The Western Washington Intensively Monitored Watersheds (IMW) study was designed to assess the cumulative effects of extensive stream habitat restoration over long durations at ‘watershed’ geographic extents. We conducted annual surveys from 2004 through 2024 at hundreds of randomly selected reaches (“sites”) throughout 10 independent watersheds (“streams”) clustered into 3 regional complexes. Our objectives were to describe cumulative change in habitat measures that are common aims of fish habitat restoration and to test for differences between watersheds with different restoration efforts. In most cases, we did not observe the expected directional changes in habitat measures summarized to whole-watershed extents nor did we observe the expected differences relative to the level of restoration effort. However, we did observe high interannual variation in fish habitat measures. The magnitude of these year-to-year fluctuations was frequently much greater than observed long-term trends. Despite physiographic differences among study watersheds, all watersheds exhibited substantial spatial variation among sites in the degree of interannual variation, indicating a widely shared pattern of within-watershed heterogeneity in habitat stability. Our results underscore the importance of anticipating and accounting for year-to-year changes in stream habitat when assessing habitat conditions and planning and implementing fish habitat restoration.

Declines in populations of Pacific salmon Oncorhynchus spp. in the Pacific Northwest have led to listings under the Endangered Species Act. One objective of current recovery efforts is the restoration of freshwater and estuarine habitats, which had been occurring prior to Endangered Species Act listing but increased dramatically afterwards. However, few listed populations are improving. We believe that there are five factors contributing to the lack of population response to habitat restoration: Not enough restoration has been done. We are not doing the right things in the right places at the right times. Ongoing habitat degradation is offsetting restoration benefits. Not enough time has passed. Monitoring has been inadequate to detect changes in Pacific salmon abundance.All factors contribute to the disappointing progress on Pacific salmon recovery, although their importance varies. Two factors are more consistently significant than the others. Resources available to address habitat damage remain insufficient. The scale of the problem is large, so the response needs to be correspondingly large to yield desired outcomes. Of equal significance is the failure of restoration programs to identify elements controlling fish production. Implementing the right projects in the right places is key to improving the outcomes of restoration.

Forested stream ecosystems involve complex physical and biotic pathways that can influence fish in numerous ways. Consequently, the responses of fish communities to disturbance can be difficult to understand. In this study, we employed a food web model that links biotic (e.g., physiology, predator-prey interactions) and abiotic (e.g., temperature, sunlight) attributes to address fish responses to changes in stream-riparian ecosystems. We modeled responses to food web dynamics in four streams, using scenarios that included responses to riparian disturbance, climate change, and shifts in top consumers. The two consumers we focused on were coastal cutthroat trout (Oncorhynchus clarkii clarkii) and sculpin (Cottus spp., collectively treated as a functional group). We found the responses to environmental changes varied by fish species and among streams, and that responses were not independent due to exploitative interspecific competition. Simulations based on long-term data indicated that coastal cutthroat trout were responsive to changes in allochthonous resources including terrestrial detritus and invertebrates, whereas sculpin were more responsive to changes to autochthonous resources that included, periphyton and aquatic invertebrates. These results may be, in part, a consequence of species-specific foraging behavior. Trout have a higher propensity to drift feed and therefore receive a substantial subsidy from terrestrial invertebrates, whereas sculpin feed mostly on aquatic insects on the streambed. Simulations of changes in summer temperature and stream discharge suggest decreased biomass of both fish species because of physiological constraints on invertebrate prey which reduce fish foraging opportunities. Exploitative competition also may be important in fish responses: when one fish taxon was removed, the other showed increased biomass. Although the pattern of simulation results was consistent across the four streams, the magnitude of change varied among streams. Streams with food webs fueled by multiple energy sources may be more resilient to changes to riparian forests and climate. Through application of a systems model, we gained insights into pathways of productivity for fish in forested stream ecosystems that provide understanding of processes that influence fish and streams, as well as implications for management of both.

