Effects of livestock exclusion on in‐stream habitat and benthic invertebrate assemblages in montane streams

1. Stream and riparian ecosystems in arid montane areas, like the interior western United States, are often just narrow mesic strands, but support diverse and productive habitats. Meadows along many such streams have long been used for rangeland grazing, and, while impacts to riparian areas are relatively well known, the effect of livestock grazing on aquatic life in streams has received less attention. 2. Attempts to link grazing impacts to disturbance have been hindered by the lack of spatial and temporal replication. In this study, we compared channel features and benthic macroinvertebrate communities (i) between 16 stream reaches on two grazed allotments and between 22 reaches on two allotments where livestock had been completely removed for 4 years, (ii) before and after the 4‐year grazing respite at a subset of eight sites and (iii) inside and outside of small‐scale fenced grazing exclosures (eight pairings; 10+ year exclosures) in the meadows of the Golden Trout Wilderness, California (U.S.A.). 3. We evaluated grazing disturbance at the reach scale in terms of the effects of livestock trampling on per cent bank erosion and found that macroinvertebrate richness metrics were negatively correlated with bank erosion, while the percentage of tolerant taxa increased. 4. All macroinvertebrate richness metrics were significantly lower in grazed areas. Bank angle, temperature, fine sediment cover and erosion were higher in grazed areas, while riparian cover was lower. Regression models identified riparian cover, in‐stream substratum, bank conditions and bankfull width‐to‐depth ratios as the most important for explaining variability in macroinvertebrate richness metrics. 5. Small‐scale grazing exclosures showed no improvements for in‐stream communities and only moderate positive effects on riparian vegetation. In contrast, metrics of macroinvertebrate richness increased significantly after a 4‐year period of no grazing. 6. The success of grazing removal reported here suggests that short‐term removal of livestock at the larger, allotment meadow spatial scale is more effective than long‐term, but small‐scale, local riparian area fencing, and yields promising results in achieving stream channel, riparian and aquatic biological recovery.