The paired catchment approach has been the predominant method for detecting the effects of disturbance on catchment-scale hydrology. Notwithstanding, the utility of this approach is limited by regression model sample size, variability between paired catchments, type II error, and the inability of locating a long-term suitable control. An increasingly common practice is to use rainfall-runoff models to discern the effect of disturbance on hydrology, but few hydrologic model studies (1) consider problems associated with model identification, (2) use formal statistical methods to evaluate the significance of hydrologic change relative to variations in rainfall and streamflow, and (3) apply change detection models to undisturbed catchments to test the approach. We present an alternative method to the paired catchment approach and improve on stand-alone hydrologic modeling to discern the effects of forest harvesting at the catchment scale. Our method combines rainfall-runoff modeling to account for natural fluctuations in daily streamflow, uncertainty analyses using the generalized likelihood uncertainty estimation method to identify and separate hydrologic model uncertainty from unexplained variation, and GLS regression change detection models to provide a formal experimental framework for detecting changes in daily streamflow relative to variations in daily hydrologic and climatic data.
Little is known on the importance of riparian areas to birds near small headwater streams in mesic forests. Progress towards understanding limiting factors that affect bird populations has been difficult because of lack of information beyond the breeding period. I compared bird assemblages between headwater riparian and upland areas throughout the post-breeding period by capturing birds using mist-nets in six paired riparian and upland locations along six headwater streams of the Trask River in northwestern Oregon. In order to assess whether birds prefer headwater riparian areas, I also examined factors affecting habitat selection by juvenile Swainson's thrushes (n=37) using radio telemetry. While riparian and upland locations had similar coarse wood volume and fruiting and tall (> 1.3 m tall) shrub cover, riparian locations had less shrub cover (< 1.3 m tall) and different shrub composition than upland locations. Total capture rate was double that of upland in riparian locations, while bird species richness was similar. Similar numbers of birds were captured in mist-nets oriented perpendicular and parallel to the stream suggesting that birds were not using riparian areas as movement corridors. Adult capture rate was greater in riparian locations than adjacent uplands while results of juvenile capture rates were ambiguous. Riparian locations supported higher capture rates of Swainson's thrushes and winter wrens than adjacent uplands.
Coastal cutthroat trout (Oncorhynchus clarkii clarkii) and cottids (Cottus spp) commonly co-occur in headwater streams in western Oregon. Little is known about the comparative trophic ecology of these species or how they respond to seasonal scarcity of resources. In this study I evaluated the seasonal variability in diets and consumption as it related to food limitation for coastal cutthroat trout and cottids. Over 340 individual diets were quantified from seasonal samples collected in May, July and September of 2008. Diet overlap was relatively low among seasons and species. Coastal cutthroat trout exhibited a more diverse diet in terms of taxonomic richness of prey and consumed both aquatic and terrestrially-derived prey, whereas cottids appeared to specialize on aquatic prey. Based on diet composition and amount consumed, all species appeared to be increasingly food limited from July to September, relative to May. However when diet composition was integrated with a bioenergetic model, coastal cutthroat trout were found to be substantially more food limited than cottids. Differences in the cost of activity between these species may explain this result. Activity costs may be higher for trout, which reside in the water column and rely on active swimming, versus cottids, which lack a swim bladder and are more benthic oriented. Results of this work suggest that cottids are dietary specialists, feeding almost exclusively on benthic prey.
In the Pacific Northwest, multiple studies have found negative effects of timber harvest on stream amphibians, but the results have been highly variable and region-specific. Over the last 30 years forest management practices have changed substantially, yet little work examines how modern forest management relates to the abundance or density of stream amphibians. I examined the influences of contemporary forest practices on Pacific giant salamanders as part of the Hinkle Creek paired watershed study. Density was positively associated with substrate, negatively associate with upstream area drained, and had a weak positive association with fish density, but I found no evidence of an effect of harvest. Pacific Northwest stream amphibians are often negatively associated with sedimentation, but the mechanism underlying this relationship is not clear. I found amphibian larvae were more visible as sediment level increased and some evidence that larvae were less visible in the presence of fish. These patterns are consistent with the hypothesis that sediment affects larval stream amphibians by increasing vulnerability to predation.
This collection of three manuscripts serves to improve methods for collecting, interpreting, and utilizing autocorrelated data from headwater stream networks. Two chapters of this work relied on a unique and comprehensive set of data which constitutes a complete census of habitat unit fish counts from 40 randomly selected headwater basins in western Oregon. The first objective of this work was to evaluate how different sampling designs captured spatial autocorrelation, given the samples were drawn from a population of spatially autocorrelated observations. The second objective was to investigate spatial autocorrelation model range parameters as measures of patch sizes. The third objective was to refine the analysis of temporally autocorrelated hydrology data from paired watershed studies. These are used to evaluate forest harvesting effects on stream biota and hydrology (i.e. fish, amphibians, insects, stream flow, and sediment yield).
