SEASONAL AND DECADAL - SCALE FORAGING HABITS OF THREE HAWAIIAN SEABIRDS By Kaycee Morra A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of Integrative Biology Doctor of Philosophy 2018 ABSTRACT SEASONAL AND DECADAL - SCALE FORAGING HABITS OF THREE HAWAIIAN SEABIRDS By Kaycee Morra This dissertation describes three lines of investigation into the foraging habits of three pelagic seabirds using either whole tissue or a mino acid - specific isotope data. First it asks whether information on foraging habits of the Hawaiian Petrel derived from isotope analysis of a primary feather grown at the beginning of the nonbreeding season (primary 1, P1) and in the middle of the nonbre eding season (primary 6, P6) is the same. Secondly, it compares long - term petrel using collagen - specific amino acid 15 N proxies for nutrient regime and trophic position. Thirdly, it evaluates seasonal changes in modern Newell's shearwater and Laysan albatross foraging habits and, for Laysan albatross, extends this seasonal analysis back a century using collagen - and feather - specific amino acid 15 N data. Chapt er one asked whether whole tissue isotope data for two feathers (P1 and P6) yielded similar information on the location and bio geochemical regime in which three colonies of Hawaiian petrel foraged. Relative to Hawaii and Lanai birds, the low P6 15 N values for Maui birds reflect foraging segregation and greater utilization of waters characterized by nitrogen fixation. There was no isotopic difference between P1 and P6, suggesting that either feather could be used to describe nonbreeding season forag ing habits. This information increases our understanding of Hawaiian petrel foraging behavior over the nonbreeding season and informs sampling protocols for conservation managers who wish to understand nonbreeding season foraging. Chapter two compared the foraging habits of modern and historical populations of three ecologically distinct species using amino acid - specific isotope analysis. The data show persistent inter - and intra - specific foraging segregation among Newell's shearwater, Laysan albatross, an d two populations of Hawaiian petrel. While our nutrient proxy showed no shift in nutrient regime use over time, a significant trophic decline occurred for Newell's shearwater and Laysan albatross within the past century , paralleling a similar trend previo usly observed in the Hawaiian petrel. This builds on current evidence of a basin - wide shift in trophic dynamics within the North Pacific Ocean. Chapter three uses amino acid - specific isotope analysis of feather and collagen to show that Newell's shearwater and Laysan albatross foraging habits (biogeochemical regime use and trophic position) differ between the breeding and nonbreeding seasons. In addition, withi n each season, each species utilize s a different foraging strategy despite the fact that they both breed on the Hawaiian Islands. While Laysan albatross have not altered their nonbreeding season foraging habits over the past century, the trophic decline they experienced occurred exclusively during the breeding season . Conservation management strategies for threatened seabirds like Newell's shearwater and Laysan albatross will require an understanding of at - sea risks as well as threats on land on seasonal timescales and may need to be individually tailored for each species. . iv This dissertation is dedicated to my Mom, who reminds me daily how much I am loved; my Dad, whose lack of concern for my future gives me confidence; my Sister, who inspires me to be courageous and compassionate; my Brother, whose strong character and gentle spirit comfort me. v ACKNOWLEDGEMENTS I owe a debt of gratitude to countless individuals for their contributions to my dissertation work. In particular, I must thank the following people. My doctoral advisor, Peggy Ostrom, for always striving toward what is best for my you ng career. Nathaniel Ostrom, for his thoughtful wisdom. Elise Zipkin and Sam Rossman for their statistical expertise, and Sam, for his willingness to teach me. My wonderful colleagues, Helen James and Anne Wiley, who always made time to offer considered an d insightful feedback. Yoshito Chikaraishi, whose guidance was essential to my success in the lab. Hasand Gandhi, who regularly offered patience and expertise in the lab and smiles a nd calm company in the office. The many conservationists and organizations working hard to protect our Hawaiian seabirds and support good research, including Cathleen Bailey, Jessie Beck, Megan Dalton, Fern DuVall, Shannon Fitzgerald, Michelle Hester, Darcy Hu, Robby Kohley, Kathleen Mi sajon, Hannah Nevins, Jay Penniman, Sheldon Plentovich, Andre Raine, Carl Schwarz, Eric Marie Van Zandt, VanderWerf, Bill Walker, Lindsay Young, The National Marine Fisheries Service, Alaska Fisheries Science Center, Pacific Islands Regional Office Fishe ries Observer Program staff and observers, the vessels and crews of the Hawaii longline fisheries that supported observers, Seabird Recovery Project. The Bird Division , National Museum of Natural History, the Bernice Bishop Museum namely, Molly Hagemann, and the California Academy of Sciences for the loan of specimens and permission to sample them. vi The many folks, like Jim, who kept the Natural Science building clean f or all of us Spartans, particularly considering the countless nights they had to clean around me. vii TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ....................... ix LIST OF FIGURES ................................ ................................ ................................ ...................... x INTRODUCTION ................................ ................................ ................................ ........................ 1 CHAPTER 1 ................................ ................................ ................................ ................................ . 2 Influence of Feather Selection and Sampling Protocol on Interpretations of Hawaiian Petrel ( Pterodroma sandwichensis ) Nonbreeding Season Foraging Habits from Stable Iso tope Analysis ................................ ................................ ................................ ................................ ......... 2 ABSTRACT ................................ ................................ ................................ ...................... 2 INTRODUCTION ................................ ................................ ................................ ............ 3 METHODS ................................ ................................ ................................ ....................... 4 Study Area ................................ ................................ ................................ ............ 4 13 - Section Protocol ................................ ................................ ............................... 4 4 - Section Protocol ................................ ................................ ................................ . 4 Barb - Sampling Protoc ol and Comparison of Protocols ................................ ........ 5 Stable Isotope Analysis ................................ ................................ ......................... 5 Statistical Analyses ................................ ................................ ............................... 5 RESULTS ................................ ................................ ................................ ......................... 6 DISCUSSION ................................ ................................ ................................ ................... 7 APPENDIX ................................ ................................ ................................ ................................ . 11 LITERATURE CITED ................................ ................................ ................................ ............... 17 CHAPTER 2 ................................ ................................ ................................ ............................... 20 Trophic Declines and Decadal - Scale Foraging Segregat ion in Three Pelagic Seabirds ............ 20 ABSTRACT ................................ ................................ ................................ .................... 20 INTRODUCTION ................................ ................................ ................................ .......... 21 METHODS ................................ ................................ ................................ ..................... 25 Sample Acquisition ................................ ................................ ............................. 25 Sample Sizes ................................ ................................ ................................ ....... 25 Sample Preparation ................................ ................................ ............................. 26 15 N Amino Acid A nalysis ................................ ................................ ................. 27 Model Selection ................................ ................................ ................................ .. 28 Statistical Analysis ................................ ................................ .............................. 29 RESULTS ................................ ................................ ................................ ....................... 30 Model Selection ................................ ................................ ................................ .. 30 Interspecific and Inter - colony Variation in Amino Acid 15 N ........................... 30 Temporal Variation in 15 N Phe and 15 N Glu - Phe ................................ .................. 31 DISCUSSION ................................ ................................ ................................ ................. 31 Modern Time Periods ................................ ................................ ......................... 31 Tempor al Variation ................................ ................................ ............................. 37 viii APPENDIX ................................ ................................ ................................ ................................ . 39 LITER ATURE CITED ................................ ................................ ................................ ............... 44 CHAPTER 3 ................................ ................................ ................................ ............................... 52 Seasonal Variation in Foraging Habits of Two Hawaiian Seabirds ................................ ........... 52 ABSTRACT ................................ ................................ ................................ .................... 52 INTRODUCTION ................................ ................................ ................................ .......... 53 METHODS ................................ ................................ ................................ ..................... 57 Sample Acquisition ................................ ................................ ............................. 57 Sample Sizes ................................ ................................ ................................ ....... 57 Sample Preparation ................................ ................................ ............................. 58 15 N Amino Acid Analys is ................................ ................................ ................. 58 Model Selection ................................ ................................ ................................ .. 59 Statistical Analysis ................................ ................................ .............................. 59 RESULTS ................................ ................................ ................................ ....................... 61 Climate Effects ................................ ................................ ................................ .... 61 Variation in 15 N Phe and 15 N Glu - Phe Betw een Feather and Bone Collagen ....... 61 Variation in 15 N Phe and 15 N Glu - Phe Bet ween Species and Time Periods ......... 61 DISCUSSION ................................ ................................ ................................ ................. 62 Modern Compari sons ................................ ................................ .......................... 62 Temporal Comparison s ................................ ................................ ....................... 64 Conclusions ................................ ................................ ................................ ......... 65 APPENDIX ................................ ................................ ................................ ................................ . 67 LITERATURE CITED ................................ ................................ ................................ ............... 71 ix LIST OF TABLES Table 1. 13 C 15 N between Hawaiian Petrel primary 6 feather 13 - section (13S), 4 - section (4S), and barb - sampling (Barb) protocols, and results of paired t - tests ( P values). Average differences are reported as absolute values. Data are from the islands of Hawaii, USA (Hawaii, Lanai and Maui islands ) ................................ ................................ ................................ ................................ ........ 14 Table 2. Parameter estimates (Est) from linear mixed models of primary 6 base and whole 13 15 N values with colony and sample type included as fixed effects and individual (Indiv) included as a random effect. Estimates are relative to the reference colony Hawaii and sample type whole feat her. The model used restricted maximum likelihood t - tests with Satterthwaite approximations to degrees of freedom. SD = standard deviation; SE = standard error ; Var = variance . Significant P values are indicated with an asterisk .................. 15 Table 3. Average differ 13 C 15 N between primary feathers 1 and 6 of Hawaiian Petrels from three colonies (Hawaii, Lanai, and Maui islands), sample sizes and results of paired t - tests ( P values). Data were obt ained using the barb - sampling protocol ................................ ................................ ................. 15 Table 4. Results of a three - way ANOVA model where 1 3 C values are a function of individuals, breeding island and/or primary 6 feather section. The three - way ANOVA model was based on data from the 4 - section protocol. Significant P values are indicated with an asterisk ............... 16 Table 5. Results of a three - way ANOVA model where 15 N values are a function of individuals, breeding colony and/or primary 6 feather section. The three - way ANOVA model was based on data from the 4 - section protocol. Significant P values are indicated with an asterisk ............... 16 Table 6. Statistical results from hierarchical modeling of collagen isotope data showing variation among groups (colony or species). Probabilities ( P ) indicate likeli 15 N Phe 15 N Glu - Phe - population and interspecific (HAPE) data for historical and modern time periods; (b) is a within species com parison of historical to modern data for NESH and LAAL. Statistical results for comparisons with Maui and a for HAPE (Ostrom et al. 2017) ..................... 43 Table 7 . Statistical results from hierarchical modeling of feather an d collagen isotope data showing variation among populations. Probabilities ( P 15 N Phe and 15 N Glu - Phe an albatross (LAAL) data; (b) shows an inter - species comparison between NESH and LAAL, and a temporal c omparison for LAAL ................................ ................................ ................................ ................................ .......... 70 x LIST OF FIGURES Figure 1. Schematic for Hawaiian Petrel primary 6 feather sampling protocols. Solid lines depict the division of the vanes into 13 1 - cm sections (labeled 1 - 13) and dashed lines indicate the locations of barbs sampled in the barb - sampling protocol. Shading indicates sections that were homogenized and analyzed in the 13 - and 4 - section protocols. ( A ) 13 - section protocol; each section was homogenized and subsampled for stable - isotope analysis. ( B ) The 4 - section protocol; sections 1 and 2, 5, 8, and 12 and 13 were homoge nized and sub - sampled. ( C ) Barb - sampling protocol; barbs from the border of sections 1 and 2 and 12 and 13 and barbs from the centers of sections 5 and 8 were combined to form a sample for isotope analysis ..................... 12 Figure 2. 