 5261.67 {S(T Indian, constant for other sites ), p (BY YR but = sites), Psi (constant /year)} 5300.59 23.138 0 0 12 5276.38 {S(yr), p (yr), Psi (yr) } 5303.52 26.069 0 0 23 5256.75 Parameters were: S = daily survival, p = detection prob ability, Psi = transition probability, BMI = body mass index, prey = average prey levels, Dist = distance between sub regions, T = linear trend in time, TT = qua regions. Lower AICc weight indicates better support for a model.

PAGE 59

59 Table 3 3. Beta parameter estimates for the multi state model w ith highest support for Red Knots wintering in the Tampa Bay Region (winters 2008 2009 and 2009 2010 ). Parameter Estimate SE LCI UCI Anna Maria daily survival 6.013789 0.331431 5.364183 6.663394 Indian Shores daily survival 1.27312 0.287554 1.83672 0.70951 Longboat daily survival 1.94492 0.348 822 1.26123 2.62861 T (linear trend in daily survival) 0.21452 0.027629 0.26867 0.16037 P rey 0.34361 0.095803 0.53138 0.15583 T*prey (interaction term) 0.026613 0.006395 0.014078 0.039148 capture probability, yr 1 1.32266 0.086013 1.49125 1.15 408 capture probability, yr 2 0.56768 0.06731 0.6996 0.43575 Inter cept 0.69139 0.87467 2.40575 1.022962 Dist Group (Indian Shores) 0.75223 0.413734 1.56315 0.058687 Dist Group (Anna Maria & Long Boat) 0.44065 0.864196 2.13447 1.253175 Dist anc e among sites 0.02858 0.028742 0.08491 0.027754 Dist*area (Indian Shores) 0.01167 0.019834 0.05055 0.027202 Dist*areas (Anna Maria & Long Boat) 0.02263 0.026364 0.0743 0.029044 The top model was S(T/state +prey+T*prey) p (By year but equal among s tates), Psi (Distance).

PAGE 60

60 Table 3 4. Results of plasma metabolite analysis conducted on plasma collected within the Tampa Bay region during late October/early November, January and late February/early March of winters 2008 2009 and 2009 2010. Metabolite Seasonal Comparison Pr > t or Pr>|Z| Triglyceride Oct Jan <0.0001* (t= 6.08, DF=192 ) Oct Mar <0.0001 (t=6.87, DF=192 ) Jan Mar 0.0971 (t=1.67, DF=192 ) OH Butyrate Oct Jan 0. 0012 (Z= 3.25 n=117) Oct Mar 0.0010* (Z=2.58, n=115) Jan Mar 0.0785 (Z=1.77, n= 162) Glycerol Oct Jan 0.0092* (Z=2.61, n=115) Oct Mar 0.0266* (Z=2.22, n=116) Jan Mar 0 .2553 (Z= 1.14, n=163) indicates significance at Tests for triglyceride were conducted using ANOVA t values and degrees of freedom (DF) are reported ; tests for the other two metabolites were conducted with a n on parametric test Z scores and sample size are reported

PAGE 61

61 Figure 3 1 Study are a Tampa Bay region. Labeled locations indicate main sites for resighting and invertebrate surveys and yellow outlines indicate sub regions (states) for resident and multi state models.

PAGE 62

62 Figure 3 2. Daily survival probabilities across season for thr ee locations in the Tampa Bay region Estimates were obtained from the top multi state model: S(T/state +prey+T*prey) p (by year but equal among states), Psi (Distance).

PAGE 63

63 Figure 3 3. Daily transition probabilities between sub regions in the Tampa Bay r egion. Indian corresponds to the Indian Shores or northern sub region, Anna corresponds to the Anna Maria or central sub region, and LongB corresponds to the Longboat or southern sub region. Estimates were obtained from the top multi state model: S(T/sta te +prey+T*prey) p (by year but equal among states), Psi (Distance).

PAGE 64

64 Figure 3 4. Plasma metabolites levels with standard error for the Tampa Bay region over the seasons. 2008 2009 and 2009 2010 data and data from all locations pooled.

PAGE 65

65 Figure 3 5. Weekly mean densities of preferred prey over the winter and spring (January 18 May 11, 2010) at three sub regions within Tampa Bay. Lines represent the model predictions from each location.

PAGE 66

66 CHAPTER 4 CONCLUSION I examined demographic and physiologic al aspects of Red Knots ( Calidris canutus rufa ) wintering in Florida. I adopted a multi scale approach in an attempt to decouple their annual cycle and differentiate between local and global (hemispheric) factors that might influence population changes. This distinction provides a stronger basis to formulate local conservation strategies. The strongest impetus for the study was the sharp decline in numbers recorded in Florida in the last 30 years. Winter counts have dropped from ~10,000 (1980s) to ~1,20 0 (2010). The relevance of this study is further strengthened because the remainder of the hemispheric population of Red Knots, which winters in South America, also exhibits a declining population trend. Assessments at multiple spatial and temporal level s are needed to develop a comprehensive At the broadest scale, I estimated annual survival of juveniles and adults, and compared them to similar estimates from the remaining populations of rufa I also t ested for differences in size adjusted body mass between Florida birds and other wintering populations as a measure of relative health. Comparisons of adult annual survival rates indicated that birds wintering in Florida survived at similar rates (0.91) a s birds from the remaining populations of rufa (0.91 0.92). Comparisons of body mass between birds wintering in Florida and South America were either similar or favored Florida birds. These findings suggest that the mechanism(s) behind the observed popul ation decline were likely acting on all populations of rufa not just the segment wintering in Florida. In addition this work and recent work at Delaware Bay leads me to suggest that adult mortality was not the primary mechanism for population declines a t

