Larval Walleye Research Articles

Bayesian hierarchical modeling of larval walleye (Sander vitreus) abundance and mortality: Accounting for spatial and temporal variability on a large river

Larval fish are extremely variable in space and time while sampling of populations is generally restricted and incomplete. However, estimates of abundance and mortality are important for understanding population dynamics, habitat quality, and anthropogenic impacts. Acknowledging and addressing variability during sampling and data analysis are imperative to producing informative estimates. A combination of spatially and temporally distributed ichthyoplankton sampling and Bayesian hierarchical and state-space modeling was used to partition variance and estimate abundance and mortality of larval walleye (Sander vitreus) in the Maumee River during 2010 and 2011. System variability and degree of sampling coverage have a direct impact on the quality of abundance estimates. Small scale factors (i.e., within site and day-to-day) accounted for the most variation in larval walleye densities, therefore sampling should concentrate on capturing these sources. Bayesian state-space modeling can improve estimates by sharing information through time, properly accounting for uncertainty, and producing probability distribution based estimates. Larval fish are highly variable and difficult to sample; however, the application of Bayesian methods during the data analysis process can lead to improved estimates of abundance and informed management actions.

Modeling the implications of multiple hatching sites for larval dynamics in the resurgent Saginaw Bay walleye population

The early life environment experienced by most larval fish is largely dependent on a combination of hatch site and water currents. Until larvae are able to swim fast enough to overcome currents, they are largely passively transported and have limited control over ambient environmental conditions, including temperature and prey availability. These factors strongly influence growth and survival of larvae, with direct consequences for subsequent recruitment. Early life survival of Saginaw Bay walleye was formerly limited by alewife predation on larvae; but following the collapse of Lake Huron alewives, the walleye population has rebounded and recruitment success may now be influenced by other factors including spawning habitat. We sought to assess the implications of successful hatching at multiple locations in Saginaw Bay, using a hydrodynamics model, particle transport model, and an individual-based bioenergetics model in series. Model results were compared to locations of young larvae collected in Saginaw Bay during 2009–2010. Results suggest that larval growth is strongly influenced by hatch date, driven by seasonal variation in temperature between sites. Larvae hatched at any location could be transported extensively within inner Saginaw Bay before reaching a sufficient size to swim independently of currents, and retention within the productive inner bay varied among years and sites. Our results indicate multiple larval walleye origins in the field, augmenting the continued production from the Saginaw River system. Successful hatching at more locations would serve to buffer walleye recruitment variation through portfolio effects, supporting arguments for more emphasis on diverse spawning habitat management and restoration.

Assessing the condition of walleye pollock Theragra chalcogramma (Pallas) larvae using muscle-based flow cytometric cell cycle analysis

Flow cytometric cell cycle analysis was used to determine the fraction of muscle cells in the S and G2 phases of the cell cycle, which were used as covariates with temperature and standard length, in a laboratory-developed model to assess the physiological condition of wild walleye pollock, Theragra chalcogramma, larvae. The assay was calibrated to the range of temperatures larvae are likely to encounter in the eastern Bering Sea, and it was sensitive to changes in condition within 3 days of starvation. The S and G2 phases of the cell cycle gave an indication of larval walleye pollock condition. Healthy larvae had a larger fraction of cells in the S phase than G2 phase, and unhealthy larvae had a larger fraction of cells in the G2 phase than the S phase. Validation tests showed that the model classified 75% to 83% of the larvae correctly. The assessment of the condition of walleye pollock larvae collected from the southeastern Bering Sea in 2007 indicated that unhealthy larvae were located on the continental shelf (6%), and this may be due in part to the coldest temperatures occurring there and less abundant prey. In the continental slope/ocean basin waters, where prey levels were higher and temperatures warmest, no larvae in unhealthy condition were found.

Gas Bladder Inflation in Walleye Fry Cultured in Turbid Water with and without a Surface Spray

AbstractIt has been demonstrated that a surface spray is needed to obtain a high percentage of gas bladder inflation (GBI) in intensively cultured larval walleyes Sander vitreus fed dry diets and that turbid‐water culture substantially enhances larval survival and growth. It was not previously known whether a surface spray is needed to obtain GBI when larvae are cultured in turbid water, which is produced by intermittently adding clay slurry to the culture tank. The present study was undertaken to test whether turbid water alone would enhance GBI without need for a surface spray. Larval walleyes were cultured with formulated feed from 2 to 28 d posthatch (dph) in tanks containing turbid water (∼55 nephelometric turbidity units) with surface spray (WS tanks) or without surface spray (WOS tanks). Larvae were stocked at a density of 20 fish/L and were given formulated feed every 5 min for 22 h/d. All husbandry and water quality conditions were similar between experimental treatments. Gas bladder inflation was 100% in fish surviving to 28 dph in WS tanks, but GBI was only 3.3% in WOS tanks. Fry survival (50% WS; 34% WOS), final length, and final weight were not significantly different between treatments. The results conclusively show that surface spray is essential for GBI in walleyes.