Forest management in riparian ecosystems can significantly alter biotic and abiotic processes in streams. Forest harvest without the retention of buffers along small streams can affect organic matter dynamics, and drive instream characteristics like trophic food webs. To investigate the extent to which differing levels of tree retention adjacent to the channel mitigated changes in organic matter dynamics, we examined coarse particulate organic matter delivery, transport, and retention, as well as canopy cover, along small streams with four harvest treatments, both before and after harvest. Our research was part of a larger effort of the Trask River Watershed Study (TRWS) in the northern Oregon Coast Range, which examined the long-term physical, chemical, and biological effects of forest management on aquatic ecosystems at multiple spatial scales. Canopy cover at reference and treatment sites prior to harvest was approximately 91%, but following harvest, mean canopy cover at treatment sites decreased with increasing harvest intensity: 3% decrease after clearcut with buffer, 14% after clearcut with leave trees, and 57% after complete clearcut. Organic matter delivery (i.e., the amount of leaf, needle, wood, reproductive, or miscellaneous litter material) to streams was dominated by leaves and varied seasonally but decreased overall after harvest with complete clearcut. Organic matter transport (i.e., the amount of material netted during a sampling period at the end of each stream reach) values fell within the observed range of reference values at all harvested sites. Organic matter retention (i.e., the amount of material sampled in plots systematically placed along the stream reach) post-harvest was dominated by woody material and was consistent with pre-harvest measurements and reference sites for all the harvested streams. The number of log jams, and the total weight of log jams, increased for all treatments after harvest, although the most dramatic increase in log jam weight (140% increase) was observed for the complete clearcut and clearcut with leave tree sites. On these five sites, our results indicated that while complete clearcutting reduced organic matter delivery, retention of even a small number of trees along these streams appeared to support litter delivery rates after harvesting comparable to before harvest. We acknowledge the limitations of results from single treatment sites; however, our results were generally consistent with studies that had full replication. As part of the Trask River Watershed study, our research adds to the comprehensive effort to understand the ecological significance of riparian buffers and forest harvest in headwater streams.

We examined riparian system responses to an extreme rainfall event on 1–4 December 2007, in eleven small watersheds (mean area—13.2 km2) from 2008–2016 at debris flow, high flood, and low flood reaches (all extended overbank flows). Macroinvertebrate responses followed expected outcomes after extreme disturbance including increasing chironomids and other multi-voltine species. A core assemblage of twenty abundant and common species-maintained populations even after debris flow (likely by recolonizing quickly) with total richness during project of 253 including 183 rare species (<0.01 total abundance) supporting an annual turnover of species from 22 to 33%. Primary disturbance changes to habitat were declines in shade and in-channel wood at all reaches, more strongly at debris flow reaches. Macroinvertebrate communities across disturbance intensities became increasingly similar after the storm. Combined effects of the flood reducing channel complexity and previous logging decreasing in-channel wood recruitment from riparian systems, limits habitat complexity. Until this feature of forested watershed streams returns, there appears to be a ceiling on reach scale aquatic biological diversity.

