JD11/19/10 slightly modified by KLC

Project Title: Development of Methods to Evaluate Effectiveness of Restoring Aquatic Organism Passage

Principals: Jason Dunham, Michael Adams, Susan Haig, and Mark Miller, USGS Forest and Rangeland Ecosystem Science Center, 3200 SW Jefferson Way, Corvallis, OR 97331; 541-750-7397, ; Bruce Hansen and Dede Olson USFS Pacific Northwest Research Station, 3200 SW Jefferson Way, Corvallis, OR 97331; 541-750-7311, ; Steve Lanigan, Module Leader, Aquatic and Riparian Effectiveness Monitoring Program, USDA Forest Service, Pacific NW Region, Resource Planning and Monitoring, POB 3623

333 SW First Ave., Portland, OR97208; 503.808.2261;

Collaborators: Kim Clarkin, USFS San Dimas Technology and Development Center; Sandra Wilson-Musser, USFS R6 Engineer; Jim Capurso, USFS R6 Fisheries Program Leader; Guillermo Giannico, Oregon State University; Fisheries; Dan Shively, Fish Passage and Habitat Partnerships Coordinator, U.S. Fish and Wildlife Service, Region 1 (OR, WA, ID, and Pacific Islands).

Background

Restoration of passage for fish and other aquatic organisms at stream-road crossings represents a major challenge, involving a potential investment of hundreds of millions of dollars nationally, and even within individual states (e.g., Oregon and Washington; GAO 2001). In recent years, passage at hundreds of stream-road crossings has been restored, primarily by replacing barrier road culverts with bridges or stream simulation culverts designed to pass all species and all life stages of aquatic life and simulate natural hydro-geomorphic processes (Stream Simulation Group, 2008). Within Pacific Northwest US alone (Oregon and Washington) millions of dollars are spent annually on projects to restore passage at stream-road crossings.

The case of the Pacific Northwest represents two questions of national significance: “Are current design standards and construction methods for stream simulation culverts adequately re-establishing passage for aquatic biota?;” and “How do we monitor and evaluate effectiveness of passage restoration?” Here we request support to contribute to a major study of the effectiveness of aquatic organism passage impairment and restoration in the Pacific Northwest region (PNW). This project has regional significance because the results will help managers in the PNW evaluate the effectiveness of their investment in passage restoration. Outside of the region, results will have national significance because results of this work can provide a model for application across the nation. The PNW is uniquely poised for this effort, due to the existence of a strong infrastructure and working relationships among diverse parties representing both government and university affiliated scientists and managers. In large part this infrastructure has been supported by strong public interest in maintaining water quality and ecosystem processes that support culturally and economically important salmon and trout in the region.

Most restoration projects in the PNW are motivated by concerns over fish passage, primarily for sensitive salmon and trout species. The benefits of culvert restoration for salmon and trout have been assessed in a small number of cases by direct observation of fish moving through replaced culverts or using them as habitat, or tracking marked fish to determine if they can pass through restored road crossings. In some cases fish were shown to move through replaced culverts, but in others, they did not move, for reasons that were not completely understood (Dave Heller, U.S. Forest Service [Retired], personal communication). Success of passage restoration for other aquatic biota and for juvenile salmonids is less well-known (Sagar et al. 2007). Furthermore, little is known about how passage restoration might facilitate invasion by a variety of nonnative species (Fausch et al. 2009), a growing threat in freshwaters of the region (Sanderson et al. 2008). Thus, while great progress toward restoring passage has been made, we still understand little about the overall effectiveness of passage restoration with respect to achieving agency goals such as maintaining viable populations of endemic species, and the diversity of aquatic life.

Our goal in this project is to improve our understanding of the effectiveness of programs for improving aquatic organism passage, by addressing the following questions:

  1. Freedom of movement. “How do individual organisms of selected species move through restored and un-restored stream-road crossings in relation to natural streams?” “How” as used here can refer to probability of crossing through structures, rates of movement, timing of movements, or gene flow as inferred from genetic analysis. These responses relate in short to “freedom of movement” of different organisms and life stages through natural streams and different types of road crossings. Impairment of movement is a key concern in restoring passage for aquatic biota.
  2. Stream simulation. “How do un-restored crossings compare to restored crossings or natural stream channels in terms of hydrological and geomorphic processes?” In other words, are current procedures for restoration of road crossings producing structures that simulate natural physical processes in streams? This question would be addressed by comparing natural channels to restored and un-restored crossings with respect to simple geomorphic features (channel morphology).
  3. Population consequences. Species may be able to move across road crossings, but is it enough? This question points to the population consequences for biota relative to the freedom of movement for different species and life stages. “Do populations upstream of stream-road crossings have different characteristics (species presence, demographic, and genetic characteristics) relative to those without stream-road crossings?” “Does passage restoration also restore population processes as indicated by these characteristics?”
  4. Protocols. “Given results of this work in the PNW, what protocols, guidelines, and recommendations can be developed for applications across the nation?” The results of this work will directly benefit managers in the PNW by allowing them to evaluate the effectiveness of stream-road crossing restoration, but will also serve a national interest in evaluating effectiveness for a much broader diversity of biota and stream habitats.

