Fish decomposition in boreal lakes and biogeochemical implications

A field study in a boreal lake using a remotely operated vehicle equipped with a camera established that falling fish carcasses did not tend to be buried in sediments after deposition. Decomposition rates of fish carcasses in three boreal lakes were experimentally assessed at different depths. In shallow waters (between 0 and 4 m), decomposition was fast (half lives, t, ranging from 40 to 230 h) and controlled by vertebrates. In deep waters (below the thermocline), decomposition was slow (t between 770 and 1,733 h) and was controlled by bacterial processes. Water temperature was a promising predictor of decomposition half-lives in freshwater. Using a novel underwater infrared camera system, we identified the daily and seasonal patterns of scavenging activity by littoral fish. Only three species displayed scavenging behavior, with creek chubs being the most active. Fast fish-mediated littoral recycling of fish carcass may explain the lack of direct observations of carcasses in lakes. Estimates of phosphorus fluxes in one of the studied lakes indicate that falling carcasses can represent a significant water-tosediment flux of nutrient. Ecologists have long recognized the role played by vertebrates in the global and regional transport and release of nutrients (Schindler et al. 2005). More recently, their role in the transport of contaminants has emerged as a fruitful field of research in aquatic ecology (Blais 2005; Blais et al. 2007). This long-range biodelivery of pollutants and nutrients is expected to be especially significant for migratory animals, such as anadromous fish. In particular, the postmortem influence of Pacific salmon on the biogeochemical dynamics of receiving freshwater ecosystems has drawn much attention (e.g., Krümmel et al. 2003; Schindler et al. 2005; Fenoglio 2005). Recently, the role of arctic seabirds on contaminant cycling has also been documented (Blais et al. 2005). Vertebrates can also influence nutrient cycling at the ecosystem scale. For instance, bears are known to transport salmon-derived nutrients from stream to land, with significant effects on the productivity of adjacent forested ecosystems (Helfield and Naiman 2006). Also, mobile avian scavengers breeding on islands in salmon nursery lakes can create hot spots of biological productivity by concentrating nutrients near their colonies (Payne and Moore 2006). In lakes not affected by such massive mortality events, the role of vertebrate scavengers as vectors of nutrients is less understood. Although fish carcasses are rarely seen at the bottom of lakes or on their shores, a number of studies on freshwater systems have reported that estimated natural mortality rates not due to predation are high, ranging between 10% and 67% per year (e.g., Craig 1984; Lorenzen 1996; Schneider 1998; Allen et al. 1998; Mills et al. 2002), with a mean of about 20–25% per year (Reznick et al. 2002). These falling carcasses constitute a water-to-sediment flux of biomass that could alter the cycling of nutrients (Parmenter and Lamarra 1991; Stevenson and Childers 2004) and bioaccumulative contaminants (Sarica et al. 2004). The fate of fish carcasses after deposition at the sediment surface of lakes is poorly documented. Several articles on the fate of large fish and whale carcasses in the deep ocean floor have been published (e.g., Premke et al. 2003; Soltwedel et al. 2003, Ruxton and Houston 2004), but they are not directly applicable to lakes. Schneider (1998) conducted the only published experimental study on the fate of fish carcasses in a lake not subjected to massive fish mortality. Although the nonpredatory mortality rates of this population ranged from 25% to 40% per year, few carcasses could be found. Schneider (1998) concluded that, in this lake, this paradox could be partly explained by the unseen decomposition of fish carcasses in deep water and the ability of resident scavengers to keep up with the supply of dead fish. However, this study did not investigate the factors influencing the decomposition rates of carcasses. The present study aims at testing some key hypotheses regarding fish decomposition in lakes in the absence of massive mortality events, to reach a better understanding of its effect on the cycling of matter. Our first hypothesis is that carcasses are not buried upon deposition and lie on sediments where they can be degraded by animals and bacteria. Such a burial would affect further degradation and likely lengthen the turnover time of nutrients in lakes. Second, we investigate the relative importance of two modes of decomposition in lakes: bacterial decomposition and vertebrate scavenging. Bacterial decomposition will presumably lead to the complete mineralization of nutrients 1 Corresponding author ([email protected]).