PAGE 67

67 least between 2005 2010, and plausible alternatives should be considered. Among these depressed reproductive or recruitment rates should be given greatest attention. These findings also suggest that despite their smaller numbers and northern wintering grounds relative to the core wintering population in South America, knots wintering in Florida should not be viewed as a marginal population. Ecological factors such as availability of food resources may explain the unequal distribution of the populations instead of trade offs in annual survival. At the local scale, I estimated daily survival and mean length of stay for each of three historically wintering locations within the grea ter Tampa Bay region These were Anna Maria, Indian Shores and Longboat. These areas spanned a range of habitat quality conditions as reflected by food resource abundance and degree of development. I also estimated transition probabilities among the three traditionally used locations to assess connectivity among them. Results indicated that daily survival was influenced by prey abundances and by location. Daily survival was highest at Anna Maria and lowest at Indian Shores. Temporal fluctuations in daily survival probabilities at Indian Shores suggested that time specific in creases in disturbance, such as spring break, could have an impact on daily survival. Transition probabilities among locations was low (< 0.08), with only one transition (from Longboat to Anna Maria) suggesting that locations could be functionally connect ed. Knots tend to reside in traditionally used areas unless conditions change markedly. Low transition probabilities, however, indicate that birds were reticent to move between traditionally used areas. As such, they are not perceived as potential alter natives or refugia by knots.

PAGE 68

68 Management Recommendations The mechanisms behind the decline of Calidris canutus rufa are not entirely understood despite the severe magnitude of decline. The parameters and scales used to assess the status of the species in F lorida did not point at local factors as the major drivers behind observed declines over the past five years However, managers in Florida should be cognizant of several patterns highlighted in this study. First, the probability of returning to Florida o n an annual basis was high (0. 70 0.81). Second, birds consistently reside in and return to historically used areas unless conditions change markedly or they are forced out. Third, low transition probabilities among traditionally used areas suggested that they are not viewed by birds as alternative habitats. The latter was exemplified by displacements recorded in Indian Shores, likely caused by spikes in human recreational activities during spring. These patterns underscore the importance of preserving the integrity of traditionally used areas, but also foster the implementation of measures to protect or buffer other areas from anthropogenic disturbance. These measures will gain importance as the coastal zone and beaches of Florida becomes increasingly segmented as a result of continued development and sea level rise. The creation of protected areas or other refugia should take into account the scale at which birds perceive the landscape. Refugia will only be effective if knots can transition to the lo cations easily. Their reluctance to travel greater distances, such as those between the three study areas, would limit the usefulness of a protected area if it is beyond the range that knots perceive as being connected to their wintering grounds.

PAGE 69

69 Recommen dations for Further Research On a hemispheric scale, obtaining estimates of reproductive success and recruitment rates is imperative. These vital parameters are emerging as the possible source of demographic vulnerability for C. c rufa and perhaps the un derlying cause of recent population declines. While juvenile to adult ratios may be useful for inference in some areas, such as Florida, they are likely inadequate for larger regions such as the wintering grounds in southern South America. Other research could also explore the effects of sub lethal factors, such as disease. While my study did not find evidence of increased adult mortality within Florida compared to South America, Florida knots seem to show a higher incidence of avian influenza antibodies than South American birds (A. Howey, pers. comm.), suggesting that they may still be vulnerable to increased disease and parasite pressure. Broader immunological studies could examine potential differences between and impacts on the wintering populations In Florida, understanding the scale at which knots consider the landscape to be connected is an important next step. My work highlighted the discrete nature of these historical wintering areas, yet the distance within these areas that knots will willing ly travel is unknown. Estimating the upper limit for this distance will be key in designing adequate protections for knots within each of these areas. Additionally, research directed at separating the effect of anthropogenic disturbance from the effect o f resource abundance on overwinter survival would be useful in determining the level of influence each factor exerts and understanding better how local factors might impinge on population demography Finally, estimating site fidelity, both within a winter and between winters, could improve our understanding of local dynamics and may help

PAGE 70

70 elucidate if habitat alterations within the wintering season are more or less detrimental than impacts on habitat quality between years.

PAGE 71

71 APPENDIX A PREY SELECTION IN WI NT ERING RED KNOTS IN T HE TAMPA BAY REGION Red Knots forage in a variety of habitats and on a variety of prey items throughout their annual cycle, from terrestrial invertebrates on the breeding grounds to intertidal invertebrates on migration and on the winte ring grounds. Within the rufa subspecies most wintering and migratory populations consume primarily small mollusks within the intertidal zone, particularly the clam genus Donax ( Harrington 2001 ). The goals of this study were: 1) document temporal pattern s in preferred prey items ( Donax mole crabs genus Emerita marine worms family Spionidae ), 2) determine prey preferences within Florida wintering knots by comparing background levels of available prey with prey levels found at foraging locations and 3 ) provide additional evidence through examination of stomach contents obtained through incidental catch mortality. Methods Study Area This study was conducted in the Tampa Bay region on the Gulf Coast of Florida. Tampa Bay is home to approximately half o Niles et al. 2008a ). Knots arrive in late October and early November and stay through late March to mid April. A small number of birds (<50), most likely juveniles, stay in the Tampa Bay area over the summer. Shore bird habitat in Tampa Bay consists of sandy beaches on barrier islands with a narrow intertidal zone. At lower tides, sand or mud flats are exposed and provide additional foraging habitat. I divided the Tampa Bay region into three main sites (Anna Maria, Indian Shores, and Longboat) which each represent distinct historical wintering areas