13 C values for Hawaiia n Petrel primary 6 (P6) feathers obtained from the 4 - section protocol. Data progress from section 1 (tip) on left to section 13 (base) on right. ( A ) 13 C for individuals. ( B 13 C (standard deviation) for each feather section from three colonies. Line style indicates breeding colony: red solid = Hawaii, green dashed = Lanai, blue dash - dot = Maui ................................ ................................ .............. 13 15 N values for Hawaiian Petrel primary 6 (P6) feathers obtained from the 4 - section protocol. Data progres s from section 1 (tip) on left to section 13 (base) on right. (A) 15 15 N (standard deviation) for each feather section from three colonies. Line style indicates breeding colony: red solid = Hawaii, gre en dashed = Lanai, blue dash - dot = Maui ................................ ................................ .............. 13 Figure 4. N* distribution, at - 15 gradient is the distributio n of N* defined and gridded by Sherwood et al. (2014), where N* = N - 1, P = phosphorus. Positive N* values are interpreted as an increase in nitrogen fixation and negative values as net denitrification (Sherwood et al. 2014). Curved line s (north of line), solid = Hawaiian petrel (Ainley et al. 1997; Simons and Hodges 1998; Awkerman et al. 2009). The data for sinking particles and sediments mark 15 N in the eastern tropical north Pacific Ocean and illustrate an isotopic gradient. As illustrated in Wiley et 15 N. Filled black points represent 15 N 15 N 15 N value ................................ ................................ .. 40 Figure 5. 15 N Phe 15 N Glu - Phe Model 15 N Phe 15 N Glu - Phe values for each population define the center of the ellipse and a covariance matrix determines the shape and orientation. Legend indicates species and time period designations. Species abbreviations are as follows: N shearwater, LAAL = Laysan albatross, HAPE = Hawaiian petrel. Time periods and sample sizes xi - - - 1998, n =9); Post - 2000 (2013 - 2016, n = 10); LAAL: Pre - 1950 (1902 - 1937, n = 11); Post - 2000 (2003 - 2014, n = 11); Maui HAPE: Foundation Period (1000 - 1400 CE; 550 - 950 y B.P., n = 5); Modern Period (1950 - - 1800 CE; 150 - 550 y B.P., n = 8); Modern Period (1950 - 2010, n = 8). The prehistoric tim with Hawaiian archaeological time periods defined by Kirch (1990) ................................ ........ 41 Figure 6. 15 N Phe 15 N Glu - Phe - - - 1998, n = 9); Post - 2000 (2013 - 2016, n = 10). Laysa n albatross time periods and sample sizes: Pre - 1950 (1902 - 1937, n = 11); Post - 2000 (2003 - 2014, n = 11) ................................ ................. 42 Figure 7. 15 N Phe 15 N Glu - Phe batross (LAAL) feather and collagen. 15 N Phe 15 N Glu - Phe values for each population define the center of the ellipse and a covariance matrix determines the shape and orientation. Legend indicates species and tissue type. Time peri ods and sample sizes (n) follow. NESH: 2013 - 2016, n = 10 for feather and collagen; LAAL: 2003 - 2014, n = 10 for feathe r, n = 11 for collagen ..................... 68 Figure 8. Density distributions 15 N Phe 15 N Glu - Phe Laysan albatross (a) feather and (b) collagen over the past 100 years . T ime periods and s ample sizes (n) follow: Pre - 1950 (1902 - 1937) n = 9 for feather, n = 11 for collagen; Post - 2000 (2003 - 2014) n = 10 for feathe r, n = 11 for collagen ................................ ................................ .............................. 69 1 INTRODUCTION Among the tapestry of top predators that forage within the Pacific Ocean, seabirds are some of the most conspicuous and wide - ranging. Although pelagic seabirds are atypical in their dependence on a terrestrial breeding colony, they spend most of their live s navigating thousands Ocean themselves and compete with seabirds, and other marine predators, for food. Scientists have recognized for decades that humans are and most remote ecosystems, and fishers have noticed rapid and drastic declines in formerly as a food source, they were dealing with the shocking realization that humans were capable of decimating such plentiful fish populations, potentially beyond recovery. Our understanding of fish stocks is often incomplete and is based on biased global marine catch statistics that offer little forewarning regarding fish abundances and distributions. Moreover, we have a limited understanding of how industrial - scale fish removal influences marine ecosystem integrity as a whole. As highly mobile top predators that occasionally be come accessible on land, seabirds offer a unique opportunity to gain valuable insight into oceanic trophic dynamics. In addition to their potential to herald marine ecosystem change, their preserved tissues can provide a retrospective glimpse into a time t hat precedes catch statistics and inform long - term temporal shifts in oceanic food webs. 2 CHAPTER 1 Influence of Feather Selection and Sampling Protocol on Interpretations of Hawaiian Petrel ( Pterodroma sandwichensis ) Nonbreeding Season Foraging Habits from Stable Isotope Analysis ABSTRACT Isotope data from Hawaiian Petrel ( Pterodroma sandwichensis ) primaries P1 and P6 were compared to determine whether foraging habits change between the beginning and middle of the nonbreeding season. P6 data did not differ between samples derived from a longitudinal and a 13 C 15 N longitudinal trends emerged, yet inter - 15 N 13 C data suggest that Hawaiian Petrels molt at low latitudes. Among 15 N values for Maui birds relative to Hawaii and Lanai birds reflect for aging segregation and differential utilization of 15 N - enriched oceanic regions. For the Hawaiian Petrel, the isotopic similarity between P1 and P6 indicates that analogous ecological interpretations can be drawn from these feathers, and similar foraging ha bits persist from the beginning to middle of the nonbreeding season. Prolonged inter - colony foraging segregation may facilitate coexistence of colonies and, together with high intra - colony foraging diversity, may reduce extinction risk for the endangered H awaiian Petrel. 3 INTRODUCTION Remiges, retrices, coverts and contour feathers have been exploited for isotopic information that has advanced our understanding of avian foraging behavior (Navarro et al . 2009; Hinke et al . 2015; Cherel et al . 2016). Increasing concern that isotopically derived ecological inferences are dependent on feather type, molt timing and sampling protocol is problematic for conservation managers who sample live individuals (Edwards et al . 2015; Cherel et al . 2016). Mor eover, sampling decisions are irreversible, emphasizing that careful consideration should be given to sampling protocols so that valuable resources are preserved for future research. 13 C values of ph ytoplankton and consumers vary inversely with latitude in the northern Pacific Ocean, and broad spatial gradients 15 N at the base of oceanic food webs are transferred to consumers with a systematic increase 15 N with trophic level (Graham et al . 201 0). The Hawaiian Petrel ( Pterodroma sandwichensis ) is an excellent species in which to study isotopic variance among feathers because there are existing data on primary 1 (P1), and the timing and sequence of primary molt is well constrained (Wiley et al . 2 010, 2012, 2013). Hawaiian Petrels, like other Procellariiformes, molt their primaries during the nonbreeding season (Pyle et al . 2011; Howell et al . 2012) and, like other species of the genera Pterodroma , their molt likely proceeds distally from the inner most to outermost primary, P1 to P10 (Warham 1996; Pyle 2008). Because P1 likely grows over the first 12 to 35 days of the 3.5 - 6 month long nonbreeding season (Simons 1985), we questioned whether P1 adequately reflects post - breeding foraging behavior. Thu s, we centered our current investigation on primary 6 (P6), which, based on estimates of feather growth rates in juvenile Hawaiian Petrels and other Procellariiformes, requires up to 100 days to grow (Sincock and Swedberg 1969; Ainley et al . 1976; Simons 1985) 4 13 15 N data from P1 and P6 provide similar information regarding nonbreeding season foraging, and to evaluate inter - and intra - feather variation. We analyzed birds from three colonies to examine whether ecological inferences were representative of the entire species. METHODS Study Area H awaiian Petrel feathers were acquired from carcasses salvag ed from colonies on Hawaii, Lanai, and Maui islands, Hawaii, USA, between 1990 and 2008. Feathers were washed ( 87:13 v/v chloroform:methanol ) , rinsed with ultrapure distilled water (E - Pure, Barnstead), dried ( 25 °C ), divided into 1 - cm long sections and sampled according to one of three protoc ols. The rachis was excluded from analysis. 1 3 - Section Protocol V anes from five P6 remiges were longitudinally divided into 13 1 - cm sections and numbered from oldest (tip, 1) to youngest (base, 13) material (Fig. 1) . M inor variation s in length w ere accounted for by adjusting the size of section 13. B arbs were weighed and cut into 3 - mm long fragments , and a 1.0 - mg homogenized 13 15 N analysis. The mass - weighted isotope value of each aliquot was used to calculate whole feather isotope averages. 4 - Section Protocol The remaining homogenates from the five P6 remiges and an additional 26 P6 feathers we re used for a 4 - section protocol . 1.0 - mg homogenates from the feather tip (sections 1 and 2 combined ), the base (sections 12 and 13 combined ), and mid - feather (sections 5 and 8 , kept separate ) were analyzed, and whole feather averages were calculated as described above (Fig. 1) . 5 Barb - Sampling Protocol and Comparison of Protocols B arbs from the base, middle, and tip of 31 P1 and 26 P6 remiges were combined into a single 1.0 - mg sample f rom each feather for isotope analysis. The number of barbs from each section was based on the distribution of mass found in other Hawaiian Petrel P1 and P6 feathers. For P1, we combined barbs from sections 1, 4 and 7 (Wiley et al . 2010). For P6, we combined barbs from the tip (0.1 mg) , section 5 (0.2 mg) , section 8 (0.3 mg) , and base ( 0.4 mg ) (Fig. 1) . We assessed whether our three sampling protocols yielded e quivalent whole feather isotope values by first comparing the mass - weight ed averages generated from the 13 - and 4 - section protocols . We then compared the 4 - section and barb - sampling protocols. Stable Isotope Analysis Feather a liquots (1.0 mg) were analyzed with an elemental analyzer (Eurovector) interfaced to an Isoprime mass s pectrometer (Elementar). Data ([R sample /R standard ] - 1) × 1 , 000, where X is 13 C or 15 N , R is 13 C/ 12 C or 15 N/ 14 N , and R standard is V - 13 15 N, respectively. Isotopically characterized muscle tissue standards were analyzed between every nine unknowns . Precision was < 13 15 N. Statistical Analyses Differences in isotope values between protocols and between P1 and P6 were evaluated with two - tailed paired t - tests . Data from the 4 - section protocol were used to evaluate the effects of colony and feather section on isotope values with three - way ANOVA models. significant difference (HSD) tests were implemented to evaluate variation among colonies and feather sections. Normalit y and heteroscedasticity were checked with n ormal quantile - quantile To determine if isotope values of white bases and whole feather averages differed , we constructed Gaussian linear mixed effects models with i sotope value as the 6 de pendent variable, and colony and sample type (feather base, whole feather average) as factors. Individual was a random effect to account for individual variation. S tatistical tests were completed in statistical package R (R Development Core Team 2016). RESULTS The 13 - difference between the 4 - section and 13 - 13 15 N) and statistically insignificant (Table 1). For all colonies, the difference between the barb - sampling and 4 - 13 15 13 C base and whole feather averages estimated by our mixed model was significant 15 N did not differ between bases and whole feather averages (Table 2). Paired P1 and P6 barb - sa mpling protocol data from 25 individuals (7 from Hawaii, 8 from Lanai, and 10 from Maui) showed no 13 15 N between P1 and P6 (Table 3). 13 C values increased from tip to base, and 13 C was equivalent across all three colonies and not a function of individual (Table 4). P6 feather section pairs were significantly 13 C , except for the section 1 and 5 comparis on (Tukey HSD P > 0.05). 15 N for individuals or colonies, and feather section 15 N 15 N was a function of colony and individual 15 N b etween Lanai and Hawaii; however, Maui P6 7 15 N was the lowest of all colonies (Tukey HSD P < 0.05 and P < 0.01 for Hawaii and Lanai, respectively; Table 5). DISCUSSION Our investigation of Hawaiian Petrel primary feathers improved our understanding of how feather selection influences isotopically derived ecological inferences. Development of P6 sampling approaches was essential in delineating isotopic variability within feathers, identifying a minimally invasive sampling protocol, and stimulating considera tions for sampling strategy. The sampling protocols used to characterize P6 whole feather averages avoid potential bias from point sampling and reflect the longest time period possible from a single feather. The minimally invasive barb - sampling protocol is relatively fast, reduces damage to museum specimens, and is unlikely to affect live bird flight ability. However, this protocol only produces a single average value for the feather. If evaluation of intra - feather isotope variation is desired, we recommend sampling along the length of the vane, similar to our 13 - and 4 - section protocols. Data from our 4 - 13 C 15 N intra - feather variation. For P1, Wiley et al 13 C values in the Hawaiian Puffinus newelli 15 N . In our results, regardless of colony, the majority of P6 feathers exhibited a longitudinal increase in 13 C , similar to P1. None of the co 15 N longitudinal trend. Of 26 P6 feathers, 15 N slopes along the length of the vanes; the remaining 20 exhibited 15 N patterns. Three factors could contribute to longitudinal isotopic in creases: melanin pigmentation, fractionation associated with feather growth, and foraging behavior changes (Wiley et al . 2010). 8 However, isotope variation cannot be confounded by melanin or fractionation for white feather base point samples that lack melan in and are not differentially influenced by growth. The observation that isotope values of white bases and whole feather averages are similar indicates that inferences regarding differences in foraging behavior among individuals and colonies can be drawn f rom our data. 13 15 N variation among seabirds (Wiley et al 15 N , rather than trophic level, is likely a dominant control of 15 N (Wiley et al . 2013). Thus, P 6 longitudinal isotopic variation and 15 N 13 C values are higher than those of seabirds that forage in the North Pacific Transition Zone (40 to 45° N) (Gould et al . 199 7), suggesting that Hawaiian Petrels molt at more southerly latitudes 15 N data all indicate that Hawaiian Petrels forage at low latitudes post - breeding. Spear et al . (1995) observ ed Hawaiian Petrels southeast of Hawaii during the nonbreeding season, and tagging data show Lanai birds south of Hawaii post - breeding (VanZandt 2012). Differential foraging in a 15 N - enriched area southeast of Hawaii or in waters whose nutrients are 15 N de pleted by nitrogen fixation may drive the inter - 15 N data (Ostrom et al . 15 N values because 15 N dur ing the nonbreeding season. Collagen amino 15 N values also suggest that Hawaii and Maui colonies differ in foraging location (Ostrom et al . 2017). An important objective was to determine whether P1 and P6 yield similar information on nonbreeding seas on foraging, when foraging habits are particularly difficult to ascertain. Given 9 that P1 and P6 do not differ isotopically, our results suggest that similar ecological information can be obtained from these feathers. Our data fortify interpretations made b y Wiley et al . (2012) regarding inter - colony foraging differences. Wiley et al . (2012) suggested that inter - colony foraging segregation may reflect niche partitioning driven by competition, and facilitate colony coexistence (Lewis et al . 2001; Navarro et al . 