Density, production, and survival of walleye ( Sander vitreus) eggs in the Muskegon River, Michigan

Walleye ( Sander vitreus) is an important sport fish in the Great Lakes that is experiencing low reproductive success after severe population declines starting in the late 1940s. In the Muskegon River, Michigan, natural reproduction of walleyes remains low and is largely supplemented by stocking. To determine factors influencing walleye reproductive success in the Muskegon River, we estimated walleye egg survival using in situ egg incubators covered with nitex screening (2003–2004) and estimated density and survival of fertilized eggs caught on furnace filter traps across different substrate types (2005–2006). We compared physical habitat suitability for walleye eggs under high and low flow scenarios. Density of walleye eggs was highest in regions of gravel/cobble substrates. Egg survival was higher in egg incubators (24–49.5%) than on furnace filter traps (2.0%), suggesting predation is an important source of walleye egg mortality in the Muskegon River. Cold water temperatures that extended developmental stage durations may also be an important source of egg mortality. The dynamic habitat suitability model predicted low suitability for eggs due to poor temperature and velocity conditions. Despite low egg survival rates, 40 million to 1 billion eggs were estimated to hatch. The low natural reproduction of walleyes in the Muskegon River is likely due to a combination of low walleye egg survival and failure of walleye larvae to reach their nursery grounds in Muskegon Lake.

Low nekton abundance in the western Bering Sea according to trawl survey data and model estimates

Complex trawl surveys were conducted in the upper epipelagic zone of the western Bering Sea and adjacent Pacific waters in the summer and fall seasons of 2002–2006. The abundance of small nekton (micronekton) was estimated using two independent methods: traditional trawling and a mathematical model of selective feeding by fish. According to the trawl data, total micronekton density varied from 1 to 158 (average 40) mg/m3 on the northwestern Bering Sea shelf and from 6 to 151 (37) mg/m3 in deep-water areas of the southwestern Bering Sea and adjacent Pacific waters. According to model calculations, micronekton density was higher—72–193 (141) mg/m3 on the shelf and 78–507 (228) mg/m3 in the deep-water part of the studied area. Both trawl and model data showed that small nekton on the northwestern shelf mostly consisted of larval and juvenile walleye pollock, as well as small fish species, such as capelin and Pacific sand lance. In the deepwater areas, mesopelagic fish and squid (northern lampfish, northern smoothtongue, and boreopacific gonate squid), which migrate to the surface at night, juvenile Atka mackerel, and shortarm gonate squid dominated among micronekton. The advantages and disadvantages of both the trawl and model methods for calculating the abundance of small fish and squid were considered. Comparison of abundance estimates for mass fish species, obtained through trawl and model methods, enabled us to analyze trawl catchability coefficients and propose a more differentiated division of micronekton into size classes than had been done earlier. A function that characterizes the dependence of the catchability coefficient (CC) on body length was offered for juvenile Atka mackerel. This equation can be also used for evaluation of CC for other fishes that have similar size and behavior.

A method for the histochemical localization of digestive enzymes in whole freeze-substituted larval fish embedded in glycol methacrylate

In order to study the localization of digestive enzymes in larval walleye (Sander vitreus vitreus), a novel method of low-temperature processing of whole, unfixed larvae and subsequent embedding in glycol methacrylate (GMA) was developed. Larval walleye aged 2–10 days post hatch were flash-frozen in liquid nitrogen and transferred into pre-chilled acetone for 12 h at a temperature of −25 °C. Acetone was gradually replaced with increasing concentrations of GMA resin monomer over a 6-h period. The polymer (100%) was further allowed to infiltrate the larvae for 36 h. Resin-embedded larvae were transferred to embedding moulds and polymerized overnight at −25 °C. Four micrometre sections were stained to identify either alkaline phosphatase, non-specific esterase, aminopeptidase M or dipeptidyl peptidase IV. All enzymes studied were readily detected and accurately localized, and no enzyme diffusion was observed. Therefore, freeze substitution combined with low-temperature GMA embedding allows the maintenance of excellent tissue morphology and accurate enzyme localization in whole larval walleye specimens from 2 to 10 days post hatch. It is recommended, however, that samples be frozen in pre-cooled fluorocarbon-based liquid coolants in order to assure optimal tissue preservation.

