The role of animals in modulating nutrient cycling [hereafter, consumer-driven nutrient dynamics (CND)] has been accepted as an important influence on both community structure and ecosystem function in aquatic systems. Yet there is great variability in the influence of CND across species and ecosystems, and the causes of this variation are not well understood. Here, we review and synthesize the mechanisms behind CND in fresh waters. We reviewed 131 articles on CND published between 1973 and 1 June 2015. The rate of new publications in CND has increased from 1.4 papers per year during 1973-2002 to 7.3 per year during 2003-2015. The majority of investigations are in North America with many concentrating on fish. More recent studies have focused on animal-mediated nutrient excretion rates relative to nutrient demand and indirect impacts (e.g. decomposition). We identified several mechanisms that influence CND across levels of biological organization. Factors affecting the stoichiometric plasticity of consumers, including body size, feeding history and ontogeny, play an important role in determining the impact of individual consumers on nutrient dynamics and underlie the stoichiometry of CND across time and space. The abiotic characteristics of an ecosystem affect the net impact of consumers on ecosystem processes by influencing consumer metabolic processes (e.g. consumption and excretion/egestion rates), non-CND supply of nutrients and ecosystem nutrient demand. Furthermore, the transformation and transport of elements by populations and communities of consumers also influences the flow of energy and nutrients across ecosystem boundaries. This review highlights that shifts in community composition or biomass of consumers and eco-evolutionary underpinnings can have strong effects on the functional role of consumers in ecosystem processes, yet these are relatively unexplored aspects of CND. Future research should evaluate the value of using species traits and abiotic conditions to predict and understand the effects of consumers on ecosystem-level nutrient dynamics across temporal and spatial scales. Moreover, new work in CND should strive to integrate knowledge from disparate fields of ecology and environmental science, such as physiology and ecosystem ecology, to develop a comprehensive and mechanistic understanding of the functional role of consumers. Comparative and experimental studies that develop testable hypotheses to challenge the current assumptions of CND, including consumer stoichiometric homeostasis, are needed to assess the significance of CND among species and across freshwater ecosystems.

Nutrient cycling is a key process linking organisms in ecosystems. This is especially apparent in stream environments in which nutrients are taken up readily and cycled through the system in a downstream trajectory. Ecological stoichiometry predicts that biogeochemical cycles of different elements are interdependent because the organisms that drive these cycles require fixed ratios of nutrients. There is growing recognition that animals play an important role in biogeochemical cycling across ecosystems. In particular, dense aggregations of consumers can create biogeochemical hotspots in aquatic ecosystems via nutrient translocation. We predicted that filter-feeding freshwater mussels, which occur as speciose, high-biomass aggregates, would create biogeochemical hotspots in streams by altering nutrient limitation and algal dynamics. In a field study, we manipulated nitrogen and phosphorus using nutrient-diffusing substrates in areas with high and low mussel abundance, recorded algal growth and community composition, and determined in situ mussel excretion stoichiometry at 18 sites in three rivers (Kiamichi, Little, and Mountain Fork Rivers, south-central United States). Our results indicate that mussels greatly influence ecosystem processes by modifying the nutrients that limit primary productivity. Sites without mussels were N-limited with -26% higher relative abundances of N-fixing blue-green algae, while sites with high mussel densities were co-limited (N and P) and dominated by diatoms. These results corroborated the results of our excretion experiments; our path analysis indicated that mussel excretion has a strong influence on stream water column N:P. Due to the high N:P of mussel excretion, strict N-limitation was alleviated, and the system switched to being co-limited by both N and P. This shows that translocation of nutrients by mussel aggregations is important to nutrient dynamics and algal species composition in these rivers. Our study highlights the importance of consumers and this imperiled faunal group on nutrient cycling and community dynamics in aquatic ecosystems.

1. In streams, the creation of nutrient-transformation hotspots by aggregated organisms may have heterogeneous and strong cumulative influences on stream nutrient dynamics. Here, we examine the potential for aggregations of freshwater mussels to create such hotspots. 2. We measured nitrogen (N) and phosphorus (P) excretion rates of six mussel species and body tissue composition of eight. We combined these data with population densities of surveyed mussel beds in the Kiamichi River, OK, to estimate reach-scale and stream-scale nutrient recycling and storage. Additionally, we estimated the temporal variability in the magnitude of mussel nutrient recycling combining volumetric excretion at a reach scale with discharge and temperature data. 3. Mussel beds constituted 1.45% of the area of the Kiamichi River. Mussel nutrient remineralisation varied greatly across beds (11.1-699.5 lmol N m À2 h À1 and 0.8-53.0 lmol P m À2 h À1 ), because of varying mussel densities. The community-wide average excretion N:P (molar) of the mussel communities was 29.57, with higher excretion N:P significantly associated with higher abundances of Actinonaias ligamentina. Total nutrient storage per bed varied two orders of magnitude (6.3-631.7 kg N and 2.3-227.5 kg P) between mussel beds. Moreover, areal nutrient storage varied among the beds (11.2-133.7 mg N m À2 , 4.1-48.9 mg P m À2 ) with the majority of nutrient storage in a long-term store, shell (c. 87% of total N storage, c. 95% of total P storage). 4. Freshwater mussels can be important to nutrient dynamics through nutrient regeneration and the creation of storage hotspots. However, the importance of nutrient remineralisation varies dramatically in response to organism patchiness, flow conditions and background nutrient concentrations.