2009). The P6 results indicate that inter - colony differences observed at the onset of the nonbreeding season are protracted. Sustained foraging segregation during the nonbreeding season shows ecological diversity, an attribute that may reduce extinctio n risk for 15 N ) may promote the viability of a small population. For Hawaiian Petrels, sampling P6 offers advantages over P1; P6 is relatively distal, acc essible and a viable option for species where molt timing and sequence are constrained. Some albatrosses simultaneously molt distally and proximally and vary the extent o f molt annually (Edwards and Ro h w er 2005). In such cases, feather selection must be de liberated with regard to the time over which foraging habits are recorded. While non - pigmented point samples avoid pigmentation - associated isotopic variance, whole feather averages reflect a longer timeframe but require more handling than point samples. Im portantly, it may be difficult to dismiss pigmentation as a confounding factor for feathers with a complicated pigmentation pattern. Isotopic inferences made from P1 and P6 suggest that foraging segregation between Hawaiian Petrel colonies persists into t he heart of the nonbreeding season, that there is large variation among individuals in foraging location, and that molt likely takes place at low latitudes. Our results strengthen evidence that Hawaiian Petrels breeding on different Hawaiian islands are ec ologically distinct (Wiley et al . 2012). Additional efforts including a well - considered study of 10 individual specialization and additional amino acid isotope analyses will be important future contributions to understand Hawaiian Petrel at - sea ecology. 11 APPENDIX 12 Figure 1. Schematic for Hawaiian Petrel primary 6 feather sampling protocols. Solid lines depict the division of the vanes into 13 1 - cm sections (labeled 1 - 13) and dashed lines indicate the locations of barbs sampled in the barb - sampling protocol. Shading indicates sections that were homogenized and analyzed in the 13 - and 4 - section protocols. ( A ) 13 - section protocol; each section was homogenized and subsampled for stable - isotope analysis. ( B ) The 4 - section protocol; sections 1 and 2, 5, 8, and 12 and 13 were homogenized and sub - sampled. ( C ) Barb - sampling protocol; barbs from the border of sections 1 and 2 and 12 and 13 and barbs from the centers of sections 5 and 8 were combined to form a sample for isotope analysis. 13 Figure 2. 13 C values for Hawaiian Petrel primary 6 (P6) feathers obtained from the 4 - section protocol. Data progress from section 1 (tip) on left to section 13 (base) on right. ( A ) Longitudinal 13 C for individuals. ( B 13 C (standard deviation) for each feather section from three colonies. Line style indicates breeding colony: red solid = Hawaii, green dashed = Lanai, blue dash - dot = Maui. 15 N value s for Hawaiian Petrel primary 6 (P6) feathers obtained from the 4 - section protocol. Data progress from section 1 (tip) on left to section 13 (base) on right. (A) 15 15 N (standard deviation) for each fe ather section from three colonies. Line style indicates breeding colony: red solid = Hawaii, green dashed = Lanai, blue dash - dot = Maui. 14 Table 1. 13 C and 15 N between Hawaiian Petrel primary 6 feather 13 - section (13S), 4 - section (4S), and barb - sampling (Barb) protocols, and results of paired t - tests ( P values). Average differences are reported as absolute values. Data are from the islands of Hawaii, USA (Hawaii, Lanai a nd Maui islands). 13 15 Sampling Protocol Sample Size Average Difference (SD) P Average Difference (SD) P 13S vs. 4S (Lanai, Maui) 5 0.0 (0.2) 0.90 0.1 (0.5) 0 .6 0 4S vs. Barb (Hawaii) 8 0.1 (0.2) 0.10 0.0 (0.2) 0.70 4S vs. Barb (Lanai) 8 0.0 (0.1) 0.60 0.2 (0.4) 0.20 4S vs. Barb (Maui) 10 0.2 (0.5) 0.30 0.4 (0.9) 0.20 15 Table 2. Parameter estimates (Est) from linear mixed models of primary 6 base and whole feather 13 15 N values with colony and sample type included as fixed effects and individual (Indiv) included as a random effect. Estimates are relative to the reference colony Hawaii and sample type whole feather. The model used restricted maximum likelihood t - tests wit h Satterthwaite approximations to degrees of freedom. SD = standard deviation; SE = standard error ; Var = variance . Significant P values are indicated with an asterisk. 13 15 Fixed Effects Est SE df t P Est SE df t P Intercept - 15.40 0.24 24 - 64.10 < 0.001* 15.60 0.68 32 23.00 < 0.001* Colony Lanai - 0.09 0.34 24 - 0.25 0.80 0.62 0.89 32 0.70 0.49 Colony Maui 0.01 0.32 24 0.03 0.98 - 2.78 0.87 32 - 3.19 0.003* Sample T ype B ase 0.22 0.08 23 2.81 0.01 * - 0.13 0.32 28 - 0.40 0.69 Random E ffect Var SD Var SD Indiv 0.44 0.66 3.25 1.80 Residual 0.03 0.16 0.42 0.65 Table 3. 13 C 15 N between primary feathers 1 and 6 of Hawaiian Petrels from three colonies (Hawaii, Lanai, and Maui islands), sample sizes and results of paired t - tests ( P values). Data were obtained using the barb - sampling protocol. 13 15 Colony Sample Size Average Difference (SD) between P1 and P6 P Average Difference (SD) between P1 and P6 P Hawaii 7 0.2 (0.3) 0.10 0.2 (1.7) 0.80 Lanai 8 0.2 (0.3) 0.10 1.8 (1.7) 0.10 Maui 10 0.3 (0.5) 0.10 0.7 (1.8) 0.30 16 Table 4. Results of a three - way ANOVA model where 1 3 C values are a function of individuals, breeding island and/or primary 6 feather section. The three - way ANOVA model was based on data from the 4 - section protocol. Significant P values are indicated with an asterisk. Model F P Individual 1.182 0.28 Colony 1.140 0.29 Feather Section 30.670 < 0.05 * Individual + Colony 0.586 0.45 Individual + Feather Section 0.716 0.40 Colony + Feather Section 0.461 0.50 Individual + Colony + Feather Section 0.090 0.77 Table 5. Results of a three - way ANOVA model where 15 N values are a function of individuals, breeding colony and/or primary 6 feather section. The three - way ANOVA model was based on data from the 4 - section protocol. Significant P values are indicated with an asterisk. Model F P Individual 9.699 < 0.05 * Colony 21.670 < 0.05 * Feather Section 0.037 0.85 Individual + Colony 0.714 0.40 Individual + Feather Section 0.050 0.82 Colony + Feather Section 0.148 0.70 Individual + Colony + Feather Section 1.090 0. 30 17 LITERATURE CITED 18 LITERATURE CITED Ainley DG , Lewis TJ, Morrell S. 1976 y Storm - Petrels. Wilson Bull. 88 : 76 - 95. Cherel YP Quillfeldt KD , Weimerskirch H. 2016 Combination of at - sea activity, geolocation and feather stable isotopes documents where and when seabirds molt. Front. Ecol. Evol . 4 : 3 . Edwards AE, Rohwer S. 2005 Large - scale patterns of molt activation in the flight feathers of two albatross species. Condor 107: 835 - 848. Edwards AE, Fitzgerald SM, Parrish JK, Klavitter JL, Romano MD. 2015 Foraging strategies of Laysan Albatross inferred from stable isotopes: impl ications for association with fisheries. PLoS One 10: e0133471. Gould P, Walker W, Ostrom P. 1997 Foods of northern fulmars associated with high - seas drift nets in the transitional region of the North Pacific. Northwest. Nat. 78: 57 - 61. Graham BS , Koch P L , Newsome SD , McMahon KW, Aurioles D. 2010 Using isoscapes to trace the movements and foraging behavior of top p redators in oceanic ecosystems. In: West JB, Bowen GJ, Dawson TE, Tu KP (eds) Isoscapes : Understanding Movement, Pattern, and Process on Earth Through Isotope Mapping . Springer Science, Dordrecht, The Netherlands, pp 299 - 318. Hinke J T , Polito MJ, Goebel ME, Jarvis S, Reiss CS, Thorrold SR, Trivelpiece WZ, Watters GM. 2015 Spatial and isotopic niche partitioning during winter in chinstrap and Adé lie penguins from the South Shetland Islands. Ecosphere 6: 1 - 32. Howell , SNG, Patteson JB, Shearwater D. 2012 Petrels, a lbatrosses, and s torm - p etrels of North America: a p hotographic g uide. Princeton University Press, Princeton, New Jersey. Lewis S, Sherratt TN , Hamer KC, Wanless S. 2001 Evidence of intra - specific competition for f ood in a pelagic seabird. Nat. 412: 816 - 819. Navarro J, F orero M , Gonzalez - Solis J , Igaul JM , Becares J, Hobson K. 2009 Foraging segregation between two closely related sh earwaters breeding in sympatry. Biol . Lett. 5: 545 - 548. Ostrom PH, Wiley AE, James HF, Rossman S, Walker WA, Zipkin EZ, Chikaraishi Y. 2017 Broad - scale trophic shift in the pelagic North Pacific revealed by an oceanic seabird. Proc. R. Soc. B: Biol. Sci. 284: 20162436 - 20162442. 19 Pyle P. 2008 Molt and age determination in Procellariiformes . Pages 248 - 260 in Identification Guide to North American Birds. P art 2 : Anatidae to Alcidae . Slate Creek Press, Point Reyes Station, C alifornia. Pyle P, Webster DL, Baird RW. 2011 Notes on petrels of the Dark - rumped Petrel complex ( Pterodroma phaeopygia/sandwichensis ) in Hawaiian waters. North Am. Birds 65 : 364 - 367. R Development Core Team. 2016 R: a language and enviro nment for statistical computing v. 3.3.2 . R Foundation for Statistical Computing, Vienna, Austria . http://www.R - project.org/ , accessed 14 November 2016. Simons TR. 1985 Biology and behavior of the endangered Hawaiian Dark - rumped Petrel. Condor 87: 229 - 245. Sincock JL, Swedberg GE. 196 Shearwater ( Puffinus puffinus newelli ), with initial observations. Condor 71: 69 - 71. Spe ar LB, Ainley DG, Nur N, Howell SNG. 1995 Population size and factors affecting at - sea distributions of four endangered Procellariids in the tropical Pacific. Condor 97: 613 - 638. VanZandt ML . 2012 Distribution and habitat selection of the endangered Hawaiian Petrel ( Pterodroma sandwichensis ), from the isla Hawaii, Hilo. Warham J. 1996 T he behavior, population biology and physiology of the p etrels. Academic Press, University of Canterbury, Christchurch, New Zealand . Wiley AE, Ostrom PH , Stricker CA , James HF, Gandhi H. 2010 Isotopic characterization of flight feathers in two pelagic seabirds: s ampling strategies for ecological studies. Condor 112: 337 - 346. Wiley AE, Welch AJ , Ostrom PH , James HF , Stricker CA , Fleischer RC , Gandhi H , Adams J, Ainley DG , Duvall F, Holmes N. 2012 Foraging segregation and genetic divergence between geographically proximate colonies of a highly mobile seabird. Oecologia 168 : 119 - 130. Wiley AE, Ostrom PH , Welch AJ , Fleischer RC , Gandhi H , Southon JR, Stafford Jr . TR , Penniman JF , Hu D , Duvall FP, Jam es HF. 2013 Millen n ial - scale isotope records from a wide - ranging predator show evidence of recent human impact to oceanic food webs. Proc . Nat l . Acad. Sci . 110 : 8972 - 8977. 20 CHAPTER 2 Trophic Declines and Decadal - Scale Foraging Segregation in Three Pelagic Seabirds ABSTRACT Oceanic ecosystems contain critical resources for humans but are increasingly at risk. Isotope values of wide - ranging seabirds offer an avenue to assess vast, remote ocean ecosystems. We investigate the fora ging habits of Newell's shearwater ( Puffinus newelli ) and Laysan albatross ( Phoebastria immutabilis 15 N proxies for nutrient regime and tropic position. We compare these data to those publishe d for the Hawaiian petrel ( Pterodroma sandwichensis ). Standard ellipses constructed from the isotope - population and interspecific foraging segregation that has pe rsisted for several decades. We found no evidence of a shift in nutrient regime used by our study species. However, a significant trophic decline occurred during the past century for Newell's shearwater and Laysan albatross similar trend observed in the Hawaiian petrel. Because our study species are broadly distributed across the North Pacific Ocean, employ distinct feeding strategies and exhibit several other divergent morphological and behavioral traits, the trophic decline suggests a pervasive shift in food web architecture within the past century, most conceivably in response to industrial fishing. 21 INTRODUCTION Vast remote open oceans, beyond the continental shelves, are of great economic and social importance (Costanza 1999). Unique insight into oceanic ecosystems can be obtained by studying wide - ranging top predators, including seabirds. Pelagic seabirds are particularly useful indicators of ecosystem change because they integrate molecular information on trophic dynam ics and biogeochemical regimes in their body tissues as they forage over large oceanic expanses (Bearhop et al. 2006; Hinke et al. 2015). This information can be accessed through stable isotope analysis of those tissues. The use of nitrogen isotopes to de lineate trophic relationships has a long history (DeNiro and Epstein 1981; Minagawa and Wada 1984) built on observations of a systematic increase of ~3 - 15 N with each trophic transfer (DeNiro and Epstein 1981; Minagawa and Wada 1984). A challenge li 15 N values respond to both trophic level and source nitrogen supplied to the base of the food web. Compound specific nitrogen isotope analysis is able to disentangle these effects (Gaebler et al. 1963; McClelland an d Montoya 2002; Chikaraishi et al. 2007; McMahon and McCarthy 2016; Ohkouchi et al. 2017). 15 N 15 N Glu ), becomes 15 N - enriched with increasing trophic level. This isotopic fractionation occurs during dea mination and other 15 15 N Phe ), fractionates to a smaller degree with each trophic transfer. 15 N Phe primarily reflects the isotope value of source nitrogen at the base of the 15 N Glu and 15 N Phe 15 N Glu - Phe ) and is used to calculate trop hic position 15 N Glu 15 N Phe )/TDF + 1, where 15 N Glu 15 N Phe difference in primary 22 producers and TDF 15 N Glu relative to 15 N Phe per trophic step (McMahon and McCart hy 2016). 15 N Phe 15 N of source nitrogen varies across the Pacific Ocean. For example, there is an isotopic gradient southeast of the Hawaiian Islands with a conspicuous localized region of 15 N - enriched waters (4 - 10 °N and 135 - 140 °W) (Fig. 4 ; 15 N gradients result from the prevalence 15 N of the primary nitrogen source of oceans, nitrate, is 5 - denitrification and phytoplankton uptake (Karl et 15 15 N at the base of the food web is transferred to consumer 15 N Phe values identify the biogeochemical regime in which seabirds forage. Spatial variation in nitrogen added to or removed from the ocean by biogeochemical processes is identified as N*. N* refers to the excess or deficit in nitrogen (N) relative to phosphorus (P) from the expected Redfield N:P stoichiometry of 16:1, where N* = N 16P + 2.9 - 1 (Deutsch et al. 2001). Whereas average marine nitrate corresponds to an N* of 0, nitrogen fixation and denitrification both alter Redfield stoichiometr y to increase or decrease N* from 0, respectively (Gruber and Sarmiento 1997; Deutsch et al. 2001). Anthropogenic atmospheric nitrogen deposition also produces positive N* values (Kim et al. 2014). However, atmospheric nitrogen deposition is greatest in co astal regions of the western North Pacific Ocean and is not likely a major influence on the food webs of pelagic seabirds that do not frequent coastal waters (Kim et al. 2014). While we cannot dismiss the possibility that atmospheric 23 nitrogen deposition af 15 N Phe 15 N Phe in terms of the relative 15 N. 15 N approach offers an important lens for understanding trophic dynamics of seabirds and the nutrient regimes on which they depend for food. Moreover, 15 N values provide valuable insight into how food web structure may have shifted over time (Hückstädt et al. 2017; Ostrom et al. 201 7; Gagne et al. 2018). The ability of amino acid specific data to expand our understanding of seabird foraging habits and develop a more comprehensive representation of oceanic food webs is facilitated by studying ecologically diverse trans - Pacific predato rs and carefully selecting tissue types. Whereas the timing of body contour molt is often poorly constrained, remiges more often harbor information related to a specific time period in the annual cycle. For example, the analysis of primary feathers provide d unique insights into the non - breeding season foraging habits of the Hawaiian petrel (Wiley et al. 2012; Morra et al. 2018). In contrast to feathers, the slow turnover time of bone collagen offers a record of isotopic information over a period of a year o r more. distribution, rather than reflecting a single season or foraging location (Rucklidge et al. 1992). We compared amino acid specific nitrogen isotope d ata from the Hawaiian petrel to those from two seabirds with distinct foraging strategies Newell's shearwater and Laysan albatross. Our data from collagen represent year - round foraging habits. The three study species grant extensive spatial coverage of the North Pacific Ocean and exhibit st ark ecological contrasts (Fig. 4 ). Newell's shearwater was once thought to be extinct on the Hawaiian Islands (Mitchell et al. 2005). Unique among our study species, Newell's shearwaters employ pursuit plunging and are ca pable of catching prey 10 meters beneath the surface (Ainley et al. 1997). Of the study species, 24 they have the most constrained marine distribution and fly the shortest distance on foraging trips during the breeding season 25% and 33% of Hawaiian petrel an d Laysan albatross tri p distance, respectively (Fig. 4 ; Spear et al. 1995; Fernandez et al. 2001; Adams and Flora 2010). In contrast to our other study species, Laysan albatross have a large and stable population (Croxall et al. 2005; BirdLife Internationa l 2017) and are not known to feed in association with tuna schools or in mixed - species flocks (Ainley et al. 1997; Simons and Hodges 1998; Awkerman et al. 2009; Ainley et al. 2014). Laysan albatross frequently scavenge, including from fishing vessels where by they are often killed as bycatch (Cousins et al. 2000), and they are extensively distributed across the North Pacific Ocean (BirdLife International 2017). 