Clearing native vegetation has pervasive effects on stream and river ecosystems worldwide. The stated aims of replanting riparian vegetation often are to restore water quality and to re-establish biotic assemblages. However reachscale restoration may do little to combat catchment-scale degradation, potentially inhibiting restoration success. Whether reinstating biodiversity is a realistic goal or appropriate indicator of restoration success over intermediate timeframes (,30 years) is currently unclear. We measured the response of aquatic macroinvertebrate assemblages to riparian replanting in a chronosequence of replanted reaches on agricultural streams in south-eastern Australia. Sites had been replanted with native vegetation 8¿22 years before the study. Indices of macroinvertebrate sensitivity did not respond to replanting over the time gradient, probably because replanting had little benefit for local water quality or in-stream habitat. The invertebrate assemblages were influenced mainly by catchment-scale effects and geomorphological characteristics, but were closer to reference condition at sites with lower total catchment agricultural land cover. Reach-scale replanting in heavily modified landscapes may not effectively return biodiversity to pre-clearance condition over decadal time-scales. Restoration goals, and the spatial and temporal scale of processes required to meet them, should be carefully considered, and monitoring methods explicitly matched to desired outcomes.