River Food Web Ecology

Studying whole-ecosystem response to large-scale environmental impacts and management programs derived by focusing on single species, particularly in Costa Rican streams and in salmon-bearing rivers of western North America.

My research in rivers follows my general focus on how environmental impacts lead to indirect effects mediated by species interactions. Because rivers play key roles in a range of societal functions and are affected by human activities both directly and through links in their surrounding watershed, this research often has more direct links to applied issues. A central theme that has evolved from this work is the importance of considering impacts from a community/ecosystem-wide perspective, rather than focusing on single targeted response variables as is often done in applied settings.

My initial work explored parallels between fish as top consumers in tropical river food webs and birds as top consumers in shoreline communities.  Aside from my fascination at seeing “aquarium fish” living in the wild, these faunas are unusual from a temperate point of view in that warm tropical water temperatures permit substantial herbivory by fish (Wootton and Oemke 1991).  Given this situation and the near absence of aquatic insects in streams of the Atlantic lowlands of Costa Rica, I carried out experiments to determine whether fish were a factor limiting algae and macrophyte biomass.  Initial experiments excluding fishes with cages found higher algal abundance in the absence of fishes (Wootton and Oemke 1991).  Subsequent experiments conducted in collaboration with Cathy Pringle (U. Georgia) further explored fish effects using a variety of experimental manipulations (cages, elevated tiles, large in-stream channels with and without fish) with complimentary strengths and weaknesses.  To our surprise, in these experiments algal biomass declined substantially in the absence of fish.  In concert with this manipulation, aquatic insect populations exploded by two orders of magnitude in the absence of fish.  In cage experiments documenting the experimental response through time, algae initially increased to high abundance in the absence of fish compared to cage controls and unmanipulated plots, then declined to lower levels as the insects increased.  Removing fish protection from experimental substrates for 30 seconds resulted in a feeding frenzy by fish and near elimination of the invertebrates, and hand-picking invertebrates from experimental substrates resulted in increased algae.  These results suggest that fish strongly suppressed insects, releasing algae from grazing pressure (a ‘trophic cascade’).  Subsequent experiments by Cathy Pringle and colleagues using electric fences failed to find any algal response, however, but did find modest increases in insects.  Why such differences in results?  Aquatic insects attained substantially different population sizes when fish were excluded across different experiments, and the net impact of fish on algae exhibits a remarkably strong negative correlation with insect response to fish, suggesting insect supply rates are an additional key variable.  These results show that fish can play very important roles in river food webs and that steps should therefore be taken to avoid detrimental impacts on their populations.  Furthermore, there may be human-health implications, because blackflies (Simuliidae), vectors for diseases such as river blindness, exhibited strong increases when fish were experimentally eliminated from the community.


The food chain-like responses to fish removal observed in my experiments and in other studies of river systems suggested to me that they might be useful models to evaluate simple versions of dynamic food web theory.  This theory is appealing because it potentially integrates a range of processes in a single framework, such as consumer-resource interactions, productivity variation and disturbance intensity.  Initial work (Wootton and Power 1993) focused on a surprising prediction from food chain theory that species at predictable positions in the food web differ in their responses in system productivity, with trophic levels alternating between strong and weak responses.  In collaboration with Mary Power (U. California Berkeley), I subdivided portions of the S. Fork Eel River in northern California and manipulated productivity levels by varying the amount of incident light into the community.  The system responded as predicted by food chain theory, generating strong increases in small predators and algae, but insignificant change in grazers, and by exhibiting alternate increases and decreases in small predators, grazers, and algae when juvenile steelhead trout (Oncorhynchus mykiss) were excluded from the system.  These results suggested that dynamic food web models can indeed be profitable for predicting direct and indirect effects of environmental impacts such as changes in productivity and species extinction on communities.  Currently I am exploring how macroecological patterns change in response to these manipulations.  Analysis indicates that herbivore and predator diversity increase with increasing productivity but other macroecological patterns, while apparent at this experimental scale, are not affected by productivity.  These results indicate that macroecological patterns may be more experimentally tractable then previously thought.

While carrying out productivity studies, we be came interested in whether the framework could also predict effects of changes in flooding disturbance within rivers.  During drought years, we noticed outbreaks of a large, predator-resistant caddisfly, Dicosmoecus gilvipes, suggesting that it might be susceptible to flooding disturbance.  After elaborating our food chain model to account for the presence of a predator-resistant, disturbance-susceptible grazer, our model predicted that reducing flood disturbance could have a negative impact on fish such as salmonids by making a large fraction of energy in the system unavailable to the food chain supporting them (Wootton et al. 1996, Power et al. 1996).  Experimental manipulations of Dicosmoecus in large in-stream channels supported key predictions of this model, causing declines in small predators and algae in the system.  Based on these results we predicted that dams would have similar effects on food web structure by regulating flow and minimizing flooding disturbance.  Surveys of regulated and unregulated rivers also followed predictions of the food web model.  These results suggest that dynamic food web models can be a useful predictive framework for synthesizing disturbance with species interactions and productivity, and indicate that a broader system-wide perspective on river management may be needed for successful restoration.  Dams have been strongly implicated in declines of Pacific salmon, and management based on a single-species perspective (fish ladders, etc.) has been implemented in response.  Such management does not offset dam impacts on other components of the food web supporting salmonids, however, and our results indicate that alternative approaches such as restoring flood disturbance to the system  may be required.

Flood disturbance may affect river function not only through the ‘classical’ food web, but also by affecting the food web that processes detritus.  Surveys of the S. Fork Eel River in drought years exhibited high abundances of Pacific lamprey larvae (Lampetra tridentata), thought to be active detritivores, whereas low densities were observed in years with strong winter floods.  To explore the implications of this impact, I experimentally removed lamprey from benthic plots and evaluated changes in detritus biomass, benthic respiration, algal productivity and invertebrate abundance. Analysis of these experiments using structural equation modeling indicate that reducing lamprey larvae caused no change in detrital biomass or invertebrate abundance, but increased benthic respiriation, presumably via reduced microbial consumption, and increased algal production, perhaps because of increases in microbial nutrient cycling.  Hence altering flooding disturbance appears to also impact the detrital processing of rivers.