Food-Web Implications for Pelagic Top Predators: from Guts and Isotopes to Models

Home , Pleuroncodes planipes, Sthenoteuthis

Food-web Implications for Pelagic Top Predators: from Guts and Isotopes to Models

Robert J. Olson Inter-American Tropical Tuna Commission La Jolla, California

Photo compliments of Dr. Frederic Menard, IRD, France Food webs and Ecosystem-based Fisheries Science

• “Ecosystem” “Ecology”: multispecies approaches to management, reduction of bycatch, including environmental factors in stock assessment models. • Ecosystem: a geographically specified system of organisms, including humans, the environment, and the processes that control its dynamics (NOAA 2005). • “The time has come for community ecology to replace population ecology as the fundamental ecological science underlying fisheries” (Mangel and Levin 2005). • Communities are assemblages of species. Interactions makes the community more than the sum of its parts. • Communities interact via the food web.

NOAA. 2005. New priorities for the 21st century: NOAA's strategic plan. NOAA, Washington, D.C. Mangel, M., and P.S. Levin. 2005. Regime, phase and paradigm shifts: making community ecology the basic science for fisheries. Phil. Trans. R. Soc. B, 360 (1453): 95-105. Why study food webs?

• Trophic structure represented in food webs is thought to be the central organizing concept in ecology (Martinez 1995). • Knowledge of pelagic food webs is still rudimentary, in many aspects. Better food-web models are needed (preferably, spatially-explicit). • Review an assortment of information about food- web research in eastern Pacific, and (less-so) on modeling efforts. Eight ecosystem characteristics

NMFS Ecosystem Principles Advisory Panel: 1. The ability to predict ecosystem behavior is limited 2. Ecosystems have thresholds and limits which, when exceeded, can effect major ecosystem restructuring 3. Once thresholds and limits have been exceeded, changes can be irreversible 4. Diversity is important to ecosystem functioning 5. Multiple scales interact within and among ecosystems 6. Components of ecosystems are linked 7. Ecosystem boundaries are open 8. Ecosystems change over time Components of ecosystems are linked

How do we determine what the important components and linkages are? Critical food-web connections.

•Keystone species •Dietary specialists •Models can help

The tools for food-web research: •Diet studies (stomach-contents analysis) •Stable isotope analysis •Compound specific stable isotope analysis (amino-acids) •Fatty acid analysis Stomach-contents analysis (species identification) (and monitoring)

V. Allain, SPC F. Galvan, CICIMAR IATTC, Manta, Ecuador Diet datafor eastern Pacific predators(’92-

Percent weight 100% 80% 60% 40% 20% 0% Julio •Julio •Felipe Galván Colleagues: Martínez Tunas Dolphins Sharks - , M, M, La Paz,CICIMAR, BCS,Mexico Billfishes ,Cumaná Venezuela DoradoRunner , Wahoo, R. ’94) Diet data formulated food web (ETP) fico ó tr Nivel – Trophic level Trophic

Olson, R.J., and G.M. Watters. 2003. A model of the pelagic ecosystem in the eastern tropical Pacific Ocean. Inter-American Tropical Tuna Commission, Bulletin 22 (3): 133-218. Yellowfin tuna stomach-contents (1990s, 2000s)

Set Locations

1990s

2000s Feeding Ecology of Surface Migrating Myctophid Fishes in the eastern Tropical Pacific

Joel Van Noord, Univ. of San Diego Jessica Redfern et al., NMFS SWFSC Trophic position: stable isotopes

15N 15 14 15 15 δ = [( N/ N) / Rstd – 1] x 1000 δ Npredator = 3.0 + δ Nprey (‰)

Isotopic fractionation – the light 14N isotope is excreted more than the heavy 15N isotope, leaving the animal enriched by 3‰ in δ15N relative to its food source. Trophic position: stable isotopes, stomach contents

18 16 14 12 10 N (‰)

15 8 δ

6 Yellowfin tuna (5-deg areas) 4 Yellowfin tuna (outside 5-deg areas) Mesozooplankton (5-deg areas) 2 Mesozoopl. (outside 5-deg areas) CSIA samples 0 -15 -10 -5 0 5 10 15 20 25 30 Latitude (degrees) Mean TP = 4.5

PFRP, B. Popp, B. Graham, C. Hannides, F. Galván, G. López, B. Fry Yellowfin trophic position (TP)