PAGE 72

72 Invertebrate Sampling refers to sampling conducted every week at the same locations to d evelop a picture of sampling at locations where birds are actively feeding. Reference sampling I sampled for invertebrates at Ft. De Soto, Shell Key, Lido Key, Longboat Key and Indian Shores in 2010. I selected these locations based on the consistent presence of foraging Red Knots in previous years. I established two reference transects at each of the five locations Each transect consisted of a fixed starting point, with fou r points total spaced ten meters apart. Transects were placed randomly along selected beaches and were at least 200 m apart. I sampled both transects weekly at each location starting the third week of January. I sampled in the surf zone, parallel to the waterline. At each point I collected sediment cores 5 cm deep and 10 cm in diameter and stored them in a lock seal plastic bag for future analysis. After sample collection, I sieved each sample with a 1 mm wide mesh sieve. I collected, identified to fam ily (or genus if possible) and counted each invertebrate in the sample. I measured common invertebrates in order to create an index of biomass. Foraging sampling When I located a flock of foraging I placed a linear transect through the foraging area, typi cally along the waterline. I took the first sample at the near edge of the flock and continued sampling every ten meters until I reached the far edge of the flock. Since flocks often abandoned feeding while I sampled, I noted the location of the near and far

PAGE 73

73 edges before we sampling. I took cores 10 cm in diameter and 5 cm deep and processed them in the same manner as described for the reference samples. Stomach contents Trapping efforts, regardless of precautions taken, inevitably result in some minimal level of mortality. During 2008 2010, trapping efforts produced 16 mortalities which were collected for later analysis under the Florida Fish and Wildlife Conservation MB745817 0 ). I collected all 16 stomachs and preserved t heir contents in ethanol. I then attempted to identify all prey items and determine the number of each prey type found in each stomach. Data Analysis In order to examine temporal trends I compared estimates of prey density using linear model (PROC MIXED, SAS 2002). For the combined preferred prey items m odel terms were region (Anna Maria, Long Boat, Indian Shores), week (16 weeks, week 8 was excluded due to lack of samples), and interaction terms (i.e., week*region, week*week, week*week*region, week*week* week). This model treated site [region] and transect [region*site] as random effects. The response variable, prey density, was log transformed prior to running the model. The Donax model differed from the overall model by including an additional interac tion term (week*week*week*region) and removing transect [region*site] from the random effects. The Emerita model included all terms from the overall model but also removed transect [region*site] from the random effects. The Spionidae model included all t he terms and random effects from the overall model but also included the additional week*week*week*region term. Differences between reference and foraging samples were analyzed using a non parametric test due to the non normal distribution of the data (PRO C npar1way SAS

PAGE 74

74 2002) Data from each group was separated by month (January May) and region within Tampa Bay and differences in means between groups was tested using the Wilcoxon two 0.05. Results Reference Sa mpling All items increased in all areas over time, although the magnitude of change for each i tem differed by area (Figures A 1 and A 2). By the end of the season Donax and Emerita were most abundant in the Longboat region while Spionidae worms were most abundant in the Indian Shores region. The Indian Shores region consistently showed lower levels of Donax throughout the season, the knots most preferred item (see Foraging Sampling Comparing reference samples to foraging samples showed few significant differences in available prey between the two types of samples, and several significant results indicated a higher presence of items in the reference samples than in th e foraging samples (Table A 1). The notable exception i s that Spionidae worms were present in significantly larger numbers in foraging samples than in reference samples at Indian Shores during every month in which foraging samples were taken (January March). No foraging samples were taken at Indian Shores dur ing April and May, since birds had largely vacated the area during these months. No foraging samples were taken in the Anna Maria area during April due to a lack of observed foraging behavior. Stomac h C ontents Donax were found in 93.8% of the stomachs, C y praeidae in 18.8%, and all other item types in 12.5% or less. Only one stomach did not contain Donax Cypraeidae a

PAGE 75

75 type of snail, was found in large quantities (38 individuals) in one stomach but all individuals were les than 2mm. After discounting ite ms smaller than 2mm, Donax represented 69.7% of all items found in the stomachs, with Crassatellidae clams being the next most frequent at 10.6% (Table A 2). Birds from the Longboat region showed a focus on Donax while birds from Ft. De Soto in the Anna M aria region had a more diverse diet, although samples sizes from both regions were small. No birds were collected from the Indian Shores region. Discussion Based on stomach sampling, Donax clams seem to be the most preferred prey item in the Tampa Bay reg ion. In addition, larger flocks regularly gathered in the Longboat and Anna Maria regions than in Indian Shores (pers. obs., resighting surveys), indicating that knots may select areas with higher Donax concentrations. However it is also clear that knots will utilize a variety of mollusks including clams other than Donax snails, and mussels. In addition, non systematic foraging observations indicate that Spionidae worms are eaten by red knots in the Tampa Bay area, and the invertebrate sampling results from Indian Shores may indicate that they provide an important alternative in a region where the normally preferred Donax are scarce. Two reasons that Spionidae were likely under represented in the stomach samples are: 1) no incidental mortality came fro m Indian Shores and 2) Spionidae are soft bodied and thus quickly digested, suggesting that they may be less likely to be detected depending on length of time from last meal for each individual bird.