Mark–recapture methods were used to examine watershed-scale survival of coastal cutthroat trout (Oncorhynchus clarkii clarkii) from two headwater stream networks. A total of 1725 individuals (‡100 mm, fork length) were individually marked and monitored seasonally over a 3-year period. Differences in survival were compared among spatial (stream segment, subwatershed, and watershed) and temporal (season and year) analytical scales, and the effects of abiotic (discharge, temperature, and cover) and biotic (length, growth, condition, density, movement, and relative fish abundance) factors were evaluated. Seasonal survival was consistently lowest and least variable (years combined) during autumn (16 September – 15 December), and evidence suggested that survival was negatively associated with periods of low stream discharge. In addition, relatively low (–) and high (+) water temperatures, fish length (–), and boulder cover (+) were weakly associated with survival. Seasonal abiotic conditions affected the adult cutthroat trout population in these watersheds, and low-discharge periods (e.g., autumn) were annual survival bottlenecks. Results emphasize the importance of watershed-scale processes to the understanding of population-level survival.
Suspended sediment and in situ turbidity data from two western Oregon streams, Oak Creek and South Fork Hinkle Creek, were used to estimate annual sediment loads for the 2006 water year (October 1, 2005 to September 30, 2006). Water samples and in situ turbidity observations were taken following the Turbidity Threshold Sampling (TTS) protocol. The annual hydrographs for Oak Creek and South Fork Hinkle Creek were divided into storms which resulted in storm-specific relationships between in situ turbidity and Suspended Sediment Concentration (SSC). In the relationship between SSC and in situ turbidity, especially for Oak Creek, there are counterintuitive value which had to be vetted out with values of laboratory turbidity, hydrograph characteristics, and hysteresis loops. Observations of in situ turbidity considered erroneous were adjusted manually with the TTS-adjuster program. The objectives of this study were to determine the efficacy of an automated turbidity adjustment program compared with a manual turbidity adjuster, and to determine the efficacy of two in situ turbidity and SSC relationships to predict annual sediment loads. Relationships between SSC and in situ turbidity were made to estimate annual sediment load for Oak and South Fork Hinkle Creeks. The SSC vs. in situ turbidity relationships were made for storm-specific time periods and for the whole water year.
The Salmon Research and Restoration Plan for the Arctic-Yukon-Kuskokwim Sustainable Salmon Initiative (AYK-SSI) recognizes the need for approaches to characterize determinants of salmon population performance at broader scales. Here we discuss data and modeling tools that have been applied in western Oregon to understand how landscape features and processes may influence salmonids in freshwater.
This dissertation is a collection of three manuscripts that serve to fill the knowledge gaps and advance methods of detecting the effects of contemporary forest harvesting in experimental catchment studies. The objective of this research was to develop change detection models using time-series records to detect and quantify the effects of forest harvesting on streamflow and sediment yield. To accomplish this, it was necessary to characterize streamflow and sediment processes at a temporal scale capable of describing daily, monthly, and seasonal dynamics following forest harvesting; increase sample sizes used to construct regression-based change detection models; and develop alternative methods to the paired-catchment approach in order to discern changes in streamflow and sediment using highly variable time-series data. The paired-catchment approach was used to detect and quantify relative changes in streamflow and sediment yield in 5 harvested catchments. The ability to detect statistically significant changes at certain time-steps was a function of accounting for all sources of variability in change detection models. In this study, we aimed to develop robust change detection models using time-series data to increase sample size and decrease false/missed detections of true treatment effects.
We assessed use and selection of cover by coastal cutthroat trout (Oncorhynchus clarkia clarkii) in six headwater streams in three watersheds in western Oregon, USA during the summer low flow period from 1 August and September 30, 2007. We tagged 1037 coastal cutthroat trout (>100 mm) with passive integrated transponder (PIT) tags across all streams. Selection of cover was analyzed by comparing characteristics of locations used for concealment by relocated fish relative to characteristics of randomly available habitat that could be used for concealment. We measured habitat characteristics for 190 relocated individual fish using cover and 797 randomly points potentially available as cover. Of the latter points, only 235 of 797 were potential cover, based on characteristics of cover actually used by fish. Coastal cutthroat trout used substrate as cover (78%) more often than all other cover types combined (22%). Availability of different cover types was variable, but overall substrate made up 92% of available cover and the remaining 8% represented all other cover types combined. Habitat characteristics measured for both used and available cover included depth at fish location (cm), surface area of cover (m2), proximity to depth of 20 cm for fish located in < 20 cm in depth, b-axis (mm) for substrate >2 mm, and distance under substrate. Each of these habitat characteristics was different for used and available cover.