Addressing these questions across Oregon and Washington will provide an important broad-scale evaluation that can be used to quantify the success of passage restoration and evaluate trade-offs in terms of economic and ecological costs and benefits of projects in different contexts (e.g., Peterson et al. 2008, Neville et al. 2009). This project was initially conceived to evaluate crossings on Federal lands; however, a watershed approach is critical when dealing with movement barriers, and the project would greatly benefit from additional funding from other agencies so that State and private lands could also be included.

To recap, this project has direct relevance and inference to the Pacific Northwest, but more importantly represents a model of collaborative effort between scientists and managers that can serve as a model nationwide. Results from this and other similar projects are essential for adaptive management of local and national passage restoration programs.

Methods

Task 1. Develop a database of culvert barriers, aquatic organism passage restoration projects, and related geographic, biotic, road, and historical information on public and private lands in selected Oregon and Washington watersheds.

We have initiated this effort in collaboration with Forest Service, Pacific Northwest Region (Region 6). Geo-referenced culverts and passage restoration projects have been identified on six forests to date: the Siuslaw, Umpqua, Wallowa-Whitman, Malheur, Umatilla and Fremont-Winema National Forests, with information scheduled to arrive from other Forests in FY 2010 and 2011 (K. MacDonald and S. Wilson-Musser, U.S. Forest Service, personal communication). We will work with the Region’s effort to geo-reference all projects throughout Region 6 on forest lands in Fiscal Year 2011. We will also work with the Oregon-Washington BLM State Office and USFWS to include their information in this database. Oregon Department of Fisheries and Wildlife, NOAA andUSDI Bureau of Reclamation culvert data sets will provide data for other public and private crossings.Priority basins for database development and field implementation will be on the Siuslaw National Forest (e.g., Nestucca, Yaquina, Alsea, Siuslaw, and smaller coastal basins) and John DayRiver (Umatilla, Malheur, and Wallow-WhitmanNational Forests). These forests and watersheds represent a very large portion of the Pacific Northwest Region, including vast tracts of BLM lands on the east and west sides of the CascadeMountain crest. Sampling in 2011 will be conducted to provide representative information from remaining landtypes in the region.

Task 2. Applydatabase to develop a retrospective study design for evaluating effectiveness of aquatic organism passage restoration in streams across the region.

The database will provide us with a “population of pipes” or culverts from which a sample can be taken to draw inference about culvert and passage conditions across the larger region. These locations will be populated with data relating to landscape conditions (stream size, elevation, valley slopes, stream gradient, drainage size, latitude, longitude, type of disturbance, time since disturbance, distance upstream of mainstem river), biotic conditions (species suspected or known to be present), and characteristics of the crossings themselves (e.g., degree of passage impairment as determined by ground measurements; e.g., streambed and particle size within culverts, K. Clarkin, personal communication; Harris 2005). This initial “filtering” of sites is meant to provide the foundation for a robust study design. More detailed on-the-ground measurements will supplement this information in locations selected for field sampling.

We will employ a retrospective study design, rather than a “before-after” comparison so we can conduct work quickly in a broad geographic area, and cover a wide range of conditions that represent diverse landscape and biotic conditions in the region. Our work will focus on comparisons of biological and selected physical responses (see Appendix 2). Measurements of these responses would be recorded above and below existing (un-restored) culvert barriers, sites with restored passage, and reference sites lacking a culvert or other structure that could impede passage (reference sites). This will require us to select sites with adequate lengths of accessible upstream and downstream reaches to sample.

Our goal would be to sample a minimum of 30 areas with specific locations representing each of three categories of stream crossings (un-restored culvert, restored, and reference section without culvert). This totals to 90 specific locations. We understand these categories of stream crossings may not be clear. For example, in reality, culverts as well as natural locations represent gradients of “passability” for different species and life stages. To account for this issue, we will measure physical characteristics of culverts and natural stream reaches during our field surveys (Appendix 2) so we can statistically adjust for these effects (e.g., physical habitat covariates).

It is also possible that we may adapt our approach to compare “above-below” locations for stream-road crossings to “above only.” In many cases culverts cross tributary streams that directly flow into a larger mainstem stream – thus a downstream reach is very limited in size or practically absent. In this case it is only possible to compare upstream locations. We have successfully employed this approach before (Neville et al. 2009). One consequence of this situation is that mark-recapture of individuals to evaluate movement is constrained by lack of adequate habitat below a culvert (e.g., a small number of individuals to mark in a stream below a culvert).

A final potential situation is the possibility that few culvert barriers remain for some stream types (e.g., larger fish-bearing channels). Thus there are very few “un-restored culvert” locations. If this is the case, we will compare two basic categories of stream reaches: natural streams and restored stream crossings. This approach also allows us to account for the influences of key factors such as stream size, time since restoration, design characteristics, and local habitat potential to support key species. In studies of passage impairment in other systems, time of isolation (since a barrier was installed) was a key variable (Morita et al. 2009), thus we expect time since restoration to be important.