PAGE 76

76 Table A 1. Comparison of reference and foraging samples from the Anna Maria, Indian Shores and Longboat regions in Tampa Bay during the winter and spring (January 18 May 11, 2010). January February March April May Anna Maria Donax 0.0084* (n=17, Z=2.64) 0.0689 (n=58, Z= 1.82) 0.2077 (n=78, Z=1.26) n/a 0.0599 (n=41, Z= 1.88) Emerita 1.0000 (n=17, Z=0) 0.4368 (n=58, Z= 0.78) 0.6641 (n=78, Z= 0.43) n/a 0.0624 (n=41, Z= 1.86) Spionidae 0.0036* (n=17, Z= 2.91) <.0001* (n=58, Z=4.01) 0.8467 (n=78, Z= 0.19) n/a 0.0345* (n=41, Z= 2.11) Indian Shores Donax 0.4160 (n=22, Z= 0.81) 0.6755 (n=41, Z=0.42) 0.3641 (n=41, Z= 0.91) n/a n/a Emerita 0.2066 (n=22, Z= 1.26) 0.7576 (n=41, Z= 0.31) 1.0000 (n=41, Z=0) n/a n/a Spionidae 0.0011* (n=22, Z=3.27) <.0001 (n=41, Z=4.80) 0.0038* (n=41, Z=2.90) n/a n/a Longbo at Donax 0.7430 (n=20, Z= 0.34) 0.1768 (n=75, Z=1.35) 0.5773 (n=114, Z=.56) 0.7852 (n=65, Z= 0.27) 0.3248 (n=35, Z= 0.98) Emerita 0.0476* (n=20, Z= 1.98) 0.2618 (n=75, Z= 1.12) 0.1170 (n=114, Z= 1.57) 0.1621 (n=65, Z=1.40) 0.9247 (n=35, Z= 0.09) Spioni dae 0.4002 (n=20, Z=0.84) 0.2424 (n=75, Z= 1.17) 0.9228 (n=114, Z= 0.10) 0.2613 (n=65, Z=1.12) 0.2539 (n=35, Z= 1.14) P values, sample sizes and Z scores for a Wilcoxon two sample, two tailed test of means from reference and foraging samples, broken down by item type. indicates significant result. n/a indicates no foraging sample available for that month.

PAGE 77

77 Table A 2. Number of items and percentage of total items by prey type found in 16 Red Knot stomachs from the Tampa Bay region. Location Dat e Donax Crassatellidae Cassidae Cypraeidae Mytilidae Spionidae Unknown snail Ft. De Soto 3/19/2008 1 6 1 Ft. De Soto 3/19/2008 2 Ft. De Soto 3/19/2008 1 2 2 Ft. De Soto 3/19/2008 3 Ft. De Soto 3/19/2008 2 Ft. De Soto 3/19/20 08 1 2 Ft. De Soto 3/19/2008 2 Ft. De Soto 3/19/2008 3 3 1 Ft. De Soto 3/19/2008 1 Ft. De Soto 3/19/2008 4 Ft. De Soto 3/19/2008 3 Ft. De Soto 3/19/2008 3 Ft. De Soto 3/19/2008 3 1 Longboat Key 10/25/2009 9 Longboat Key 1/7/2010 6 Longboat Key 1/7/2010 4 FREQUENCY 46 7 4 5 1 2 1 % all items 69.7 10.6 6.1 7.6 1.5 3.0 1.5

PAGE 78

78 Figure A 1. Weekly mean densities of all preferred prey over the winter and spring (January 18 May 11, 2 010) at three sub regions within Tampa Bay. Lines represent the model predictions from each location.

PAGE 79

79 Figure A 2. Weekly mean densities of Donax Emerita and Spionidae over the winter and spring (January 18 May 11, 2010) at three sub regions wit hin Tampa Bay. Lines represent the model predictions from each location.

PAGE 80

80 APPENDIX B PREY AVAILABILITY AN D SELECTION IN MIGRA TORY RED KNOTS IN TH E CEDAR KEY REGION The vast majority of rufa red knots pass through Delaware Bay on their northbound migration to the breeding grounds ( Harrington 2001 ). Here they feed almost exclusively on horseshoe crab ( Limulus polyphemus ) eggs before completing their final non stop flight to the Arctic breeding grounds ( Tsipoura and Burger 1999 ). Florida supports a healthy breeding population of horseshoe crabs, particularly in the Cedar Key region of the Gulf coast. Although breeding densities do not reach the super abundance found in Delaware Bay, local breeding concentrations can be high ( Gerhart 2007 ). M igratory red kn ots are known to pass through the Cedar Key region, located on the north central Gulf Coast of Florida, during the peak horseshoe crab spawning season. Thus this study was designed with two goals: 1) determine what the availability of eggs within reach of shorebirds is compared to the number laid and 2) examine the importance of horseshoe crab eggs and larvae to migratory knots in Florida. Methods Study Area The Cedar Key area consists of large swaths of saltwater marsh and oyster beds, interspersed with s mall islands with sandy beaches. Despite the relative limited size of sandy beaches, the shallow waters and low wave action make this region an ideal place for spawning horseshoe crabs. This area has some of the highest spawning crab densities in Florid a during the spring months (March May) ( Gerhart 2007 ). It also provides habitat for wintering and migrating shorebirds. Within the region I chose two

PAGE 81

81 sites, McClamory Key and Atsena Otie, which provided both good habitat for crab spawning and potential f oraging habitat for red knots ( Figure B 1 ). Reference S ampling I sampled each site weekly, except when weather did not permit access to the islands, during the spring months (late February to mid May). I sampled each location for three years during 2008 2 010. Within each of the sites I chose transect locations at random. In the first year, I chose two transects at Atsena Otie and one transect at McClamory Key. Due to the small amount of suitable habitat I was unable to place two transects on McClamory K ey. Severe erosion destroyed one section of beach at Atsena Otie between the first and second years, and thus Atsena Otie was also reduced to one transect for year two and three. Each transect was 50 m long x 3 m wide along the hide tide line to encompass the area where horseshoe crabs lay their eggs (Collazo et. al 2002). I subdivided each transect into five quadra n ts (10 m long x 3 m wide), and each quadra n t into 30 plots (1 m x 1 m); each quadra n t and plot were numbered ( Figure B 2 ). I sampled three p lots from each reference site weekly. To ident ify specific plots to sample, I used a random number generator to select quadra n t and plot numbers; individual plots were not resampled during subsequent sampling periods. I sampled one core with a 10 cm diam eter and 25 cm deep within each plot and subdivided each core into a shallow (0 5 cm) and deep (5 25 cm) sample. Additionally, I walked a perpendicular line from the transect starting at the plot down to the waterline and took an additional shallow core ( 0 5 cm) since this is the area used most frequently for foraging by shorebirds. Using this method allowed us to compare what was available to the birds at the high high tide line I stored each sample in a lock se al plastic bag

