May 2010 Nislow proposal

A hierarchical approach to assess the effectiveness of aquatic organism passage

Principal Investigator - Andrew Whiteley – Assistant Professor, Department of Natural Resources Conservation, UMASS and Conservation Geneticist, USDA Forest Service Northern Research Station, Amherst, MA 01003

Co – Principal Investigators - Benjamin H. Letcher, Ecology Section Leader, USGS-BRD CAFRC, Turners Falls, MA; Keith H. Nislow, Fish and Wildlife Habitat Relationships Team Leader, USDA Forest Service Northern Research Station, Amherst, MA 01003

Cooperators – Mark Hudy, National Aquatic Ecologist, USDA Forest Service, Harrisonburg, VA; Steven Roy and Dan McKinley, Green Mountain National Forest, Rutland, VT; Mark Prout, White Mountain National Forest, Campton, NH; Michael Owen, Monongahela National Forest; Nick Schmal, USFS Region 9; John Magee, New Hampshire Fish and Game Department, Concord, NH; Kim Lutz, The Nature Conservancy – Connecticut River Program, Northampton, MA, Colin Apse, TNC Eastern Regional Office

Background and Justification

Individual passage provides definitive evidence that the removal or replacement of a barrier has been effective in achieving AOP goals. This evidence can be obtained in a number of different ways. Two indirect methods hold some promise. In situations where species are only present below barriers, detection of occupancy above the barrier after its removal or modification can be taken as evidence for successful passage. However, this approach is limited to specific situations (species present below barriers but not above) and the power of the approach is strongly limited by probability of detection. For example, low probabilities of detectioncan result in ‘false positives’ where detection of occupancy following barrier removal is falsely attributed to improved passage, but in reality the species had been already present, but not detected. In addition, in situations where species are present above and below a potential barrier, differences in abundance may indicate lack of effective aquatic organism passage, and elimination or reduction of these above/below differences may signal re-establishment of passage following barrier removal. However, natural variation in the abundance makes the detection of the barrier ‘signal’ challenging, requiring robust datasets and good estimates of variation in this relationship. Both these approaches have the advantage of being amenable to standard fish sampling techniques of the kind typically used by management agencies.

In contrast to these indirect measures of passage, direct evidence for individual passage can be accomplished via tagging studies. However, because movement is highly episodic and only a small fraction of the population may move over any given time interval, traditional tagging studies frequently have high probabilities of not detecting movement when it actually occurs, while still incurring substantial time and personnel costs (for example requiring at least two visits to each monitored site, at least four for a before-after replacement design). New technology, particularly the development of passive integrated transponder (PIT) tags and stationary readers can increase probability of detection and reduce the number of sample bouts, but equipment costs and requirements make these techniques impractical for most standard monitoring situations. Genetics techniques may provide a solution. Standard genetic approaches that assess differences among populations lack resolution to detect individual movement at an ecological time scale in many situations. However, we have recently developed genetic methods that appear (from simulations based on data from wild populations) to be able to detect movements with a high level of reliability and on an ecological time scale. This approach takes advantage of the fact that most stream-dwelling species have point distributions of reproduction (for example salmonid and lamprey redds, sculpin and stream salamander nests) and that dispersal of siblings from these point sources can be quantified from a genetic survey and subsequent assignment of individuals (parents and offspring) to family groups. The feasibility of family assignments has been tested in the field at two long-term study sites, which have demonstrated that standard electrofishing surveys, such as currently undertaken by management personnel, are sufficient to assign a large proportion of the individuals in a population. Further, given the straightforward collection of samples, decreasing costs of genetic sample analyses, and institutional support for data analysis and reporting (via the joint USFS/USGS/UMASS conservation genetics laboratory in Amherst), this approach promises to provide management a cost-effective method to monitor effectiveness in the context of aquatic organism passage. Further, because these genetic analyses can eventually be used to provide estimates of effective population size and allelic diversity, they have the potential to contribute to the ultimate goal of monitoring effects on population resilience and viability.