YFT δ15N = 13-16‰

Copepods δ15N = 6-12‰

ΔYFT-COP = 4.0 – 7.6 ‰ Gladis Lopez-I., CICIMAR, Mexico

TP ≈ 4.3 - 5.3, spanning ~ 1 trophic level

B. Popp, UH B. Graham, UH F. Galvan-Magana, CICIMAR C. Lennert-Cody, IATTC PFRP δ15N of Amino Acids specific (eds.), Stable Stable (eds.), Magaña, and Magaña, B. - Yellowfin tuna – eastern tropical Pacific

- compound , from (“Trophic” AA) Ibarra, F. Galván F. Ibarra, - albacares Thunnus

TL 4.5 Academic Press, Terrestrial Ecology Series, San Diego: Diego: San Series, Ecology Terrestrial Press, Academic Bulk white muscle amino acids. In Dawson, T.E., and R.T.W. Siegwolf R.T.W. and amino acids. In Dawson, T.E.,

proteinaceous (“Source” AA) 190. nitrogen isotope nitrogen analysis of - Elsevier Change. of Ecological as Isotopes Indicators 173 - Fry. 2007. Insight into the trophic ecology of yellowfin tuna, tuna, yellowfin of ecology trophic the into Insight 2007. Fry. Popp, B.N., B.S. Graham, R.J. Olson, C.C.S. Hannides, M.J. Lott, G.A. López Lott, M.J. Hannides, R.J. Olson, C.C.S. Graham, B.S. B.N., Popp, E-W shift in trophic position in ETP

TP shift≈1 Lipids as Dietary Tracers Traditional techniques problematic, e.g. gut content analysis  Prey species have unique lipid / fatty acid compositions  Many fatty acids readily transferred from prey to predator with minimal modification  Constituent fatty acids therefore represent, to some extent, a temporal integration of diet  Can be quantitative and allows temporal integration (cf gut content analysis)  Signature fatty acids: combinations of fatty acids preserved as they pass up the food chain  Complements other approaches * Jock Young, CSIRO Eight ecosystem characteristics

• Nature is seldom linear, and often unpredictable (Francis et al. 2007) . • Ecosystem resilience depends on “stability domain” of existing food web: how broad is it, how resistant is it to change, how close is it to reorganizing? (Francis et al. 2007) Models are required. • How should components of the food web be represented in models? • Can models highlight key areas for field/lab studies?

Francis, R.C., M.A. Hixon, M.E. Clarke, S.A. Murawski, and S. Ralston. 2007. Ten commandments for ecosystem-based fisheries scientists. Fisheries, 32 (5): 217-233. Taxonomy in models fico ó tr Nivel – Trophic level Trophic

(Olson, R.J., and G.M. Watters. 2003. A model of the pelagic ecosystem in the eastern tropical Pacific Ocean. Inter-American Tropical Tuna Commission, Bulletin 22 (3): 133-218.) Functional groups in models

Epipelagic Epi

Epi-Meso

Meso Epi- Bathy

Meso

Bathy Meso-Bathy

Bathy Qualitative analysis of Pacific Ocean predators

20 N

South-Western Pacific Ocean Central-Eastern Pacific Ocean 5.0 77 5.0 76 169 106 75 124 74

43 44 115 121 165 65 170 122 110 168 134 77 152 4.5 33 109 108 159 31 45 157 32 160 154 126 125 4.5 74 75 76 124 106 169 121 115 44 211 85 198 127 173 215 66 70 65 122 165 168 149 172 83 155 163 156 157 114 108 112 126 167 120 161 23 219 37 36 24 12 151 14 2 13 68 4.0 26 4.0 210 212 174 215 116 177 119 117 127 209 38 5 4 147 123 3 113 28 20 18 197 193 37 192 219 196 191 195 25 69 153 27 15 35 34 52 3.5 57 60 51 59 58 64 56 194 178 Food webs 3.5 238 204 153 27 118 63 180 179 64 98 95 89 91 88 79 53 81 80 82 47 78 9 6 103 203 199 201 202 200 206 207 205 3.0 3.0 95 142 49 100 9396 140 92 72 105 102 94 23690 39 71 9 composed of 200+ 229 143 226 223 239 231 235 239 145 234 2.5 2.5 41 taxa level Trophic 245 242 222 247 237 244 246 217 216 2.0 2.0 182 241 232 185 228

1.5 1.5 183

1.0 1.0 243

Aggregated food webs composed of 24 nodes with similar predator prey relationships Can models highlight research needs? Sensitivity analysis of ETP Ecopath model