15 N Phe and 15 N Glu - Ph e from bone collagen of Newell's shearwater and Laysan albatross. We used our 15 N Phe 15 N Glu - Phe as an indicator of trophic position differences within and among our study species. We compared data extending 15 N records from the Hawaiian petrel. We also determined the probability that Newell's shearwater and Laysan albatross changed their foraging habits (i.e. the nutrient regime they associate with o r their trophic position) during the recent past. In addition to providing insight into seabird foraging dynamics, our study offers a more comprehensive understanding of ecosystem integrity in the North Pacific Ocean over a time when marine ecosystems expe rienced increased threats. 25 METHODS Sample Acquisition Samples dating from 2001 to present were collected from salvaged carcasses. Acquisition of Salvaged Newell's shearwaters predominantly consist of birds found dead after grounding by light attraction or and Laysan albatross from the Hawaii longline fisheri es between 2003 and 2014. Samples from prior to 2000 are from museum study skins from collections housed at the National Museum of Natural History, the Bernice Bishop Museum, and the California Academy of Sciences. Samples from after 2000 are designated as the modern time period in each species. Sample Sizes Samples from modern and historical specimens were derived from after hatch - year birds. Hawaiian petrels were aged previously by Wiley et al. (2013), and we determined Newell's shearwater ages using the color and shape of primaries (Pyle 2008). For historical Laysan albatross, we referred to the age designation indicated in the museum collection. Modern Laysan albatross were aged based on bursa size; the absence of a bursa indicates the bird was likely g reater than 4 years of age (Broughton 1994). We obtained bone samples from 22 Laysan albatross divided evenly between two time periods corresponding to before and after the onset of industrialized fishing in the North Pacific Ocean: pre - 1950 (1902 - 1937; th e historical sample) and post - 2000 (2003 - 2014; the modern sample). Of the 11 historical Laysan albatross, 8 were from Laysan Island, the remaining 3 were from Lisianski Island, the Aleutians West Census, and ided 24 samples into three time periods. The 26 oldest available museum samples were from 1964 - 1966 (n = 5); we compared those to two more recent time periods: 9 birds from 1983 - 1998 and 10 birds from 2013 - 2016. Newell's shearwaters prior to 2000 originated f Pacific Ocean within 500 miles of the Hawaiian Islands (n = 3). Ancient and modern Hawaiian petrel data are from Ostrom et al. (2017) and include Maui individuals from the Foundation (1000 - 14 00 CE; 550 - 950 y B.P., n = 5) and Modern (1950 - 2010, n = 7) time periods and birds - 1800 CE; 150 - 550 y B.P., n = 8) and Modern (1950 - 2010, n = 8) time periods. Hawaiian petrel time periods are defined by Kirch (1990) and represent archaeological periods in the development of human societies in the Hawaiian Islands. Sample Preparation Collagen was isolated and purified at Michigan State University according to Stafford et al. (1988) as modified by Wile y et al. (2013). Modern and ancient bone fragments (50 - 200 mg) were scraped with a razor blade and rinsed with ultrapure distilled water (E - pure, Barnstead). Clean bone fragments were demineralized with multiple changes of quartz - distilled 1 N hydrochloric acid. The demineralized samples were soaked in 0.05 N potassium hydroxide overnight to remove humate contaminants and the resulting collagen was lyophilized. Collagen was gelatinized with 0.05 N hydrochloric acid in a 105°C oven for 1 - 8 hours and passed t hrough stored frozen prior to analysis. Gelatin (0.5 - 1.2 mg) was hydrolyzed in 0.5 mL of quartz - distilled 12 N hydrochloric acid in a 105°C oven for approximat ely 20 hours. Lipids were removed from the resultant filtrate with n - hexane/dichloromethane (3:2, v/v), and evaporated to dryness in methanol under a gentle N 2 27 stream at 50°C. Amino acids in the lipid extracted hydrolysate were esterified and acylated with N - pivaloyl/isopropyl (NP/iPr) derivatization (Chikaraishi et al. 2009). Samples were esterified with thionyl chloride/2 - propanol (1:4, v/v) at 105°C for 2 hours then acylated with pivaloyl chloride/dichloromethane (1:4, v/v) at 105°C for 2 hours. The amin o acid derivatives were extracted with n - hexane/dichloromethane (3:2, v/v) and stored at 25 o C. 15 N Amino Acid A nalysis The nitrogen isotopic composition of individual amino acids was determined by gas chromatography/combustion/isotope ratio mass spectro metry using an Isoprime isotope ratio mass spectrometer (IRMS; Elementar, UK) coupled to a 7890 gas chromatograph (GC; Agilent Technologies, USA) via a combustion and reduction furnace. Combustion and reduction were performed in a glass capillary tube with CuO, NiO, and Pt wires at 950°C. The amino acids were injected on column at 250 °C and separated on a BPX - 5 capillary column (60 m x 0.32 mm inner programmed as fol lows: initial temperature 40°C for 2 min, ramp of 10°C min - 1 to 280°C and hold for 10 min, ramp of 10°C min - 1 to 325°C and hold for 25 min. Carrier gas (He) flow through the GC column was 1.6 ml min - 1 . The CO 2 and H 2 O generated in the combustion furnace we re removed from the sample stream using a liquid nitrogen trap. 15 N a = [( 15 N/ 14 N sample / 15 N/ 14 N standard ) - 1] x 10 3 relative to the standard, atmospheric N 2 . Accuracy was evaluated by daily analysis of external reference mixtures consisting of NP/iPr derivatives of several isotopically characterized amino acids (Gly, Val, Leu, Pro, Asp, Met, Glu, Phe). standar 28 the two. Nitrogen isotope analyses for Laysan albatross and Newell's shearwater were conducted at Michigan State University. The analyses for Hawaiian petrel samples from Ostrom et al. (2017) were performed by Yoshito Chikaraishi at the Japan Agency for Marine - Earth Science and Technology using a Delta - plus XP IRMS (Thermo Fisher Scientific) coupled to a 6890 GC (Agilent Technologies) via combustion and reduction furnaces. A blind inter - laboratory comparison on five of our Laysan albatross samples showed no influence of analyst or laboratory. Between laboratories, the averag 15 N Glu and 15 N Phe Model Selection We evaluated the effect of large scale climatic phenomena (i.e. El Niño Southern Oscillation, 15 N Phe 15 N Glu - Phe data by including the - - 2000) as 15 N Phe 15 N Glu - Phe were evaluated separately. Based on an estimate that avian bone collagen has a half - life of approximately 6 months (Hobson and Clark 1992), we assigned MEI classifications based on the average NOAA MEI ranking for the 18 months prior Information Criterion corrected for finite sample sizes (AICc). The model with the lowest AICc value contained year as the only independent variable, so we did not include MEI in our model . All historical Laysan albatross samples were from an MEI neutral period, eliminating the need for evaluating the influence of MEI on the historical data. Modern Laysan albatross were 29 from MEI neutral or moderate La Niña conditions. We asked if 15 N Phe a nd 15 N Glu - Phe differed between these two conditions with a two - tailed unpaired t - test that assumed homoscedasticity. Statistical Analysis W e developed a hierarchical method that account s for analytical variation and reduces type 1 errors in hypothesis testing . The method consists of two sub - models. The first, the observational model, estimates variation associated with the measurement process using replicate samples . The second , the ecologica l model, estimates population mean s and covariance between 15 N Phe and 15 N Glu - Phe given uncertainty due to analytical error. The o bservational m odel is and the e cological m odel is w here is the data vector indexed by individual (i), replicate (j), time bin (t), and species (s), and consisting of isotope values 1:r, where r is the number of amino acid variables used, is the mean vector of individual i, in time bin t for species s . is the covariance matrix (dimensions r by r) associated with the total analytical error shared across all individuals, time bins, and species. is the population mean vector for time t and species s and , the covariance matrix associated with naturally occurring isotopic variation , was assumed constant within (but not between) time bin t and species s . The model parameters were estimated in a Bayesian framework using the program JAGS (Plummer 2003) interfaced to R (R Development Core Team 2013). Minimally informative priors were used for estimated parameters (Rossman et al. 2016). The model was fit in JAGS using a Markov Chain Monte Carlo for 100,000 iterations with a 10,000 iter ation burn in and three chains. The posterior distributions were thinned at a rate of saving one iteration in every three. Convergence was ensured through monitoring traceplots and Rhat values (Gelman and Hill 2007). The probability that two parameters wer e different was calculated by summing the 30 number of posterior estimates in which one parameter was larger than the other dividing by the 15 N Phe and 15 N Glu - Phe differ among modern and historical colonies and species, and declined over time. We also used posterior estimates to visually represent the data and associated uncertainty. Standard ellipses characterize the foraging habits of each population. Estimated p opulation 15 N Phe 15 N Glu - Phe means define the center of each standard ellipse and a covariance matrix determines the shape and orientation (Jackson et al. 2011; Rossman et al. 2015). We generated probability density distributions for Newell's shearwa 15 N Phe and 15 N Glu - Phe values, similar to those for the Hawaiian petrel in Ostrom et al. (2017). RESULTS Model Selection In addition to time, we recognize isotope values may be influenced by MEI , however, there was no significan t difference in Laysan albatross 15 N Phe or 15 N Glu - Phe between MEI neutral and moderate La Ni ñ a conditions ( 15 N Phe t = 0.22, df = 9, P = 0.83; 15 N Glu - Phe t = 0.72, df = 9, P = he model with the lowest AIC c value (39.94, P = 0 .002 0 ) incorporated year as the only predictor . Interspecific and Inter - Colony Variation in Amino Acid 15 N Our results indicate that all three species as well as the Hawaiian petrels breeding on two separate islands likely have significantly different 15 N Glu - Phe means in modern as well as historical time periods ( P a; Fig. 5 ), although the historical time period of each Hawaiian petrels likely have a lower 15 N Phe mean in the historical time period ( P = 0.95) but not 31 in the modern time period ( P = 0.70). In modern and historical time periods, both Hawaiian 15 N Phe mean than Hawaiian petrels from the island of Hawai i and Laysan albatross ( P that Hawaiian petrels fr om the island of Hawai i have a lower 15 N Phe mean than Laysan albatross in the historic al time period ( P = 0.995). Temporal Variation in 15 N Phe and 15 N Glu - Phe The probability that 15 N Phe declined in Newell's shearwaters over the past 50 years was low (P = 0.27) (Table 6 b; Fig. 6 ). There was a 0.79 probability that 15 N Phe declined in Laysan albatross over the past 100 years. We did not consider this difference to be ecologically or statistically significant. Our results indicated at least a 0.97 probability that 15 N Glu - Phe declined in Newell's shearwater and Laysan albatr oss over the past 50 and 100 years, respectively. DISCUSSION Modern Time Periods The two Hawaiian petrel populations occupy unique isotopic niches despite the close proximity of their breeding colonies and their high degree of morphological similarity (Fi g 5 ; Wiley et al. 2012, 2013; Ostrom et al. 2017). 15 N Phe mean of the Maui Hawaiian petrel colony indicates these birds tend to utilize waters characterized by nitrogen fixation that are associated with positive N* ( Fig. 5 ; Ostrom et al. 2017 ). In contrast , less time foraging in biogeochemical regimes supplied by nitrogen fixation than those from Maui, shown by their higher mean 15 N Phe . Instead, Hawaiian petrels from Hawai i Island may frequent regions where nutrient sources at the base of the food web are influenced by average 32 marine nitrate or denitrification and/or substantial phytoplankton uptake. Such regions are characterized by lower N* values than the waters frequented by the Maui population. The possi bility that H awaiian petrel 15 N Phe is strongly influenced by differences in nutrient regime is supported by observation and tagging data. Adams and Flora (2010) tracked a breeding adult from Maui making a large counterclockwise loop over the northeast Pac ific, which consists of positiv e and negative N* waters (Fig. 4 ). During the nonbreeding season, Hawaiian petrels are predominantly distributed south and southeast of Hawaii ( Spear et al. 1995; VanZandt 2012), overlapping an area of 15 N - enriched water ( Alt abet and Francois 1994) . Differential foraging within a 15 N gradient could drive inter - colony spatial foraging segregation in Hawaiian petrels. Interestingly, Hawaiian petrels from Maui appear to occupy a higher trophic position than their Hawai i Island co nspecifics (Fig. 5 ; Ostrom et al. 2017). This may indicate that food chains supplied by nitrogen fixation are longer than those supplied by nitrate . Multiple factors may contribute to the observed intra - species trophic disparity. Individuals from Maui are slightly larger than those from the Hawai i colony in some linear dimensions (e.g. culmen and tarsus length , 8% and 10% larger, respectively; Judge et al. 2014) suggesting that Maui Hawaiian petrels are likely capable of consuming larger food item s, thereby elevating their trophic position. I t is uncertain whether the magnitude of the bill and body size difference s between Hawaiian petrel populations is large enough to confer a difference in trophic position. Although body size is often thought to correlate with prey size, some studies find that a 1000 - fold range in seabird body mass does not explain diet variation (Ballance et al. 2001). Thus, we considered that Hawaiian petrels from Maui and Hawai i Islands might specialize on distinct suites of prey 33 items under the species - level generalist umbrella. Niche partitioning with respect to trophic position and nutrient regime may facilitate population viability. I n addition to intra speci fic foraging segregation, we observed isotopic segregat ion between species. The Newell's shearwater standard ellipse does not overlap with those of the Hawaiian petrel colonies (Fig. 5 ) . The 15 N Phe mean of Newell's shearwaters is similar to that of Maui Hawaiian petrel s , likely resulting from a similar relian ce on oceanographic regions characterized by nitrogen fixation. Observational data show that Newell's shearwaters are dispersed across the isotopic gradient extending east from the area of 15 N - enriched waters southeast of the Hawaiian Islands (Spear et al. 1995 ). This could contribute to The 15 N Glu - Phe data suggest that Newell's shearwater s occupy a lower trophic position than either of the two Hawaiian petrel population s. Newell' s shearwater diet is poorly known. Ainley et al. provided the first report on Newell's shearwater diet in 2014, using stomach contents from carcasses of recently fledged individuals. The diet was dominated by squid (e.g. ommastrephids) with few fish, mainl y flying fish. This appears to be similar to Hawaiian petrel diet but with limited data it is unclear how much the relative proportion of diet items differs between species (Spear et al. 1995). If not a result of diet differences, t he apparent trophic sepa ration between Hawaiian petrels and Newell's shearwaters may be explained by a difference in body size, given that Newell's shearwaters have a slightly shorter wingspan and lower body mass (Ainley et al. 1997; Simons and Hodges 1998). However, a s discussed above, body size difference s may not be sufficient to explain the observed divergence in trophic