PAGE 82

82 for future analysis. After sample collection, I sieved each sample with a 1 mm wide mesh sieve. I collected, identified to family (or genus if possible) and counted ea ch invertebrate in the sample. Foraging S ampling For the second methodol ogy I placed a linear transect along the foraging area, typically along the waterline. When I located a flock of foraging knots I took the first sample at the near edge of the flock and continued sampling every ten meters until I reached the far edge of t he flock. Since flocks often abandoned feeding while I sampled, I noted the location of the near and far edges before sampling. I took cores 10 cm in diameter and 5 cm deep and processed them in the same manner as described for the reference samples. Red Knot Surveys I surveyed shorebirds in the Cedar Key region at week ly intervals from late February to mid May. Each survey was completed within a day. Survey locations included the Cedar Key Municipal Beach, Atsena Otie, Seahorse Key, North Key (until su itable habitat eroded away in 2009), Rattlesnake Key, McClamory Key, Derrick Key and parts of the Shell Mound area (Figure B 1). Surveys were conducted by boat and/or land, depending on beach topography and accessibility, using two observers when possible to verify species identification and count. Once a flock with red knots was identified I quantified each species then scanned red knots for leg band s and recorded band information (flag codes and color combinations) Data Analysis Availability of horses hoe crab eggs/larvae (i.e. differences between shallow and deep cores) and differences between reference and foraging samples were analyzed

PAGE 83

83 using a non parametric test due to the non normal distribution of the data (PROC npar1way SAS 2002) Data from eac h group was separated by month (Mar May) and year and differences of means between groups were tested using the Wilcoxon two 0.05. Results Overall densities of horseshoe crab eggs/larvae in the reference sample s available to birds were low ( 5.61 eggs/larvae per sample or less, Table B 1 ) although individual samples could contain up to 62 eggs/larvae in shallow samples and 841 eggs/larvae in deep samples Availability also varied from year to year and site to s ite. In 2008 more eggs were available on average at McClamory Key than at Atsena Otie, while in 2009 this trend was reversed. In 2010 there were no eggs available to the birds at all at either site. Eggs buried at depths between 5cm an d 25cm were signif icantly more numerous than eggs available to the birds in April of 2009 and 2010 ( Table B 2 ). Otherwise densities at any depth were too low to show significant differences. Foraging samples had significantly more horseshoe crab eggs/larvae during April an d May of 2008 ( p=0.0004 (Z=3.56, n= 50) and p=0.0002 (Z=3.76, n=28) respectively ) but not in April of 2010 since there were no eggs/larvae in either reference or foraging samples (n=29) Foraging data is sparse however, with no data available for 2009 or March of any year. Most of these areas covered by this study were sandy beaches where red knot and horseshoe crab habitat would overlap. However the Cedar Key area supports extensive mud flats at low tide and it is likely that knots were utilizing this h abitat. The inaccessibility of these flats made it nearly impossible to survey or sample them although knots were observed foraging on flats in the Shell Mound area during 2010 (pers. obs.).

PAGE 84

84 Peak counts for bird surveys in the area ranged from 0 269 with overall peak counts of 121 in 2008, 100 in 2009, and 269 in 2010. L arge numbers were only present from mid April to early May each year. Red knots were only found in the area encompassing McClamory Key, Derrick Key and the Shell Mound area. A total of 28 banded individuals were seen over the three years. Discussion The results above demonstrate that although red knots may preferentially choose horseshoe crabs eggs and larvae for food when opportunistically available horseshoe crab eggs and larvae likel y do not serve as an important food item for red knots migrating out of or through Florida. The lack of observed foraging data for multiple months during all three years suggests that knots are regularly using inaccessible mud flats instead of sandy beach es for their foraging. Based on the biology of the horseshoe crab, one can conclude that no eggs or larvae would be available on these flats. Peak counts of knots in the Cedar Key region are very small compared to peak counts at other migratory sites in northeast Florida and southeast Georgia (P. Leary, pers. comm.) suggesting that it is not a primary stopover for migratory knots in Florida In fact, the peak counts and band resights in Cedar Key constitute only a fraction of the entire southwest Florida knot population (~27 0 birds versus ~800 birds) indicating that In addition, the majority o f banded knots (i.e. 26 out of 28 ) resighted in Cedar Key come from the southwest Florida area an d none were banded outside of the United States, demonstrating that the availability of horseshoe crab eggs does not attract interest from other migratory populations. Furthermore (but low level) presence in the Cedar Key region eve n during low spawning years and the complete absence of

PAGE 85

85 knots observed on Seahorse Key, the island with the highest horseshoe spawning densities in the area (J. Brockmann, pers. comm.) provides evidence that knots in this region are selecting for factors other than the horseshoe crab. Although I did not determine what these other factors may be, this conclusion makes sense given that the highly fluctuating levels of horseshoe crab eggs available from year to year would not allow migratory birds to rely on them as a consistent food source. Migratory birds often rely on the predictability of food sources at migratory stopovers (Bairlein and Gwinner 1994) and this resource would be unsuitable for this purpose.