Large marlinsLarge Marlins (5.5) Small sharksSmall Sharks (5.4) Small marlinsSmall Marlins (5.4) Toothed whalesToothed Whales (5.4) LargeLarge bigeye Bigeye Tuna (5.3) Spotted dolphinsSpotted Dolphins (5.2) Large sharksLarge Sharks (5.2) Large wahooLarge Wahoo (5.1) Large swordfishLarge Swordfish (5.1) Large sailfishLarge Sailfish (5.1) PursuitPursuit birds Birds (4.9) Small mahimahiSmall Mahimahi (4.8) Small sailfishSmall Sailfish (4.8) Large mahimahiLarge Mahimahi (4.8) Large Largeyellowfin Yellowfin Tuna (4.8) MesopelagicMesopelagic dolphins Dolphins (4.8) Small wahooSmall Wahoo (4.7) SmallSmall bigeye Bigeye Tuna (4.7) Small Smallyellowfin Yellowfin Tuna (4.7) SkipjackSkipjack Tuna (4.7) CephalopodsCephalopods (4.6) Small swordfishSmall Swordfish (4.6) Misc. piscivoresMisc. Piscivores (4.5) BluefinBluefin tuna Tuna (4.5) Auxis spp.Auxis spp. (4.1) Baleen whalesBaleen Whales (4.1) RaysRays (3.9) GrazingGrazing birds Birds (3.9) Sea turtlesSea Turtles (3.8) CrabsCrabs (3.6) Cephalopods Misc. mesopelagicMisc. Mesopelagic fishes Fishes (3.6) FlyingfishesFlyingfishes (3.6) Auxis spp. Misc. epipelagicMisc. Epipelagic fishes Fishes (3.3) Secondary consumersSecondary Consumers (3.0) Primary consumersPrimary Consumers (2.0) ProducersProducers (1.0)

0 20 40 60 80 100 120 140 160 Index of Sensitivity Eight ecosystem characteristics

• Jumbo (Humboldt) squid range expansion • Are tunas effective biological samplers of the middle trophic levels*? Indicator species in stomach contents? – Squid consumption by tunas has increased over time – Decadal changes in yellowfin tuna diet composition

* Generalist predators (opportunistic), high energy requirements, food limited, range widely, prey size-predator size ranges widely Pelagic ommastrephid squids (e.g. Dosidicus gigas): Ecosystem indicators?

Olson, R.J., M.H. Román-Verdesoto, and G.L. Macías-Pita. 2006. Bycatch of jumbo squid Dosidicus gigas in the tuna purse-seine fishery of the eastern Pacific Ocean and predatory behaviour during capture. Fish. Res. 79(1-2): 48-55. Percent frequency of cephalopods in the stomach contents of yellowfin tuna in the eastern Pacific Ocean

100

Unidentified Octopus 80 Squid All cephalopods

20 Percent frequency of occurrence frequency Percent PFRP, F. Galvan, N. Bocanegra, V. 0 Alatorre, J. Martinez, 1955-1960 1969-1972 1992-1994 2003-2005 F. Alverson

Hunsicker, Essington, Olson, Duffy. Manuscript in prep. “Evidence of increased cephalopod production in a large marine ecosystem.” Decadal variation in yellowfin tuna diet composition

Classification tree analysis • Nonparametric: relationships between variables that may include: nonlinearity, high order interactions, lack of balance, missing values • Combinations of explanatory variables used to explain variation of a response variable (prey groups % weight), by repeatedly splitting the data into groups that are as homogenous as possible • Each possible value for each explanatory variable is considered as a potential candidate split • The candidate split which provides the largest decrease in impurity, or minimizes the misclassification rate, is chosen to split the data into two subgroups • Procedure is repeated with each subgroup until no significant decrease in impurity is possible, resulting in a terminal node (leaf). • 10-fold cross-validation used to prune trees (1-SE Rule) • The proportions in each category are represented in each leaf Classification tree model constructed using the Diet library of R written by Petra Kuhnert, CSIRO Classification tree analysis

Response variable Explanatory variables (18 dominant prey groups) Cephalopods • Argonauta spp. • Dosidicus gigas • Year • Sthenoteuthis oualaniensis Crustaceans • Quarter of year • Pleuroncodes planipes • Purse-seine set time of day • Portunidae family • Other Crustaceans • Latitude Fishes Longitude • Cetengraulis mysticetus • • Engraulis mordax • SST • Phosichthyidae family • Myctophidae family • Yellowfin size • Exocoetus spp. • Yellowfin sex • Other Exocoetids • Oxyporhamphus micropterus • Yellowfin stomach fullness • Carangidae family • Purse-seine set type • Auxis spp. • Scomber japonicus • Cubiceps spp. • Lactoria diaphana Yellowfin tuna stomach sample locations

