EFFECT OF SMALL-SCALE TURBULENCE ON FEEDING RATES OF LARVAL COD AND HADDOCK IN STRATIFIED WATER ON GEORGES-BANK

A set of vertically stratified MOCNESS tows made on the southern flank of Georges Bank in spring 1981 and 1983 was analyzed to examine the r elationship between larval cod and haddock feeding success and turbule nt dissipation in a stratified water column. Observed feeding ratios ( mean no. prey larval gut(-1)) for three size classes of larvae were co mpared with estimated ingestion rates using the Rothschild and Osborn (Journal of Plankton Research, 10, 1988, 465-474) predator-prey encoun ter rate model. Simulation of contact rates requires parameter estimat es of larval fish and their prey cruising speeds,density of prey, and turbulent velocity of the water column. Turbulent dissipation was esti mated from a formulation by James (Estuarine and Coastal Marine Scienc e, 5, 1977, 339-353) incorporating both a wind and tidal component. La rval ingestion rates were based on swallowing probabilities derived fr om calm-water laboratory observations. Model-predicted turbulence prof iles generally showed that dissipation rates were low to moderate (10( -11)-10(-7) W kg(-1)). Turbulence was minimal at or below the pycnocli ne (approximate to 25 m) with higher values (1-2 orders of magnitude)n ear the surface due to wind mixing and at depth due to shear in the ti dal current near bottom. In a stratified water column during the day, first-feeding larvae (5-6 mm) were located mostly within or above the pycnocline coincident with their copepod prey (nauplii and copepodites ). The 7-8 mm larvae were most abundant within-the pycnocline, whereas the 9-10 mm larvae were found within and below the pycnocline. Feedin g ratios were relatively low in early morning following darkness when the wind speed was low, but increased by a factor of 2-13 by noon and evening when the wind speed doubled. Comparison of depth-specific feed ing ratios with estimated ingestion rates, derived from turbulence-aff ected contact rates, generally were reasonable after allowing for an a verage gut evacuation time (4 h), and in many cases the observed and e stimated values had similar profiles. However, differences in vertical profiles may be attributed to differential digestion time, pursuit be havior affected by high turbulence, vertical migration of the larger l arvae, an optimum light level for feeding, smaller-scale prey patchine ss, and the gross estimates of turbulence. Response-surface estimation of averaged feeding ratios as a function of averaged prey density (0- 50 m) with a minimum water-column turbulence value predicted that 5-6 mm larvae have a maximum feeding response at the highest prey densitie s (> 30 prey l(-1)) and lower turbulence estimates (<10(-10) W kg(-1)) . The 7-8 mm and 9-10 mm larvae also have a maximum feeding response a t high prey densities and low turbulence, but it extends to lower prey densities (>10 prey l(-1)) as turbulence increases to intermediate le vels, clearly showing an interaction effect. In general, maximum feedi ng ratios occur at low to intermediate levels of turbulence where aver age prey density is greater than 10-20 prey l(-1). Copyright (C) 1996 Elsevier Science Ltd