position. T he 15 N Glu - Phe differences might also be driven by interactions in mixed - species feeding flocks. Although foraging in large diverse aggregations c an improve foraging success , evidence also suggests that intense interference competition can occur, especially among 34 seabirds employing distinct foraging methods (Schreiber and Burger 2001) . For example, surface - seizers, like Hawaiian petrel s , might reduc e access to prey by diving birds, like Newell's ( Schreiber and Burger 2001 ). Laysan albatross appear to occupy a distinct niche position . Their 15 N Phe mean indicates little reliance on food webs supplied by nitrogen fixation (Fig 5 ). Additionally, Laysan albatross appear to occupy the lowest trophic position of the three species. While our 15 N Glu - Phe data may seem surprising given the large size of this spe cies , recent data from contour feathers and remiges place Laysan albatross at the same or lower trophic position than smaller species (e.g. Bulweria bulwerii ; Gagne et al. 2018 ). While it is difficult to make comparisons between year - round data from collagen and short - term data from feather, data from both tissues emphasize a gap in our knowledge of Laysan albatross trophic dynamics or an unexplored factor 15 N of amino acids. Diet studies suggest Laysan albatross consume mostly squid, but because they are conducted at colonies or on a specific age class, they provide a limited understanding of diet (Harrison et al. 1983). In addition, stomach content analyses could be compromised by more rapid digestion of soft - bodied organisms that are likely to have a low trophic position (Gould et al. 1997) . Such bias would result in an overestimation of their trophic position , a possibility suggested by Gould et al. (1997 ). Alternatively, we considered whether modern Laysan albatross have low 15 N Glu - Phe because they represent a specific subset of the population (salvaged from fisheries). However, as discussed below, historical Laysan albatross also exhibit low average 15 N Glu - Phe , indicating that our results are not simply an attribute associated with one subset of the population. Our data force us to think broadly about the biology of our study species and factors that control 15 N Glu - Phe . Diet quality and mode of nitr ogen 35 excretion are two factors that we considered owing to their impact on 15 N discrimination and, thus, trophic discrimination factors (McMahon and McCarthy 2016 ). If TDF differs acutely among species, inter - specific 15 N Glu - P he variation may not solely reflect trophic position. This is because trophic position is a function of TDF. T he degree of 15 N enrichment of Glu b etween a consumer and its diet, TDF Glu , may be related to diet quality (i.e. amino acid imbalance and protein content; Bradley et al. 201 5; Chikaraishi et al. 2015; McMahon et al. 2015; Nielsen et al. 2015; McMahon and McCarthy 2016 ). If h igh trophic level organisms have protein rich diets or diets with amino acid compositions similar to their own bodies, they will exhibit low TDF Glu . Thus, i f, relative to Hawaiian petrel s, Laysan albatross consume a higher quality diet , they could have a lower TDF Glu , which would result in the relatively low 15 N Glu - Phe values we observed. Because our study species are generalist predators with some overlap in the prey they consume, it seems unlikely that differences in TDF account for the 15 N Glu - Phe differences we observe. However, a better understanding of the prey, protein, and amino a cid composition of our seabird diets will be important in assessing the influence of diet quality on 15 N Glu - Phe . Alternatively, or in addition to diet composition, metabolic effects might influence our 15 N Glu - Phe data. As a consequence of kinetic isoto pe effects, the product of a reaction is expected to have a lower isotope value than its substrate. Such isotopic discrimination occurs during nitrogen transformation reactions. Because our study species belong to the same taxonomic order, we do not antici pate large differences in amino acid transformation pathways and associated discrimination against 15 N. Instead, differential synthesis of amino acids or loss of excretory nitrogen could produce variation in 15 N Glu , particularly because glutamic acid has a central role in transamination and deamination. In birds, the primary excretory product of 36 nitrogen metabolism is uric acid. Two of the three nitrogen atoms in uric acid derive from glutamine and, ultimately, glutamic acid , a substrate in glutamine formation. Thus, the loss of 14 N during the formation and 15 N Glu . Uric acid can be derived from exogenous or endogenous proteins. But, we do not expect dietary protein to be in excess for the l ong - ranging seabirds we studied. Production of uric acid from endogenous protein is related to the extent of protein catabolism, and differences in water balance influence uric acid abundance regardless of origin (Jenni and Jenni - Eiermann 1998; Battley et al. 2000; Gerson and Guglielmo 2011; Braun 2015) . Catabolism of protein increases under various circumstances such as prolonged flight, breeding, stress, molt, and nutrient limitation , which differ among our study species (Mori and George 1978; Robin et al. 1987, Cherel et al. 1988; Pearcy and Murphy 1997; Jenni and Jenni - Eiermann 1998; Viblanc et al. 2017 ). Of the three study species, Hawaiian petrel s have been observed to fly farthest on a sing le foraging trip (10,000 km ; Spear et al. 1995; Fernández et al. 2001; Adams and Flora 2010). B irds often increase reliance on protein catabolism during extended flights (Gerson and Guglielmo 2011). Laysan albatross often skip a year of breeding whereas Hawaiian petrel shearwater s breed annually (Edwards and Rohw er 2005). Moreover, Laysan albatross is the only study species that does not undergo a complete annua l primary molt (Edwards and Rohw er 2005). Additionally, our study species exhibit differences in wing loading and flight type (Carey 2012) . A plethora of differences among our study species could affect their nitrogen budgets, leading to differential protein 15 N Glu fractionation. However, it is unlikely that the temporal decline in 15 N Glu - Phe observed for all three species was a consequence of identical and simultaneous changes in protein catabolism. 37 Temporal Variation O ur long - term 15 N isotope chronologies reveal that foraging segregation among modern seabird colonies and species is not a recent phenomenon, but has endured over the course of several decades. Moreover, the relative position of each isotopic niche is similar between historic al and modern time periods (Fig. 5 ) . 15 N Phe provide s no evidence of a shift in that, like the Hawaiian petrel (Wiley et al. 2013; Ostrom et al. 2017) , Newell's shearwater and Laysan albatr oss experienced a significant decline in troph ic position within the past century . The latter finding is in contrast to the findings of Gagne et al. (2018). This disparity could be related to the fact that Gagne et al. (2018) analyzed feathers that, unlike bone, represent only a short portion of the i O ur study species represent two families and diverse foraging strategies , and grant different oceanographic characteristics, such as se a surface temperature and salinity (Schreiber 1984; Spear et al. 1995; Block et al. 2011). The discovery of a trophic decline in three ecologically distinct pelagic predators indicates a remarkably pervasive shift in food web architecture across the North Pacific Ocean. Climate change is a widely accepted driver of oceanic ecosyst em change (Cheung et al. 2012; Woodworth - Jefcoats et al. 2013 ), however, we failed to find an effect of climate on feeding habits. This is similar to the results of Gagne et al. (2018) for Laysan albatross. A nother major anthropogenic force potentially capable of causing a widespread pelagic trophic decline, fishing , likely impacts modern seabird niches . With industrial fisheries, humans have the power to modify v ast oceanic ecosystems to a degree once thought impossible. There is mounting evidence that fisheries 38 compete with marine predators for food and alter distributions as well as abundances of marine organisms ( Tasker et al. 2000; Myers and Worm 2003; Ward an d Myers 2005; Polovina and Woodworth - Jefcoats 2013 ). The amino acid data may signify threats to marine ecosystem integrity, but they also offer important information to seabird conservationists by illuminating at - sea behavior. This is particularly critical for the Hawaiian petrel and Newell's shearwater, which have experienced drastic population declines in recent decades (Raine et al. 2017). The number of Newell's 13, largely due to collisions with power lines, light attraction, and predation by introduced predators (Raine et al. 2017). While many of the threats seabirds face on land are well - characterized , seabirds typically spend 90% of their lives at sea (Ballanc e et al. 2001) . Conservation managers cannot assess at - sea risks without sufficient inform ation regarding pelagic distribu tions and foraging habits . Our results enhance the perspective, brought forth in other studies, that biogeochemical regimes and trophi c dynamics of the North Pacific Ocean have recently changed (Wiley et al. 2013; Kim et al. 2014; Sherwood et al. 2014; Ostrom et al. 2017). The emerging challenge is to characterize how a changing oceanographic environment influences population viability, an important consideration given that no species is exempt from the possibility of extinction. 39 APPENDIX 40 Figure 4 . N* distribution, at - 15 values of sinking particles and core top sediments. The color gradient is the distribution of N* defined and gridded by Sherwood et al. (2014), where N* = N - 1, P = phosphorus. Positive N* values are interpreted as an increase in nitroge n fixation and negative values as net denitrification (Sherwood et al. 2014). Curved lines (north of line), solid = Hawaiian petrel (Ainley et al. 1997; Simons and Hodges 1998; Awkerman 15 N in the eastern tropical north Pacific Ocean and illustrate an isotopic gradient. As illustrated in Wiley et al. (2012), the isotopic gradient 15 N. Filled 15 N 15 N 15 N value. 41 Figure 5. 15 N Phe and 15 N Glu - Phe 15 N Phe and 15 N Glu - Phe values for each population define the center of the ellipse and a covariance matrix determines the shape and orientation. Legend indicates shearwater, LAAL = Laysan albatross, HAPE = Hawaiian petrel. Time periods and sample sizes - 1966, n - - 1998, n =9); Post - 2000 (2013 - 2016, n = 10); LAAL: Pre - 1950 (1902 - 1937, n = 11); Post - 2000 (2003 - 2014, n = 11); Maui HAPE: Foundation Period (1000 - 1400 CE; 550 - 950 y B.P., n = 5); Modern Period (1950 - e Expansion Period (1400 - 1800 CE; 150 - 550 y B.P., n = 8); Modern Period (1950 - with Hawaiian archaeological time periods defined by Kirch (1990). 42 Figure 6. Density distributions of 15 N Phe and 15 N Glu - Phe - - - 1998, n = 9); Post - 2000 (2013 - 2016, n = 10). Laysan albatross time periods and sample sizes: Pre - 1950 (1902 - 1937, n = 11); Post - 2000 (2003 - 2014, n = 11) . 43 Table 6 . Statistical results from hierarchical modeling of collagen isotope data showing variation among groups (colony or species). Probabilities ( P 15 N Phe 15 N Glu - Phe - population and interspecific Hawaiian petrel (HAPE) data for historical and modern time periods; (b) is a within species comparison of historical to modern data for NESH and LAAL. Statistical results for comparisons with Maui and (Ostrom et al. 2017). 44 LITERATURE CITE D 45 LITERATURE CITED Adams J, Flora S. 2010 Correlating seabird movements with ocean winds: linking satellite telemetry with ocean scatterometry. Mar. Biol. 157 : 915 - 929. Ainley DG, Telfer TC, Reynolds MH. 1997 Newell's Shearwater ( Puffinus newelli ), version 2.0. In : Rodewald PG (ed) The Birds of North America . Cornell Lab of Orn ithology, Ithaca, New York, USA. Ainley DG, Walker W A , Spencer GC, Holmes ND. 2014 Puffinus newelli in Hawaiian waters. Mar. Ornithol . 44 : 69 - 72 . Altabet M, Francois R . 1994 Sedimentary nitrogen isotopic ratio as a recorder for surface ocean nitrate utiliz ation. Glob. Biogeochem. Cycles 8: 103 - 116. Awkerman JA , Anderson DJ, Whittow GC. 2009 Laysan Albatross ( Phoebastria immutabilis ), version 2.0. In : Rodewald PG (ed) The Birds of North America . Cornell Lab of Orn ithology, Ithaca, New York, USA. Ballance LT, Pitman RL, Reilly SB. 1997 Seabird community structure along a productivity gradient: importan ce of competition and energetic constraint. Ecol. 78 : 1502 - 1518 . Ballance LT , Ainley DG, Hunt Jr. GL. 2001 Seabird Foraging Ecology. In: Steele JH , Thorpe SA, Turekian KK (eds ) Encyclopedia of Ocean Sciences. Academic Press, London, 5: 2636 - 2644. Battley PF , Piersma T, Dietz MW, Tang S , Dek inga A , Hulsman K . 2000 Empirical evidence for differential organ reductions during trans oceanic bird flight. Proc. R . Soc. B: Biol . Sci. 267 : 191 - 195. Bearhop S, Phillips RA, McGill R, Cherel Y , Dawson DA, Croxall JP. 2006 Stable isotopes indicate sex - specific and long - term individual foraging specialization in diving seabirds. Mar. Ecol. Prog. Ser. 311: 157 - 164. BirdLife International. 2 017 IUCN Red List for birds. Downloaded from http://www.birdlife.org on 10/12/2017. Block B A , Jonsen ID , Jorgensen S J , Winship A J, Shaffer SA, Bograd SJ , Hazen EL, Foley DG, Breed GA, Harrison AL, Ganong JE, Swithenbank A, Castleton M, Dewar H, Mate BR, Shillinger GL, Schaefer KM, Benson SR, Weise MJ, Henry RW, Costa DP. 2011 Tracking apex marine predator movements in a dynamic ocean. Nat. 475 : 86 - 90. Bolnick DI, Svanbäck R, Fordyce JM, Yang LH, Davis J M, Hulsey CD , Forister ML . 2003 The ecology of indiv iduals: incidence and implications of individual specialization. Am. Nat. 46 161 : 1 - 28. Bradley CJ, Wallsgrove NJ, Choy CA, Drazen JC, He therington ED, Hoen DK, Popp BN. 2015 Trophic position estimates of marine teleosts using amino acid compound specific isotopic analysis . Limnol. Oceanogr: Methods 13: 476 - 493. Braun EJ . 2015 Osmoregulatory systems of birds . In : Scanes CG (ed) Sturkie's Avian Physiology, 6 th edn. Academic Press, San Diego, California, USA, pp 285 - 300 . Broughton JM . 1994 Size of the bursa of Fabricius in relation to gonad size and age in Laysan and Black - footed Albatrosses. Condor 96 : 203 - 207. Carey C. 2012 Avian energetics and nutritional ecology . Springer Science and Business Media , New York, USA. Casciotti KL , Trull TW, Glover DM, Davies D. 2008 Constraints on nitrogen cycling at the subtropical North Pacific Station ALOHA from isotopic measurements of nitrate and particulate nitrogen. Deep. Res. Part II : Trop. Stud. Oceanogr. 55: 1661 - 1672 . Cherel Y, Robin JP, Walch O , Karmann H, Netchitailo P , Le Maho Y . 1988 Fasting in king penguin. I. Hormonal and metabolic changes during breeding. Am . J . Physiol. - Regul . , Integr . Comp . Physiol . , 254 : R170 - R177. Cheung WW, Meeuwig JJ, Feng M, Harvey E, Lam VW, Langlois T , Sl awinski D, Sun C, Pauly D. 2012 Cli mate - change induced tropicalisation of marine communities in Western Australia . M ar. Freshw . Res . 63 : 415 - 427. Chikaraishi Y, Kashiyama Y, Ogawa N O , Kitazato H, Ohkouchi N. 2007 Metabolic control of nitrogen isotope composition of amino acids in macroalgae and gastropods: implications for aquatic food web studies. Mar. Ecol. Prog. Ser. 342 : 85 - 90. Chikaraishi Y, Ogawa NO, Kashiyama Y, Takano Y, Suga H, Tomitani A, Miyashita H, Kitazato H, Ohkouchi N. 2009 Determinati on of aquatic food - web structure based on compound - specific nitrogen isotopic composition of amino acid s. Limnol. Oceanogr: Methods 7: 740 - 750. Chikaraishi Y, Steffan SA, Takano Y, Ohkouchi N. 2015 D iet quality influences isotopic discrimination among amino acids in an aquatic vertebrate. Ecol. Evol. 5: 2048 - 2059. Costanza R . 1999 The ecological, economic, and social importance of the oceans. Ecol . Econ. 31 : 199 - 213 . Cousins KL, Dalzell P, Gilman E. 2000 Managing pelagic longline - albatross interactions in the North Pacific Ocean. Mar. Ornithol. 28: 159 - 174. 47 Croxall JP, Silk JR, Phillips RA, Afanasyev V, Briggs D R. 2005 Global circumnavigations: tracking year - round ranges of nonbreeding albatrosses. Sci . 