Effect of Turbidity Duration on Culture of Walleye Larvae

AbstractAlthough turbid water is an essential environmental requirement for intensive larviculture of walleyes Sander vitreus on formulated feed, research has not determined the duration of turbidity required to achieve optimal benefits for survival and growth while reducing cannibalism. We conducted a replicated, single‐factor analysis to determine the effect duration of turbidity on these performance indicators at 28 d posthatch (dph). Four treatments, each with three replicate 150‐L tanks, involved culture of larvae in turbid water (84–125 nephelometric turbidity units [NTU]) from 3 dph to 7, 14, 21, or 28 dph. After culture under turbid conditions for the specified period, the fish were then cultured in clear water until 28 dph (i.e., fish in the 28‐dph treatment received no clear‐water culture). At 28 dph, substantial differences were observed in survival and growth among treatments. Fish survival for the 7‐ and 14‐dph treatments (1% and 2%, respectively) was significantly less than that for the 21‐ and 28‐dph treatments (11% and 14%, respectively). There was a significant difference in total length at 28 dph among the 7‐, 14‐, and 21‐dph treatments but not between the 21‐ and 28‐dph treatments. Fish in the 28‐dph treatment were 177% heavier than those in the 14‐dph treatment. A shift in phototaxic response at about 20 dph further reduced the need for turbid‐water culture to prevent clinging behavior in walleye larvae. Cannibalism (2–5%) was observed in all treatments between 7 and 13 dph but not thereafter. The study substantiates benefits of turbid‐water culture for larval walleyes and demonstrates that duration of turbidity should extend through 28 dph to maximize survival and growth but that it is less important for preventing cannibalism after 13 dph and clinging behavior after 20 dph.

Evaluation of Walleye Embryo Survival and Larval Viability after Iodine Treatment

AbstractSecuring health and avoiding viral disease transmission are of paramount importance for stocking and management of walleyes Sander vitreus. Iodine compounds are widely used as nonselective antiviral substances; however, their adverse effects on embryo survival are less known. Our experiments investigated how maternal attributes influence the relationship between toxicity of iodine (expressed as concentrations and duration of treatment) and survival of walleyes after fertilization and during early larval growth. Eleven female and four male walleyes were collected in April 2007 from the Maumee River, Perrysburg, Ohio, and the gametes were transported unfertilized (2 h) to the laboratory. After fertilization and tannic acid treatment, embryos were exposed for 15, 30, and 180 min to iodine concentrations of 0 (control), 50, 100, 200, 400, and 800 mg/L. The calculated survival of 50% of the embryos at the eyed stage corresponded to an iodine concentration of 380 mg/L for the exposure duration of 15 min (i.e., the LD50; linear relationship). An increase in exposure time (to 30 or 180 min) significantly decreased embryo survival. Progeny from individual females varied in sensitivity to iodine treatment, particularly with respect to the high concentration (800 mg/L), but high survival in controls did not correspond to better resistance to iodine treatments. No significant differences were observed in larval walleye size or swim bladder inflation rates among various iodine exposure groups within 15 d after treatment.

Evaluation of Replacement Feeds for the Fry Feed Kyowa for Larval Walleyes

AbstractBecause Fry Feed Kyowa (FFK) was the standard and the only commercially manufactured starter feed suitable for intensive culture of larval walleyes Sander vitreus, it became imperative to evaluate other commercial feeds after FFK importation to the USA was banned to prevent the spread of bovine spongiform encephalopathy. A project was undertaken at the Rathbun Fish Culture Research Facility (Rathbun), Moravia, Iowa, and Iowa State University (ISU), Ames, using a single‐factor experimental design with Lansy and Epac feeds (INVE Aquaculture, Inc., Ogden, Utah) and Gemma feed (Skretting, Vancouver, British Columbia). Walleyes were cultured at both sites from 3 to 28 d posthatch to evaluate the feeds. Performance indicators included presence of feed in gut, gas bladder inflation, survival, total length (mm), weight (mg), growth rate, and incidence of deformities. At both sites, survival was 2.6–3.6 times greater for fish given Gemma than for fish given the INVE feeds; the difference was significant at Rathbun but not at ISU. Length, weight, and growth rate at both sites were also significantly greater for Gemma‐fed walleyes than for INVE‐fed fish. Weight advantage produced by Gemma was 192% at ISU and 534% at Rathbon. Performance measures of fish given Gemma were similar to those of fish fed FFK in previous years.

