MacKenzie 2000 Ocean-Acta

OCEANOLOGICA ACTA VOL. 23 N 4

Turbulence, larval fish ecology and fisheriesrecruitment: a review of field studiesBrian R. MacKENZIE *

Danish Institute for Fisheries Research, Department of Marine Ecology and Aquaculture,Kavalergarden 6, DK-2920 Charlottenlund, Denmark

Revised 11 February 2000; accepted 24 February 2000

Abstract Fish recruitment varies widely between years but much of this variability cannot be explained bymost models of fish population dynamics. In this review, I examine the role of environmental variability on fishrecruitment, and in particular how turbulence affects feeding and growth of larval fish, and recruitment in entirepopulations. One of the main findings is that field studies show contrasting effects of turbulence on feeding,growth and mortality rates in nature and on recruitment. Coincident and multiple variations in ecosystemprocesses, lack of understanding of how some of these processes (e.g. larval diet composition, feeding behaviour,growth rates, prey patchiness) respond to turbulence, and unavoidable sampling artifacts are mainly responsiblefor this result. Upwelling as well as frontal processes appear important for larval fish growth and survival, andturbulence levels vary both within and across these features. Process-oriented studies of some of theseinteractions could provide more definitive links between turbulence and biological responses in future. 2000Ifremer:CNRS:IRD:Editions scientifiques et medicales Elsevier SAS

turbulence : larval fish : recruitment : upwelling : front

Resume Turbulence, ecologie des larves de poissons et recrutement des pecheries : synthe`se des etudes de terrain.Le recrutement des poissons est tre`s variable dune annee a` lautre, mais sa variabilite reste inexpliquee dans laplupart des mode`les de dynamique des populations. Le present travail porte sur le role de la variabiliteenvironnementale dans le recrutement du poisson, en particulier leffet de la turbulence sur lalimentation et lacroissance des larves, et sur le recrutement des populations. Les etudes de terrain montrent que la turbulence ades effets contrastes sur lalimentation, sur les taux de croissance et de mortalite dans la nature, ainsi que surle recrutement. Ce resultat est attribue a` des variations concomittantes et multiples dans les mecanismes delecosyste`me, a` notre mauvaise connaissance de la manie`re dont ces processus (composition de lalimentationlarvaire, mode dalimentation, taux de croissance, dispersion des proies) repondent a` la turbulence, et auxartefacts inevitables de lechantillonnage. Lupwelling et les phenome`nes frontaux semblent importants pour lacroissance et la survie des larves de poissons, et lintensite de la turbulence varie a` la fois a` linterieur et au-dela`de ces zones. Des etudes consacrees au mecanisme de certaines de ces interactions devraient conduire a` desrelations plus generales entre la turbulence et les reponses biologiques. 2000 Ifremer:CNRS:IRD:Editionsscientifiques et medicales Elsevier SAS

turbulence : larve de poisson : recrutement : upwelling : front

1. INTRODUCTION

Turbulence in the sea is an important agent of ecosys-tem variability and has received increasing attentionfrom the oceanographic community in the last 2030years. This attention has resulted from numerous* Correspondence and reprints: [email protected]

357 2000 Ifremer:CNRS:IRD:Editions scientifiques et medicales Elsevier SAS

PII: S0399-1784 (00 )00142-0:FLA

B.R. MacKENZIE : Oceanologica Acta 23 (2000) 357375

field, laboratory and theoretical studies that havedemonstrated how turbulent mixing can influencebiological and physical processes in the ocean. Manyof these studies have reported changes in primary andsecondary plankton production, plankton patchiness,and pelagic food web structure as a consequence ofvariability in the intensity of turbulent mixing [2, 32,68, 89, 90, 135]. Since the intensity of turbulence inthe ocean is coupled to climatic and hydrographicprocesses such as storms, tides, upwellings, and buoy-ancy inputs (e.g. solar heating, freshwater runoff;[139]), variations in weather and climate will havedirect and indirect effects on biological processes [6,61].

Many fisheries oceanography studies have addressedhow feeding and growth rates in larval fish vary inrelation to different physical processes and in differ-ent types of hydrographic environments [28, 52, 77].These processes and environments presumably havedifferent levels of turbulent dissipation and frequen-cies of input of kinetic energy. A large number ofother studies have compared the population dynam-ics of entire fish populations to turbulence-relatedvariables such as wind speed [29, 112, 136]. Thisreport will summarize the literature concerning re-sponses of fish at the individual and population levelto turbulence in natural environments.

2. METHODS

2.1. Literature compilation

The published, referreed fisheries oceanographic liter-ature was examined for papers describing larval fishfeeding, growth, mortality and recruitment in relationto estimated turbulence (e.g. wind- or tidal-inducedmixing) in the sea. Modelling and laboratory studiesof these processes have been excluded, except in caseswhere they may help interpret the field observations.In order to synthesize the findings, the overall rela-tionship of the biological response to variations inturbulence was recorded from each study. In mostcases, studies reported biological responses via multi-variate statistical analyses that accounted for influ-ences of other important variables such as foodconcentration (for feeding and growth responses) orspawning stock biomass (recruitment responses).

Responses to turbulence were categorized accordingto the type of relationship observed: positive, nega-tive, dome-shaped (increasing response at low-inter-mediate turbulence and decreasing response atintermediate-high turbulence) and independent (nodifferences in biological responses at different turbu-lent dissipation rates).

2.2. Definition of turbulence

It is appropriate to specify how the term turbulencewill be used in this review. Turbulence representsthose motions which transfer kinetic energy on alarge scale (e.g. up to tens of metres, as in the case oftidally mixed water columns) to a small scale (e.g.centimetresmillimetres). At the smaller scales, vis-cosity dampens the mechanical motions and the re-maining kinetic energy is dissipated as heat [135]. Theresulting motion at small scales can be induced bymany different physical processes, including surfaceand internal wave-breaking, convective mixing andinteractions of currents with bottom topography (e.g.tides). Hence the processes that generate mixing andturbulence differ greatly, but the characteristics ofturbulent motion itself (i.e. the consequence of thephysical forcing) are universally described by theory[135, 149]. Since processes such as upwelling, stormsand frontal circulations all generate turbulent motionin the water column, this review will address howthese processes affect young fish stages and fishrecruitment.

3. RESULTS

3.1. General

The studies that have been used in this literaturesurvey are listed together with the biological re-sponses in table I (e.g. feeding, growth) and table II(recruitment). In table I, the studies are groupedaccording to the complexity of the biological re-sponse to turbulence: gut content analyses as indica-tors of recent feeding activity; nucleic acid studies asindicators of recent somatic growth; otolith growthand abundance and mortality (survival) indices.

358


Table I. Effects of turbulence on larval fish captured in the sea. Note that in many studies, turbulence was directly or inversely proportional to in situprey concentrations, and that in other studies, prey concentrations were not measured. For these reasons, the table must be interpreted cautiously. Allturbulence estimates are based on indirect estimates (e.g. from wind speed), except for [36] which are based on ADCP-derived Richardson numbers,and [46] which is based on both wind-based and direct measurements of turbulent dissipation rate. Table entries reflect increasing complexity of thebiological process related to turbulence. For studies near tidal fronts, turbulence is assumed to increase with hydrographic situation from stratified tofrontal to mixed [56, 138, 139].

Feeding:GrowthPrey Relationship to Response toSpecies SiteAuthorsTurbulenceTurbulence Relationship Investigated

Gut contents studies:Decreased following storm Gut contents before:after decreased CaliforniaNorthernLasker [75]

stormanchovy CurrentNorthern Gut contents at calm andOwen et al. [119] More prey at calm stations California

turbulent stations Currentthan turbulent stationsanchovyVaried between stations. Gut contents (no. of prey)CodSundby and Fossum [145] Lofoten, NorwayAccounted for in analyses. and wind speed 26 m s1

Varied between stations. Gut contents (no. of prey)CodSundby et al. [144] Lofoten, Norwayand wind speed 210 mAccounted for in analyses.s1

Walleye pollock Larval feeding and stormsBailey et al. [4] Prey concentrations lower Shelikof Strait,during:after storms Gulf of AlaskaVaried between stations. Gut contents andCodLough and Mountain 82] (weak) Georges Bank

wind:tidal turbulenceAccounted for in analyses.Haddock Varied between stations.Lough and Mountain [82] Gut contents and (weak) Georges Bank

Accounted for in analyses. wind:tidal turbulenceGut contents (no. of prey)Varied between stations. (weak) Scotian Shelf,CodMcLaren et al. [95]

CanadaAccounted for in analyses. and wind speedAnchovy Prey concentrationConway et al. [20] Gut contents before:after No. Adriatic

stormdecreased after storm.Newfoundland,Biovolume and prey sizeDower et al. [36] Varied between stations. , Radiated shanny

Accounted for in analyses. of gut contents and Canadaturbulence

Stable within samplingHerringCox et al. [23] Gut contents at different BlackwaterEstuary, UKperiods. points in tidal cycle.

Somatic growth rate studies:HaddockBuckley and Lough [14] RNA:DNA content inPrey higher at stratified site Georges Bank

stratified and mixed waterthan mixed siteMore zooplankton at Growth rates highest inSpratMunk [104] Se. North Seastratified and frontal mixed and frontal water.stations than mixed.

