Trophic transfer of metals along freshwater food webs: Evidence of cadmium biomagnification in nature

We conducted a study with cadmium (Cd) and copper (Cu) in the delta of San Francisco Bay, using nitrogen and carbon stable isotopes to identify trophic position and food web structure. Cadmium is progressively enriched among trophic levels in discrete epiphyte-based food webs composed of macrophyte-dwelling invertebrates (the first link being epiphytic algae) and fishes (the first link being gobies). Cadmium concentrations were biomagnified 15 times within the scope of two trophic links in both food webs. Trophic enrichment in invertebrates was twice that of fishes. No tendency toward trophic-level enrichment was observed for Cu, regardless of whether organisms were sorted by food web or treated on a taxonomic basis within discrete food webs. The greatest toxic effects of Cd are likely to occur with increasing trophic positions, where animals are ingesting Cd-rich prey (or food). In Franks Tract this occurs within discrete food chains composed of macrophyte-dwelling invertebrates or fishes inhabiting submerged aquatic vegetation. Unraveling ecosystem complexity is necessary before species most exposed and at risk can be identified. Ecosystems are threatened by a steadily increasing number of pollutants that cause adverse effects. Yet the link between metal exposure and effects in aquatic organisms remains poorly known, likely because biological responses differ among species, metals, physicochemical conditions, and exposure routes. For example, marine crustaceans appear more sensitive to metals accumulated from food rather than from their aqueous environment (Hook and Fisher 2001a,b). This implies that ingested metals are likely to cause toxicity not only at the base of food webs but also in top consumers if the assimilated metals build up through the food web (e.g., Cabana and Rasmussen 1994). However, controversy surrounds the question of metal biomagnification, defined here as the progressive accumulation of chemicals with increasing trophic levels (Leblanc 1995). For instance, Gray (2002) concluded that metal biomagnification is an exception rather than a rule among metals and metalloids. Reinfelder et al. (1998) suggested that trophic transfer potential (TTP) could be described from the biodynamic parameters weight-specific ingestion rate, assimilation efficiency (AE) and rate constant of loss. Among metals, organic mercury is the most likely to biomagnify because organisms efficiently assimilate methylmercury and very slowly elim1 Corresponding author ([email protected]). Acknowledgments The technical assistance of F. Parchaso, G. E. Moon, M. K. Shouse, and K. Higgins with sample collections is acknowledged. We thank L. Grimaldo and E. Santos (DWR) for their assistance with fish sampling. Guidance with ICP-MS analysis by B. R. Topping is recognized. We thank C. B. Lopez (USGS) and L. Hare (INRS-ETE) for their assistance in identifying algae and chironomids, respectively. Critical comments from B. R. Topping and J. K. Thompson are appreciated. Funding was provided by the U.S. Geological Survey Toxic Substances Research Program; with some support to S. N. Luoma from a WJ Fulbright Distinguished Scholar Award. M.-N. Croteau was supported by postdoctoral fellowships from Fonds de Recherche sur la Nature et les Technologies (Québec) and NSERC Canada. inate it in proportion to biomass (Mason et al. 1996; Reinfelder et al. 1998). There are theoretical reasons to suspect that selenium (Se), Cd, and perhaps even silver could biomagnify under some circumstances, but only Se has been carefully evaluated in the field (Stewart et al. 2004). Unambiguous evaluations of metal biomagnification in nature are rare because metal concentrations in whole-body prey are often compared with those in predator’s specific tissues without knowledge of bioaccumulation processes, feeding relationships, and trophic status (Reinfelder et al. 1998; Gray 2002). Here we study two metals that can be both toxic (i.e., they commonly appear on government priority-substances lists, e.g., the U.S. Environmental Protection Agency), but that contrast in their biological functions and perhaps their potential for biomagnification. Copper (Cu) can act as essential micronutrient (e.g., Sunda and Huntsman 1995). However, little is known about its trophic transfer potential mainly because the lack of a suitable radioisotope prevented quantification of AE and loss-rate constants (until recently; Croteau et al. 2004). In contrast, Cd has no known biological use in animals (although it may substitute for zinc in certain enzymes in phytoplankton; Lane and Morel 2000). Cadmium might biomagnify if consumers efficiently assimilate and slowly lose it (Reinfelder et al. 1998; Wang 2002); however, this has not been demonstrated directly in the laboratory or in nature. First, we address the question as to whether Cd and Cu concentrations differ, and to what degree, among species collected from the same habitat and at the same time. We then ask whether feeding relationships could be used to explain at least some of those differences; if different types of food webs transfer metals differently; and lastly, whether Cd and Cu differ in food web transfer. The study of trophic transfer is also limited by the difficulty of discriminating food webs and accurately ascribing trophic position to organisms. Stable isotope ratios of carbon (13C : 12C; d13C) and nitrogen (15N : 14N; d15N) are now recognized as powerful tools to provide time-integrated evalu-