Forest harvesting has the potential to alter water quality and quantity and change the physical character and biotic composition of streams in the forested catchments over many years. Impacts and recovery are assessed in terms of stream water clarity (fortnightly), temperature (continuous during summer), and stream habitat and biota (in summer and at the end of winter). Data collected for this study are described for the winter 2014 and summer 2015 surveys (13 sites), and incorporates findings from earlier surveys (since 1992) from streams that differ in size and catchment harvest intensity. New harvest occurred upstream of only one site, Wa<20, over the 2014-15 monitoring year. The area harvested was 2 ha, representing 18 % of the catchment upstream of Wa<20. Stream bank erosion responses to local and upstream harvesting varied among sites, being more prevalent and greatest where channel morphology was meandering, where riparian buffers had been removed, where sites were located in catchments with unconsolidated soil structures and where harvest debris altered flow patterns during storms. Channel width increases were observed at some smaller harvested sites (<50 ha catchment area) where the banks were steep and the riparian soils were unconsolidated and vulnerable to erosion. These increases persisted for approximately three years. Channel widths also increased at some larger sites following major storm events. Increased water widths at some sites probably reflected increased catchment water yields (and hence stream discharge) as evapotranspiration declined post-harvest. Deposits of sand/silt increased following a lag of between 1 and 3 years were observed after harvest at sites in catchments with unconsolidated soils. Little change in amount of deposited fine material was apparent at sites where soil disturbance during the logging process was minimal. Water clarity responses in the years after logging varied among sites. At most sites, harvest reduced annual median water clarity by around 35 % in the year after logging adjacent to the study reach. Harvest impacts on clarity were apparent for up to 12.5 years after harvest at one site with exceptional pre-harvest clarity, but for most other sites clarity reductions lasted less than three years. The largest post-harvest reduction in annual median clarity was 49 % in a single year (exceeding the MfE guideline for the protection of water clarity from point source discharges) and occurred the year after catchment harvest adjacent to the study reach. Water temperatures at all small sites (< 50 ha) increased following harvest, but did not exceed the Waikato Regional Plan standard maximum of 25 °C. Increases in average summer temperature (relative to reference and/or pre-harvest) at all but one small site were within the 3 °C change permitted in the Plan (that does not specify the assessment regime for compliance assessment). Results for these small sites suggest that slash left suspended above the channel and riparian vegetation regrowth after harvesting were providing adequate shade to prevent exceeding the standards. At large sites, where upstream catchment area was greater than 50 ha, maximum summer temperatures between December and March increased by up to 8.5 °C following harvest, and both mean and maximum summer temperatures remained above reference conditions for up to 10.5 years post-harvest. Increases of > 3 °C in mean water temperature were observed for a maximum of 5 years at the large sites. Summer temperatures greater than the 25 °C limit in the Waikato Regional Plan were apparent at one large site 10.5 years after local riparian harvesting. High re-aeration in these moderately steep, fast-flowing, streams generally prevented decaying slash in the water from causing significant dissolved oxygen (DO) depletion. DO concentration reductions were observed locally where large amounts of slash were deposited in one small stream post-harvest. The reductions were relatively short-lived, however, and did not persist for more than a month in this Coromandel system where frequent high flows disperse deposited slash. Epilithon (attached algae, also known as “periphyton”) responses were most pronounced in medium (20-50 ha) to large streams (> 50 ha) where logging reduced shade and increased temperatures. Spikes in biomass at the large sites were often observed every 2-4 years, and surpassed MfE guidelines at one site 9.5 years after onsite riparian harvesting. The spikes at the large sites are likely to occur when a long time has passed since the last significant flood and growing conditions post-harvest (light, temperature, nutrients) are favourable for growth. In contrast, epilithon responses were least in small (<50 ha) streams where shade provided by slash, topographic features and regrowth of riparian vegetation reduced light levels and temperatures. At small sites where increases in epilithon biomass were observed after logging, regrowth of riparian vegetation took approximately 4 years to reduce biomass to pre-harvest levels. Across all sites, epilithon spikes following harvest occurred after a 1-2.5 year lag period. Invertebrate taxon richness at large sites appeared more resistant to change than at sites <50 ha in the 2-3 years immediately following riparian and upstream harvest, but typically trended downwards 2-3 years post-riparian harvest for extended periods (10.5 years plus). At the small sites harvest had little impact on taxon richness where large amounts of slash were left suspended over the channel. At the small sites where only small amounts of slash remained after harvest, taxon richness tended to decline for between 1.5 and 2 years but typically improved to the pre-harvest range within 2.5-5 years. In contrast to taxon richness, invertebrate abundance at many sites was observed as substantial increases either immediately or shortly after local riparian harvest and disturbance. At some sites, increases were observed up to 12.5 years post-logging adjacent to the study reach. The invertebrate Index of Biotic Integrity (IBI) showed shifts in invertebrate communities associated with upstream logging, particularly with harvesting of riparian trees alongside the sampling reach. At large sites, harvest adjacent to the study reach typically reduced IBI from “non-impaired/slightly impaired” to “moderately impaired”. Whilst some improvements were observed subsequently, impacts on invertebrate communities were apparent up to 12.5 years post-riparian harvest. At all the small sites, IBI values declined from “non-impaired/slightly impaired” to “slightly/moderately impaired” immediately following harvest adjacent to the study reach. Recovery to “non-impaired” followed a period of between one and three years after riparian logging before further declines to “slightly/moderately impaired”. Further recovery to pre-harvest condition was apparent between 3 and 7 years after riparian logging. The length of the intermediate recovery phase and subsequent post-harvest recovery may be related to the extent and longevity of slash, which provided protection to the stream following harvest. Similarly, multivariate analysis of invertebrate community composition showed different levels of change amongst the monitoring sites; most notably, significant shifts in community composition at sites when logging impact coincided with severe storm disturbance. In streams of all sizes, riparian buffers helped reduce harvest impacts on physical, water quality and biological stream variables. Several notable observations were made over the 2014-15 monitoring year. Epilithon biomass remained relatively high at site W, 14.5 years after site shade removal during logging. Maximum summer water temperatures exceeded the 25 °C Significant Indigenous Fisheries and Fish Habitat Standard at site Wi on 5 occasions and at site WaL on 7 occasions. Site 108<50 was the only site to exceeded the 3 °C change (relative to pre-harvest) for maximum summer temperatures. Elevated levels of the proportion of sand in the stream bed have continued at site Wa<20. Prominent improvements in invertebrate IBI occurred at sites W, 105<50, 105<20, 108<20, Wa<20, 108<50 and 81<50, after 13, 7, 6.5, 6, 5.5, 4.5 and 4.5 years since riparian harvest, respectively.