Set Locations

1990s

2000s Proportion Relative Importance Classification Tree Analysis: Variable

0.0 0.2 0.4 0.6 0.8 1.0 Full = yellowfin stomach fullness = SA Lat = l set type, Qtr = quarter, Time = set time of day, FL = yellowfin fork length, atitude, Lon = longitude, SST = Sea Surface SSTSealongitude, atitude,Temperature,= Lon = year, YR= Importance Rankings Predictor variable b-DG: Dosidicus gigas squids R2 = 0.45 c-SO: Sthenoteuthis oualaniensis e-PP: Pleuroncodes planipes crustacean g-CM: Cetengraulis mysticetus h-EM: Engraulis mordax N of 17.3°N i-Phos: Phosichthyidae family m-OM: Oxyporhamphus micripterus fishes n-Car: Carangidae family o-Aux: Auxis spp. p-SJ: Scomber japonicus q-Cub: Cubiceps spp. r-LD: Lactoria diaphana Yellowfin tuna stomach sample locations

North of Latitude 17.3°N

Set Locations

1990s

2000s Yellowfin tuna diet composition at first split

35% north of latitude 17.3N south of latitude 17.3N 30%

25% 2SE) ± 20%

15%

10% Mean %Weight ( %Weight Mean

0% b-DG: Dosidicus gigas squids R2 = 0.45 c-SO: Sthenoteuthis oualaniensis e-PP: Pleuroncodes planipes crustacean g-CM: Cetengraulis mysticetus h-EM: Engraulis mordax i-Phos: Phosichthyidae family m-OM: Oxyporhamphus micripterus fishes n-Car: Carangidae family o-Aux: Auxis spp. South of 4.8° S p-SJ: Scomber japonicus q-Cub: Cubiceps spp. r-LD: Lactoria diaphana Yellowfin tuna stomach sample locations

North of Latitude 17.3°N

Set Locations

South of Latitude 4.8°S 1990s

2000s Yellowfin tuna diet composition

70%

60% Jumbo squid North of Latitude 4.8°S 50% South of Latitude 4.8°S 2SE)

± 40%

30% Anchoveta mean %W ( 20%

10%

0% Yellowfin tuna stomach sample locations

South of Latitude 17.3°N

North of Latitude 4.8°S Set Locations

1990s

2000s b-DG: Dosidicus gigas squids R2 = 0.45 c-SO: Sthenoteuthis oualaniensis e-PP: Pleuroncodes planipes crustacean g-CM: Cetengraulis mysticetus h-EM: Engraulis mordax i-Phos: Phosichthyidae family m-OM: Oxyporhamphus micripterus fishes n-Car: Carangidae family o-Aux: Auxis spp. p-SJ: Scomber japonicus q-Cub: Cubiceps spp. r-LD: Lactoria diaphana Yellowfin tuna diet composition at year split

Yellowfin tuna diet composition at time split (south of latitude 17.3 N, north of latitude 6.1 S, SST ≥23.42 45%

40% 1992-1994 diet composition 2003-2005 diet composition 35% Auxis spp.

2SE) 30%

± Mesopel. fishes 25% Epipel. fishes

20% Jumbo squid

15% Mean %Weight ( 10%

0% Summary

• Research on pelagic food webs is progressing; should be encouraged. • Chemical tracer methods are providing insight (SIA, AA-CSIA, fatty acids). • Stomach contents analyses are still necessary (monitoring indicator prey species). • Better ecosystem (food web) models are needed (spatially-explicit). Identify critical food-web connections. • Food-web models should depict taxonomic ecosystem components (indicator species). • Epipelagic ecosystems appear to change over time. • Recommendation: low-level, well-designed, continuous stomach sampling of tunas, biological samplers, to monitor changes. Acknowledgements

• Pelagic Fisheries Research Program and John Sibert, Univ. Hawaii • NOAA Fisheries STAR Project, Lisa Ballance, G. Watters, and many others • IATTC observers and staff in Ecuador and Mexico • Brian Popp, B. Graham, N. Wallsgrove, E. Gier, J. Tanimoto, T. Rust, A. Carter, Isotope Biogeochemistry Laboratory, Univ. Hawaii • Felipe Galván-Magaña, G. López-Ibarra, N. Bocanegra-Castillo, V. Alatorre-Ramírez, CICIMAR, La Paz Mexico • Tim Essington, M. Hunsicker, Univ. Washington • C. Lennert-Cody, M. Maunder, L. Duffy, M. Román-Verdesoto, C. Patnode, G.L. Macías-Pita, IATTC • B. Fry, Louisiana State Univ., Baton Rouge • V. Allain, Secretariat of the Pacific Community, New Caledonia • Jock Young, Jeff Dambacher, CSIRO • Jim Kitchell, Univ. Wisconsin