Walleye pollock Prey concentrations lowerBailey et al. [4] RNA content and storms Shelikof Strait,Gulf of Alaskaduring:after storms

SardineNakata et al. [113] RNA:DNA content inHigher in front than Kuroshio region,Higher incentral Japan.frontoffshore stratified area. frontal and offshore

stratifed zone.Prey concentration approx. RNA:DNA before vs.Cod Georges BankLough et al. [81]Same before and after after stormstorm.Prey highest in front, lower RNA:DNA highest inClemmesen et al. [19] Anchovy Argentinian shelf

stratified waterin mixed and stratified waterPrey concentration approx.HaddockLough et al. [81] RNA:DNA before vs. Georges Bankincreased

after stormSame before and afterstorm.Prey concentration unrelated RNA:DNA and windSardine So. PortugalChicharo et al. [18]to wind speed. speed

359


Table I. (Continued)

Response to SitePrey Relationship toSpeciesAuthors Feeding:GrowthTurbulence TurbulenceRelationship Investigated

Otolith growth studies:Mid-Atlantic Bight,Not described. Otolith growth and stormsMaillet et al. [88] AtlanticUSAmenhaden

DomeOtolith growth and wind W. of ScotlandVaried between stations.HerringGallego et al. [47]No effect on growth. speedVaried between stations. No. North SeaOtolith growth and wind Gallego et al. [46] HaddockNo effect on growth. speed

Abundance & mortality studies:California CurrentNot described Larval mortality rate and Peterman and Bradford [121] Northern

storm frequencyanchovyNot describedButler [15] Larval mortality rate andSardine California Current

wind-based water columnstability (Lasker events)

Bailey and Macklin [5] Walleye pollock Not described Larval mortality rate and Shelikof Strait, Gulfstorms of Alaska

Shelikof Strait, GulfPrey concentrations lowerBailey et al. [4] Walleye pollock Larval survival rate andduring:after storms of Alaskastorms

NorthernCury et al. [26] California CurrentNot described DomeLarval abundance andupwellinganchovy

3.2. Feeding and growth

There were a total of 22 studies that compared feeding(N 11) and growth (N 11) to variations inturbulence. The responses to turbulence are diverse(positive, negative, dome-shaped, no effect) for bothvariables (table I). Food concentrations in many casesco-varied with variations in turbulence or among thedifferent hydrographic situations. What is perhapssurprising is that some of the studies investigating gutcontents as measures of feeding activity [82, 95, 144,145] have used similar species (cod, haddock larvae)and similar sampling and larval analysis (gut contents)methods but have not produced consistent results.

3.3. Mortality

Four out of six studies showed that larval mortalityrates appear to be highest in windy conditions, whichare partly related to turbulent disssipation rates. How-ever this pattern must be treated cautiously becausethree of the four studies with this relationship involveone species in one region: several years of intensivestudy in the Gulf of Alaska have shown that walleyepollock mortality rates are higher in windier years orfollowing windy periods than in calm years [4, 5].

One other study showed that larval mortality rates werehigher in years with strong winds (northern anchovy:[121]) and one study found no relationship betweenlarval mortality rates and a wind-based stability index(larval sardine mortality and Lasker events; [15]).

3.4. Recruitment

Recruitment studies typically used wind speed as aproxy indicator of the amount of turbulence in thesurface mixing layer. One of the more consistentresponses involved clupeids in upwelling regions.Nearly all of these populations responded to turbu-lence in a dome-shaped fashion (see also [7] for review;table II). Relationships of populations in other typesof ecosystems to wind estimates of turbulence weremuch more variable (table II).

4. DISCUSSION

4.1. General

Biological responses of individual fish larvae and fish

360


populations to variations in turbulence are inconsis-tent, especially when results from many studies arecompiled and analysed together. All four types ofturbulence relationship are evident both for process-oriented field studies of larval ecology and for thestatistical studies investigating recruitment dynamics

of entire populations. These observations suggest thatin some studies other processes affecting these re-sponses vary by larger amounts and have biggerimpacts on the responses than the variations in turbu-lent dissipation rates. In addition, strong covariationin some variables (e.g. patchiness and turbulent

Table II. Effects of wind speed (e.g. expressed as indices of turbulence, upwelling, storms) on fish recruitment. Note that in some cases thestatistical analyses have not adjusted the recruitment series for interannual differences in spawning stock biomass which can cause trendsin recruitment [112].

Species Spawning Stock BiomassAuthors Wind Index Relationship SiteConsidered in Analysis?

PeruDomeAnchoveta wind-based turbulenceCury and Roy yesestimate[25]

Sardine yes upwelling DomeCury and Roy Peru-California[25]

noCury and Roy MoroccoDomeSardine wind-based turbulenceestimate[25]wind-based water column CaliforniaSardineButler [15] yesstability (Lasker events)

Ware and Sardine yes wind speed Dome West coast, N.AmericaThomson

[156]no upwelling DomeNorthern CaliforniaRoy et al. [129]

anchovy CurrentMoroccoDomeupwellingyesRoy et al. [129] Sardine

yesCod Artco-Norwegianturbulence (wind speed3)Ottersen and Sundby [117] cod

Shelikof Strait,yesMegrey et al. wind speed (age 0 recruits)Walleye pollock Gulf of Alaska[97]

Whiting no wind speed3 North SeaSvendsen et al.[147]

wind speed3Herring no North SeaSvendsen et al.[147]

no North Seawind speed3HaddockSvendsen et al. [147]

North Seawind speed3noSvendsen et al. Cod[147]

North Seano wind speed3Svendsen et al. Saithe[147]

North Seano wind speed3Svendsen et al. Sandeel[147]

wind speedno Bay of BiscayAnchovyBorja et al. [12]

Anchovy S. Benguelawind-based upwelling estimateHutchings [59] noCurrent

Plaice no wind speed (age 0 recruits) Nielsen et al. Kattegat[114]

Sprat yesDaskalov [29] wind speed Black SeaBlack SeadomeDaskalov [29] wind speedyesAnchovyBlack SeaHorse mackerel Daskalov [29] wind speed3yes

wind speed Black SeaDaskalov [29] yesWhiting

361


dissipation rates; advection and wind mixing) andvarious sampling and measurement errors (to be de-scribed below) will also contribute to the observedvariability between studies. The biological response tovariations in turbulence within a study may thereforerepresent the response to the given state of the entireecosystem (e.g. prey distribution, zooplankton com-munity composition) rather than the response to asingle component process (e.g. food availability) withinthe ecosystem. Lastly the aggregate findings describedhere are also partly due to the effects of combiningstudies conducted at different ranges of variables (e.g.prey concentrations, turbulent dissipation rates). Someof these factors will be discussed in more detail below.

The next two sections use two of the findings by Doweret al. [35] as starting point: there is no direct evidenceof increased encounter rates between larvae and preyin field settings (p. 200); and there is no field evidencewhich shows increased growth of larval fish resultingfrom the effects of turbulence [35]. It will be suggestedbelow that many studies that have been used toevaluate these ideas and processes in the field may notand could not be designed to address these issues withhigh statistical power, and new alternative studiescould increase our process-based understanding ofhow turbulence affects larval fish feeding and growth.

4.2. Factors contributing to inconsistent feeding,growth and mortality relationships to turbulence

4.2.1. Gut contents as a measure of lar6aprey encounterrateSeveral of the studies which investigated encounter andfeeding rate responses to turbulence relied almostexclusively on analyses of the contents of larval guts.However gut contents may not always be an accurateindex of larvalprey encounter rates. The prey con-tained within a gut represents the endproduct of severalprevious behavioural events (e.g. encounter, pursuit,attack), all of which must be successful in order for theprey to be ingested. If some of these components areunsuccessful then the assumption that gut contentsrepresent encounter probability is no longer valid. Forexample, pursuit and attack success vary with larvaland prey behaviour [53, 57, 103, 105, 107]. Both theory[64, 71, 87, 93] and experiments [85] show that pursuitsare unlikely to be as successful in turbulent water as

in calm water. Evaluation of encounter rates due tovariations in turbulence probably requires direct be-havioural observations [57, 84, 105].

4.2.2. Wind conditions only pro6ide approximationsof turbulent dissipation rates or turbulent 6elocitiesA second factor that contributes to the inconsistentnature of the findings in table I is that almost withoutexception no study listed in table I has directly mea-sured turbulent dissipation rates using microstructureprofilers [33]. Only one group of investigators [46] hasincluded direct dissipation rate measurements usingthese methods in their sampling program. These mea-surements were then used to develop a site- andcruise-specific relationship between turbulent dissipa-tion rate, wind speed and depth that was applied inotolith growth rate analyses. Turbulent dissipationrates in most of the remaining studies have beenestimated from general relationships to wind speed,which provide a reasonable estimate of turbulentdissipation rate in parts of the surface mixing layer [86,115]. However such estimates are much less certainthan direct measurements [48, 86, 115] and in any casecannot provide reliable estimates of the turbulentdissipation rates at and below pycnoclines [48, 82]where some larvae may feed [75, 82]. Improved sam-pling and modelling [48] could reduce uncertaintiesassociated with wind-based turbulence estimates.