Estimating abundance of spatially aggregated populations: comparing adaptive sampling with other survey designs

The main goal in estimating population abundance is to maximize its accuracy and precision. This is difficult when the survey area is large and resources are limited. We implemented a feasible adaptive sampling survey applied to an aggregated population in a marine environment and compared its performance with five classical survey designs. Specifically, larval walleye pollock (Theragra chalcogramma) in the Gulf of Alaska was used as an example of a widespread aggregated population. The six sampling designs included (i) adaptive cluster, (ii) simple random, (iii) systematic, (iv) systematic cluster, (v) stratified systematic, and (vi) unequal probability. Of the five different adaptive estimators used for the adaptive cluster design, the modified Hansen–Hurwitz performed best overall. Of the six survey designs, the stratified systematic survey provided the best overall estimator, given there was accurate prior information on which to base the strata. If no prior information was available, a systematic survey was best. A systematic survey using a single random starting point with a simple random estimator performed as well as and sometimes better than a systematic cluster survey with two starting points (clusters). The adaptive cluster survey showed no advantages when compared with these two designs and furthermore presented substantial logistical challenges.

Growth and survival of age-0 walleye (Sander vitreus): interactions among walleye size, prey availability, predation, and abiotic factors

We examined the importance of prey availability, predation, and abiotic factors in determining growth and survival of age-0 walleye (Sander vitreus) across 15 Illinois reservoirs during 7 years. Multiple life stages were examined by stocking walleye at three different size groups: larval (6 mm total length (TL)), small (46 mm TL), and large (100 mm TL). Factors affecting growth and survival of walleye varied depending on walleye size. Growth of small and large walleye increased with benthic invertebrate density. Temperature had a positive effect on larval and small walleye growth but a negative effect on large walleye growth. Prey availability was an important factor for walleye survival across all size groups, whereas temperature affected only larval and large walleye. Juvenile centrarchid density had a negative effect on larval walleye survival, presumably caused by predation. Our best predictive models explained substantial variation in survival for larval (97%), small (57%), and large (83%) walleye. We also explained a high proportion of variation in growth of large (98%), small (55%), and larval (52%) walleye. Our study demonstrates the importance of examining multiple life stages to predict growth and survival and leads to a better understanding of walleye recruitment and recommendations for stocking strategies.

Determining Site-Specific Toxicity of Copper to Daphnids and Fishes in a Brown-Water Ecosystem

ABSTRACT We examined the influence that wetland brown water had on the toxicity of copper to aquatic species in Swamp Creek basin, Wisconsin. Copper was added to Swamp Creek water and to laboratory-prepared reconstituted water and tested for toxicity using daphnids (Ceriodaphnia dubia and Ceriodaphnia reticulate), larval fathead minnows (Pimephales promelas), and larval walleye (Stizostedion vitreum). The toxicities of copper in Swamp Creek water, which contained dissolved organic carbon (DOC), were significantly lower than in water with no DOC. Therefore in successive experiments. using cultured C. dubia as the test organism, we experimentally reduced DOC in Swamp Creek water. We established that increased concentrations of DOC and hardness increased the odds of survival of C. dubia in the presence of potentially lethal concentrations of copper. In the Swamp Creek watershed which contains brown water and in watersheds where DOC concentrations are substantial, DOC, and hardness could be the principal factors in developing site-specific criteria involving copper and perhaps other metals.

Patterns in diet reveal foraging site fidelity of short-tailed shearwaters in the southeastern Bering Sea

