Visual sensitivity of deepwater fishes in Lake Superior.

The predator-prey interactions in the offshore food web of Lake Superior have been well documented, but the sensory systems mediating these interactions remain unknown. The deepwater sculpin, (Myoxocephalus thompsoni), siscowet (Salvelinus namaycush siscowet), and kiyi (Coregonus kiyi) inhabit low light level environments. To investigate the potential role of vision in predator-prey interactions, electroretinography was used to determine visual sensitivity for each species. Spectral sensitivity curves revealed peak sensitivity at 525 nm for each species which closely corresponds to the prevalent downwelling light spectrum at depth. To determine if sufficient light was available to mediate predator-prey interactions, visual sensitivity was correlated with the intensity of downwelling light in Lake Superior to construct visual depth profiles for each species. Sufficient daytime irradiance exists for visual interactions to approximately 325 m for siscowet and kiyi and 355 m for the deepwater sculpin during summer months. Under full moon conditions, sufficient irradiance exists to elicit ERG response to light available at approximately 30 m for the siscowet and kiyi and 45 m for the deepwater sculpin. Visual interactions are therefore possible at the depths and times when these organisms overlap in the water column indicating that vision may play a far greater role at depth in deep freshwater lakes than had been previously documented.

Food resource effects on diel movements and body size of cisco in north-temperate lakes.

The movement patterns and body size of fishes are influenced by a host of physical and biological conditions, including temperature and oxygen, prey densities and foraging potential, growth optimization, and predation risk. Our objectives were to (1) investigate variability in vertical movement patterns of cisco (Coregonus artedi) in a variety of inland lakes using hydroacoustics, (2) explore the causal mechanisms influencing movements through the use of temperature/oxygen, foraging, growth, and predation risk models, and (3) examine factors that may contribute to variations in cisco body size by considering all available information. Our results show that cisco vertical movements vary substantially, with different populations performing normal diel vertical migrations (DVM), no DVM, and reverse DVM in lakes throughout Minnesota and northern Wisconsin, USA. Cisco populations with the smallest body size were found in lakes with lower zooplankton densities. These smaller fish showed movements to areas of highest foraging or growth potential during the day and night, despite moving out of preferred temperature and oxygen conditions and into areas of highest predation risk. In lakes with higher zooplankton densities, cisco grew larger and had movements more consistent with behavioral thermoregulation and predator avoidance, while remaining in areas with less than maximum foraging and growth potential. Furthermore, the composition of potential prey items present in each lake was also important. Cisco that performed reverse DVM consumed mostly copepods and cladocerans, while cisco that exhibited normal DVM or no migration consumed proportionally more macro-zooplankton species. Overall, our results show previously undocumented variation in migration patterns of a fish species, the mechanisms underlying those movements, and the potential impact on their growth potential.

Forecasting the spread of invasive rainbow smelt in the Laurentian Great Lakes region of North America.

Rainbow smelt (Osmerus mordax) have invaded many North American lakes, often resulting in the extirpation of native fish populations. Yet, their invasion is incipient and provides the rationale for identifying ecosystems likely to be invaded and where management and prevention efforts should be focused. To predict smelt presence and absence, we constructed a classification-tree model based on habitat data from 354 lakes in the native range for smelt in southern Maine. Maximum lake depth, lake area, and Secchi depth (surrogate measure of lake productivity) were the most important predictors. We then used our model to identify lakes vulnerable to invasion in three regions outside the smelt's native range: northern Maine (52 of 244 lakes in the non-native range), Ontario (4447 of 8110), and Wisconsin (553 of 5164). We further identified a subset of lakes with a strong potential for impact (potential-impact lakes) based on the presence of fish species that are affected by rainbow smelt. Ninety-four percent of vulnerable lakes in the non-native range in Maine are also potential-impact lakes, as are 94% and 58% of Ontario and Wisconsin's vulnerable lakes, respectively. Our modeling approach can be applied to other invaders and regions to identify invasion-prone ecosystems, thus aiding in the management of invasive species and the efficient allocation of invasive species mitigation and prevention resources.