PAGE 86

86 Table B 1. Mean horseshoe crab egg/larvae co unts by site for 2008 2010. Year Month Site M ean std 2008 March AO 0 0 MC 0.58 1.64 April AO 0.06 0.24 MC 0.44 0.78 May AO 0.08 0.29 MC 0.83 1.34 2009 March AO 0 0 MC 0.50 0.84 April AO 5.67 14.52 MC 3.44 14.61 May AO 0.50 1.24 MC 0 0 2010 March AO 0 0 MC 0 0 April AO 0 0 MC 0 0 May AO 0 0 MC 0 0 Means and standard deviations were obtained from reference samp les taken during each month from March to May. Cores were 5 cm deep and had a 10 cm diameter. For site, AO = Atsena Otie and MC = McClamory Key.

PAGE 87

87 Table B 2. Differences in horseshoe crab eggs/larvae buried compared to eggs available. 2008 2009 2010 March 0.5562 (Z= 0.59, n=81) 0.9425 (Z=0.07, n=18) 1.000 (Z=0.0, n=36) April 0.3761 (Z=0.89, n=54) 0.1945 (Z =1.30, n=55) *0.0470 (Z=1.99, n=36) May 0.3425 (Z=0.95, n=36) *0.0412 (Z=2.04, n=36) 0.1753 (Z=1.36, n=36) P values, Z scores, and sample sizes for a Wilcoxon two sample, two tailed test of means from deep and shallow samples, broken down by month and ye ar. indicates significant result.

PAGE 88

88 Figure B 1. Map of Cedar Key region with major resighting and invertebrate spawning locations.

PAGE 89

89 Figure B 2. Example invertebrate reference sampling plot. Two cores (0 5 cm and 5 25 cm) would be taken at the point and a third core (0 5 cm) would be taken at the waterline.

PAGE 90

90 LIST OF REFERENCES Acevedo Seaman D A C G Guglielmo, R W Elner, and T D Williams. 2006. Landscape scale physiology: site differences in refueling rates indicated by plasma metabolite a nalysis in free living mi gratory sandpipers. Auk 123: 563 574. Alerstam, T. and A. Lindstrm. 1990. Optimal bird migration: the relative importanc e of time, energy, and safety. Pages 331 351 in E. Gwinner, editor. Bird Migration: th e physiology and ecophys iology. Springer, Berlin, Germany. Atkinson P W A.J. Baker K.A. Bennett N.A. Clark J.A. Clark K.B. Cole A.D. Dey S. Gillings P.M. Gonzalez K. Kalasz C.D.T Minton L.J. Niles R.A. Robinson and H.P. Sitter s 2007. Rates of mass gain and energy deposition in Red Knot on their final spring staging site is both time and condition depend ent. Journal of Applied Ecology 44:885 895 Bairlein F. and E. Gwinner. 1994. Nutritional mechanisms and temporal control of migratory energy accumulation in birds Annual Review of Nutrition 14:187 215 Baker A J P.M. Gonzalez, T. Piersma L.J. Niles I.L. Serrano P.W. Atkinson N.A. Clark, C.D.T. Minton M.K. Peck and G. Aarts 2004. Rapid population declines in red knots: fitness consequences of decreased ref ueling rates and late arrival in Delaware Bay. Proceedings of the Royal Society of London, Series B. 271:875 882. Baker A J P.M. Gonzalez I.L. Serrano W.R.T. Junior, M. A. Efe, S. Rice, V. Amico, M.C. Rocha, and M.E. Echave. 2005 Asses sment of th e wintering area of red k nots in Maranho, northern Brazil in February 2005 Wader Study Group Bull. 107: 10 18 Barker, R.J. 1997. Joint modeling of live recapture, tag resight, and tag recovery data. Biometrics 53:666 677. Bart, J., C. Ke pler, P. Syke s, and C. Bocetti. 1999. Evaluation of mist net sampling as an index to produ Auk 116:1147 1151. Belisle, M. 2005. Measuring landscape connectivity: The challenge of behavioral landscape ecology. Ecology 86:1988 1995. Belth off, J.R. and S.A. Gauthreaux. 1991. Partial migration and differential winter distribution of house finches in the eastern United States. Condor 93:374 382. Brownie, C., K.H. Pollock. 1985. Analysis of multiple c ap ture recapture data using band recovery m ethods Biometrics 41: 411 420

PAGE 91

91 Burnham, K.P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information theoretic approach. 2nd Edition. Springer Verlag, New York, New York, USA. 488 pp. Cerasale D J and C.G. Guglielm o 2006. Dietary effects on prediction of body mass changes in birds by plasma metabolites. Auk. 123:836 846. Collazo, J.A., D.A. O'harra, and C.A. Kelly. 2002. Accessible habitat for shorebirds: factors influencing its availability and conservation impli cations. Waterbi rds 25 (Special Publication 2): 13 24. Dietz, M.W., S. Jenni Eiermann, and T. Piersma. 2009. The use of plasma metabolites to predict weekly body mass change in red knots. Condor 111:88 99. Dinsmore, S.J., and J.A. Collazo. 2003. The infl uence of body condition on local apparent survival of spring migrant sanderlings in coastal North Carolina. Condor 105 :465 473. Farmer, A.H., and A. H. Parent. 1997. Effects of the landscape on shorebird movements at spring migration stopovers. Condor 99:6 98 707. Farmer, A.H. and J. A. Weins. 1999. Models and reality: time energy trade offs in Pectoral Sandpiper ( Calidris melanotos ) migration. Ecology 80:2566 2580. Fernandez, G., H. de la Cueva, N. Warnock, and D.B. Lank. 2003. Apparent survival rates of W estern Sandpiper (Calidris mauri) wintering in Nort hwest Baja California, Mexico. Auk 120:55 61. Gentle, L. K. and A. G. Gosler. 2001. Fat reserves and perceived predation risk in the Great Tit, Parus major. Royal Society of London Proceedings, Biologi cal Sciences 268:487 491. Gerhart, S.D. 2007. A review of the biology and management of horseshoe crabs, with emphasis on Florida populations. Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute TR 12. 24 p. Guglielmo C G D.J. Cerasale and C. Eldermire 2005. A field validation of plasma metabolite profiling to assess refueling performance of migratory birds. Physio logical and Biochemical Zoology 78:116 125. trinsic and intrinsic sources of variation in plasma lipid metabolites of free living western sandpipers (Calidris mauri). Auk 119:437 445. Haramis G M W A Link, P.C. Osenton D.B. Carter R.G. Weber N.A. Clark M.A. Teece and D.S. Mizrahi 2007. Sta ble isotope and pen feeding trial studies