Methods

We propose to move forward with the development and application of a hierarchical approach to apply this range of techniques in a combined field and modeling simulation study incorporating planned culvert replacements in the eastern region. We will use simulation results to design optimal sampling protocols (sample area extent, spatial distribution and minimum number of genetic samples) and focus on fish species representing three swimming guilds:

Small, demersal fishes (darters, sculpins) – e.g. Slimy sculpin (Cottus cognatus)

Small midwater species (small minnows and dace, YOY salmonids) – e.g. Blacknose dace (Rhinichthys atratulus) YOY Brook trout

Large midwater species (overyearling salmonids) – e.g. Overyearling Brook trout

These guilds are common to most small stream fish communities, which should allow these results to be widely applicable across the country. Genetic markers (microsatellites) for species that are common representatives of these guilds (or their close relatives) have been developed. Genetic analyses, sibship and parentage assignment, and identification of probable spawning locations are described in detail in Hudy et al. (in press).

Scope of Workand Deliverables

Scope of Work

The study will take place in sites located within three national forests (Green Mountain National Forest, White Mountain National Forest, Monongahela National Forest) one state forest (Nash Stream State Forest, NH), and one location on private land (West Brook, MA).

In each site, surveys will be conducted at passable and non-passable culverts (25-30 total per NF)will be conducted. To sample fish in each survey, we will generally follow the design employed in a study of the Greenbriar watershed in the Monongahela National Forest (Nislow et al. in preparation), where abundance of all species will be measured at multiple plots upstream and downstream of predicted passable and impassable road crossings for occupancy and demographic analyses (see enclosed draft manuscript). Probability that a road crossing currently is a barrier will be determined by the Coffman coarse filter score – all candidate crossings have already been surveyed using this method. In a subset of these sites (4-6 per forest) including all sites where culvert replacements are currently being planned we will collect genetic samples and run sibship and parentage analyses to apply the sibship genetics method for assessing passage effectiveness. Finally, in two sites (Nash Stream and West Brook), in addition to occupancy, demographic, and genetics analysis, we will use pit-tags and stationary antennas to monitor passage.

We already have two years of surveys on 31crossings in the MNF. On this forest we will therefore only sample at the subset of sites where we will collect genetic samples and conducts sibship analysis. A preliminary set of crossings on the GMNF and WMNF has been identified (see enclosure) Sites with planned culvert replacements will be sampled in both Year 1 (pre-replacement) and Year 2 (post-replacement); all other sites will be sampled in either Year 1 or Year 2.

To analyze field data on occupancy and abundance we will use mixed linear models (following Nislow et al. in prep). The primary objective will be to determine whether:

1)There an interaction between location (upstream vs. downstream) and predicted passage such that abundance and occupancy differences are more likely to be observed when culverts are predicted to be barriers

2)Upstream and downstream differences in abundance and occupancy when culverts are predicted to be barriers will not be present after barriers are removed

To analyze the genetics data we will use standard laboratory procedures to extract DNA and genotype individual fish, then use software developed in our laboratory to assign individuals to sibship and family groups. Concordant with the field study, we will continue to conduct simulations based on existing data, and incorporating data gathered in the context of this study which will further refine our estimates of probability of dispersal detection under alternative scenarios.

Budget and Justification

The project budget has three main items:

Full-time postdoctoral scientist = $110K

The postdoctoral scientist, advised by the co-p.i.s, will be the point person for the project. The postdoc will coordinate all field activities, manage all the data, and take the lead on models and simulations with direction and training from the p.i. and co-p.i.s. They will also be responsible for coordination with local and regional managers, and along with the co-p.i.s will coordinate with the effectiveness monitoring working group on the progress and the outcomes of the project on a regular basis. Also, working with the co-p.i.s, they will prepare reports and manuscripts.

Full-time technical assistant = $70K

The technical assistant, supervised directly by the postdoc, will assist in all field and laboratory work.

Laboratory and field supplies = $60K

The majority of the supply budget will go to the purchase of pit tags, genetics supplies and standard sampling consumable supplies.

Travel (to and from field sites) = $10K

Total Budget = $250K

This budget request will be substantially leveraged via support for P.I. salaries, equipment, and substantial investment in software and analytical tools which are uniquely positioned to accomplish our objectives.

Deliverables

Results of the study will be published in appropriate scientific journals, and communicated at relevant regional and national meetings. In addition, a synthesis of these results will be used to develop a formalprotocol for effectiveness monitoring which will be summarized in a USFS General Technical Report.