4.2.3. Estimated prey concentrations do not alwaysrepresent those experienced by fish lar6aeEstimating the true concentration and taxonomic:sizecomposition of zooplankton experienced by fish larvaeduring feeding periods is problematic due to samplinglimitations. Most of the studies in table I must employvarious assumptions about the prey field which maynot be valid. These assumptions include: i) allzooplankton species and sizes within a given size rangeare suitable prey; ii) concentrations of all species andsizes within the depth layer where zooplankton concen-trations are measured are homogeneous [152]; iii)larvae actively feed within this same depth layer andnot elsewhere in the water column where food, lightand turbulence conditions might be different.

It is unlikely that these assumptions are fully validduring most field studies. Regarding the first assump-tion, larval diets are often composed of taxa whichdiffer significantly from those present in the water

362


column [37, 41, 65, 107, 132]. Some taxa or sizegroups are therefore more vulnerable to larval fishpredation than others [53].

The second assumption that zooplankton distribu-tions in the sea are homogeneous is also unlikely,given that zooplankton are patchily distributed atsmall scales (B 12 m) even in moderately turbulentenvironments [31, 45, 118]. If larvae encounter andcan exploit such patches, then the depth-averagedconcentration will underestimate the concentrationexperienced by larvae during feeding periods. In gen-eral the role of patchiness on larval feeding rates isunclear, with some authors demonstrating importanteffects for some species [75], and other authors show-ing that other species can feed and grow at low foodconcentrations [10, 40, 42, 43, 84, 100, 101, 103, 105,108]. Other evidence shows that feeding and growthrates in fish larvae in nature are frequently food-limited [17, 19, 81, 113, 141, 144, 146, 150]. Feedingrates in these studies may have been higher if larvaewere able to locate higher concentrations of prey.Some suggestions for clarifying the role of patchinesson larval fish feeding will be discussed below.

Moreover, and related to the third assumption statedabove, the depth layers within the photic zone withinwhich larvae feed are not known, but probably de-pend on local food concentration [75], light condi-tions [39, 50] and turbulence [55]. Larval verticaldistributions vary with all three of these variables [55,109, 116]; during vertical movements larvae may en-counter zooplankton patches and taxonomic:sizecompositions of prey which differ from the depth-averaged conditions within the depth layer used toestimate food concentrations. Measured feeding re-sponses may therefore reflect gut fillings at layersother than those at which food concentration andturbulent dissipation rates were measured. This willalso contribute to the variability observed in table I.

Considerations such as these suggest that few studieshave accurately measured the true prey field experi-enced by feeding fish larvae.

4.2.4. Lar6al growth rates may not show the sameresponse to turbulent dissipation rates as lar6alfeeding beha6iourThe effects of turbulence on growth may not beidentical to their effects on encounter, pursuit or

ingestion [35]. Levels of turbulent dissipation thatyield optimal growth may be at different magnitudesthan those which optimise ingestion rate and theincrease (or decrease) in growth due to variations inturbulence (for a given prey concentration) may bedifferent from the amplitude of variation in feeding-related processes (e.g. encounter, ingestion). Differ-ences in the responses of growth and ingestion toturbulence will depend on how assimilation andmetabolic rates are influenced by turbulence, which ispartly related to activity levels (see also [35]). Bothfish larvae and other taxa display different activitylevels in calm and turbulent water [84, 134] so it isreasonable to assume that metabolic rates and growthrates will display different functional responses toturbulence than encounter or ingestion rates. Theseissues have not been addressed but should be investi-gated further.

4.2.5. Food concentrations may not ha6e limitedingestion or growth ratesIn some of the studies in table I, it is possible thatprey concentrations were high enough to ensure highfeeding and growth rates. In these cases turbulencewill not have a positive effect on ingestion rates eventhough encounter rates are promoted by turbulence[127]. Turbulence will in these cases have an insignifi-cant, or a negative, influence on ingestion rates due toreduced pursuit success [85, 127].

4.3. Suggestions for future studies

4.3.1. Influence of turbulence on lar6al acti6ity andgrowth ratesResolving how larval fish growth rates vary withexposure to turbulence may be a suitable topic forboth experimental and field evaluation. Definition ofsuch relationships could lead to a better understand-ing of how survival probability might be affected byclimatic and hydrographic variability [127]. Stateddifferently, are the surviving individuals of a cohort asubset of offspring which have experienced someoptimal level of turbulence that enabled high growthor survival rates? These hypotheses could perhaps beevaluated from otolith studies of survivors hatch-dates [21, 96, 99] and daily growth rates, particularly

363


if the relationship between otolith and somaticgrowth rates can be defined and modelled accurately[102, 159].

Changes in larval and juvenile (e.g. 0-group) preysearch behaviour could also affect the functionalresponse of growth rates to turbulence. For examplein some moderately turbulent situations, it may bebioenergetically more efficient for cruise-searchinglarvae or juveniles to greatly reduce their swimmingspeeds while searching or even behave as pause-travelor ambush predators. This strategy would reducemetabolic costs due to lower swimming activity [11,74], while allowing the mechanical energy of turbu-lent motion to advect prey within visual range [84,143]. The net result could be higher growth rates andefficiencies than if the fish maintained an active cruisebehaviour.

Habitats that might be suitable for employing anadaptive search strategy would have to be moderatelyenergetic (e.g. tidally mixed regions, other areaswhere bottom flow interacts with bathymetry). Oneexample of fish using such a strategy appears to becoral reef fish occupying the windward and highlyturbulent side of coral reefs [51]. Fish along parts ofthe Great Barrier Reef effectively form a wall ofmouths by orienting upstream against the mainorientation of wind- and tidally-driven currents andfeeding on drifting zooplankton as they are advectedonto the reef [51]. The hypothesis that larval andjuvenile fish actively adjust their search behaviour(and reduce foraging costs) when feeding in turbulentor other flow situations where advection increasesprey encounter rates could be investigated by directin situ observations and video-recording of larval andjuvenile behaviour [66, 78, 79].

4.3.2. Lar6al diet compositionThe difficulty in understanding variations in larvaldiets limits progress in linking feeding and growthrates to food availability. Without improved knowl-edge a precise definition of food or prey is notpossible. Food concentration variables in many casestherefore cannot be configured accurately (regardingspecies or sizes) in statistical analyses. The factorscontrolling diet composition depend on behaviouraltraits of both predator and prey [36, 53, 72, 107], andsmall scale water motion [36, 72, 153]. However thisknowledge is still insufficient to allow accurate gener-

alisations and predictions of dietary taxonomic andsize composition in many wild larvae when exposedto different species mixtures and concentrations [52,120]. This situation could perhaps be overcome infuture if behavioural studies (experimental, field, the-ory) of predatorprey interactions at the individuallevel could resolve how dietary composition dependson factors such as larval-prey behaviours and smallscale water motion [72, 73, 153].

4.3.3. Importance of small scale (12 m) preypatches for fish lar6aeThe ability of larvae to locate and respond to patcheshas not received much experimental investigation,and field approaches to resolve the issue are not yetable to sample fish larvae and their prey simulta-neously at the small scales of larval search and feed-ing behaviour. Given that patches do occur in bothturbulent and other environments, it is possible thatlarvae might encounter a patch due to turbulentmotion in a manner analogous to that described forthe role of turbulence on encounter rates betweenindividual planktonic prey and predators [127] andbetween individual microzooplankton and patches ofbacteria feeding on microclouds of dissolved organicmatter [126]. Encounter rates of fish larvae withpatches could be quite high because the relative ve-locity due to turbulence will be much larger at thespatial scale of these patches (e.g. 50100 cm) than atthe spatial scale of larvae encountering a single prey(ca. 1 reactive distance or 1 body length [38, 154]).

The contribution of patchiness to feeding and growthrates in nature remains therefore unclear, but couldperhaps be clarified experimentally following the ap-proaches of Hunter and Thomas [58] and Price [123].These two studies created patches of phytoplanktonin respectively small aquaria and large mesocosmsand then recorded the search behaviour of predators(anchovy larvae and euphausiids) in and near thepatches. In both cases, when the predators locatedthe patches, the search behaviour changed to main-tain proximity to the patch. Notably the role ofpatchiness on feeding and growth in fish larvae hasbeen investigated frequently using models [9, 30, 160],but many of the findings of these models still requiretesting.

One experimental approach that could be used to testthe hypothesis that larvae can locate and ingest prey

364


in patches in situ would be to conduct similar experi-ments to those of Hunter and Thomas [58] and Price[123] in the field with wild larvae and zooplankton,and record the behaviours with modern video-record-ing camera-ROV technology [66, 151] or by diversrecords [78, 79]. In particular, Kils [66] has recordedand described quantitatively direct feeding by wild0-group herring in a dense patch of zooplankton inKiel Bay. Leis and colleagues have shown that it ispossible to capture and release wild fish larvae, andto follow and record undisturbed swimming be-haviour of such larvae sufficiently long to producequantifiable and statistically significant results. Ap-plying these techniques to feeding behaviour studiesis a promising and as yet essentially untried way toinvestigate whether larvae can exploit small patchesof zooplankton prey in field situations. Similar meth-ods could also be used to test other aspects of larvalfeeding behaviour (reactive distances, encounter, pur-suit, attack) which are usually only investigated inlaboratory aquaria [84, 85, 103].

