Interpreting stable isotopes in food webs: Recognizing the role of time averaging at different trophic levels

Lake Tanganyika, East Africa, has a simple pelagic food chain, and trophic relationships have been established previously from gut-content analysis. Instead of expected isotopic enrichment from phytoplankton to upper level consumers, there was a depletion of 15N in August 1999. The isotope signatures of the lower trophic levels were an indicator of a recent upwelling event, identified by wind speed and nitrate concentration data, that occurred over a 4-d period several days prior to sampling. The isotope structure of the food web suggests that upwelled nitrate is a nutrient source rapidly consumed by phytoplankton, but the distinctive signature of this nitrate is diluted by time averaging in the upper trophic levels. This time averaging is a consequence of the fact that the isotopic signature of an organism is related to variable nitrogen sources used throughout the life of the organism. This study illustrates the importance of recognizing differences in time averaging among trophic levels. Stable isotopes are becoming a standard analytical tool in food web ecology. Differences in carbon and nitrogen isotope ratios between consumers and their diet provide information on energy flows, nutrient sources, and trophic relationships. Typically, carbon provides information on the primary energy source (e.g. benthic vs. pelagic photosynthesis), while nitrogen allows discrimination among trophic levels. Relative enrichment with increasing trophic level often allows a better interpretation of dietary relationships than gut-content analysis alone because isotope ratios record material that is actually assimilated (Michener and Schell 1994). Recent field studies have shown an average enrichment of 0.05‰ 6 0.63 d13C and 3.49‰ 6 0.23 d15N (Vander Zanden and Rasmussen in press) and are similar to the commonly used trophic fractionation values of 1‰ d13C and 3.4‰ d15N (Michener and Schell 1994). The ease of stable isotope analyses makes them an appealing tool in ecology, but care must be taken in interpretation of the results (Gannes et al. 1997). An isotopic ratio of an organism is usually understood to represent its diet, but it should be remembered that this isotopic ratio is also time specific, representing an average ratio related to tissue turnover rate and the life of the organism. The pelagic food web in Lake Tanganyika, East Africa, provided an excellent example of how stable isotopic analyses may not be straightforward. Lake Tanganyika is a deep (mean 570 m; max 1470 m), large (mean width 50 km; length 650 km) lake located a few degrees south of the equator. The lake is permanently stratified and is anoxic below approximately 150 m. Lake Tanganyika has a simple pelagic food web, and trophic relationships have been established previously from gut-content analysis (Fig. 1)(Coulter 1991). The zooplankton are dominated by the copepod Tropodiaptomus simplex, a major dietary component of the two clupeid fish species, Stolothrissa tanganicae and Limnothrissa miodon. The upper trophic level is represented by four Lates species. Of these, Lates angustifrons and Lates mariae rely on nearshore fishes as a food source throughout their lives. Lates microlepis spend their larval and juvenile stages near shore and recruit to the pelagic as adults, whereas Lates stappersi has a fully pelagic life cycle. Thus, the most abundant species of the pelagic food web form a linear food ’chain’ from phytoplankton to the copepod T. simplex to the zooplanktivorous Stolothrissa to the predatory L. stappersi. This linear food chain should be apparent as a distinct sequential enrichment in the carbon and nitrogen isotopes with increasing trophic level (Michener and Schell 1994). Methods—Our study took place near Kigoma, Tanzania, during the dry season. Food web samples were collected approximately 15 km offshore (48509S, 298299E) during the night of 1 August 1999. Phytoplankton were collected using vertical tows with a 50-mm mesh net (n 5 6). Samples were filtered through 100-mm mesh to remove zooplankton, then collected on 0.45-mm glass-fiber filters and rinsed with 0.01 N HCl and distilled water. Zooplankton were caught using vertical tows with a 100-mm mesh net (n 5 4). They were placed in filtered lake water for 2 h to clear gut contents then collected on 0.45-mm glass-fiber filters and rinsed with 0.01 N HCl and distilled water. Fish specimens of Stolothrissa (n 5 6), Limnothrissa (n 5 5), Lates stappersi juveniles (n 5 4), and Lates stappersi (n 5 4) were obtained from local fishermen who were fishing adjacent to the site when the plankton samples were collected. For all species except Stolothrissa, we collected a section of white muscle tissue from behind the dorsal fin. As Stolothrissa were too small to obtain a large muscle sample, we used the entire body after removing the head, tail, and viscera. Fish samples were washed with distilled water, dried, and homogenized before analysis. All samples were dried at 508C and stored wrapped in aluminum foil. Samples were analyzed at the University of Waterloo Environmental Isotope Lab on an isochrom continuous flow stable isotope mass spectrometer (Micromass) coupled to a Carla Erba Elemental Analyzer (CHNS-O EA1108). The isotope ratios are expressed in delta notation with respect to deviations from standard reference material (Pee Dee belemnite carbon and atmospheric nitrogen). Standard error is 0.2‰ for carbon and 0.3‰ for nitrogen. Statistical analyses were done using JMP IN (SAS Institute). Isotope structure of the food web—Overall, the isotope structure of the pelagic food web was not that of a linear food chain (Fig. 2). There was a general trend of carbon enrichment but a depletion of nitrogen with trophic level increase. The enrichment in d13C across the lower food web from phytoplankton to Stolothrissa was larger than would