The Calapooia River in western Oregon supports a small winter steelhead trout (Oncorhynchus mykiss) population and historically supported spring Chinook salmon (Oncorhynchus tshawytscha). Early timber harvesting removed the riparian forest, and log transportation practices simplified the channel. Those disturbance legacies continue to affect fish habitat by limiting shade and channel complexity, complicating conservation efforts. To evaluate juvenile salmonid rearing potential, macroinvertebrate drift, thermal regime and physical habitat were measured at eight sites in 24km of the upper river during late summer baseflow. Overall physical habitat was simple, with few functioning instream structures or pools. During the 22-day drift study, flows declined and maximum site stream temperatures ranged from 23.1 degrees C at the lower end to 16.4 degrees C 24km upstream. Macroinvertebrate drift concentrations ranged from 0.7-13.7 ind. m(-3) with biomasses from 0.02-1.23mg m(-3). Drift concentration biomass was higher upstream (P=0.006) than downstream and declined overall (P<0.001) during the study. Drift biomass was dominated by five taxon groups - Baetis tricaudatus, Calineuria californica, Hesperoperla pacifica, Simulium spp., and Chironomidae, which were 65% of total biomass. During twilight, total biomass and biomass of B. tricaudatus, Simulium spp., and Chironomidae (both larvae and adults) were higher. Total drift declined dramatically over the study period owing to decreases in drift concentration and a 58% decline in discharge, greatly reducing overall drift and available food resources for juvenile-rearing salmonids. The upper catchment, both with cooler temperatures and higher food availability, provided the best conditions for juvenile anadromous salmonids to survive late summer conditions. Conservation consequences of climate change-induced alterations in flow and temperature may further affect habitat quality for juvenile salmonids in this catchment in the coming decades.

The Campos grasslands of southern South America have been affected by long-term livestock grazing and increasingly widespread afforestation, but the effects of multiple disturbances have not been well studied, especially at multiple spatiotemporal scales. To determine vegetation response to these disturbances, we constructed grazing exclosures bisecting the forest-grassland ecotone at five sites and measured species richness in nested plots for four treatments (grazed forest, ungrazed forest, grazed grassland, and ungrazed grassland) over 18 months. Species-area curves revealed greater small-scale species richness in grasslands than forests but greater rates of species accumulation in forests and thus little difference in richness between habitats at larger scales. Temporally, small-scale richness differed among sample dates, but rates of species accumulation did not. The interactive effect of habitat and management on total species richness varied with times since the exclosures were built. Grazing and afforestation interactively affect the vegetation of the Campos, as plants respond to changes in both disturbances; over the duration of this study, the removal of grazing decreased species richness in grasslands but increased richness in forests. Our results illustrate that the response of plant species richness to grazing and afforestation varies by the nature of the disturbance mechanism, and that multiple disturbances have complex interactive effects on vegetative responses.

Leaves and needles are an important energy source for streams. However, relatively little is known about factors that determine width of the area contributing litter to streams. We assessed the relative effect of wind speed, wind direction, litter type, tree height and riparian topography on litter delivery to streams. Wind speed effect on litter travel distance was determined by releasing needles (Douglas-fir) and leaves (red alder) from mature and young tree canopies over a range of wind velocities. Short-term litter collections were conducted to determine the effect of wind speed on litter fall. Litter travel distance increased with increasing wind speed and increasing tree height. At low wind speeds leaves traveled further than needles. Litter fall increased with wind speed. These relationships were combined with literature values for seasonal rates of litter production and an annual wind speed and direction record from a riparian area in western Washington to evaluate the relative effect of various factors on litter delivery area width. We found that width of the contributing area for needles was about 35% greater in riparian stands supporting mature conifer trees than at sites with young trees. Increasing riparian area slope from 0 to 45 increased width of contributing area by 71-95%, depending on litter type and stand age. Doubling measured wind speed increased contribution zone width 67-82%. Estimated buffer width required to capture 95% of annual litter input ranged from about 14 m to over 25 m under the conditions which we evaluated. The variety of factors influencing litter delivery area width and the spatial and temporal variation in these factors indicates that the common practice of employing fixed-width buffers to protect stream-riparian interactions cannot be consistently effective. (C) 2016 Elsevier B.V. All rights reserved.