4.3.4. Roles of swimming beha6iour and turbulent dissi-pation on zooplankton patch formation and dispersionA major gap in the zooplankton patchiness literaturewhich is potentially relevant to larval fish feeding isthe interaction between turbulence and zooplanktonswimming behaviour in producing and maintainingpatches, especially at spatial scales which are mostrelevant to larval fish search behaviour (i.e. B 50100 m). Few field studies have yet considered in threedimensions the dynamics of patch dispersion or for-mation at scales B 100 m that occurs during majorchanges in turbulent dissipation rates (e.g. stormpassage, diel tidal cycles). Strongly swimmingzooplankton should be able to maintain themselves inpatches or swarms under higher turbulent dissipationrates than weaker swimmers; expressed alternatively,aggregations of strong swimming zooplankton shouldhave greater persistence against dispersion (assuminga stimulus to remain aggregated) than aggregationsof weak swimmers. While the potential importance ofbehaviour on patchiness has been acknowledged formany years [1, 83, 118, 122, 142], swimming be-haviour has, with few exceptions [30, 44], not beenquantitatively included in recent models of zooplank-ton patchiness [1, 122]. Investigation of non-equi-librium situations in the field at smaller scales than in

the past will probably add considerable insight intopatch dynamics and their utilization by fish larvae.

Studies to address this topic could include transectacoustic or particle counting studies [24, 31, 45, 83,158] in which patch characteristics (e.g. size andshape of patches; concentration of prey inside andoutside patches) are directly related by statisticalanalyses or modelling to organism swimming ability:buoyancy and water motion variables (e.g. turbulentdissipation rates, shear, relative velocities). Ad-ditional approaches could include laboratorybehavioural investigations of patch formation:dispersion during different regimes of turbulent dis-sipation and using zooplankton species that are keylarval fish dietary components (e.g. copepod naupliiand copepodites). These studies could contribute to asize- (or swimming behaviour-) based modellingframework that could enable one to describe andpredict zooplankton distributions in nature undervarying combinations of swimming behaviour andturbulent dissipation. Such a framework is presentlylacking.

4.4. Fisheries recruitment and interannual variationsin turbulent mixing

4.4.1. Upwelling zonesSome of the strongest evidence for the influence ofwind-forcing (and consequently turbulence) on fish isat the population level. Bakun, Cury, Roy and col-leagues have shown that recruitment in sardine andanchovy populations in nearly all the major up-welling systems around the world is related to wind-based upwelling or turbulence indices in adome-shaped fashion (table II). This result is rathersurprising given the wide range of effects turbulencehas on plankton production and ecosystem structureand functioning. Nevertheless in nearly all cases,maximal recruitment is obtained at a seasonally orannually averaged wind speed of ca. 6 m s1. Inaddition analyses by other colleagues investigatingfish recruitment in upwelling areas have identifiedsimilar patterns (Pacific sardine [156]; Bay of Biscayanchovy [12]).

The oceanographic mechanisms responsible for theseresults involve the production of plankton in thespawning and nursery areas [25]. The mechanisms are

365


Figure 1. Schematic relationship between recruitment and upwelling and turbulence for pelagic clupeoids in major upwelling systemsaround the world. Below the figure are oceanographic processes associated with limiting or excessive upwelling and turbulence which mayaffect growth and survival of larvae and pelagic 0-groups. Modified after Cury and Roy [25].

based on some general principles of planktonproduction and distribution in the sea (nutrient:lightlimitation of phytoplankton production, patchinessof phyto- and zooplankton). These earlier ideas maynow have to be extended to include more recentinterpretations of the roles of upwelling circulationon advection:retention of eggs and larvae [128] andof turbulence on larval feeding behaviour [7, 84, 85,127, 145].

The strong relationship between upwelling intensityand recruitment in the various clupeid stocks, and theinferred oceanographic mechanisms responsible forthese patterns suggests that conditions for growthand survival of the larvae are the key to above-average recruitment in these stocks. Since it isbelieved that faster-growing larvae are probably theones most likely to survive to recruitment [54, 96],and since some but not all studies [3, 21] have foundpositive links between growth or condition during thefirst few months of life and subsequent recruitment[16], one might hypothesize that condition, growth

and survival rates of clupeid larvae and 0-groupsmight also have a dome-shaped relationship toupwelling:turbulence (figure 1). Few time-series arepresently available to test this hypothesis but newtime-series of larval growth or condition in relationto upwelling and turbulent mixing could support themechanisms believed responsible for statisticalcorrelations between recruitment and upwelling:turbulence.

The consistent influence of upwelling on clupeidrecruitment in different ecosystems suggests thatupwelling and turbulence in these regions may havesimilar effects on other local fish species (e.g. hakes,jack mackerel, rock- and flatfishes). To the authorsknowledge a multi-species comparison of recruitment[34, 147] for non-clupeid species within individualupwelling systems or among different systems has notyet been conducted. In contrast, conditions which aresuboptimal for clupeid recruitment occur frequently[25]. This type of environment may be favourable forsome species whose life history may be adapted to

366


years when turbulent mixing and upwelling are tooweak or too strong for optimal clupeid recruitment.

4.4.2. Shelf ecosystemsThere are some other examples of wind conditionsaffecting recruitment (table II). In a comprehensivestatistical study of several species in the North Sea,significant positive relationships between recruitmentand the cube of wind speed were found for herring,cod and saithe in the North Sea, but no effect ofwind speed on recruitment was found for haddock orsandeel; whiting recruitment was negatively related towind speed [147]. Wind speed was also shown toinfluence recruitment in several fish species in theBlack Sea (sprat, anchovy, hake, horse mackerel;[29]). In both sets of analyses, wind speeds weremeasured during the months when eggs and larvaewere in the water column and the relationships ob-tained are believed to represent impacts on survivalduring the early life history stages [29, 147].

Few studies have directly considered the influence oftidal mixing on fish recruitment, although the size ofherring stocks in the northwest Atlantic is positivelycorrelated to size of the larval nursery areas whichare located in tidally mixed regions [62]. Tidal mixingitself however is unlikely to influence interannualvariations in recruitment because tidal mixing is ahighly predictable environmental signal to which fishlife histories have presumably evolved [140]. How-ever, tides do affect stratification of the watercolumn, and in years having different buoyancy in-puts (e.g. heat, freshwater runoff), the onset of strat-ification and frontal positions will vary fromlong-term mean dates and locations [13]. This varia-tion will affect plankton production processes andpotentially also recruitment.

The empirical evidence for such couplings is weak butcan be addressed with field work (e.g. comparisons oflarval and 0-group growth and survival in yearshaving different buoyancy inputs). Notably the statis-tical analyses [29, 34, 136, 147] obtained significantcorrelations between recruitment and stratification-related variables (e.g. heat content, stability, watertemperature, river runoff). Lastly, ocean stability andits effects on lower trophic level production may alsobe the key mechanism controlling decadal-scale fluc-tuations in North Pacific salmon stocks [8, 49].

4.4.3. Role of ju6enile mortality onen6ironment-recruitment relationshipsThe hypothesis that fish recruitment is related toturbulent mixing assumes that turbulence affects thesurvival probability of the larval and early juvenilestages, and that recruitment trends are independentof processes during the late juvenile or benthic stages.In many fish populations this latter assumption is notvalid because juvenile mortality can be so high andvariable that it can obscure time trends in larvalmortality [98]. In these cases environmental variablesthat affect larval mortality rates will not explainsignificant amounts of variation in recruitment. Theinconsistent responses of recruitment to turbulenceobserved in table II may reflect varying degrees ofjuvenile mortality between and within studies.

4.4.4. SummaryAside from the clupeids residing in upwelling areas,there is little evidence that recruitment depends onwind speed or turbulence in a consistent manneracross species or geographic regions. The remainingstudies were all from temperate-boreal shelf ecosys-tems where interactions between oceanographic pro-cesses and fish recruitment still remain unclear [28,52] and appropriate environmental variables cannotbe scaled properly [8]. Relationships in these areaswere positive, negative or insignificant, and some didnot account for interannual differences in egg pro-duction or spawning biomass which can contribute totime trends in recruitment [92, 112].

4.5. Similarities between frontal zones and upwellingregions

The results of Cury and Roy [25] and Roy et al. [129]indicate that environmental factors associated withwind (i.e. turbulent mixing, upwelling intensity, trans-port, lower trophic level production) and the spawn-ing stock have significant impacts on fish recruitment.This relationship means that different levels of re-cruitment can be produced per unit spawningbiomass (figure 2), depending on the ability of theenvironment to produce new fish (i.e. carrying capac-ity). No other comparable effect of an environmentalvariable has yet been detected for any other fishspecies complex in widely separated ecosystems. Inthis way, the clupeid recruitment-upwelling relation-ship is unique at the present time.

367


In this context it is relevant to ask whether there areother similar dominant production-related processesthat might affect recruitment in temperate-borealecosystems in the way that upwelling systems do forclupeids. Three such processes seem obvious candi-dates. First, the spring and autumn diatom bloomshave long been considered to be key features aroundwhich most fish species in temperate-boreal systemshave evolved to reproduce [27, 28]. Some evidence ofsuch a linkage is available [131] but in general therelationship between spring primary or secondaryproduction and recruitment is uncertain [67, 77, 130].As a result, recruitment models used in stock assess-ment do not include this or generally any otherenvironmental variable [112].

