Identifying monitoring information needs that support the management of fish in large rivers.

Management actions intended to benefit fish in large rivers can directly or indirectly affect multiple ecosystem components. Without consideration of the effects of management on non-target ecosystem components, unintended consequences may limit management efficacy. Monitoring can help clarify the effects of management actions, including on non-target ecosystem components, but only if data are collected to characterize key ecosystem processes that could affect the outcome. Scientists from across the U.S. convened to develop a conceptual model that would help identify monitoring information needed to better understand how natural and anthropogenic factors affect large river fishes. We applied the conceptual model to case studies in four large U.S. rivers. The application of the conceptual model indicates the model is flexible and relevant to large rivers in different geographic settings and with different management challenges. By visualizing how natural and anthropogenic drivers directly or indirectly affect cascading ecosystem tiers, our model identified critical information gaps and uncertainties that, if resolved, could inform how to best meet management objectives. Despite large differences in the physical and ecological contexts of the river systems, the case studies also demonstrated substantial commonalities in the data needed to better understand how stressors affect fish in these systems. For example, in most systems information on river discharge and water temperature were needed and available. Conversely, information regarding trophic relationships and the habitat requirements of larval fishes were generally lacking. This result suggests that there is a need to better understand a set of common factors across large-river systems. We provide a stepwise procedure to facilitate the application of our conceptual model to other river systems and management goals.

Can data from disparate long-term fish monitoring programs be used to increase our understanding of regional and continental trends in large river assemblages?

Understanding trends in the diverse resources provided by large rivers will help balance tradeoffs among stakeholders and inform strategies to mitigate the effects of landscape scale stressors such as climate change and invasive species. Absent a cohesive coordinated effort to assess trends in important large river resources, a logical starting point is to assess our ability to draw inferences from existing efforts. In this paper, we use a common analytical framework to analyze data from five disparate fish monitoring programs to better understand the nature of spatial and temporal trends in large river fish assemblages. We evaluated data from programs that monitor fishes in the Colorado, Columbia, Illinois, Mississippi, and Tallapoosa rivers using non-metric dimensional scaling ordinations and associated tests to evaluate trends in fish assemblage structure and native fish biodiversity. Our results indicate that fish assemblages exhibited significant spatial and temporal trends in all five of the rivers. We also document native species diversity trends that were variable within and between rivers and generally more evident in rivers with higher species richness and programs of longer duration. We discuss shared and basin-specific landscape level stressors. Having a basic understanding of the nature and extent of trends in fish assemblages is a necessary first step towards understanding factors affecting biodiversity and fisheries in large rivers.

Early detection monitoring for larval dreissenid mussels: how much plankton sampling is enough?

The development of quagga and zebra mussel (dreissenids) monitoring programs in the Pacific Northwest provides a unique opportunity to evaluate a regional invasive species detection effort early in its development. Recent studies suggest that the ecological and economic costs of a dreissenid infestation in the Pacific Northwest of the USA would be significant. Consequently, efforts are underway to monitor for the presence of dreissenids. However, assessments of whether these efforts provide for early detection are lacking. We use information collected from 2012 to 2014 to characterize the development of larval dreissenid monitoring programs in the states of Idaho, Montana, Oregon, and Washington in the context of introduction and establishment risk. We also estimate the effort needed for high-probability detection of rare planktonic taxa in four Columbia and Snake River reservoirs and assess whether the current level of effort provides for early detection. We found that the effort expended to monitor for dreissenid mussels increased substantially from 2012 to 2014, that efforts were distributed across risk categories ranging from high to very low, and that substantial gaps in our knowledge of both introduction and establishment risk exist. The estimated volume of filtered water required to fully census planktonic taxa or to provide high-probability detection of rare taxa was high for the four reservoirs examined. We conclude that the current level of effort expended does not provide for high-probability detection of larval dreissenids or other planktonic taxa when they are rare in these reservoirs. We discuss options to improve early detection capabilities.

A survey of benthic sediment contaminants in reaches of the Columbia River Estuary based on channel sedimentation characteristics.

While previous studies have documented contaminants in fish, sediments, water, and wildlife, few specifics are known about the spatial distribution of contaminants in the Columbia River Estuary (CRE). Our study goal was to characterize sediment contaminant detections and concentrations in reaches of the CRE that were concurrently being sampled to assess contaminants in water, invertebrates, fish, and osprey (Pandion haliaetus) eggs. Our objectives were to develop a survey design based on sedimentation characteristics and then assess whether sediment grain size, total organic carbon (TOC), and contaminant concentrations and detections varied between areas with different sedimentation characteristics. We used a sediment transport model to predict sedimentation characteristics of three 16km river reaches in the CRE. We then compartmentalized the modeled change in bed mass after a two week simulation to define sampling strata with depositional, stable, or erosional conditions. We collected and analyzed bottom sediments to assess whether substrate composition, organic matter composition, and contaminant concentrations and detections varied among strata within and between the reaches. We observed differences in grain size fractions between strata within and between reaches. We found that the fine sediment fraction was positively correlated with TOC. Contaminant concentrations were statistically different between depositional vs. erosional strata for the industrial compounds, personal care products and polycyclic aromatic hydrocarbons class (Indus-PCP-PAH). We also observed significant differences between strata in the number of detections of Indus-PCP-PAH (depositional vs. erosional; stable vs. erosional) and for the flame retardants, polychlorinated biphenyls, and pesticides class (depositional vs. erosional, depositional vs. stable). When we estimated mean contaminant concentrations by reach, we observed higher contaminant concentrations in the furthest downstream reach with a decreasing trend in the two upstream reaches. Contaminant survey designs that account for sedimentation characteristics could increase the probability that sampling is allocated to areas likely to be contaminated.