307 : 249 - 250. DeNiro MJ, Epstein S. 1981 Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta 45: 341 - 351. Deutsch C, Gruber N, Ke y RM, Sarmiento JL, Ganachaud A. 2001 Denitrification and N 2 fixation in the Pacific Ocea n. Glob. Biogeochem. Cycles 15: 483 - 506. Edwards AE, Rohwer S. 2005 Large - scale patterns of molt activation in the flight feathers of two albatross species. Condor 107: 835 - 848. Edwards AE, Fitzgerald SM, Parrish JK, Klavitter JL, Romano MD. 2015 Foraging s trategies of Laysan alba tross inferred from stable isotopes: implications for association with f isheries . PLoS One 10 : e0133471 . Fernández P, Anderson DJ , Sievert PR , Huyvaert KP. 2001 Foraging destinations of three low - latitude albatross (Phoebastria) species. J . of Zool . 254: 391 - 404. Gaebler OH , Choitz HC, Vitti TG, Vukmirovich R. 1963 Significance of N 15 excess in nitrogenous compounds of biological origin. Can . J. of B iochem . and P hysiol . 41 : 1089 - 1097 . Gagne TO, Hyrenbach KD, Hagemann ME, Van Houtan KS. 2018 Trophic signatures of seabirds suggest shifts in oceanic ecosystems. Sci. Adv. 4: eaao3946. Gelman A , Hill J. 2007 Data analysis using regression and multilevel/hierarchical models. Cambr idge University Press, New York, USA. Gerson AR, Guglielmo CG . 2011 Flight at low ambient humidity increases protein catabolism in migratory birds. Sci . 333 : 1434 - 1436 . Gould P, Walker W , Ostrom P. 1997 Foods of northern fulmars associated with high - seas drift nets in the transitional region of the North Pacific. Northwest. Nat. 78: 57 - 61 . Graham BS, Koch PL, Newsome SD, McMahon KW, Aurioles D. 2010 Using isoscapes to trace the movements and foraging behavior of top predators in oceanic ecosystems. In: West JB, Bowen GJ, Dawson TE, Tu KP (eds ) Isoscapes : Understan ding Movement, Pattern, and Process on Earth Through Isotope Mapping. Springer Scie nce, Dordrecht, The Netherlands, pp 299 - 318 . Gruber N, Sarmiento JL. 1997 Global patterns of marine nitrogen fixation and denitrification. Glob. Biogeochem. Cycles 1 1: 235 - 266 . 48 Hinke JT, Polito MJ , Goebel ME, Jarvis S, Reiss CS, Thorrold SR, Trivelpiece WZ, Watters GM . 2015 Spatial and isotopic niche partitioning duri ng winter in chinstrap and Adé lie penguins from the South Shetland Islands. Ecosphere 6: 125. Hobson KA, Clark RG . 1992 Assessing avian diets using stable isotopes I: Turnover of 13 C in tissues. Condor 94 : 181 - 188 . Hückstädt LA, McCarthy MD, Koch PL, Costa DP . 2017 What difference does a century make? Shifts in the ecosystem structure of the Ross Sea, Antarctica, as evidenced from a sentinel species, the Weddell seal. Proc. R. Soc. B 284: 20170927 . Jackson AL, Inger R, Parnell AC, Bearhop S . 2011 Comparing isotopic niche widths amon g and within communities: SIBER Stable Isotope Bayesian Ellipses in R. J . Anim. Ecol . 80 : 595 - 602 . Karl D, Letelier R, Tupas L, Dore J, Christian J, Hebel D. 1997 The role of nitrogen fixation in biogeochemical cycling in the subtropical North P acific Ocean. Nat. 388: 533 - 538. Kim IN, Lee K, Gruber N, Karl DM, Bullister JL, Yang S, Kim T W. 2014 Increasing anthropogenic nitrogen in the North Pacific Ocean. Sci . 346 : 1102 - 1106. Kirch PV. 1990 The evolution of sociopolitical complexity in prehistoric Hawaii: An assessment of the archaeological evidence . J. World. Prehist. 4: 311 - 345. Jenni L, Jenni - Eiermann S . 1998 Fuel supply and metabolic constraints in migrating birds. J . Avian Biol . 29: 521 - 528 . J udge SW, Hu D , Bailey CN . 2014 Comparative analyses of Hawaiian Petrel Pterodroma sandwichensis morphometrics . Mar . Ornithol . 42 : 81 - 84 . McClelland JW, Montoya JP . 2002 Trophic relationships and the nitrogen isotopic composition of amino acids in plankton. Ecol . 83 : 2173 - 2180 . McMahon KW, Thor rold SR, Elsdon TS, Mccarthy MD. 2015 Trophic discrimination of nitrogen stable isotopes in amino acids varies with diet quality in a ma rine fish. Limnol. Oceanogr. 60: 1076 - 1087. McMahon KW, McCarthy, M D . 2016 15 N fractionation: mechanisms, implication s, and applications for trophic ecology. Ecosphere 7 : e01511 . Minagawa M , Wada E. 1984 Stepwise enrichment of 15 N along food chains: further evidence and the relation between 15 N and animal age . Geochim. Cosmochim. Acta 48: 1135 - 1140. Mitchell C, Ogura C, Meadows D W, Kane A, Strommer L, Fretz S, Leonard D, McClung A . 2005 Department of Land and Natural Resources, Honolulu, Hawaii , USA. 49 Mori J G, George JC . 1978 Seasonal changes in serum levels of certain metabolites, uric acid and calcium in the migratory Canada goose (Branta canadensis interior). Comp . Biochem . Physiol . Part B: Comp . Biochem . 59 : 263 - 269. Morra KE, Wiley AE, James HF, Dutcher A, Ostrom PH. 2018 Influence of feather selection and sampling protocol on interpretations of Hawaiian Petrel ( Pterodroma sandwichensis ) nonbreeding season foraging habits from stable isotope analysis. Waterbirds 41: 93 - 100. Myers RA, Worm B. 2003 Rapid worldwide depletion of predatory fish communities. Nat. 423: 280 - 283 . Nielsen JM, Popp BN, Winder M. 2015 Meta - analysis of amino acid stable nitrogen isotope ratios for estimating trophic position in m arine organisms. Oecologia 178: 631 - 642. Ohkouchi N, Chikaraishi Y, Close HG, Fry B, Larsen T, Madigan DJ, McCarthy MD, McMahon KW, Nagata T, Naito YI, Ogawa NO, Popp BN, Steffan S, Takano Y, Tayasu I, Wyatt ASJ, Yamaguchi YT, Yokoyama Y. 2017 Advances in the application of amino acid nitrogen isotopic analysis in ecological and biogeochemical studies. Org. Geochem. 113: 150 - 174. Ostrom PH, Wiley A E, Jam es HF, Rossman S, W alker WA, Zipkin EF, Chikaraishi Y. 2017 Broad - scale trophic shift in the pelagic North Pacific revealed by an oceanic seabird. Proc . R. Soc. B 284: 20162436 - 20162442. Pearcy SD, Murphy ME . 1997 Essential amino acid metabolism in molting and non - molting sparrows in different nutritional states. Comp . Biochem . Physiol . Part A: Physiol . 118 : 1157 - 1163 . Plummer M. 2003 JAGS: a program for analysis of Bayesian graphical models using Gibbs Sampling. Proceedings of the Third International Workshop on Distributed Statistical Computing. R Project for Statist ical Computing, Vienna, Austria. Polovina J J, Woodworth - Jefcoats PA . 2013 Fishery - induced changes in the subtropical Pacific pelagic ecosystem size structure: observations and theory. PLoS One 8 : e62341 . Pyle P. 2008 Molt and age determination in Procellariiformes. In: Identification Guide to North American Birds, Part 2. Slate Creek Press, Point Reyes Station, CA, USA, pp 248 - 260 . R Development Core Team. 2013 R: a language and environment for statistical computing. R Foundation for Statis tical Computing, Vienna, Austria. Raine AF, Holmes ND, Travers M, Cooper BA, Day RH. 2017 Declining population trends of 119: 405 - 415. 50 Robin JP, Cherel Y, Girard H , Géloen A, Le Maho Y . 1987 Uric acid and urea in relation to protein catabolism in long - term fasting geese. J . Comp . Physiol . B 15 7 : 491 - 499 . Rossman S, Ostrom PH, Stolen M, Barros NB, Gandhi H, Stricker CA, Wells R S. 2015 Individual specialization in the foraging habits of female bottlenose dolphins living in a trophically diverse and habitat rich estuary. Oecologia 178 : 415 - 425. Rossman S, Ostrom PH, Gordon F , Zipkin EF . 2016 Beyond carbon and nitrogen: guidelines for estimating three - dimensional isotopic niche s pace. Ecol. and Evol. 6: 2405 - 2413. Rucklidge GJ, Milne G, McGaw BA, Milne E, Robins SP . 1992 Turnover rates of dif ferent collagen types measured by isotope ratio mass spectrometry. Bioc him. Biophys. Acta 1156: 57 - 61. Schreiber R W . 1984 Tropical seabird biology . Cooper Ornithological Society, Kansas, USA. Schreiber EA, Burger J . 2001 Biology of marine birds . CRC Press , Florida, USA. Sherwood OA, Guilderson TP, Bat ista FC, Schiff JT, McCarthy MD. 2014 Increasing subtropical North Pacific Ocean nitrogen fixation since the Litt le Ice Age. Nat. 505: 78 - 81. Sigman DM, Mccorkle DC, Francois R, Fisc her G. 2000 The 15 N of nitrate in the Southern Ocean: nitrogen cycling and circulation in the ocean interior. J. Geophys. Res. 105 : 19599 - 19614. Simons TR, Hod ges CN. 1998 Hawaiian Petrel ( Pterodroma sandwichensis ), version 2.0. In : Rodewald PG (ed) The Birds of North America . Cornell Lab of Ornithology, Ithaca, New York, USA. Spear LB, Ainley DG, Nur N, Howell SNG. 1995 Population size and factors affecting at - sea distributions of four endangered Procellariids in the tropical Pac ific. Condor 97: 613 - 638. Stafford TW, Brendel K, Duhamel RC. 1988 Radiocarbon, 13 C and 15 N analysis of fossil bone: Removal of humates with XAD - 2 resin. Geochim . Chosmochim . Acta 52: 2257 - 2267 . Tasker ML, Camphuysen CJ, Cooper J, Garthe S, Montevecchi WA, Blaber SJ. 2000 The impacts of fishing on marine birds . ICES J. Mar. Sci. 57 : 531 - 547. VanZandt ML. 2012 Distribution and habitat selection of the endangered Hawaiian Petrel ( Pterodroma sandwichensis ) D d issertat ion, University of Hawaii at Hilo, Hawaii, USA. Viblanc VA, Schull Q, Cornioley T, Stier A , Ménard JJ, Groscolas R, Robin J P . in press An integrative appraisal of the hormonal and metabolic changes induced by acute stress using king penguins as a model . Gen . C omp . E ndocrinol . 51 Ward P, Myers RA. 2005 Shifts in open - ocean fi sh communities coinciding with the commencement of commercial fi shin g. Ecol. 86 : 835 - 847. Wiley AE , Welch AJ , Ostrom PH, James HF , Stricker CA, Fleischer RC, Gandhi H, Adams J, Ainley DG , Duvall F, Holmes N. 2012 Foraging segregation and genetic divergence between geographically proximate colonies of a highly mobile seabird. Oecologia 168 : 119 - 130. Wiley AE, Ostrom PH, Welch AJ, Fleischer RC, Gandhi H, Southon JR, Stafford Jr. TR, Penniman JR, Hu D, Duvall FP, James HF. 2013 Millennial - scale isotope records from a wide - ranging predator show evidence of recent human impact to oceanic food webs. Proc . Na tl. Acad. Sci. 110: 8972 - 8977. Woodworth Jefcoats PA, Polovina JJ, Dunne JP, Blanchard JL . 2013 Ecosystem size structure response to 21st century climate projection: large fish abundance decreases in the central North Pacific and increases in the California Current. Glob. Chang. Biol. 19 : 724 - 733. 52 CHAPTER 3 Seasonal Variation in Foraging Habits of Two Hawaiian Seabirds ABSTRACT Pelagic seabirds confront many challenges as they course vast expanses of the open ocean in search of food. We evaluated seasonal differences in foraging habits of two ecologically d istinct species, Newell's shearwater and Laysan albatross, and for Laysan albatross, extended this seasonal analysis back 100 years. Our assessment of foraging habits relied on amino acid 15 N proxies for nutrient regime use ( 15 N Phe ) and trophic position ( 15 N G lu - P he ). Seasonal differences were revealed by contrasting data from feather grown during the nonbreeding season and data from collagen , which incorporates an average of foraging habits over a year or more. Our results show that Newell's shearwater and Laysan albatross foraging habits differ between breeding and nonbreeding seasons. Despite the fact that both species breed on the Hawaiian Islands, they use different foraging strategies depending on the season. While the nonbreeding season foraging ha bits of Laysan albatross have persisted over the past century, the seabirds experienced a trophic decline that was exclusive to the breeding season. In addition to addressing threats on land, conservation management strategies for threatened seabirds, including Newell's shearwater and Laysan albatross, will require an understandi ng of at - sea risks on seasonal timescales and may need to be tailored for individual species. 53 INTRODUCTION About 90% of spatially and temporally variable oceanographi c conditions, often while navigating thousands of kilometers in search of food (Ba l l ance et al. 2001). The foraging strategies seabirds employ to overcome the challenges of a pelagic lifestyle are difficult to characterize. Pelagic seabirds are inaccessibl e for most of their lives . T hey feed far from land, they remain entirely at sea during the nonbreeding season and, in many cases, during the first several years of life, as well ( Ballance et al. 2006 ). The importance of investigating at - sea behavior is rec ognized, particularly in the context of recent oceanographic change (Lewison et al. 2012; Wiley et al. 2013; Ostrom et al. 2017; Gagne et al. 2018). However, our understanding of how flexible seabird foraging strategies are in response to oceanographic cha nges within a single year and over several decades is incomplete . Such flexibility could include shifts in feeding location or dietary characteristics, like trophic position. Without an understanding of both breeding and nonbreeding season foraging, it is unclear whether oceanographic changes differentially affect seabirds over the course of the annual cycle. Studies of pelagic seabird foraging ecology have relied on stomach contents, satellite tagging and stable isotope analysis (Cherel 2008; Cook et al. 2013; Conners et al. 2015 ). Diet and satellite tagging involve handling live birds sometimes repeatedly. Su ch approaches may not be desirable, especially for endangered or threatened species. Yet, tagging studies can offer insights into feeding locations and strategies if parameters like flight speed, turning rates, and area - restricted - searching are monitored ( Adams and Flora 2010 ). While stomach content analyses provide information on prey ingested, they are influenced by differential digestion, reflect only short - term diet, and are often conducted on a specific subset of the population that 54 can be easily acces sed, such as chicks or adults on breeding grounds (Harrison et al. 1983; Ainley et al. 2014). The analysis of carbon and nitrogen isotopes values offers information on trophic level and foraging location (Cherel 2008; Wiley et al. 2013;). Because these ana lyses may be performed on the tissues of salvaged birds or museum specimens, they do not compromise live animals and reveal unique information complementary to stomach content analyses and tagging studies. There are numerous examples of the application of nitrogen isotope analysis in seabird foraging ecology, particularly for the identification of trophic level ( Farmer and Leonard 2011; Jaeger et al. 2013; Edwards et al. 2015; Hinke et al. 2015 ). These studies rely on the observation that 15 N systematically increases ~3 - ( Minagawa and Wada; Young et al. 2010 ). Several studies have taken advantage of the fact that different tissue s record foraging information over discrete time tur nover time (Young et al. 2010; Wiley et al. 2012). For example, bone collagen incorporates an average of foraging habits over a year or more, while feathers reflect foraging during their growth, on the scale of days to weeks (Rucklidge et al. 1992; Wiley e t al. 2010) . Regardless of tissue type, w 15 N values can be challenging to interpret because they are influenced However, c ompound specific nitrogen isoto pe analysis can separate these source and trophic effects (McClelland and Montoya 2002; McCarthy et al. 2007; Chikaraishi et al. 2007; Popp et 15 15 N Phe ), undergoes min imal fractionation during trophic transfer, remaining faithful to the isotope value of source nitrogen (e.g. nitrate, ammonium) at the base of the food web. As such, 15 N Phe 15 N uch as 55 15 N Glu ), becomes 15 N - enriched with increasing trophic level. This 15 N enrichment occurs during metabolic reactions (such as d eamination ) that discriminate against 15 N . The 15 N of Glu and Phe functions as a trophic proxy ( 15 N Glu - Phe ) and is used to calculate trophic position (TP): TP = ( 15 N Glu 15 N Phe )/TDF + 1 , where is the difference between 15 N Glu and 15 N Phe in primary producers and TDF is the trophic discrimination factor, or the net elevation between 15 N Glu and 15 N Phe per trophic step (McMahon and McCarthy 2016). Because of the similarity in TDF between feather and bird muscle, and muscle and collagen , the 15 N Glu - Phe of feather and collagen are comparable (Chikaraishi et al. 2014; Hebert et al. 2016; McMahon and McCarthy 2016; Blanke et al. 2017) . As a consequence, we can evaluate seasonal differences in trophic dynamics using flight feathers molted post - breeding and collagen that is assimilated year - round . 15 N Phe 15 N of source nitrogen varies with biogeochemical processes across the Pacific Ocean. The primary nitrogen 15 N of 5 - ( Alt abet and Francois 1994 ) . While d enitrification and phytoplankton uptake elevate th e 15 N of nutrients, food webs influenced by nitrogen fixation for example, waters near the Hawaiian Islands have lower values than those supplied by average marine nitrate ( Karl et al. 1997; Sigman et al. 2000; Casciotti et al. 2008). Spatial variation in these different nutrient regimes produces isotopic gradients within the north Pacific Ocean (Altabet and Francois 1994; Graham et al. 2010). One such gradient located southe ast of the Hawaiian Islands with a conspicuous localized region of 15 N - enriched waters around 4 - 10 °N and 135 - 140 °W . 