A second phenomenon which might be related torecruitment in some temperate-boreal species couldinvolve frontal zones [28]. Frontal zones occuraround banks, on continental shelves where tidalmixing is strong, and at continental shelf breaks [76].In the case of tidal fronts, hydrographic processesresult in a balance between mixing (winds, tides) andstratification processes (heating, freshwater runofffrom land masses; [139]). These processes have beenparameterized into predictive models that describethe locations and seasonal durations of the presenceof tidally-generated fronts in shelf seas [13, 139].Tidal fronts represent areas of intermediate mixingintensity [56, 138] and are transition zones betweenfully mixed and stratified water masses [139]. Forexample, direct measurements of turbulent dissipa-tion rate at the front and mixing zones on GeorgesBank showed that dissipations in the frontal zonewere ca. 10-fold less than those in the mixed area [56]and ADCP current measurements revealed large vari-ability across a frontal zone in the North Sea [106].

The oceanographic and production characteristics ofmany frontal zones [76] bear some similarities withthose of upwelling zones. In particular, frontal circu-lations mix nutrients into the photic zone whichstimulates primary production by large phytoplank-ters. Higher rates of primary production and higherphytoplankton abundance stimulate production atsecondary trophic levels so that concentrations ofconsumers (e.g. copepods) increase [70, 113]. In addi-tion, circulation near the frontal zone can advect andretain plankton to the front [44, 76]. Hence somepatchiness of plankton can be expected, encounterrates between trophic levels will be increased due toturbulence, and pursuit success of encountered preywill still be higher than in fully mixed water masses.In these ways, frontal regions share many key plank-ton production features with upwelling areas.

Frontal areas are also important larval and juvenilefish habitats. A large number of studies have shownthat larval fish in temperate-boreal ecosystems aremore abundant in such zones than surrounding watermasses [22, 104, 110, 111, 124, 133, 141, 148]. Themechanism by which larvae accumulate at such re-gions is unclear. Nevertheless, the nutritional condi-tion of larvae and pelagic 0-groups captured infrontal zones is often significantly higher [19, 113,141] or more variable [80] than in larvae captured inneighbouring water masses. The improved conditionof these larvae may give them a survival advantage intimes of food limitation and is believed to be relatedto the generally higher levels of plankton abundanceand production found in frontal regions [69, 70, 76,113, 148]. Persistence of both the frontal feature andlarvae in its vicinity appear to enable the larvae tofeed and grow at faster rates. However, in situationswhere frontal zones are ephemeral the coupling oflarval distributions or growth:condition to frontalplankton production and abundance may be lessevident or manifested as increased variability [80].

The enhanced abundance of fish larvae in frontalregions, coincident with both relatively high levels ofprey and moderate turbulence, suggests that frontalregions may be important nursery areas for the pro-duction of new recruits [60, 110, 111]. Given thatfrontal formation and positions depend on climaticvariables (e.g. heating, freshwater input), their impacton larval fish growth and survival could vary betweenyears. If this is true, differences in recruit production

Figure 2. Schematic of how population egg production and envi-ronmental processes might interact to determine recruitment inclupeoid populations in upwelling ecosystems. As depicted, carry-ing capacity for recruit production depends on variations inupwelling or turbulence.

368


at frontal zones may be a factor determining recruit-ment in some temperate-boreal species [60, 110, 141,148]. This hypothesis remains to be formally testedfor most species occupying frontal areas. Given thepotential similarities between frontal and upwellingtypes of plankton production, and the documentedinfluence of upwelling intensity on clupeid recruit-ment [7], this hypothesis may well be supported byempirical data for some species. The type of dataneeded to test this hypothesis could include time-series of recruitment, plankton and larval abundancesat the frontal zone, characteristics (e.g. turbulentdissipation rates, geographic size, seasonal duration,nutrient fluxes) of the frontal zone, and temporal andspatial variability in larval growth rates or conditionindices.

However, there are areas (Baltic Sea, Newfoundlandshelf, Scotian Shelf) where fish stocks are (were)successful but which appear to have no major persis-tent frontal structures (sensu [139]) around whichearly life-history stages of fishes and their prey (andpredators?) aggregate. In these cases, other mecha-nisms of recruit production must be available. One ofthese might involve a greater reliance of fish onvertical gradients in plankton production, particu-larly in the vicinity of pycnoclines [125]. These can beproductive, prey-rich and moderately turbulent re-gions [63], and occupy large horizontal areas ofstratified seas. As such, they may represent anotherimportant habitat responsible for recruit productionand survival among species inhabiting shelf regions oftemperate-boreal seas.

Interestingly, many sardine and anchovy populations(e.g. Japan, Mediterranean, Black Sea, southwest At-lantic shelf) reproduce in shelf or shelf-break areas [7,29, 113, 157] and estuaries [94] in addition to up-welling areas. Recent investigations of Black Seaanchovy [29], Argentinian-Brazilian anchovy [19, 137]and Japanese sardine [113, 155, 157] indicate thatthese populations are sensitive to environmental vari-ability. The fact that these species and genera repro-duce in upwelling, shelf-tidal and estuarine regionssuggests that there are some fundamental similaritiesin the structure and functioning of the ecosystems towhich adults of these species expose their offspring.This observation supports the possibility that frontalzones and:or pycnoclines may serve functionally sim-

ilar roles for larval growth and survival as the majorupwelling zones, at least in terms of pelagic food webinteractions. A comparative analysis of sardine:an-chovy recruitment variability in relation to environ-mental variables (e.g. winds, tides, thermoclines,fronts, and turbulence), reproductive biology andspawning biomass could be useful for understandinghow these stocks survive in different habitats (up-welling vs. shelf-tidal regions), and how other speciesliving in shelf areas might be influenced by wind andturbulence conditions.

5. SYNTHESIS AND CONCLUSIONS

One of the main findings from this review is thatthere is no clear relationship across systems betweenturbulent mixing and field estimates of larval fishfeeding, growth and mortality, and fish recruitment.The diversity of the response to turbulence that hasbeen observed so far makes it difficult to infer howlarval fish and fish populations might respond toturbulence in nature. Reasons for this finding includea failure to sample directly the processes of interest(i.e. those directly responsible for variations ingrowth or mortality), and the covariation of manyecosystem processes as water columns become more(less) turbulent [91, 135].

It is evident that the field studies surveyed here havemainly investigated population-level responses (e.g.changes in growth rates by entire populations; re-cruitment) to turbulence, or more accurately, to mix-ing-induced differences in the state of entire pelagicecosystems (e.g. their food web structure, prey distri-bution, zooplankton community compositions, etc.).The observed response within each study representsan integrated response to the effect of turbulence onthe structure and functioning of the local pelagicecosystem. The roles of component processes (e.g.changes in prey abundance due to increased turbu-lence) on a biological response (e.g. recruitment) arenot precisely known in a functional or mechanisticsense, are difficult to resolve and perhaps may not bethe critical issue in some contexts [7]. What is perhapsmost critical in these cases, at least for understandingcauses of fluctuations in fish biomass, is the popula-tion-level response to the overall effect of turbulenceon the ecosystem.

369


It is also evident that these same field studies areunlikely to increase our understanding of how pro-cesses (e.g. encounter rate, patch utilization, predatoractivity levels and metabolic rates) related to feedingor growth of individuals might vary with turbulence.This type of individual-based process informationcould nevertheless be useful in interpreting feedingand growth rates, diets and distributions of fish in thefield. Investigations of individual-based processes re-quire different approaches from those used in mostfield studies considered here. These approaches mustinclude methods where the responses and forcingvariables (e.g. dissipation rate, prey concentrationand type) of interest are not influenced by otherprocesses and where they can be reliably measuredand quantified.

Acknowledgements

The author thanks the organizers of the Symposium PNDR,Ifremer-Nantes (89 December, 1998) for their invitation to partic-ipate in the meeting. Travel funds for participation in the meetingwere kindly provided by Ifremer. Thanks to Thomas Kirboe,Peter Munk, Mike St. John and Andy Visser (DIFRES) fordiscussions and two anonymous reviewers for helpful suggestionsand comments to an earlier draft. Partial research funding supportwas provided by grants from the European Unions FAIR pro-gramme (Contract nos. FAIR 98-3959: STORE and FAIR98-4122: STEREO).

REFERENCES

[1] Abraham E.R., The generation of plankton patchiness byturbulent stirring, Nature 391 (1998) 577580.

[2] Alcaraz M., Turbulence and copepods: grazing, behavior andmetabolic rates, in: Marrase C., Saiz E., Redondo J.M.(Eds.), Lectures on plankton and turbulence, Scientia Marina61 (Suppl.1) 1997, pp. 177195.

[3] Bailey K.M., Brodeur R.D., Hollowed A.B., Cohort survivalpatterns of walleye pollock, Theragra chalcogramma, in She-likof Strait, Alaska: a critical factor analysis, Fish. Oceanogr.5 (1996) 179188.

[4] Bailey K.M., Canino M.F., Napp J.M., Spring S.M., BrownA.L., Contrasting years of prey levels, feeding conditions andmortality of larval walleye Theragra chalcogramma in thewestern Gulf of Alaska, Mar. Ecol. Prog. Ser. 119 (1995)1123.

[5] Bailey K.M., Macklin S.A., Analysis of patterns in larvalwalleye pollock Theragra chalcogramma survival and windmixing events in Shelikof Strait, Gulf of Alaska, Mar. Ecol.Prog. Ser. 113 (1994) 112.

