**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**

Using stable isotopes to assess potential sea lamprey damage to Lake Superior fishes

by:

Chris J. Harvey¹, Mark P. Ebener², Carolyn K. White¹, Justin M. Fox³, Timothy E. Essington⁴, Charles P. Madenjian⁵, James F. Kitchell³, and Phillip S. Levin¹

¹Northwest Fisheries Science Center, NOAA Fisheries, 2725 Montlake Blvd. E., Seattle, WA 98112

²Chippewa Ottawa Resource Authority, 179 W. Three Mile Rd., Sault Sainte Marie, MI 49783

³Center for Limnology, University of Wisconsin, 680 N. Park St., Madison, WI, 53706-1482

⁴School of Aquatic and Fishery Sciences, University of Washington, Box 355020, Seattle, WA 98195-5020

⁵Great Lakes Science Center, USGS, 1451 Green Rd., Ann Arbor, MI 48105-2807

ABSTRACT:

Assessing the impact of sea lamprey (Petromyzon marinus) on the fish community of the Great Lakes requires understanding the spatial variability and ontogeny of their diets. This information would reveal which host species they parasitize in space and time, thus allowing us to estimate the likelihood of host mortality and, hence, impact to the community. We analyzed carbon and nitrogen stable isotope ratios (d¹³C and d¹⁵N) and conducted bioenergetics and stable isotope dynamic modeling in order to estimate sea lamprey diets throughout Lake Superior. The initial step was to experimentally validate the trophic fractionation of stable isotope ratios between host and sea lamprey. We allowed newly metamorphosed sea lamprey to feed ad libitum in three host treatments—lake trout (Salvelinus namaycush), lake whitefish (Coregonus clupeaformis) or white sucker (Catostomus commersoni)—and measured trophic fractionation (= the change in stable isotope ratios associated with assimilation of diet into a consumer’s tissues) from host blood to sea lamprey muscle. Fractionation of d¹⁵N ≈ 0‰, and was -1 to -3‰ for d¹³C; these values were significantly lower than the general literature would predict for either isotope ratio. The mechanism by which these lower-than-expected fractionation rates arose remains unclear. Further complicating interpretation, when we attempted to apply the experimental fractionation rates to our field studies, the results were untenable and we were forced to use fractionation values from the literature throughout the field and modeling portions of the study. In the field portion of the study, we analyzed stable isotope ratios of sea lamprey muscle and host blood from six regions of Lake Superior. Post-larval sea lamprey, spawning sea lamprey and hosts were collected in agency monitoring programs, and parasitic sea lamprey were captured by commercial fisheries through a bounty program. Stable isotope analysis implied that sea lamprey were feeding primarily on upper trophic level species, including lean lake trout, siscowet lake trout (S. namaycush namaycush) and burbot (Lota lota), which is generally in agreement with wounding data collected in the previous two decades. However, in Ontario waters (particularly Black Bay), sea lamprey occupied a lower trophic level, apparently feeding on hosts such as lake whitefish and suckers. In the first modeling portion of the study, we used three techniques to derive regional, quantitative estimates of sea lamprey diets from the stable isotope data. The first technique is a growth-based stable isotope (GBSI) model that was used to estimate sea lamprey diet ontogeny by fitting model output to observed relationships between sea lamprey mass and stable isotope ratios. The second technique is the widely used IsoSource model, a Monte Carlo tool that algebraically generates plausible diets given the stable isotope signatures of consumer and prey. The third technique, developed and presented here for the first time, uses Bayesian methods and likelihood principles to estimate diets (means and variances) from stable isotope ratios. We call this method B-SIDE (Bayesian Stable Isotope-based Diet Estimator). There was considerable agreement among the three methods: all three found that apex predators (lean lake trout, siscowet lake trout, burbot) comprised 60-90% of sea lamprey diet in most areas. In Black Bay, however, all methods estimated that lake whitefish were the most important host. Each method found that suckers were more important in sea lamprey diets in western Lake Superior than in eastern waters. The B-SIDE method produced the highest (and most statistically robust) variance estimates, and the IsoSource method the lowest. The temporal GBSI method indicates an ontogenetic shift from smaller coregonines to larger hosts. In the second modeling portion of the study, we integrated regional sea lamprey and host stable isotope data and regional sea lamprey growth data into an individual-based model (IBM) of sea lamprey feeding in order to generate host-specific estimates of sea lamprey blood consumption and sea lamprey-induced mortality. The IBM included a bioenergetics and stable isotope dynamic routine, a host attachment routine and a host detachment routine. The IBM was fitted to seasonal sea lamprey growth data by least squares, and attachment probabilities to six host types (lean lake trout, siscowets, lake herring, lake whitefish, Chinook salmon and longnose sucker) were adjusted to fit sea lamprey stable isotope signatures. In most regions, lean and siscowet lake trout were the most important source of blood consumption for sea lamprey, but far more lake herring and lake whitefish were killed because of unfavorable lamprey:host size ratios. Fewer lean lake trout deaths resulted than would have been expected from a model in which lean lake trout were the only hosts; this implies that a diverse host community offers some buffering against sea lamprey mortality, and siscowet lake trout may be especially effective buffers because their relatively large size results in lower direct mortality. In Black Bay, the IBM predicted that all blood consumption and mortality were allocated between lake herring and lake whitefish, with considerable mortality. The multi-host IBM outputs should be interpreted cautiously because it remains in a developmental stage until host size preferences and host detachment probabilities can be established for hosts other than lake trout, and until a better fitting routine can be implemented that takes advantage of stable isotope data. Nonetheless, the model sheds light on the effects of sea lamprey on “alternative hosts” of high economic value (i.e., coregonines and salmon) and of lesser economic value but high ecological importance as buffer species (siscowet lake trout, suckers).