Our database will allow us to troubleshoot these issues and adapt our study approach to clearly address the questions posed herein. The main constraint determining our design will be locations of stream crossings that have been restored and have followed AOP prescriptions.

Task 3. Implement effectiveness monitoring at selected sites to evaluate alternative measures of organism responses within the study design – based on forthcoming summary of methods for evaluating the effectiveness of aquatic organism passage (FS-WO contract with USGS; Dunham PI).

Our initial focus will be to determine culvert passage responses based onfour classes of methods:1) mark and recapture of individuals to assess movement through stream-road crossings or reference sites; 2) demography as indicated by the presence of different life stages for selected species; 3) occupancy modeling to estimate probabilities of detection and presence for selected species and/or life stages; and 4) genetic analyses of connectivity across the road crossing (e.g., Neville et al. 2009). Each of these methods provides unique perspectives on the responses of biota to passage impairment or restoration.

Mark and recapture and demography are the two methods implemented most commonly in field applications. These approaches can indicate effectiveness, but also come with some critical assumptions. For example, if mark and recapture fails to detect movement at stream crossings, lack of movement could be due to a variety of factors unrelated to “passability” such as insufficient numbers of marked individuals (low relocation or recapture probability) or insufficient frequency, duration or timing of monitoring to detect movement.

Demography is usually considered in terms of total abundance of individuals or life stages. Changes in demography associated with changes in passage should be expected, as abundance is a product of movement (emigration and immigration) as well as other processes (births and deaths). The expected direction of change in abundance can be difficult to predict, however, because 1) these four processes (birth, death, immigration, emigration) are operating simultaneously and can be impossible to isolate, and 2) influences of migratory life history variability. The latter is important because restoration of passage may lead to increased rates of emigration and lower abundance, or alternatively increases in immigration and increases in density. These effects depend on the tendency of fish to migrate, which can vary dramatically for many species.

Occupancy modeling is useful for species that show variability in occupancy in relation to passage conditions. Unlike movement, occupancy can be determined in a single visit. Unlike demography, sampling for occupancy is much less intensive and expensive. Occupancy does not provide detailed information, but can be an extremely powerful and cost-effective method for analyzing multiple species and population impacts of passage restoration and impairment.

Genetic analyses, like occupancy, require only a single site visit for sample collections. Indirect inferences about individual movement and effective population sizes from genetic markers are widely applied and can provide insights that are impossible to reveal through studies of individual movement, demography, or occupancy (e.g., Neville et al. 2006). Genetic analyses can also detect influences such as cryptic hybridization (Neville et al. 2009) that other methods cannot detect. Costs of genetic analyses are comparable to other methods, but applications are limited to species with enough individuals to provide sufficient sample sizes (e.g., 30-50 individuals).

In conclusion, these very brief examples illustrate the complementary nature of each approach to evaluating biotic responses to passage or connectivity. Each method offers potentially unique insights into understanding biotic responses, and newer approaches (occupancy and genetics) offer some distinct advantages that can be evaluated in the study proposed herein. In addition to biotic responses, selected physical variables will also be measured at sites to be used in evaluation of physical effects of stream crossings as well as for covariates in analyses of species’ responses (Appendix 2).

Because mark-recapture and genetic analyses are more intensive than other methods (occupancy, demography, physical habitat) described herein, we will evaluate these responses for a selected subset of species and sites. We will conduct an evaluation of occupancy, demography, and physical responses at all locations. The species we chose to study in more detail for mark-recapture and genetics will be determined by management priority, species prevalence, and expected species abundance – the latter to be determined in part by the database we are assembling.

In general, we would like to more intensively study at least one salmon or trout species that is widespread and commonly impacted by stream-road crossings. On forests where we have geo-referenced culvert information, for example, this would be represented by coastal cutthroat trout Oncorhynchusclarkii (e.g., Siuslaw National Forest in 2011) or redbandO. mykiss and bull trout Salvelinusconfluentus (e.g., John Day River in 2011). We would also like to select a non-salmonid species, potentially a stream-living amphibian, such as the coastal giant salamander, Dicamptodontenebrosus or coastal tailed frog Ascaphustruei, and potentially others (e.g., sculpin or crayfish). The non-salmonid species will be selected based on their different modes of movement and expected dispersal abilities. For example, sculpins (Cottus, sp.) are small-bodied species that lack a swim bladder. We would expect these species to be less likely to pass through culverts than stronger swimmers, such as salmon and trout (LeMoine 2007).

Task 4. Evaluate alternative methods for monitoring effectiveness of aquatic organism passage based on study results. See appendix 2 for more detailed discussion of alternative methods.

Based on results of our work, we will be able to evaluate the efficacy of the four classes of approaches for evaluating the responses of biota to passage impairment or restoration. Our evaluation will focus on the questions posed in the beginning of this proposal relating to individual and population responses of native species, as well as invasion by nonnative species (Neville et al. 2009; Fausch et al. 2009). Which of these methods provides the best indication of the degree of passage impairment or restoration and its effects? How are the answers complementary, or perhaps even contradictory? Which are most useful in terms of practical applications for land management and regulatory agencies?