[6] Bakun A., Global climate change and intensification ofcoastal ocean upwelling, Science 247 (1990) 198201.

[7] Bakun A., Patterns in the ocean: Ocean Processes andMarine Population Dynamics. Cal. Sea Grant System,NOAA and Centro de Investigaciones Biologicas delNoroeste, La Paz, BCS Mexico, 1996, 323.

[8] Baumann M., The fallacy of the missing middle: physics -fisheries, Fish. Oceanogr. 7 (1998) 6365.

[9] Beyer J.E., Nielsen B.F., Predator foraging in patchy envi-ronments: the interrupted Poisson process (IPP) model unit,Dana 11 (1996) 65130.

[10] Blom G., Svasand T., Jrstad K.E., Ottera H., Paulsen O.I.,Holm J.C., Comparative survival and growth of two strainsof Atlantic cod (Gadus morhua) through the early life stagesin a marine pond, Can. J. Fish. Aquat. Sci. 51 (1994)10121023.

[11] Boisclair D., Sirois P., Testing assumptions of fish bioen-ergetics models by direct estimation of growth, consumption,and activity rates, Trans. Amer. Fish. Soc. 122 (1993) 784796.

[12] Borja A., Uriarte A., Egana J., Motos L., Valencia V.,Relationships between anchovy (Engraulis encrasicolus) re-cruitment and environment in Bay of Biscay, Sci. Mar. 60(1998) (1967) 179192.

[13] Bowers D.G., Simpson J.H., Mean position of tidal fronts inEuropean shelf-sea, Cont. Shelf Res. 7 (1987) 3544.

[14] Buckley L.J., Lough R.G., Recent growth, biochemical com-position, and prey field of larval haddock (Melanogrammusaeglefinus) and Atlantic cod (Gadus morhua) on GeorgesBank, Can. J. Fish. Aquat. Sci. 44 (1987) 1425.

[15] Butler J.L., Mortality and recruitment of Pacific sardine,Sardinops sagax caerulea, larvae in the California current,Can. J. Fish. Aquat. Sci. 48 (1991) 17131723.

[16] Campana S.E. Year-class strength and growth rate in youngAtlantic cod Gadus morhua, Mar. Ecol. Prog. Ser. 135 (1196)2126.

[17] Canino M.F., Bailey K.M., Incze L., Temporal and geo-graphic differences in feeding and nutritional condition ofwalleye pollock larvae Theragra chalcogramma in ShelikofStrait, Gulf of Alaska, Mar. Ecol. Prog. Ser. 79 (1991)2735.

[18] Chicharo M.A., Chicaro L., Valdes L., Lopez-Jamar E., ReP., Estimation of starvation and diel variation of the RNA:DNA ratios in field-caught Sardina pilchardus larvae off thenorth of Spain, Mar. Ecol. Prog. Ser. 164 (1998) 273283.

[19] Clemmesen C., Sanchez R., Wongtschowski C., A regionalcomparison of the nutritional condition of SW Atlanticanchovy larvae, Engraulis anchoita, based on RNA:DNAratios, Arch. Fish. Mar. Res. 45 (1997) 1743.

[20] Conway D.V.P., Coombs S.H., Smith C., Feeding of an-chovy Engraulis Encrasicolus larvae in the northwesternAdriatic Sea in response to changing hydrobiological condi-tions, Mar. Ecol. Prog. Ser. 175 (1998) 3549.

370


[21] Coombs S.H., Conway D.V.P., Morley S.A., Halliday N.C.,Carbon content and nutritional condition of sardine larvae(Sardina pilchardus) off the Atlantic coast of Spain, Mar.Biol. 134 (1999) 367373.

[22] Coombs S.H., Giovanardi O., Conway D., Manzueto L.,Halliday N., Barrett C., The distribution of eggs and larvaeof anchovy (Engraulis encrasicolus) in relation to hydrogra-phy and food availability in the outflow of the river Po, ActaAdriatica 38 (1997) 3347.

[23] Cox C.J., Harrop R., Wimpenny A., Feeding ecology ofherring (Clupea harengus) larvae in the turbid BlackwaterEstuary, Mar. Biol. 134 (1999) 353365.

[24] Currie W.J.S., Claereboudt M.R., Roff J.C., Gaps andpatches in the ocean: a one-dimensional analysis of plank-tonic distributions, Mar. Ecol. Prog. Ser. 171 (1998) 1521.

[25] Cury P., Roy C., Optimal environmental window and pelagicfish recruitment success in upwelling areas, Can. J. Fish.Aquat. Sci. 46 (1989) 670680.

[26] Cury, P., Roy, C., Mendelssohn, R., Bakun, A., Husby,D.M., Parrish, R.H., Moderate is better: Exploring nonlinearclimatic effects on the Californian northern anchovy (En-graulis mordax), in: Beamish R.J. (Eds.), Climate Changeand Northern Fish Populations, 1995, pp. 417424.

[27] Cushing D.H., A difference in structure between ecosystemsin strongly stratified waters and in those that are only weaklystratified, J. Plankton Res. 11 (1989) 113.

[28] Cushing D.H., Population production and regulation in thesea: a fisheries perspective. Univ. Press, Cambridge UK,1995, 354.

[29] Daskalov G., Relating fish recruitment to stock and physicalenvironment in the Black Sea using generalized additivemodels, Fish. Res. 41 (1999) 123.

[30] Davis C.S., Flierl G.R., Wiebe P.H., Franks P.J.S., Mi-cropatchiness, turbulence and recruitment in plankton, J.Mar. Res. 49 (1991) 109151.

[31] Davis C.S., Gallager S.M., Solow A.R., Microaggregationsof oceanic plankton observed by towed video microscopy,Science 257 (1992) 230232.

[32] Denman K.L., Gargett A.E., Biological-physical interactionsin the upper ocean: The role of vertical and small scaletransport processess, Ann. Rev. Fluid Mech. 27 (1995) 225255.

[33] Dewey R.K., Crawford W.R., Gargett A.E., Oakey N.S., Amicrostructure instrument for profiling oceanic turbulence incoastal bottom boundary layers, J. Atmos. Ocean. Techn. 4(1987) 288297.

[34] Dippner J.W., Recruitment success of different fish stocks inthe North Sea in relation to climate variability, Ger. J.Hydrogr. 49 (1997) 277293.

[35] Dower J.F., Miller T.J., Leggett W.C., The role of microscaleturbulence in the feeding ecology of larval fish, Adv. Mar.Biol. 31 (1997) 169220.

[36] Dower J.F., Pepin P., Leggett W.C., Enhanced gut fullnessand an apparent shift in size selectivity by radiated shanny

(Ul6aria subbifurcata) larvae in response to increased turbu-lence, Can. J. Fish. Aquat. Sci. 55 (1998) 128142.

[37] Economou A.N., Food and feeding ecology of five gadoidlarvae in the northern North Sea, J. Cons. Int. Explor. Mer47 (1991) 339351.

[38] Evans G., The encounter speed of moving predator and prey,J. Plankton Res. 11 (1989) 415417.

[39] Fiksen ., Utne A.C.W., Aksnes D.L., Eiane K., HelvikJ.V., Sundby S., Modelling the influence of light, turbulenceand ontogeny on ingestion rates in larval cod and herring,Fish. Oceanogr. 7 (1998) 355363.

[40] Folkvord A., Ystanes L., Johannessen A., Moksness E.,RNA:DNA ratios and growth of herring (Clupea harengus)larvae reared in mesocosms, Mar. Biol. 126 (1996) 591602.

[41] Fox C.J., Harrop R., Wimpenny A., Feeding ecology ofherring (Clupea harengus) larvae in the turbid BlackwaterEstuary, Mar. Biol. 134 (1999) 353365.

[42] Frank K., Leggett W.C., Coastal water mass replacement: itseffect on zooplankton dynamics and the predatorprey com-plex associated with larvae capelin (Mallotus 6illosus), Can.J. Fish. Aquat. Sci. 39 (1982) 9911003.

[43] Frank K.T., Leggett W.C., Effect of prey abundance and sizeon the growth and survival of larval fish: an experimentalstudy employing large volume enclosures, Mar. Ecol. Prog.Ser. 34 (1986) 1122.

[44] Franks P., Sink or swim: accumulation of biomass at fronts,Mar. Ecol. Prog. Ser. 82 (1992) 112.

[45] Gallager S., Davis C.S., Epstein A.W., Solow A., BeardsleyR.C., High-resolution observations of plankton spatial distri-butions correlated with hydrography in the Great SouthChannel, Georges Bank, Deep-Sea Res. II 43 (1996) 16271663.

[46] Gallego A., Heath M.R., Basford D.J., MacKenzie B.R.,Variability in growth rates of larval haddock in the northernNorth Sea, Fish. Oceanogr. 8 (1999) 7792.

[47] Gallego A., Heath M.R., McKenzie E., Cargill L.H., Envi-ronmentally induced short-term variability in the growthrates of larval herring, Mar. Ecol. Prog. Ser. 137 (1996)1123.

[48] Gargett A.E., Theories and techniques for observing tur-bulence in the ocean euphotic zone, Sci. Mar. 61 (1997)2545.

[49] Gargett A.E., The optimal stability window: a mechanismunderlying decadal fluctuations in North Pacific salmonstocks?, Fish. Oceanogr. 6 (1997) 109117.

