SCTB17 Working Paper BIO–1 Diet of yellowfin tuna in different areas of the western and central Pacific. Valérie ALLAIN Oceanic Fisheries Programme Secretariat of the Pacific Community Noumea, New Caledonia July 2004 2 Introduction To develop ecosystem approaches of fisheries management it is important to take into account species interactions and underlying ecosystem dynamics. Assessing the impact of fisheries and environmental effects on the ecosystem implies a good comprehension of this system. Predation induces an important mortality in the ecosystem that is often higher than fishery mortality, and determining trophic interactions between species is a major step towards a better understanding and modeling of the ecosystem dynamic. A large sampling programme has been implemented in the western and central Pacific to collect samples and determine the diet of the top predators of the pelagic ecosystem. The area covered by the study is extended and the samples are collected more or less close to the equatorial warm pool and on a large east-west transect. An hypothesis to explain the important tuna production in the warm pool where primary production is low, is that the tuna move extensively to feed at the edge of the warm pool, and particularly in the upwelling area in the east. It is then suspected that different food webs support the tuna production in the different areas of the Pacific. To examine this hypothesis, the diet of the yellowfin is analyzed to try to detect differences in the different areas of the western and central Pacific using stomach content data and stable isotope analysis. Methods Sampling programme, sampling protocol Stomach samples are collected from target fishes (tunas) and bycatch species by observers from the different national observer programmes in the area. Since the beginning of the programme in January 2001, 49 sampling trips have been done, 39 on longline boats and 10 on purse seine vessels. Sixteen sampling trips were organised in French Polynesia, 11 in New Caledonia, three in Federated States of Micronesia, three in Papua New Guinea, 14 in Solomon Islands and two in Cook Islands. For further details on the sampling programme, the sampling protocols and the stomach examination protocol check SCTB16 - BBRG-6 http://www.spc.int/OceanFish/Html/SCTB/SCTB16/bbrg6.pdf and http://www.spc.int/OceanFish/Html/TEB/EcoSystem/index.htm. Stomach examination Classical procedure is used to analyze the stomachs: -Fullness coefficient is determined according to a scale from 0 (empty) to 4 (full) (see caption of Figure 4 p.7 for details). If baits are present, they are removed to determine the fullness coefficient. However fullness coefficient is a subjective parameter which determination will vary with the examiner, so another way to evaluate and quantify the stomach fullness is to calculate the stomach content weight – predator weight ratio (= stomach content weight * 100 / predator weight, in %). The weight of the predator is estimated from the measured length and using a length-weight converting factor (in this case for the yellowfin: a=0.00001908 and b=2.9776199, based on the measure of 329 fish from Regional Observer Data). -Preys are sorted by species or group, identified at the lowest taxonomic level, a digestion state is attributed (from 1 to 4, see details in caption of Figure 6 p.8), development state is determined when possible (larvae, juvenile, adult), they are counted, weighted and measured. The number of baits, the presence of parasites, the number of cephalopod beaks, gladius and otoliths are recorded. It is important to note that some species are underestimated due to a high digestion rate and/or because of a lack of specific structure that could help for identification. The typical jaws of Tetraodontiformes, Alepisaurus, Gempylidae/Trichiuridae, the dorsal spine of Monacanthidae, 3 Balistidae, the photophores of Myctophidae, Ommastrephidae (squids) facilitate identification of these species, even when they are in an advanced state of digestion. Data on habitat and type in Table 5 (p.15) are extracted from Fishbase (http://www.fishbase.org/home.htm) and FAO species identification guide. For data analysis, frequency of occurrence (%F), percentage of number (%N) and percentage of weight (%W) were calculated by taxon, cumulating all the data from the same area. Frequency of occurrence of an item is the number of stomachs where this item is present divided by the number of non-empty stomachs. Percentages of number and weight are the respective number and weight of the taxon studied divided by total number or weight of this taxon of all the samples, by area. Stable isotope analysis Chemical elements can exist in both stable and unstable (radioactive) forms. Most elements of biological interest (including C, H, O, N, and S) have two or more stable isotopes (atoms of the same atomic number but different atomic