The importance of multiple processes and instream factors to aquatic biota has been explored extensively, but questions remain about how local spatiotemporal variability of aquatic biota is tied to environmental regimes and the geophysical template of streams. We used an individual-based trout model to explore the relative role of the geophysical template versus environmental regimes on biomass of trout (Oncorhynchus clarkii clarkii). We parameterized the model with observed data from each of the four headwater streams (their local geophysical template and environmental regime) and then ran 12 simulations where we replaced environmental regimes (stream temperature, flow, turbidity) of a given stream with values from each neighboring stream while keeping the geophysical template fixed. We also performed single-parameter sensitivity analyses on the model results from each of the four streams. Although our modeled findings show that trout biomass is most responsive to changes in the geophysical template of streams, they also reveal that biomass is restricted by available habitat during seasonal low flow, which is a product of both the stream’s geophysical template and flow regime. Our modeled results suggest that differences in the geophysical template among streams render trout more or less sensitive to environmental change, emphasizing the importance of local fish–habitat relationships in streams.

Stream temperature changes as a result of forest practices have been a concern in the Pacific Northwest for several decades. As a result of this concern, stream protection requirements for forest lands were first adopted in the early 1970s and have become progressively more stringent. While there have been multiple studies examining the effects of stream protection buffers on water temperature, there are few studies examining temperature patterns over long periods on intensively managed forests. Water temperature in the upper Deschutes River watershed, Washington has been monitored since 1975 and represents one of the longest studies of water quality on managed forests in the Pacific Northwest. This data record, collected from basins of varying sizes, has enabled us to examine the combined effects of hydro-climatic patterns and forest management on stream temperature. Effects of harvest conducted prior to buffer regulations were clearly identifiable and most pronounced on smaller streams. We were not able to detect any response on larger channels to more recent timber harvest where riparian buffers were required. This analysis also emphasizes that it is critical to account forchanging climate when examining long-term temperature patterns. We found that in many cases the temperature improvements associated with more stringent buffer requirements implemented over the last 35 years in the Deschutes watershed have been offset by warming climatic conditions.

Human activities have resulted in novel ecosystems around the globe, but many of these ecosystems are poorly understood, particularly when multiple disturbances interact to create novel combinations of species. The Campos ecoregion of South America is one of the most extensive grasslands globally, yet is unique in the fact that there has been little historical evidence of herbivory or fire that traditionally maintain grassland ecosystems. Livestock were introduced several centuries ago, and afforestation began several decades ago. Introducing afforestation to an already grazed system can further alter light conditions, soil characteristics, and moisture regimes, which may result in decreased species richness, increased plant species introductions, and altered abundances of grasses, herbs, and woody species. We examined agroforestry management cycle effects on vegetation species richness, frequency, and composition in Eucalyptus grandis plantations, studying five phases of the management cycle (grassland, young forest, mid‐stage forest, old forest, and post‐harvest). At each of 25 sites, we recorded species presence in 40 quadrats (1 m 2 ) and categorized species by origin (native or introduced) and life form (annual graminoid, annual herb, perennial graminoid, perennial herb, woody, or fern). Impacts on species richness and composition were greatest in mid‐stage forests. The proportion of introduced species did not differ by management phase. Life form categories responded strongly but differently to management phases. Our research suggests that afforestation initially reduces vegetation richness and alters composition from that in grasslands, although there is considerable recovery following thinning and tree harvest. Factors that enhance recovery in this system include the presence of disturbance‐adapted vegetation, proximity to remnant grassland patches, and the short rotation cycles of plantation forests. Further research is needed to understand the effects of multiple rotations, effects of afforestation on taxa other than plants, and consequences of landscape fragmentation across the Campos. While our research suggests that afforestation has created a novel ecosystem that is a variant of the system that evolved under long‐term grazing, the extent to which this new condition will evolve over time with the continued presence of grazing and repeated forest management cycles remains unknown.