[50] Grnkjr P., Wieland K., Ontogenetic and environmentaleffects on vertical distribution of cod larvae in the BornholmBasin, Baltic Sea, Mar. Ecol. Prog. Ser. 154 (1997) 91105.

[51] Hamner W.M., Jones M.S., Carleton J.H., Hauri I.R.,Williams D., Zooplankton, planktivorous fish, and watercurrents on a windward reef face: Great Barrier Reef, Aus-tralia, Bull. Mar. Sci. 42 (1988) 459479.

[52] Heath M.R., Field investigations of the early life historystages of marine fish, Adv. Mar. Biol. 28 (1992) 2174.

371


[53] Heath M.R., The role of escape reactions in determining thesize distribution of prey captured by herring larvae, Env.Biol. Fish. 38 (1993) 331344.

[54] Heath M.R., Gallego A., Bio-physical modelling of the earlylife stages of haddock in the North Sea, Fish. Oceanogr. 7(1998) 110125.

[55] Heath M.R., Henderson E.W., Baird D.L., Vertical distribu-tion of herring larvae in relation to physical mixing andillumination, Mar. Ecol. Prog. Ser. 47 (1988) 211228.

[56] Horne E.P.W., Loder J.W., Naimie C.E., Oakey N.S., Tur-bulence dissipation rates and nitrate supply in the upperwater column on Georges Bank, Deep-Sea Res. II 43 (1996)16831712.

[57] Hunter J.R., Swimming and feeding behaviour of larvalanchovy, Engraulis mordax, Fish. Bul. 70 (1972) 821834.

[58] Hunter J.R., Thomas G.L., Effect of prey distribution anddensity on the searching and feeding behaviour of larvalanchovy, in: Blaxter J.H.S. (Eds.), The early life history offish, Springer-Verlag, 1974, pp. 559574.

[59] Hutchings L., Barange M., Bloomer S.F., Boyd A.J., Craw-ford R.J.M., Huggett J.A., Kerstan M., Korrubel L.J., DeOliviera J.A.A., Painting S.J., Richardson A.J., ShannonL.J., Schulein F.H., Van der Lingen C.D., Verheye H.M.,Multiple factors affecting South African anchovy recruitmentin the spawning, transport and nursery areas, S. Afr. J. Mar.Sci. 19 (1998) 211225.

[60] ICES, Report of the ICES:GLOBEC Cod and ClimateChange AGGREGATION Workshop, Charlottenlund, Den-mark, 2224 August 1994, ICES CM 1994:A:10 (1994)113.

[61] IGBP Global Ocean Ecosystem Dynamics (GLOBEC) Im-plementation Plan. IGBP Report 47; GLOBEC Report 13,IGBP Secretariat, Royal Swedish Academy of Sciences,Stockholm, 1999, 207 p.

[62] Iles T.D., Sinclair M., Atlantic herring: stock discretenessand abundance, Science 215 (1982) 627633.

[63] Incze L.S., Aas P., Ainaire T., Distributions of copepodnauplii and turbulence on the southern flank of GeorgesBank: implications for feeding by larvae cod (Gadusmorhua), Deep-Sea Res. II 43 (1996) 18551873.

[64] Jenkinson I.R., A review of two recent predation-rate mod-els: the domeshaped relationship between feeding rate andshear rate appears universal, ICES J. Mar. Sci. 52 (1995)605610.

[65] Kane J., The feeding habits of co-occurring cod and haddocklarvae from Georges Bank, Mar. Ecol. Prog. Ser. 16 (1984)920.

[66] Kils U., The ecoSCOPE and dynIMAGE: microscale toolsfor in situ studies of predatorprey interactions, Arch. Hy-drobiol. Beih (Ergebn. Limnol.) 36 (1992) 8396.

[67] Kirboe T., Pelagic fisheries and spatio-temporal variabilityin zooplankton productivity, Proc. 4th Intl. Conf. Copepoda,Bull. Plankton Soc. Japan, Spec. Vol 1991 Spec. vol. (1991).

[68] Kirboe T., Turbulence, phytoplankton cell size, and thestructure of pelagic food webs, Adv. Mar. Biol. 29 (1993)172.

[69] Kirboe T., Johansen K., Studies of a larval herring (Clupeaharengus) patch in the Buchan area. IV. Zooplankton distri-bution and productivity in relation to hydrographic features,Dana 6 (1986) 3751.

[70] Kirboe T., Munk P., Richardson K., Christensen A.,Paulsen H., Plankton dynamics and larval herring growth,drift and survival in a frontal area, Mar. Ecol. Prog. Ser. 44(1988) 205219.

[71] Kirboe T., Saiz E., Planktivorous feeding in calm andturbulent environments with emphasis on copepods, Mar.Ecol. Prog. Ser. 122 (1995) 135145.

[72] Kirboe T., Saiz E., Visser A.W., Hydrodynamic signalperception in the copepod Acartia tonsa, Mar. Ecol. Prog.Ser. 179 (1999) 97111.

[73] Kirboe T., Visser A.W., Predator and prey perception incopepods due to hydromechanical signals, Mar. Ecol. Prog.Ser. 179 (1999) 8195.

[74] Krohn M.M., Boisclair D., Use of a stereo-video system toestimate the energy expenditure of free-swimming fish, Can.J. Fish. Aquat. Sci. 51 (1994) 11191127.

[75] Lasker R., Field criteria for survival of anchovy larvae: therelation between inshore chlorophyll maximum layers andsuccessful first feeding, Fish. Bull. 73 (1975) 453462.

[76] LeFevre J., Aspects of the biology of frontal systems, Adv.Mar. Biol. 23 (1986) 163299.

[77] Leggett W.C., de Blois E., Recruitment in marine fishes: is itregulated by starvation and predation in the egg and larvalstages?, Neth. J. Sea Res. 32 (1994) 119134.

[78] Leis J.M., Carson-Ewart B.M., In situ swimming speeds ofthe late pelagic larvae of some Indo-pacific coral-reef fishes,Mar. Ecol. Prog. Ser. 159 (1997) 165174.

[79] Leis J.M., Sweatman H.P.A., Reader S.E., What the pelagicstages of coral reef fishes are doing out in blue water:daytime field observations of larval behavioural capabilities,Mar. Freshw. Res. 47 (1996) 401411.

[80] Lochmann S.E., Taggart C.T., Griffin D.A., ThompsonK.R., Maillet G.L., Abundance and condition of larval cod(Gadus morhua) at a convergent front on Western Bank,Scotian Shelf, Can. J. Fish. Aquat. Sci. 54 (1997) 14611479.

[81] Lough R.G., Caldarone E.M., Rotunno T.K., BroughtonE.A., Burns B.R., Buckley L.J., Vertical distribution of codand haddock eggs and larvae, feeding and condition instratified and mixed waters on southern Georges Bank, May1992, Deep-Sea Res. II 43 (1996) 18751904.

[82] Lough R.G., Mountain D.G., Effect of small-scale turbu-lence on feeding rates of larval cod and haddock in stratifiedwater on Georges Bank, Deep-Sea Res. II 43 (1996) 17451772.

[83] Mackas D.L., Boyd C.M., Spectral analysis of zooplanktonspatial heterogeneity, Science 204 (1979) 6264.

372


[84] MacKenzie B.R., Kirboe T., Encounter rates and swim-ming behavior of pause-travel and cruise larval fish predatorsin calm and turbulent laboratory environments, Limnol.Oceanogr. 40 (1995) 12781289.

[85] MacKenzie B.R., Kirboe T., Larval fish feeding and turbu-lence: a case for the downside, Limnol. Oceanogr. 45 (2000)110.

[86] MacKenzie B.R., Leggett W.C., Wind-based models for esti-mating the dissipation rates of turbulent energy in aquaticenvironments: empirical comparisons, Mar. Ecol. Prog. Ser.94 (1993) 207216.

[87] MacKenzie B.R., Miller T.J., Cyr S., Leggett W.C., Evidencefor a dome-shaped relationship between turbulence and lar-val fish ingestion rates, Limnol. Oceanogr. 39 (1994) 17901799.

[88] Maillet G.L., Checkley D.M. Jr., Storm-related variation inthe growth rate of otoliths of larval Atlantic menhadenBre6oortia tyrannus : a time series analysis of biological andphysical variables and implications for larva growth andmortality, Mar. Ecol. Prog. Ser. 79 (1991) 116.

[89] Mann K.H., Lazier J.R.N., Dynamics of Marine Ecosystems:Biological-Physical Interactions in the Oceans, Blackwell Sci-ence Ltd., 1996, 394 p.

[90] Margalef R., Life-forms of phytoplankton as survival alter-natives in an unstable environment, Oceanol. Acta 1 (1978)493509.

[91] Marrase C., Saiz E., Redondo J.M., Lectures on planktonand turbulence, Scientia Marina 61 (1997).

[92] Marshall C.T., Kjesbu O.S., Yaragina A., Solemdal P., Ull-tang ., Is spawner biomass a sensitive measure of theproductive and recruitment potential of the North-east Arc-tic cod?, Can. J. Fish. Aquat. Sci. 55 (1998) 17661783.

[93] Matsushita K., Possible importance of turbulence at thespatial scale of a larval fish visual field on feeding success,Arch. Hydrobiol. Beih (Ergebn. Limnol.) 36 (1992) 109121.