weights due to a different number of neutrons), with the lightest of these present in much greater abundance than the others. Among stable isotopes the most useful as biological tracers are the heavy isotopes of carbon and nitrogen. These two elements are found in the earth, the atmosphere, and all living things. Each has a heavy isotope (13C and 15N) with a natural abundance of ~1% or less and a light isotope (l2C and 14N) that makes up all of the remainder, in the case of nitrogen, or virtually all in the case of carbon (carbon also has a radioactive isotope, 14C.). Studies examining stable isotopes at or near natural abundance levels are usually reported as delta, a value given in parts per thousand or per mil (‰). Delta values are not absolute isotope abundances but differences between sample readings and one or another of the widely used natural abundance standards which are considered delta = zero (e.g. air for N, Pee Dee Belemnite for C). For more information on isotopes: http://www.uga.edu/~sisbl/stable.html. The nitrogen isotopic composition of marine fauna is particularly sensitive to trophic level (Fry, 1988). δ15N isotope values are used to estimate the trophic level occupied by the tunas, other predators, their forage prey species, and plankton. It implies to know the δ15N values of the component at the base of the food chain (phytoplankton) that has a trophic level of 1 by definition. The trophic level of the other components is then calculated assuming an increase in δ15N of 3.4‰ per trophic level: δ15N of herbivorous zooplankton = δ15N of phytoplankton + 3.4‰. The same rule applies to δ13C with an increase of 1‰ per trophic level. If the isotope values of the base of the food web are not known the trophic level cannot be calculated but the isotope values give an idea of the relative positions of the different components in the food web. The carbon isotopic composition of phytoplankton and consumers often reflects the algal sources of production, with high δ13C values associated with rapid growing diatoms characteristic of upwellings and blooms (Popp et al., 1998; 1999). δ13C values are used to identify different sources of primary production, and to distinguish rapid production associated with upwelling. The combination of δ15N and δ13C will serve to map different regions of primary and secondary production in the Pacific Ocean. Although there are few preliminary isotope data for fisheries in the equatorial Pacific, we anticipate clear regional signals. We expect high δ15N associated with the eastern Pacific due to denitrification, low δ15N in the western Pacific and gyres associated with nitrogen fixation (Deutsch et al., 2001), and we expect high δ13C associated with large diatoms in upwelling regions, lower δ13C in low productivity areas. This kind of isotope cartography has been successful in the North Pacific and in the Atlantic in studies of POM (Particulate Organic Matter) and zooplankton (Wu et al., 1997; Perry et al., 1999; Waser et al., 2000), and has been recently suggested as an appropriate way to use isotopes in the open ocean (Rost et al., 2002). The isotope analysis is conducted on muscle and liver of individual fish. Samples are analyzed with a mass spectrometer in the stable isotope biogeochemical laboratory at the University of Hawaii. Precision and accuracy of isotopic determinations of laboratory standards has been better than ±0.1‰ over the last three years of operation of the equipment. 4 Characteristics of the samples For this study, 195 yellowfin tuna have been sampled so far, they come from different areas in the Pacific (Figure 1). Isotope analysis was conducted on 37 of them while the 195 stomachs were examined. MICRO PAPUA NC POLY Figure 1. Positions of the yellowfin tuna sampled for the study in the western and central Pacific. MICRO: Micronesia includes EEZ from Marshall Islands, Federated States of Micronesia, Nauru, Kiribati (Gilbert Islands) and nearby international waters; NC: New Caledonia EEZ; PAPUA: includes EEZ from Papua New Guinea and Solomon Islands; POLY: Polynesia includes French Polynesia and Cook Islands. Most of the samples come from New Caledonia (82) and Polynesia (52), 39 were sampled in Micronesia and 22 in Papua New-Guinea – Solomon Islands (Table 1). The fish were caught on longline except for the samples from Micronesia and for some samples from PAPUA that were caught by purse seine (Table 1). Length distribution of the yellowfin varies according to the area and it is mainly linked to the gear used (Figure 2). Samples caught by purse seine in MICRO are between 30 and 60 cm with a mean at 62 cm (Table 1); fish caught on longline are larger: most of them are between 110 and 130 cm (mean = 125 cm) in NC, between 110 and 120 cm (mean = 102 cm) in PAPUA and at 90 cm (mean = 97 cm) in POLY. The percentage of females is close to 50% in the samples of NC and POLY (Table 1).