PAGE 92

92 confirm the value of horseshoe crab Limulus polyphemus eggs to spring migrant shorebirds in Delaware Bay. Journal of A vian Biology 38:367 376. Harrington, B. A. 2001. Red Knot (Calidris canutus), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America Online: http://bna.birds.cornell.edu/bna/species/563 Harrington, B.A., J.M. Hagan, and L.E. Leddy. 1988. Site fidelity and survival differences between two groups of new world red knots (Calidris canutus). Auk 105: 439 445. Jahn A E D.J. Levey, K.G. Smith 2004. Reflections across hemispheres: a system wide approach t o new world bird migration. Auk 121:1005 1013. Karpanty, S.M ., J.D. Fraser, J. Berkson, L.J. Niles A. Dey, and E.P. Smith 2006. Horseshoe crab eggs determine red knot distribution in Delaware Bay. Journal of Wildlife Management 70:1704 1710. Kelly J F V. Atudorei Z.D. Sharp and D.M. Finch 2002. Warb ler migration from analyses of hydrogen stable isotope ratios. Oecologia 130:216 221. Kendall, W.L. and J. D. Nichols. 2004. On the estimation of d ispersal and movement of birds. Condor 106:720 731. Ketterson, E.D. and V. Nolan Jr. 1976. Geographic variat ion and its climatic correlations in the sex ratio of eastern wintering Dark eyed Juncos ( Junco hyemalis hyemalis ). Ecology 57:679 693. Lauber M ., and J. Langenberg 1998. Natural history and rehabilitation of the Sandhill Crane (Grus canadensis). Journ al of Wildlife Management 21 :3 19. Leyer, J., B. Spaans, M. Camara, and Piersma T. 2006. Small home ranges and high Mauritania. Journal of Ornithology 147:376 384. Lyons, J. E. and S. M. Haig. 1995. Fat content and stopover ecology of spring migrant Semipalmated Sa ndpipers in South Carolina. Condor 97:427 437. Marra P P. 2000. The role of behavioral dominance in structuring patterns of habitat occupancy in a migrant bird duri ng the nonbreed ing season. Behavioral Ecology 11:299 308. Martin, K. 1995. Patterns and mechanisms for age dependent reproduction and survival in birds. American Zoologist 35: 340 348.

PAGE 93

93 McGowan C.P, J.E. Hines J. D. Nichols J. E. Lyons, D. R. Smith K. S. Kalasz L.J. Niles, A.D. Dey N. A. Clark P. W. Atkinson C.D.T. Minton, W. Kendall 2011 Demographic consequences of migratory stopover: linking red knot survival to horseshoe crab spawning abundance Ecosphere 2:art69 Morrison R I G N.C. Davidson a nd T. Piersma 2005. Transformations at high latitudes: why do Red Knots bring body stores to the breeding grounds? Condor 107:449 457. Morrison, R. I. G., R.K. Ross, and L.J. Niles. 2004. Declines in wintering populations of red knots in sout hern South A merica. Condor 106: 6 0 70. Morrison, R.I.G., and B.A. Harrington 19 92. The migration system of the Red Knot Calidris canutus rufa in the New World. Wader Study Group Bull. 64 (Supp.): 71 84. Myers, J., R. Morrison, P. Antas, B. Harrington, T. Lovejoy, M. Sallaberry, S. Senner, and A. Tarak. 1987. Conservation strategy for migratory species. American Scientist 75:19 26. Myers J P J.L. Maron and M. Sallaberry. 1986. Going to extremes: Why do Sanderlings migrate to the Neotropics? Neotropical Ornitholog y, Ornithological Monographs No. 36:520 534. Niles, L.J. J. Bart, H.P. Sitters, A.D. Dey, K.E. Clark, P. W. Atkinson, A.J. Baker, K.A. Bennett, K.S. Kalasz, N. A. Clark, J. Clark, S. Gillings, A.S. Gates, P.M. Gonzalez, D.E. Hernandez, C.D.T. Minton, R.I. G. Morrison, R.R. Porter, R.K. Ross, and C.R. Veitch. 2009. Effects of horseshoe crab h arvest in Delaware Bay on Red Knots: Are harvest r estr ictions working? BioScience 59: 153 164. Niles L J H.P. Sitters A.D. Dey P.W. Atkinson A.J. Baker R. Carmona K.E. Clark N.A. Clark C. Espoza P.M. Gonzalez B.A. Harrington D.E. Hernandez K.S. Kalasz R. Matus C.D.T. Minton R.I.G. Morrison M.K. Peck W. Pitts R.A. Robinson I.L. Serrano 2008a. Update to the status of the Red Knot Calidris canutus in t he Western Hemisphere. US Fish and Wildlife Service, Ecological Services, Region 5, Pleasantville, New Jersey. Niles, L. J., H.P. Sitters, A.D. Dey P.W. Atkinson, A J. Baker, K.A. Bennett, R. Carmona, K.E. Clark, N.A. Clark, C. Espoza, P.M. Gonzalez, B.A. Harrington, D.E. Hernandez, K.S. Kalasz, R.G. Lathrop, R.N. Matus, C.D.T. Minton, R.I.G. Morrison, M.K. Peck, W. Pitts, R.A. Robertson and I.L. Se rrano. 2008b Status of the Red Knot in the Western Hemisphere. Studies in Avian Biology No. 36 Niles L J A.D. Dey A.S. Gates T. Mckee, W. Pitts M.K. Peck. 2008c Investigation of Red Knot Wintering Population of the Florida Gulf Coast 2006 2008: Fi rst Interim Report (Report on Red Knot Surveys in Florida 2006 2008). Prepared for Florida