15 N Phe values of 15 N of source nitrogen, they reveal the biogeochemical regime s on which seabirds depend f or food. 56 15 N Glu - Phe and 15 N Phe of feathers can resolve seasonal foraging habits when the timing and sequence of molt is known. Relative to body contours, the molt of remiges is often better constrained (Warham 1996). For example , remiges can provide foraging information related to a specific stage of the annual cycle (Warham 1996). As Procellariiformes, our study species Newell's shearwater ( Puffinus newelli) and Laysan albatross ( Phoebastria immutabilis ) molt their ten primaries during the nonbreedin g season ( Warham 1996 ; Edwards and Rohwer 2005 ). Primary molt in Newell's shearwaters takes place annually . As with other shearwaters, it likely begins at the innermost primary (P1) and proceeds distally (Warham 1996 ). Laysan albatross flight feather molt is complicated with one of four molt patterns classified as large, medium, or small in scope occurring annually (Edwards and Rohw er 2005). Only the outer three primaries are necessarily molted every year and of these , primary 10 (P10) is always molted last (Langston and Rohwer 1996). Isotope analysis of primary feathers that are replaced annually near the end of the nonbreeding season will paint the best picture of post - breeding foraging habits, when the birds are not tied to the colony. Thus, for Newell's shearwater, we analyzed primary 9 (P9); for Laysan albatross, we selected P10 for analysis . In this study, we compare modern nonbreeding season to year - round foraging habits using 15 N Glu - Phe and 15 N Phe of primary feathers and bone collagen , respectively , for two ecologically distinct species Newell's shearwater and Laysan albatross. Given that isotopic differences between feather and collagen must derive from the breeding season, we can draw inferences about breeding season foraging. Additionally, we evaluate Laysan albatross season al foraging habits over the past century to determine whether the trophic decline occurred and if it was associated with a particular stage of the annual cycle. 57 METHODS Sample Acqu isition Samples dating from 2003 to present were collected from salvaged carcasses and are designated as the modern time period in each species. Salvaged Newell's shearwaters were acquired from birds fo und dead after grounded by light attraction or killed by introduced predators. Salvaged Laysan albatross were collected from the Hawaii longlin e fisheries between 2003 and 2014. Laysan albatross samples from prior to 1950 make up the historical time period . These samples are from museum study skins from collections housed at the National Museum of Natural History and the California Academy of Sciences . Sample Sizes Samples from all specimens were derived from after hatch - year birds. We determined Newell's shearwater ages using the color and shape of primaries (Pyle 2008). For historical Laysan albatross, we referred to the age designation indicated in the museum collection. Modern Laysan albatross were aged based on bursa size; the absence of a bursa indica tes the bird was at least four years old and likely of breeding age (Broughton 1994). Newell's shearwater samples were t aken fr om the base of primary 9 (n=10); Laysan albatross samples were from the base of primary 10 (n=19) . Laysan albatross samples were divided between two time periods corresponding to before and after the onset of industrialized fishing in the North Pacific Ocean: historical (1902 - 1937, n=9) and modern (2003 - 2014, n=10). Of the historical samples , 6 individuals were from La ysan Island and the remaining 3 were from Lisianski Island, the Aleutians West Census, and Midway Islands. For both study species, bone collagen from the same individuals was previously extracted and prepared as described in Chapter 2. 58 Sample Preparation Barbs (~0.5mg) were plucked from the base of the feather with forceps, washed (87:13 v/v chloroform:methanol), rinsed with ultrapure distilled water (E - Pure, Barnstead), and dried (25°C). Cleaned feather barbs were cut into thirds and hydrolyzed in 0.5 mL of quartz - distilled 12 N hydrochloric acid in a 105 °C oven for approximately 20 hours. Lipids were removed from the resultant filtrate with n - hexane/dichloromethane (3:2, v/v), and evaporated to dryness in methanol under a gentle N 2 stream at 50°C. Amino acids in the lipid extracted hydrolysate were esterified and acylated with N - pivaloyl/isopropyl (NP/iPr) derivatization (Chikaraishi et al. 2009). Samples were esterified with thionyl chlor ide/2 - propanol (1:4, v/v) at 105 °C for 2 hours then acylated with p ivaloyl chloride/d ichloromethane (1:4, v/v) at 105 °C for 2 hours. The amino acid derivatives were extracted with n - hexane/dichloromethane (3:2, v/v) and stored at - 25 o C. 15 N Amino Acid A nalysis The nitrogen isotopic composition of individual amino acids was determined by gas chromatography/combustion/isotope ratio mass spectrometry using a n Isoprime isotope ratio mass spectrometer (IRMS; Elementar, UK) coupled to a 7890 gas chromatograph (GC; Agilent Technologies, USA) via a combustion and reduction furna ce. C ombustion and reduction were performed in a glass capillary tube with C uO, NiO, and Pt wires at 950°C. The amino a cids were injected on column at 250 °C and separated on a BPX - 5 capillary column (60 m x 0.32 mm inner GE Analytical Science, USA). The GC oven temperature was programmed as follows: initial temperature 40°C for 2 min, ramp of 10°C min - 1 to 280°C and hold for 10 min, ramp of 10°C min - 1 to 325°C and hold for 25 min. Carrier gas (He) flow through the GC colum n was 1.6 ml min - 1 . The CO 2 and H 2 O generated in the combustion furnace was removed from the sample stream using a liquid nitrogen trap. 59 15 N a = [( 15 N/ 14 N sample / 15 N/ 14 N standard ) 1] x 10 3 relative to the standard, atmospheric N 2 . Accuracy was evaluated by daily analysis of external standard mixtures consisting of NP/iPr derivatives of several isotopically characterized amino acids (Gly, Val, Leu, Pro, Asp, Met, Glu, Phe). Reproducibility o d the average of the two. Model Selection We considered the effect of large scale climatic phenomena (i.e. El Niño Southern Oscillation, ENSO ) on our 15 N Phe and 15 N Glu - Phe data . All historic al Laysan albatross samples were from an MEI neutral period, eliminating the need for evaluating the influence of MEI on these data. Modern Laysan albatross samples were from MEI neutral (n=7) or moderate La Niña ater samples were from MEI neutral conditions (n=8), with only 2 from moderate El Niño conditions. Within each species, we asked if 15 N Phe and 15 N Glu - Phe differed between MEI conditions with a two - tailed unpaired t - test that assumed homoscedasticity. S tatistical Analysis W e developed a hierarchical method that account s for analytical variation and reduces type 1 errors in hypothesis testing . The method consists of two sub - models. The first, the observational model, estimates variation associated with the measurement process using replicate samples . The second , the ecological model, estimates population mean s and covariance between 15 N Phe and 15 N Glu - Phe given uncertainty due to analytical error. The o bservational m odel is 60 and the e cological m odel is w here is the data vector indexed by individual (i), replicate (j), time bin (t), and species (s), and consisting of isotope values 1:r, where r is the number of amino acid isotope variables used, is the mean vector of individual i, in time bin t for species s . is the covariance matrix (dimensions r by r) associated with th e total analytical error shared across all individuals, time bins, and species. is the population mean vector for time t and species s and , the covariance matrix associated with naturally occurring isotopic variation , was assumed constant within (but not between) time bin t and species s . T he model parameters were estimated in a Bayesian framework using the program JAGS (Plummer 2003) interfaced to R (R Development Core Team 2013 ). Minimally informative priors were used for estimated parame ters (Rossman et al. 2016). The model was fit in JAGS using a Markov Chain Monte Carlo for 10 0,000 iteration s with a 10 ,000 iteration burn in and three chains. The posterior distributions were thin ned at a rate of saving one iteration in every three . Conve rgence was ensured through monitoring traceplots and Rhat values ( Gelman and Hill 2007). The probability that two parameters were different was calculated by summing the number of posterior estimates in which one parameter was larger than the other and div iding by the total number of posterior estimates. Specifically, w e report probabilities that 15 N Phe and 15 N Glu - Phe differ between tissue types and between species, and for Laysan albatross, that they declined over time. We also used posterior estimates of 15 N Phe and 15 N Glu - Phe to develop standard ellipses and probability density distributions. 15 N Phe and 15 N Glu - Phe means define the center of each standard ellipse and a covariance matrix determines the shape and orientation 61 (Jackson et al. 2011; Rossman et al. 2015). Standard ellipses characterize the foraging habits of modern Newell's shearwater and Laysan albatross by tissue type. The p robability density distribution s were generated for modern and historical Laysan albatross feather and collagen 15 N Phe and 15 N Glu - Phe values. RESULTS Climate Effects T here was no significant difference in Laysan albatross 15 N Phe or 15 N Glu - Phe between MEI neutral and moderate La Niña conditions ( 15 N Phe t = 1.9, df = 8, P = 0.10; 15 N Glu - Phe t = 1.4, df = 8, P 15 N Phe or 15 N Glu - Phe between MEI neutral and moderate El Niño conditions ( 15 N Phe t = 0.13, df = 8, P = 0.90; 15 N Glu - Phe t = 2.1, df = 8, P = 0.06). Variation in 15 N Phe and 15 N Glu - Phe Between Feather and Bone C ollagen Our results had higher 15 N Phe and lower 15 N Glu - Phe means in feather re lative to bone collagen ( Table 7 a, P > 99% ; Fig. 7 ). Modern Laysan albatross had lower 15 N Phe and higher 15 N Glu - Phe means in feather re lative to bone collagen ( Table 7 a, P ; Fig. 7 ). Historically, Laysan albatross feather 15 N Phe means were lower than those of bone collagen ( P > 99%) but it is unl ikely that feather 15 N Glu - Phe means were higher than those of bone collagen (Table 7 a , P = 60% ; Fig. 8 ). Variation in 15 N Phe and 15 N Glu - Phe B etwe en Species and Time P eriods Modern Laysan albatross feather had lower 15 N Phe and higher 15 N Glu - Phe means relative to the ther (Table 7 b, P ; Fig. 7 ). There was an 81% probability that 15 N Phe declined in Laysan albatross feather over the past 100 years and only a 47% probability 62 of a decline for 15 N Glu - Phe (Table 7 b ; Fig. 8 ). We did not consider either of these differences to be ecologically or statistically significant. DISCUSSION Our comparison of amino acid nitrogen isotope data from feather and collagen reveals that two ecologically distinct species, Newell's shearwater and Laysan albatross, both alter their foraging habits during their annual cycles in a species - specific manner. Furthermore, our data support a trophic decline for Laysan albatross over the past century, a phenomenon that appears to be isolated to a particular season of the annual cycle. Modern Comparisons In the case of Newell's shearwater, trophic position durin g the nonbreeding season is lower than during the breeding season. We are able to ascertain this because the high 15 N Glu - Phe of collagen (year - round data) relative to feather (nonbreeding season data) must derive from the breeding season. The observation that s eabirds often feed their chicks different food items than they eat year - round may impose the unique foraging strategy associated with the breeding season ( Hodum and Hobson 2000; Le Corre 2003; Cherel 2008; Cherel et al. 2008; Young et al. 2010). If high trophic level prey offer a nutritional benefit, they may be important for satisfying the energetic demands of chick rearing (Gutowsky et al. 2009). In addition to seasonal trophic differences, chick provisioning may drive seasonal variation in Newell's shearwater foraging location. In comparison to the nonbreeding season, Newell's shearwater 15 N Phe values indicate a greater reliance on food webs supp lied by nitrogen fixation during the breeding season. This may indicate that breeding adults forage extensively in the waters surrounding the Hawaiian Islands, where primary productivity is significantly 63 supported by nitrogen fixation (Kim et al. 2014). O b servational data show Newell's shearwaters less dispersed from Hawaii when they are colony - bound , perhaps to facilitate frequent visits to the burrow (Spear et al. 1995). In contrast, the large variation in nonbreeding season 15 N Phe values is consistent with dissemination across a broad range that overlaps with a nitrogen isotopic gradient extending east from 15 N - enriched waters southeast of Hawaii . While Newell's shearwaters are year - long residents of the eastern tropical Pacific, Laysan a lbatross occupy a marine range that encompasses the vast majo rity of the No rth Pacific Ocean (BirdLife 2018 ). However, several tracking studies reveal that Laysan albatross distribution contracts substantially during the early stages of the breeding season , when frequent visits to the colony are required (Fernandez et al. 2001; Hyrenbach et al. 2002; Young 2009; Conners et al. 2015; Gutowsky et al. 2015; Thorne et al. 2015). We considered whether this contracted range could influence Laysan albatross 15 N G lu - Phe data. Laysan albatross that breed in the Hawaiian Islands are atypical among albatrosses in their reliance on tropical waters during the breeding season, where resources are less abundant and patchier than in higher latitudes (Ainley 1977; Seki and Polovina 2001; Conners et al. 2015). Moreover, their constricted breeding season marine range increases the potential for competition (Ashmole 1963; Birt et al. 1987; Schreiber 2001). For example, although Laysan and black - footed albatrosses exhibit spatia l segregation at sea for most of the year, during the early breeding season the two species overlap significantly (Fernandez et al. 2001; Fischer et al. 2009; Kappes 2009; Kappes et al. 2010). Perhaps competition in suboptimal low latitudes during an energ etically demanding stage of the annual cycle contributes to Laysan albatross relying on lower trophic level prey. Substantial seasonal differences in the marine distribution of Laysan albatross may explain seasonal variation in nutrient regime use. Laysan albatross 15 N Phe values are lower in the 64 nonbreeding season than the yearly average , suggesting a high reliance on waters supported by nitrogen fixation post - breeding . While our understanding of regional variation in nitrogen fixation within the North Pa cific Ocean continues to emerge, our data suggest that Laysan albatross are able to access regions of high nitrogen fixation when they are released from their central place foraging strateg y and expand their marine range ( Shiozaki et al. 2010; Young et al. 2010 ; Luo et al. 2012 ). In addition to illuminating intra - species seasonal foraging differences, our study also permits assessment of inter - species differences in foraging habits within each season. Feather data suggest that, relative to Laysan albatross , Newell's shearwater s occupy a lower trophic position and rely less on food webs supplied by nitrogen fixation during the nonbreeding season. However, during the breeding season, collagen data reveal that both of these relationships are reversed. At this time, Newell's shearwaters appear to occupy a higher trophic position and rely more on food webs supplied by nitrogen fixation than Laysan albatross. Thus, while both species exhibit flexibility in their foraging strategies throughout the annual cycle, thi s flexibility manifests itself differently in each species and yields inter - species foraging differences that are season - specific. The dynamic relationship between Newell's shearwater and Laysan albatross foraging habits would have escaped notice if we had not analyzed both feather and collagen. Temporal Comparisons Flexibility in Laysan albatross foraging strategies may be a particularly beneficial trait given that they appear to have experienced a trophic decline over the past century. Because this troph ic