The short-tailed shearwater Puffinus tenuirostris is an apex predator in the southeastern Bering Sea ecosystem. During 1997 to 1999, a period of great variability in the Bering Sea, we used a multi-pronged approach to study transfer of carbon and nitrogen to short-tailed shearwaters through analysis of stomach contents of birds collected while foraging, and stable isotope and fatty acid com- position of tissues from shearwaters and their prey. Two conclusions result from these 3 analyses of feeding history. First, short-tailed shearwaters demonstrated localized differences in diet with respect to sampling location and season, indicating that shearwaters feed in discrete locations long enough (several weeks) to reflect regional differences in prey availability. Second, elevated δ 15 N levels (~1 to 2‰) in the liver of shearwaters in the fall of 1997 and 1998 were likely a response to nutritional stress and overturn of nutrients in the tissues of birds or elevated isotope levels in tissues of prey, rather than an increase in the trophic level of the diet. Over the 3 yr period, shearwater diets switched from adult euphausiids Thysanoessa raschii and T. inermis (usually females with spermatophores) taken in spring to an increase in the amount of fish (Pacific sandlance Ammodytes hexapterus), juvenile Gadi- dae fishes and larval walleye pollock Theragra chalcograma particularly in fall (1997 excepted). For each location and year, there was a consistent trend in isotopic values, with an increase of 3 to 5‰ for δ 15 N and 1 to 3‰ for δ 13 C between shearwaters and their primary prey. The resolution of our sam- pling techniques indicates little movement of shearwater flocks among sample locations for periods of up to several weeks.

Modeling Effects of Forest Cover Reduction on Larval Walleye Survival in Lake Erie Tributary Spawning Basins

An integrated modeling approach is used to link land use to river discharge, and then to survival of larval walleye that hatch in northern Ohio streams draining into Lake Erie (USA). First, to link land use and river discharge, the parameters of a simple hydrologic model are statistically related to watershed landscape attributes, including forest cover. One such relationship allows estimation of the change in daily river discharge that could result from a reduction in basin-scale forest cover. Second, to represent the river discharge-larval survival link, we reexamine a dataset from Mion and others to propose a relationship between daily flow velocity, water temperature, and walleye larval survival. Together, these linked models provide estimates of the reduction in larval survival due to reduction in forest cover, along with the uncertainty of those estimates. For the Grand River watershed, decreasing forest cover from 45.2 to 30% is projected to reduce average larval survival by about 45%. In the adjacent Chagrin River, dropping cover from 62.5 to 30% reduces survival by almost 60%. The greater rate of reduction of survival in the Chagrin River as forest levels fall is explained by a relatively greater increase in storm flows for the Chagrin, due to more frequently saturated soils. Therefore, forest preservation in the Chagrin River watershed is projected to be more effective in preserving walleye larval tributary habitat.

Effect of Diet Processing Method and Ingredient Substitution on Feed Characteristics and Survival of Larval Walleye, Sander vitreus

Abstract Two methods were developed for the production of larval fish diets. The first method, microextrusion marumerization (MEM), has been tested in laboratory feeding trials for many years and produces particles that are palatable and water stable. The second method, particle‐assisted rotational agglomeration (PARA), produced diets that have lower density than diets produced by MEM. Each method was used to produce diets in the 250‐ to 400‐ and 400‐ to 700‐μm range and compared with a reference diet (Fry Feed Kyowa* [FFK]) for feeding larval walleye in two experiments. The effect of substituting 4% of the fish meal with freeze‐dried artemia fines was also investigated. In the first experiment, 30‐d survival was greater (P < 0.05) for fish fed a diet produced by PARA without Artemia (49.1.0%) than for fish fed the same diet produced by MEM (27.6%). The addition of Artemia to a diet produced by MEM did not increase survival of larval walleye. Fish fed the reference diet had 24.4% survival. In the second experiment, there was an effect of both processing method and Artemia supplementation, and an interaction of these effects, on survival. Fish fed a diet produced by PARA without Artemia supplementation had 48.4% survival, and fish fed the same diet produced by MEM had only 19.6% survival. Inclusion of 4% freeze‐dried Artemia improved (P < 0.04) survival of fish fed MEM particles but not those fed PARA particles. Fish fed FFK had greater weight gain than fish fed other diets in both experiments. Data indicate that the PARA method of diet processing produces smaller, lower density particles than the MEM process and that diets produced by the PARA process support higher survival of larval walleye with low capital and operating costs.

Effects of Zooplankton Density on Survival of Stocked Walleye Fry in Five Pennsylvania Reservoirs

ABSTRACT Over the past several years, management agencies in the United States have stocked billions of larval walleyes to augment natural reproduction, yet success of these programs has generally been limited by poor first-year survival. We conducted a two-year field study of five impoundments in western and central Pennsylvania to determine if seasonal zooplankton density was related to first year survival of stocked walleye larvae. Using sampling methods that could be easily replicated by state fisheries management agencies, we assessed the magnitude and timing of spring zooplankton blooms in relation to first-year survival of stocked walleye larvae. Our findings showed that although zooplankton production may vary widely among study waters and across years, fry survival was highest when the combined density of large-bodied clasdocerans and adult copepods exceeded 100 organisms/L. When peak zooplankton densities were less than 100/L, few walleye fry survived.