PAGE 94

94 Fish and Wildlife Conservation Commission, Division of Habitat and Species Conservation. Niles L J A.D. Dey N.J. Douglass J.A. Clark N.A. Clark A.S. Gate s B .A. Harrington M.K. Peck and H.P. Sitters 2006. Red Knots wintering in Florida: 2005/6 expedition. Wader Study Group Bull. 111: 86 99. Noy Meir, I. 1973. Desert Ecosystems: environment and producers. Annual Review of E cology and Systematics 1973.4: 25 51 Oppel S and A.N. Powell 2010. Age specific survival estimates of king eiders derived from satellite telemetry. Condor 12 :323 330. Piersma T. 1997. Do global patterns of habitat use and migration strategies co evolve with relative investments in im munocompetence due to spatial variation in parasite pressure? Oikos 80 :623 631. Piersma T. 2007. Using the power of comparison to explain habitat use and migration strategies of shorebirds worldwide. Journal of Orn ithology 148 (Suppl 1):S45 S59. Rice S M J.A. Collazo M.W. Alldredge B.A. Harrington and A.R. Lewis. 2007. Local annual survival and seasonal residency rates of Semipalm ated Sandpipers in Puerto Rico. Auk 124:1397 1406. Skagen, S. K. 1997. Stopover ecology of transitory populations: the case of migrant shorebirds. Ecological Studies 125:244 269. Skage n, S.K., and F.L. Knopf. 1994 Residency patterns of migrating sandpipers at a mid continental stopover. Condor 96:949 958. Smith, D.R., J.D. Nichols, J.A. Collazo. 2008. Adaptive Managemen t of Horseshoe Crabs and Shorebirds in Delaware Bay: Predictive Modeling and Implementation. North Carolina Fish and Wildlife Cooperative Research Unit, Research Work Order 155, Raleigh, NC. Taft, O.W., and S. M. Haig. 2006a. Landscape context mediates influence of local food abundance on wetland use by wintering shorebirds in an agricultural valley Biological Conservation 128:298 307. Taft, O.W., and S M. Haig. 2006b. Importance of wetland landscape structure to shorebirds wintering in an agricultural valley. Landscape Ecology 21:169 184. Tarr, N. M., T. R. Simons, and K. H. Pollock. 2010. An Experimental Assessment of Vehicle Disturbance Effects on Migratory Shorebirds. Jour nal of Wildlife Management 74 :1776 1783.

PAGE 95

95 Tsipoura N and J. Burger 1999. Shorebird diet during spring migration stopover on Delaware Bay. Condor 101:635 644. v a n d er Veen, I. T. and L. E. Sivars. 2000. Causes and consequen ces of mass loss upon predator encounter: Feeding interruption, stress or fit for flight? Functional Ecology 14:638 644. van Gils, J.A., S.R. de Rooij, J. van Belle, J. van der Meer, A. Dekinga, T. Piersma, and R. Drent. 2005a. Digestive bottleneck affec ts foraging decisions in red knots Calidris canutus: I. Prey choice. Journal of Applied Ecology 74:105 119. van Gils, J.A., A. Dekinga, B. Spaans, W.K. Vahl, and T. Piersma. 2005b. Digestive bottleneck affects foraging decisions in red knots Calidris canu tus: I. Patch choice and length of working day. Journal of Applied Ecology 74:105 119. van Gils, J.A., and T. Piersma. 2004. Digestively constrained predators evade the cost of interference competition. Ecology 73:386 398. Warnock, N., and M. A. Bishop. 1998. Spring stopover ecology of migrant Western Sandpipers. Condor 100:456 467. Weber, L. and S. Haig. 1996. Shorebird use of South Carolina managed and natural coastal wetlands. Journal of Wildlife Management 60:73 82. White G C and K.P. Burnham 1999. Program MARK: survival estimation for populations of marked ani mals. Bird Study Supplement 46: 120 138. Williams T D C.G. Guglielmo O. Egeler C.J. Martyniuk 1999. Plasma metabolites provide information on mass change over several days in capti ve western sandpipers. Auk 116: 994 1000.

PAGE 96

96 BIOGRAPHICAL SKETCH Amy C. Schwarzer was born and raised in the Berkshire Mountains of western Massachusetts. There she cultivated an interest in wildlife in nature through a variety of opportunities provided by family camping trips, the Girl Scouts, the local Audubon feeders. After moving to Florida with her family she finished high school at the Pine View School in Osprey. She attended the University of Florida an d gradua ted with a Bachelor of Arts in environmental s cience (policy concentration) from the School of Natural Resources in 2003. From 2000 2007 she worked in a variety of biological field positions, with a heavy emphasis on birds and conservation biology She rejoined the School of Natural Resources in 2007 and received her Master of Science in interdisciplinary e cology in August 2011.