Restoration of riparian vegetation is widely recognized as a tool in stream rehabilitation, but information on whether local riparian characteristics can mitigate the effects of catchment-level stressors on in-stream processes is limited. We measured community metabolism in 21 streams in the Canterbury region of New Zealand along 2 independent gradients of agricultural intensity and riparian cover (from closed canopied to open canopied) to assess relative effects of landscape and local factors on stream trophic state. We measured stream metabolism with the single-station open-channel diel O 2 method. We found a correlation between gross primary production (GPP) and ecosystem respiration (ER), indicating a gradient of trophic states across sites. Streams were strongly heterotrophic with P ∶R values varying from 0.01 to 0.25. GPP and ER increased with % agriculture and % macrophyte cover, but decreased with % shade from riparian vegetation. Hierarchical partitioning analysis indicated that % agriculture was the only landuse variable to have a significant independent effect on GPP and ER. Among local variables, % shade and % macrophyte cover had significant independent effects on GPP. Percent shade was the only local variable to have a significant independent effect on ER. Percent shade had a stronger effect on both GPP and ER than did % agriculture, and a trade-off exists between the importance of agricultural and forest cover on stream metabolism at different spatial scales. Our results highlight the role of local riparian conditions in controlling trophic state and the importance of riparian buffers as a tool to mitigate eutrophication in streams and rivers.

Abstract. Traditional productivity-diversity theory predicts that eutrophication will result in greater species richness due to increased resources at the bottom of the food web. However, few studies on the effects of increasing ecosystem productivity on biological communities have included responses at multiple trophic levels. We hypothesized that the effect of eutrophication on species richness would vary between different trophic levels due to shifts in community composition and trophic interactions. To investigate the mechanisms driving productivity-richness relationships, we constructed food webs for 18 streams across a eutrophication gradient on South Island, New Zealand, and measured productivity, water quality, and habitat characteristics in each stream. A principal components analysis yielded two orthogonal axes of eutrophication: one associated with macrophytes, benthic substrate, and gross primary productivity (GPP) and the other with catchment area, temperature, and algal standing stock (chlorophyll a). Surprisingly, the majority of community response variables, especially defended primary consumer abundance and biomass, were more strongly associated with the macrophyte/habitat axis. The lack of change in abundance and declines in biomass and average mass of predatory invertebrates and fish with increasing eutrophication, despite increases in prey abundance, suggest that energy was not being passed up the food chain to higher trophic levels. The predominance of defended producers (macrophytes) and consumers (snails and microcrustacea) in eutrophic streams likely served as trophic bottlenecks; both are largely inedible by higher trophic levels. Consequently, our results suggest that managers need to focus on preserving food web linkages and energy flow, as well as biodiversity, in eutrophic systems.