**The title, authors, and abstract for this completion report are provided below.  For a copy of the completion report, please contact the GLFC via e-mail or via telephone at 734-662-3209**

Quantitative Tools to Predict Sea Lamprey Production Based on Habitat: Prioritizing Dam Removal and Control Decisions

S. Adlerstein¹ and E. Silverman ²

¹University of Michigan, School of Natural Resources and Environment, 3010 Dana Bldg., Ann Arbor, 48109

²USGFWS, Division of Migratory Bird Management, 11510 American Holly Drive, Laurel, MD 20708

Abstract

We developed models to understand distributions of sea lamprey larvae (total density, as well as <50 mm – considered young of the year – 50-100 mm, and >100 mm) and sea lamprey larval habitat in 29 Michigan and 20 Huron tributaries located in the Michigan Lower Peninsula. We compiled and evaluated larval and habitat data collected from 1999 to 2004 during Quantitative Larval Assessment Surveys (QAS) extracted from the GLFC Empirical Stream Treatment Ranking system, and information from databases maintained by the U.S. Fish and Wildlife Service, the agency responsible for conducting lamprey surveys. We also used data from the Michigan Department of Natural Resources Landscape-based Ecological Classification System for River Valley Segment Classification (MI-DNR-VSEC) database, the Michigan Rivers Inventory database, the MI-DNR Michigan Dam database and Digital Maps combined with US Geological Survey Digital Line Graph. Our analyses included investigation of relationships between sea lamprey larval density and habitat, and between larval habitat (as defined by QAS) and landscape characteristics, considering variability at several spatial scales. We used GIS to link data based on geographic location and to generate explanatory variables. We developed models for larval densities by stream and basin and for proportion of suitable and preferred larval habitat using mixed linear and modern regression techniques and used explanatory variables measured during QAS, taken from other databases, or derived from models. Specific objectives were to identify factors that contributed to the variation of larval densities and habitat in streams and develop relationships between larval densities and habitat characteristics. Mean densities of total larvae in high quality habitat, measured by QAS surveys conducted at least three years after lampricide treatment in the corresponding reach, were around 10 and 13 larvae per 15m2 plot in Michigan and Huron tributaries, respectively, and of young of the year were around 2 and 5 per plot, for all years treatment was not conducted in the same reach. Counts in plots reached over 100 larvae. Our stream-specific models for densities of small, medium, large and total larvae explained about 10 to 90 % of the variation within tributaries. About half of the models explained more than 50% of the density variation and with the best predictions for Michigan tributaries. Basin models explained 24 to 47%. Best predictors were habitat type, years since lampricide treatment, valley segment hydrology, and distances to upstream spawning habitat and to the river mouth. Densities in preferred habitat were on average double those in acceptable habitat and generally higher in groundwater than in runoff stream segments, depending on spatial arrangement. Densities were higher close to spawning habitat and varied with distance to the river mouth depending on the stream and size of the larvae. The strength and sign of relationship with spawning habitat varied among streams and depended on larval size; for example, the effect was stronger for young of the year larvae. Densities typically decreased with stream depth, suggesting an effect from sampling gear performance. Acceptable sea lamprey larval habitat made up a large proportion of the surveyed streams and was more prevalent in Michigan streams, with an average of 80% and 60% for Michigan and Huron, respectively; about 20% of sampled habitat was classified as preferred in both systems. Our habitat models explained 43% of the variability in larval sea lamprey acceptable habitat and 28% in preferred habitat and indicated that the proportion of preferred quality habitat for lamprey larvae increased with decreasing estimates of particle size (a function of stream power) and with increasing percentage of the channel bank covered by forest. We estimated that 40% of the currently inaccessible VSECs (damned) have a greater than 50% probability of providing suitable habitat for sea lamprey larvae, should dams be removed. Our models can be applied to predict lamprey productivity potential in tributaries to Lakes Michigan and Huron in streams where no QAS have been conducted, such as above impermeable dams and provide input into sea lamprey assessment methods. Our results are limited by the data available, which came from streams where QAS were performed, i.e, streams with high larval densities that were often treated with lampricide. Random samples of habitat and larval densities are necessary for a more refined analysis.