Modelling the effects of dam removal on migratory walleye (Sander vitreus) early life-history stages

Many dams in the USA have outlived their intended purpose and an increasing number are being considered for removal. Yet, quantitative studies of the potential physical, biological and ecological responses are needed to assess dam removal decisions. In this paper, the responses of migratory walleye (Sander vitreus) to increased spawning habitat availability as a result of dam removal was studied by comparing scenarios with and without a high-head dam in the Sandusky River (Ohio), a major tributary to Lake Erie. A conceptual, ecological model was proposed to define the relationship between hydrodynamics and walleye spawning, egg hatching, larval drift and survival. A mathematical, ecological model of the early life-history stages was then developed and coupled with time series of depth and velocity predictions over the spawning grounds from a 1-D hydrodynamic model. Model simulations were run for 1984–1993 for both the with- and without-dam scenarios to assess the potential benefit of dam removal. The simulation results demonstrated that velocity, depth and water temperature are major factors influencing adult walleye spawning success. Without the dam, 10 times the amount of spawning habitat would be available for walleye to spawn. This increase in spawning habitat area resulted in up to five times the total egg deposition and seven times the larval output to the nursing grounds, based on the assumption that 5% of the walleye population of Lake Erie migrated up the Sandusky River to spawn. We concluded that the spawning habitat in the current condition (with the dam) is limiting and additional spawning habitat upstream could significantly increase the number of larval walleye drifting to Lake Erie. The model sensitivity analysis showed that the number of walleye migrating up the river in spring is the dominant factor for larval recruitment to the lake. Copyright © 2006 John Wiley & Sons, Ltd.

Spatial Patterns Emphasize the Importance of Coastal Zones as Nursery Areas for Larval Walleye in Western Lake Erie

Lake Erie walleye Sander vitreus exhibits significant interannual variability in year-class strength. Recent research revealed the importance of larval growth and survival rates in determining walleye year-class strength in western Lake Erie, indicating that spatial and temporal overlap of larvae with good habitat conditions (e.g., abundant prey, warm waters) promoted walleye growth and survival. To assess the spatial overlap between walleye larvae and habitat parameters (water depth, temperature, water clarity, prey density) in western Lake Erie, we evaluated the spatial distribution of walleye larvae and these habitat parameters with intensive sampling at 30 to 36 sites during spring 1994–1999. We analyzed spatial relationships among pelagic walleye larvae and various habitat attributes using a geographic information system and principal components analysis. Larval walleye density was consistently highest at nearshore sites during all years and showed a high degree of spatial overlap with high ichthyoplankton density, and warm water temperatures. Larval walleye density was negatively associated with water depth and water clarity. Two principal components represented 79.6% of the total variability in site attributes. Principle components analysis supported our spatial analysis by graphically separating sites into distinct groups based on larval walleye density and habitat attributes. These analyses indicated that similar relationships between larval distribution and habitat attributes occur each year, emphasizing the importance of nearshore coastal zones as nursery areas for walleye.

Environmental factors influencing larval walleye pollock Theragra chalcogramma feeding in Alaskan waters

This study examines potential interactions among the environmental variables likely to affect larval walleye pollock Theragra chalcogramma feeding in the sea. Walleye pollock larvae were sampled from Shelikof Strait, Gulf of Alaska, and from the eastern Bering Sea, with corresponding environmental data. Variables used in our study were time spent feeding, seawater temperature, light, prey density, wind speed and standard length of the larvae. We applied an additivity test to detect the presence of potential interactions among these variables and adopted an expansion of gen- eralized additive models (GAMs) that allowed the inclusion of interaction terms in a non-parametric regression analysis. After testing all possible 2-way interactions among these variables, we found 4 significant terms: (1) time feeding-standard length, (2) temperature-light, (3) light-wind speed and (4) prey density-standard length. The most influential interaction term was the light-wind speed interaction, which caused a decrease of the model-generalized cross-validation (GCV, whereby lower values indicate more parsimonious models) from 0.0402 (for a model with all variables but no interaction term) to 0.0290 (for a model with all variables and this interaction), and an increase of the model explained variance by 7% (R 2 = 0.89 versus 0.82). This result indicates that the effect of wind speed (turbulence) on larval walleye pollock feeding is dependent upon the amount of light avail- able. This may be due to vertical movements in the water column by both walleye pollock larvae and their prey in response to turbulence.