decline was evident from collagen but not feather, it seems to be associated solely with the breeding season, which presents unique challenges. Trophic declines have been linked to reproductive decline in other seabirds ( Gutowsky et al. 2009 ). Reproduct ion is an acutely 65 energetically expensive phase of the annual cycle when breeders must balance foraging efficiency with rate of energy delivery to the chick (Weimerskirch et al. 2003). Energy requirements may not be fulfilled even if low trophic level prey may have higher energy content than their high trophic level counterparts . If low trophic level prey have a small body size, higher energy content is unlikely to compensate for the increased foraging effort required to catch many smaller organisms in patc hy low - latitude waters (Gutowsky et al. 2009). Moreover, catching numerous small prey may contribute to increased trip duration by breeders, which can negatively influence reproductive success ( Thorne et al. 2015 ). Beyond the potential consequences for Laysan albatross reproduction, the trophic decline we identified has broader implications. As top predators, Laysan albatross act as sentinels of food web dynamics because they integrate trophic information up their fo od web. Additionally, if their trophic decline is a result of oceanographic change over the past century, any number of marine predators may be similarly affected, an important consideration as we evaluate the consequences of climate change and fisheries o n pelagic ecosystems. Conclusions Identifying how foraging strategies of two ecologically distinct species Newell's shearwater and Laysan albatross differ and vary over a year empowers conservation efforts . Our results imply that management strategies mus t be tailored to individual species and require addressing threats both on land and at sea. Moreover, because foraging strategies vary over several decades, the ability to look back in time is essential for securing a future for marine species. Laysan alba tross appear to be particularly vulnerable to oceanographic change during the breeding season, perhaps because their ability to seek out alternate food sources is hindered . This vulnerability could have severe implications for popul ation viability, especia lly in various lo ng - 66 lived, low fecundity species, including Laysan albatross and Newell's shearwater. As humans continue to increase competition with seabirds for food, the ability of marine species to exploit alternative resources will be essential. 67 APPENDIX 68 Figure 7 . 15 N Phe 15 N Glu - Phe 15 N Phe 15 N Glu - Phe values for each population define the center of the ellipse and a covariance matrix determines the shape and o rientation. Legend indicates species and tissue type. Time periods and sample sizes (n) follow. NESH: 2013 - 2016, n = 10 for feather and collagen; LAAL: 2003 - 2014, n = 10 for feather, n = 11 for collagen. 69 Figure 8. Density distributions of model 15 N Phe 15 N Glu - Phe Laysan albatross (a) feather and (b) collagen over the past 100 years . T ime periods and s ample sizes (n) follow: Pre - 1950 (1902 - 1937) n = 9 for feather, n = 11 for collagen; Post - 2000 (2003 - 2014) n = 10 for feathe r, n = 11 for collagen 70 Table 7 . Statistical results from hierarchical modeling of feather and collagen isotope data showing variation among populations. Probabilities ( P 15 N Phe and 15 N Glu - Phe inter - species comparison between NESH and LAAL, and a temporal comparison for LAAL. 71 LITERATURE CITED 72 LITERATURE CITED Adams J, Flora S. 2010 Correlating seabird movements with ocean winds: linking satellite telemetry with ocean scatterometry. Mar. Biol. 157 : 915 - 929. Ainley DG, Telfer TC, Reynolds MH. 1997 Newell's Shearwater ( Puffinus newelli ), version 2.0. In : Rodewald PG (ed) The Birds of North America. Cornell Lab of Ornithology, Ithaca, New York, USA. Puffinus newelli in Hawaiian waters. Mar. Ornithol. 44 : 69 - 72. Altabet M, Francois R. 1994 Sedimentary nitrogen isotopic ratio as a recorder for surface ocean nitrate utilization. Glob. Biogeochem. Cycles 8 : 103 - 116. Ashmole NP. 1963 The regulation of numbers of tropical oceanic birds. Ibis 103b : 458 - 473. Ballance LT, Ainley DG, Hunt Jr. GL. 2001 Seabird Foraging Ecology. In : Steele JH, Thorpe SA, Turekian KK (eds) Encyclopedia of Ocean Sciences . Academic Press, London, 5 : 2636 - 2644. Ballance LT, Pitman RL, Fiedler PC. 2006 Oceanographic influences on seabirds and cetaceans of the eastern tropical Pacific: a review. Prog. Oceanogr. 69 : 360 - 390. BirdLife International. 2018 IUCN Red List for birds. Downloaded from http://www.birdlife.org on 1/12/2018. for prey depletion by a seabird. Mar. Ecol. Progr. Ser. 40: 205 - 208. Blanke CM, Chikaraishi Y, Takizawa Y, Steffan SA, Dhara mpal PS, Vander Zanden MJ. 2017 Comparing compound - specific and bulk stable nitrogen isotope trophic discrimination factors across multiple freshwater fish species and diets. Can. J. Fish. Aquat. 74 : 1291 - 1297. Broughton JM. 1994 Size of the bursa of Fab ricius in relation to gonad size and age in Laysan and Black - footed Albatrosses. Condor 96 : 203 - 207. Casciotti KL, Trull TW, Glover DM, Davies D. 2008 Constraints on nitrogen cycling at the subtropical North Pacific Station ALOHA from isotopic measurements of nitrate and particulate nitrogen. Deep. Res. Part II: Trop. Stud. Oceanogr. 55 : 1661 - 1672. Cherel Y. 2008. Isotopic niches of emperor and Adélie penguins in Adélie Land, Antarctica. Mar. Biol . 154 : 813 - 821. 73 Cherel Y, Le Corre M, Jaquemet S, Menard F, Richard P, Weimerskirch H. 2008 Resource partitioning within a tropical seabird community: new information from stable isotopes. Mar. Ecol. Prog. Ser . 366 : 281 - 291. Chikaraishi Y, Kashiyama Y, Ogawa NO, Kitazato H, Ohkouchi N. 2007 Metabolic control of nitrogen isotope composition of amino acids in macroalgae and gastropods: implications for aquatic food web studies. Mar. Ecol. Prog. Ser . 342 : 85 - 90. Chikaraishi Y, Ogawa NO, Kashiyama Y, Takano Y, Suga H, Tomita ni A, Miyashita H, Kitazato H, Ohkouchi N. 2009 Determination of aquatic food - web structure based on compound - specific nitrogen isotopic composition of amino acids. Limnol. Oceanogr: Methods 7 : 740 - 750. Chikaraishi Y, Steffan SA, Ogawa NO, Ishikawa NF, Sasaki Y, Tsuchiya M, Ohkouchi N. 2014 High - resolution food webs based on nitrogen isotopic composition of amino acids. Ecol. Evol. 4 : 2423 - 2449. Conners MG, Hazen EL, Costa DP, Shaffer SA. 2015 Shadowed b y scale: subtle behavioral niche partitioning in two sympatric, tropical breeding albatross species. Mov. Ecol. 3 : 28. Cook TR, Lescroel A, Cherel Y, Kato A, Bost C - A. 2013 Can foraging ecology drive the evolution of body size in a diving endotherm? PLoS One 8 : e56297. Edwards AE, Rohwer S. 2005 Large - scale patterns of molt activation in the flight feathers of two albatross species. Condor 107 : 835 - 848. Edwards AE, Fitzgerald SM, Parrish JK, Klavitter JL, Romano MD. 2015 Foraging strategies of Laysan albatross inferred from stable isotopes: implications for association with fisheries. PLoS One 10 : e0133471. Farmer RG, Leonard ML. 2011 Long - term feeding ecology of Great Black - backed Gulls ( Larus marinus ) in the northwest Atlantic: 110 years of feather isotope data. Can. J. Zool. 89 : 123 - 133. Fernández P, Anderson DJ, Sievert PR, Huyvaert KP. 2001 Foraging destinations of three low - latitude albatross (Phoebastria) species. J. Zool. 254 : 391 - 404. Fischer KN, Suryan RM, Roby DD, Balogh GR. 2009 Post - breeding season distribution of black - footed and Laysan albatrosses satellite - tagged in Alaska: inter - specific differences in spatial overlap with North Pacific fisheries. Biol. Conserv. 142 : 751 - 760. G agne TO, Hyrenbach KD, Hagemann ME, Van Houtan KS. 2018 Trophic signatures of seabirds suggest shifts in oceanic ecosystems. Sci. Adv. 4 : eaao3946. Gelman A, Hill J. 2007 Data analysis using regression and multilevel/hierarchical models. Cambridge University Press, New York, USA. 74 Graham BS, Koch PL, Newsome SD, McMahon KW, Aurioles D. 2010 Using isoscapes to trace the movements and foraging behavior of top predators in oceanic ecosystems. In : West JB, Bowen GJ, Dawson TE, Tu KP (eds ) Isoscapes : Understanding Movement, Pattern, and Process on Earth Through Isotope Mapping . Springer Science, Dordrecht, The Netherlands, pp 299 - 318. Gutowsky SE, Janssen MH, Arcese P, Kyser TK, Ethier D, Wunder MB, Bertram DF, Tranquilla LM, Lougheed C, Norris DR. 2009 Concurrent declines in nestling diet quality and reproductive success of a threatened seabird over 150 years. Endanger. Species Res. 15 : 247 - 254. Gutowsky SE, Leonard ML, Conners MG, Shaffer SA, Jonsen ID. 2015 Individual - level variation and higher - level interpretations of space use in wide - ranging species: an albatross case study of sampling effects. Front. Mar. Sci . 2 : 93. Harrison CS, Hida TS, S eki MP. 1983 Hawaiian seabird feeding ecology. Wildl. Monogr. 85 : 3 - 71. - Lyons C, Rattner BA, Wallsgrove N. 2016 Amino acid specific stable nitrogen isotope values in avian tissues: insights from captive American Kestrels and wild Herring Gulls. Environ. Sci. Technol. 50 : 12928 - 12937. Hink e JT, Polito MJ, Goebel ME, Jarvis S, Reiss CS, Thorrold SR, Watters GM. 2015 Spatial and isotopic niche partitioning during winter in chinstrap and Adélie penguins from the South Shetland Islands. Ecosphere . 6 : 1 - 32. Hodum PJ, Hobson KA. 2000 Trophic rel ationships among Antarctic fulmarine petrels: insights into dietary overlap and chick provisioning strategies inferred from stable - isotope ( 15 N and 13 C ) analyses. Mar. Ecol. Prog. Ser. 198 : 273 - 281. Hyrenbach KD, Fernández P, Anderson DJ. 2002 Oceanogr aphic habitats of two sympatric North Pacific albatrosses during the breeding season. Mar. Ecol. Prog. Ser. 233 : 283 - 301. Jackson AL, Inger R, Parnell AC, Bearhop S. 2011 Comparing isotopic niche widths among and within communities: SIBER Stable Isotope Bayesian Ellipses. R. J. Anim. Ecol. 80 : 595 - 602. Jaeger A, Jaquemet S, Phillips RA, Wanless RM, Richard P, Cherel Y. 2013 Stable isotopes document inter - and intra - specific variation in feeding ecology of nine large southern Procellariiformes. Mar. Ecol. Prog. Ser. 490 : 255 - 266. Kappes M. 2009 Comparative foraging ecology and energetics of albatrosses, PhD dissertation, Uni versity of California, Santa Cruz, California, USA. Kappes MA, Shaffer SA, Tremblay Y, Foley DG, Palacios DM, Robinson PW, Bograd SJ, Costa DP. 2010 Hawaiian albatrosses track interannual variability of marine habitats in 75 the North Pacific. Prog. Oceanog r. 86 : 246 - 260. Karl D, Letelier R, Tupas L, Dore J, Christian J, Hebel D. 1997 The role of nitrogen fixation in biogeochemical cycling in the subtropical North Pacific Ocean. Nature 388 : 533 - 538. Kim IN, Lee K, Gruber N, Karl DM, Bullister JL, Yang S, Kim TW. 2014 Increasing anthropogenic nitrogen in the North Pacific Ocean. Science 346 : 1102 - 1106. Langston NE, Rohwer S. 1996 Molt - breeding tradeoffs in albatrosses: life history implications for b ig birds. Oikos 76 : 498 - 510. Le Corre M, Cherel Y, Lagarde F, Lormée H, Jouventin P. 2003 Seasonal and inter - annual variation in the feeding ecology of a tropical oceanic seabird, the red - tailed tropicbird Phaethon rubricauda. Mar. Ecol. Prog. Ser. 255 : 289 - 301. Lewison R et al. 2012 Research priorities for seabirds: improving seabird conservation and management in the 21 st century. Endanger. Species Res. 17 : 93 - 121. Luo YW et al. 2012 Database of diazotrophs in global ocean: Abundance, biomass, and nitrogen fixation rates. Earth Syst. Sci. Data 4 : 47 - 73. McCarthy MD, Benner R, Lee C, Fogel ML. 2007 Amino acid nitrogen isotopic fractionation patterns as indicators of heterotroph y in plankton, particulate, and dissolved organic matter. Geochim. Cosmochim. Acta 71 : 4727 - 2744. McClelland JW, Montoya JP. 2002 Trophic relationships and the nitrogen isotopic composition of amino acids in plankton. Ecol . 83 : 2173 - 2180. McMahon KW, 15 N fractionation: mechanisms, implications, and applications for trophic ecology. Ecosphere 7 : e01511 . Minagawa M, Wada E. 1984 Stepwise enrichment of 15 N along food chains: further evidence and the relation between 15 N and animal age. Geochim. Cosmochim. Acta 48 : 1135 - 1140. Ohkouchi N et al. 2017 Advances in the application of amino acid nitrogen isotopic analysis in ecological and biogeochemical studies. Org. Geochem. 113 : 150 - 174. Ostrom PH, Wiley AE, James HF, Rossman S, Walker WA, Zipkin EF, Chikaraishi Y. 2017 Broad - scale trophic shift in the pelagic North Pacific revealed by an oceanic seabird. Proc. R. Soc. B. 284 : 20162436 - 20162442 . Plummer M. 2003 JAGS: a program for analysis of Bayesian graphical models using Gibbs Sampling. Proceedings of the Third International Workshop on Distributed Statistical Computing. R Project for Statistical Computing, Vienna, Austria. 76 Popp BN, Graham BS, Olson RJ, Hannides CCS, Lorr MJ, Lopez - Ibarra G, Galvan - Magana F, Fry B. 2007 Insight into the trophic ecology of yellowfin tuna, Thunnus albacares , from compound specific nitrogen isotope analysis of proteinaceous amino acids. In : Dawson TE, Siegwolf RTW (eds ) Stable isotopes as ind icators of ecological change . Elsevier, New York, USA, pp 173 - 190. Pyle P. 2008 Molt and age determination in Procellariiformes. In : Identification Guide to North American Birds, Part 2 . Slate Creek Press, Point Reyes Station, California, USA, pp 248 - 260. R Development Core Team. 2013 R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rossman S, Ostrom PH, Stolen M, Barros NB, Gandhi H, Stricker CA, Wells RS. 2015 Individual specialization in the foraging habits of female bottlenose dolphins living in a trophically diverse and habitat rich estuary. Oecologia 178 : 415 - 425. Rossman S, Ostrom PH, Gordon F, Zipkin EF. 2016 Beyond carbon and nitrogen: guidelines for estimating th ree - dimensional isotopic niche space. Ecol. Evol. 6 : 2405 - 2413 . Rucklidge GJ, Milne G, McGaw BA, Milne E, Robins SP. 1992 Turnover rates of different collagen types measured by isotope ratio mass spectrometry. Biochim. Biophys. Acta 1156 : 57 - 61. Schre iber EA, Burger J. 2001 Biology of marine birds . CRC Press, Florida, USA. Seki M, Polovina J. 2001 Biological enhancement at cyclonic eddies tracked with GOES thermal imagery in Hawaiian waters. Geophys. Res. Lett. 28 : 1583 - 158 6. Shiozaki T, Furuya K, Kodama T, Kitajima S, Takeda S, Takemura T, Kanda J. 2010 New estimation of N 2 fixation in the western and central Pacific Ocean and its marginal seas. Glob. Biogeochem. Cycles 24 : GB1015. Sigman DM, Mccorkle DC, Francois R, Fischer G. 2000 The 15 N of ni trate in the Southern Ocean: nitrogen cycling and circulation in the ocean interior. J. Geophys. Res. 105 : 19599 - 19614. Spear LB, Ainley DG, Nur N, Howell SNG. 1995 Population size and factors affecting at - sea distributions of four endangered Procellari ids in the tropical Pacific. Condor 97 : 613 - 638. Thorne LH, Hazen EL, Bograd SJ, Foley DG, Conners MG, Kappes MA, Kim HM, Costa DP, Tremblay Y, Shaffer SA. 2015 Foraging behavior links climate variability and reproduction in North Pacific albatrosses. Mov. Ecol . 3 : 27. Warham J. 1996 The behavior, population biology and physiology of the petrels. Academic Press, University of Canterbury, Christchurch, New Zealand. 77 Weimerskirch H, Ancel A, Caloin M, Zahariev A, Spagiari J, Kersten M, Chastel O. 2003 F oraging efficiency and adjustment of energy expenditure in a pelagic seabird provisioning its chick. J. Anim. Ecol. 72: 500 - 508. Wiley AE, Ostrom PH, Stricker CA, James HF, Gandhi HF. 2010 Isotopic characterization of flight feathers in two pelagic seabi rds: Sampling strategies for ecological studies. Condor 112 : 337 - 346. Wiley AE, Welch AJ, Ostrom PH, James HF, Stricker CA, Fleischer RC, Gandhi H, Adams J, Ainley DG, Duvall F, Holmes N. 2012 Foraging segregation and genetic divergence between geographic ally proximate colonies of a highly mobile seabird. Oecologia 168 : 119 - 130. Wiley AE et al. 2013 Millennial - scale isotope records from a wide - ranging predator show evidence of recent human impact to oceanic food webs. Proc. Natl. Acad. Sci. 110 : 8972 - 8977. Young LC. 2009 Foraging ecology, population genetics and risk of fisheries bycatch for the Laysan albatross ( Phoebastria immutabilis ). PhD dissertation, University of Hawaii, Hawaii, USA. Young HS, McCauley DJ, Dirzo R, Dunbar RD, Shaffer SA. 2010 Niche partitioning among and within sympatric tropical seabirds revealed by stable isotope analysis. Mar. Ecol. Prog. Ser. 416 : 285 - 294.