Afforestation, the practice of planting trees where they did not occur recently, is a common practice around the globe. In recent decades, afforestation has begun to occur in the Campos grasslands of South America. Exotic species are used for these plantations and potential encroachment of these trees into remaining grassland areas is a concern. We quantified the germination and establishment of loblolly pine (Pinus taeda) in and adjacent to five plantations in northern Uruguay. Transects were placed in grazed areas and in areas where grazing was experimentally excluded. Areas were sampled semi-annually for 2 years to examine germination (seedling density in Spring) and establishment (seedling density in Autumn). In grasslands, seedlings were present only near the forest edge (<10 m). Germination densities were significantly higher in forests than grasslands, and establishment densities tended to be higher in forests and ungrazed areas. Protection from livestock increased seedling establishment, particularly in the first year after the exclosures were built. However, the mechanisms controlling encroachment differ with grazing management: in ungrazed grasslands, the dense cover of herbaceous and grass species likely prevent establishment whereas in grazed grasslands, the intensive livestock grazing prevents tree establishment. Our results suggest that loblolly pine encroachment into adjacent grasslands is relatively unlikely, although further work is necessary to examine the long-term survival and growth of those individuals that do manage to survive in this habitat. Finally, similar studies should be undertaken with other species used to afforest these grasslands. (C) 2013 Elsevier B.V. All rights reserved.

We compared stream ecosystem responses to two types of disturbances: flood and debris flows. A large storm in February 1996 disturbed four similarly sized sub‐watersheds of the Calapooia River, in the Cascade Mountains of Oregon, USA. All sub‐watersheds had flood and in two, 31 and 81% of the perennial channel, a debris flow affected the channel. For 8 years, we used a suite of approaches: stream temperature, nutrient regime, periphyton and macroinvertebrate assemblages, and resident trout abundance and habitat to evaluate the persistence of instream impacts. Differences in stream temperatures and nitrate‐nitrogen concentrations were evident, probably, due to the removal of vegetation and modification of riparian soils at the debris flow sites. Instream biological responses varied. After the event, fish, including trout, were rare with no fish at the debris flow sites. Within 6 years, trout densities ( Oncorhynchus mykiss and Oncorhynchus clarki ) were similar and young‐of‐the‐year trouts were common. In contrast, periphyton and macroinvertebrate assemblages differed. Periphyton biomass was lower and nitrogen‐fixing periphyton was more abundant at the debris flow sites. Macroinvertebrate assemblage diversity was higher at the debris flow sites due to fewer dominant taxa. Macroinvertebrate functional feeding groups also differed with fewer gatherers and more scrapers at the debris flow sites. Debris flow impacts related to loss of riparian canopy will probably persist until mature red alder stands are re‐established along stream‐reaches affected by debris flows to provide nitrogen input and shade. Copyright © 2011 John Wiley & Sons, Ltd.

Riparian plant communities along small streams occupy a small proportion of the total landscape but can provide disproportionally large ecological, social, and economic benefits. We examined plant communities at 25 study sites along small fish-bearing streams in temperate managed forests of the Pacific Northwest spatially as a function of distance from stream and temporally by assessing a chronosequence of stand ages: young (31-51 years), mature (52-70 years), and old (> 100 years). We identified three distinct vegetation communities based on species cover and richness in shrub and herb layers: riparian (0-9 m), transitional (10-29 m), and upslope (30-80 m); 12 species were indicators of these vegetation communities. For tree species, basal area increased with stand age. Shrub species cover and richness were greatest in old stands, but herb species richness was highest in young stands. Composition varied with stand age; 15 species were indicators of these differences in composition. These results, together with information on successional and wetland status, suggest that plant communities on small fish-bearing streams reflect geomorphic and fluvial settings but also follow successional patterns found in natural forests. These stands will become some of the primary unharvested, older forests within the managed forest landscape and provide insights for effective riparian management on sites impacted by historical management practices prior to the regulations requiring riparian buffers.

Drifting invertebrates from small headwater streams are a food subsidy for fishes and other downstream consumers in larger streams. This subsidy can become especially important for stream fishes during summer due to high metabolic demand caused by elevated water temperatures. The source length in small streams contributing these drifting invertebrates to fish bearing habitat has not been established. We conducted an experiment to determine if summer drift delivered by headwater streams to larger channels is influenced by contributions originating more than 100 m upstream. We conducted this experiment on tributaries of the Calapooia River, Oregon, USA. We blocked drift with 250 pm mesh nets in three streams 100 m upstream of our sample nets continuously for 48 d. We also sampled five streams without blocking nets. Four 24 h samples of drift were collected prior to the placement of blocking nets and every two weeks afterwards at each stream. Overall mean abundance and mean biomass of drift did not differ between blocked and unblocked streams for any sampling period. There was a temporal trend at all sites of declining invertebrate abundance and biomass. We found no treatment effect for any of the most common insect orders or for the 16 genera that comprised at least 1 % of the total organisms captured. The majority of invertebrate drift is of local origin in headwater streams during summer baseflow.

This study examines watershed patterns of riparian vegetation, shade, and stream temperature eight years after extreme storm events triggered numerous debris torrents throughout the Pacific Northwest. We examined twelve impacted streams in two western Oregon watersheds: the Calapooia River in the western Cascades and the Williams River in the Coast Range. Red alder (Maus rubra) and willow (Salix spp.) were the dominant species on debris torrented areas in both watersheds. Post-disturbance vegetation recovery was significant in both watersheds, impacting shade and stream temperatures. However, red alder density, basal area, and height were significantly greater along streams in the Williams River watershed than along streams in the Calapooia River watershed. Willow density, basal area and height were similar between the watersheds. Stream shading levels mirrored red alder growth, with greater average shading in the Williams River watershed. The greater shade translated into lower summer maximum stream temperatures and maximum diurnal stream temperature fluctuations in the Williams River as compared to the Calapooia River watershed. Minimum stream temperatures were not different between the two watersheds. The rapid re-growth of red alder along the Williams River watershed ultimately lead to a rapid decline in maximum summer stream temperatures for that watershed compared to the Calapooia River watershed. The location where the disturbance occurred had an important role in determining the rate and pathway of stream recovery. (C) 2011 Elsevier B.V. All rights reserved.

Elevated landslide rates in forested landscapes can adversely impact aquatic habitat and water quality and remove and/or degrade soil resources required for forest regeneration. As a result, understanding the associations between management actions, natural factors, and landslide rates is important information needed for land managers. An unusual and powerful storm in early December, 2007, caused record flooding and thousands of landslides across southwest Washington and northwest Oregon, USA, and provided a rare opportunity to examine the effects of both natural factors and forest management practices on landslide density. Landslide inventory data were collected from both aerial photos and systematic field surveys to provide a broad survey database that was used to develop estimates of landslide density and to examine associations between landslide density, precipitation, topography, and forest stand age across a 152,000ha forested landscape in the Willapa Hills, Washington. We estimated the probability of detecting landslides on aerial photos for six strata defined by forest stand age and a broad range of rainfall intensity, expressed as percent of the 100-year, 24-h, maximum rainfall. Key findings are that landslide detection probability decreased with increasing stand age, but was similar across rainfall intensities. The overall fraction of field-detected landslides that were not detected on 1:12,000-scale aerial photos was 39%. Very few landslides occurred in the 0–100% of 100-year rainfall category, regardless of stand age or slope gradient class. At higher rainfall intensities, significantly higher landslide densities occurred on steep slopes (>70% gradient) compared to lower gradient slopes, as expected. Above ∼150% of 100-year rainfall, the density of landslides was ∼2–3 times larger in the 0–5 and 6–10 year stand age categories than in the 11–20, 21–30, 31–40, and 41+ categories. The effect of stand age was strongest at the highest rainfall intensities. Our results demonstrate that ground-based landslide inventory data are required in order to correct for detection bias from aerial photos, develop reasonable estimates of landslide density across environmental gradients such as rainfall magnitude and topography, and make unbiased interpretations of relationships between forest management associations and landslide occurrence.