[94] McFadzen I.R.R., Franceschini G., The nutritional condi-tion of larvae of anchovy (Engraulis encrasicolus L.) in theoutflow of the River Po (Northern Adriatic), Acta Adriatica38 (1997) 4964.

[95] McLaren I.A., Avendano P., Taggart C.T., Lochmann S.E.,Feeding by larval cod in different water masses on WesternBank, Scotian Shelf, Fish. Oceanogr. 6 (1997) 250265.

[96] Meekan M.G., Fortier L., Selection for fast growth duringthe larval life of Atlantic cod Gadus morhua on the ScotianShelf, Mar. Ecol. Prog. Ser. 137 (1996) 2537.

[97] Megrey B.A., Bograd S.J., Rugen W.C., Hollowed A.B.,Stabeno P.J., Macklin S.A., Schumacher J.D., IngrahamW.J., Jr., An exploratory analysis of associations betweenbiotic and abiotic factors and year-class strength of Gulf ofAlaska walleye (Theragra chalcogramma) in: Beamish R.J.(Eds.), 1995, pp. 227243.

[98] Mertz G., Myers R.A., Estimating the predictability of re-cruitment, Fish. Bul. 93 (1995) 657665.

[99] Methot R.D., Seasonal variation in survival of larval north-ern anchovy, Engraulis mordax, estimated from the age dis-tribution of juveniles, Fish. Bull. 81 (1983) 741750.

[100] Moksness E., Validation of daily increments in the otolithmicrostructure of Norwegian spring-spawning herring (Clu-pea harengus L.), ICES J. Mar. Sci. 49 (1992) 231235.

[101] Moksness E., iestad V., Interaction of Norwegian spring-spawning herring larvae (Clupea harengus) and Barents Seacapelin (Mallotus 6illosus) in a mesocosm study, J. Cons. Int.Explor. Mer. 44 (1987) 3242.

[102] Mosegaard H., Svedang H., Taberman K., Uncoupling ofsomatic and otolith growth rates in Arctic charr (Sal6elinusalpinus) as an effect of differences in temperature response,Can. J. Fish. Aquat. Sci. 45 (1988) 15141524.

[103] Munk P., Foraging behaviour and prey size spectra of larvalherring Clupea harengus, Mar. Ecol. Prog. Ser. 80 (1992)149158.

[104] Munk P., Differential growth of larval sprat Sprattus sprat-tus across a tidal front in the eastern North Sea, Mar. Ecol.Prog. Ser. 99 (1993) 1727.

[105] Munk P., Foraging behaviour of larval cod (Gadus morhua)influenced by prey density and hunger, Mar. Biol. 122 (1995)205212.

[106] Munk P., The concentration of larval cod and its prey in thezone of a hydrographic front, ICES CM 1995:Q: 23 (1995)110.

[107] Munk P., Prey size spectra and prey availability of larval andsmall juvenile cod, J. Fish Biol. 51 (1997) 340351.

[108] Munk P., Kirboe T., Feeding behaviour and swimmingactivity of larval herring (Clupea harengus) larvae in relationto density of copepod nauplii, Mar. Ecol. Prog. Ser. 24(1985) 1521.

[109] Munk P., Kirboe T., Christensen V., Vertical migrations ofherring, Clupea harengus, larvae in relation to light and preydistribution, Env. Biol. Fishes 26 (1989) 8796.

[110] Munk P., Larsson P.O., Danielssen D.S., Moksness E.,Larval and small juvenile cod Gadus morhua concentrated inthe highly productive areas of a shelf break front, Mar. Ecol.Prog. Ser. 125 (1995) 2130.

[111] Munk P., Larsson P.O., Danielssen D.S., Moksness E.,Variability in frontal zone formation and distribution ofgadoid fish larvae at the shelf break in the northeasternNorth Sea, Mar. Ecol. Prog. Ser. 177 (1999) 221233.

[112] Myers R.A., When do environmentrecruitment correlationswork?, Rev. Fish Biol. Fish. 8 (1998) 285305.

[113] Nakata K., Zenitani H., Inagake D., Differences in foodavailability for Japanese sardine larvae between the frontalregion and the waters on the offshore side of Kuroshio, Fish.Oceanogr. 4 (1995) 6879.

[114] Nielsen E., Bagge O., MacKenzie B.R., Wind-induced trans-port of plaice (Pleuronectes platessa) early life-history stagesin the Skagerrak-Kattegat, J. Sea Res. 39 (1998) 1128.

[115] Oakey N.S., Elliott J.A., Dissipation within the surfacemixed layer, J. Phys. Oceanogr. 12 (1982) 171185.

373


[116] Olla B.L., Davis M.W., Ryer C.H., Sograd S.M., Be-havioural determinants of distribution and survival in earlystages of walleye pollock, Theragra chalcogramma : a syn-etheis of experimental studies, Fish. Oceanogr. 5 (1996)167178.

[117] Ottersen G., Sundby S., Effects of temperature, wind andspawning stock biomass on recruitment of Arcto-Norwegiancod, Fish. Oceanogr. 4 (1995) 278292.

[118] Owen R.W., Microscale and finescale patchiness of smallplankton in coastal and pelagic environments, J. Mar. Res.47 (1989) 197240.

[119] Owen R.W., Lo N.C.H., Butler J.L., Theilacker G.H., Al-varino A., Hunter J.H., Watanabe Y., Spawning and survivalpatterns of larval northern anchovy, Engraulis mordax, incontrasting environments - site-intensive study, Fish. Bull. 87(1989) 673688.

[120] Pepin P., Penney R.W., Patterns of prey size and taxonomiccomposition in larval fish: are there general size-dependentmodels?, J. Fish Biol. 51 (Supplement A) (1997) 84100.

[121] Peterman R.M., Bradford M.J., Wind speed and mortalityrate of a marine fish, the northern anchovy (Engraulis mor-dax), Science 235 (1987) 354356.

[122] Powell T.M., Okubo A., Turbulence, diffusion and patchi-ness in the sea, Phil. Trans. R. Soc. Lond. B 343 (1994)1118.

[123] Price H.J., Swimming behaviour of krill in response to algalpatches: a mesocosm study, Limnol. Oceanogr. 34 (1989)649659.

[124] Richardson K., Heath M., Pihl N.J., Studies of a larvalherring (Clupea harengus L.) patch in the Buchan area. I. Thedistribution of larvae in relation to hydrographic features,Dana 6 (1986) 110.

[125] Richardson K., Visser A.W., Pedersen F.B., Subsurface phy-toplankton blooms fuel pelagic production in the North Sea,J. Plankton Res. (submitted).

[126] Rothschild B.J., Haley P.J. Jr., Cai D., Influence of physicalforcing on carbon microclouds of dissolved organic matterand nutrients in the ocean, J. Plankton Res. 21 (1999)12171230.

[127] Rothschild B.J., Osborn T.R., Small-scale turbulence andplankton contact rates, J. Plankton Res. 10 (1988) 465474.

[128] Roy C., An upwelling-induced retention area off Senegal: amechanism to link upwelling and retention processess, S.Afr. J. Mar. Sci. 19 (1998) 8998.

[129] Roy C., Cury P., Kifani S., Pelagic fish recruitment successand reproductive strategy in upwelling areas: environmentalcompromises, S. Afr. J. Mar. Sci. 12 (1992) 135164.

[130] Runge J.A., Should we expect a relationship between pri-mary production and fisheries? The role of copepod dynam-ics as a filter of trophic variability, Hydrobiologia 167:168(1988) 6171.

[131] Runge J.A., Castonguay M., De Lafontaine Y., RinguetteM., Beaulieu J.-L., Covariation in climate, zooplanktonbiomass and mackerel recruitment in the southern Gulf of St.Lawrence, Fish. Oceanogr. 8 (1999) 139149.

[132] Runge J.A., de Lafontaine Y., Characterization of thepelagic ecosystem in surface waters of the northern Gulf ofSt. Lawrence in early summer: the larval redfish-Calanus-mi-croplankton interaction, Fish. Oceanogr. 5 (1996) 2137.

[133] Sabates A., Maso M., Effect of a shelf-slope front on thespatial distribution of mesopelagic fish larvae in the westernMediterranean, 2 Deep-Sea Res. 7A (1990) 10851098.

[134] Saiz E., Alcaraz M., Free-swimming behaviour of Acartiaclausi (Copepoda: Calanoida) under turbulent water move-ment, Mar. Ecol. Prog. Ser. 80 (1992) 229236.

[135] Sanford L.P., Turbulent mixing in experimental ecosystemstudies, Mar. Ecol. Prog. Ser. 161 (1997) 265293.

[136] Shepherd J.G., Pope J.G., Cousens R.D., Variations in fishstocks and hypotheses concerning their links with climate,Rapp. P.-V. Reun. Cons. Int. Explor. Mer 185 (1984) 255267.

[137] Sieg A., A study on the histological classification of the insitu nutritional condition of larval south-west Atlantic an-chovy, Engraulis anchoita Hubbs and Marini, 1935, Arch.Fish. Mar. Res. 46 (1998) 1936.

[138] Simpson J.H., Crawford W.R., Rippeth T.P., CampbellA.R., Cheok J.V.S., The vertical structure of turbulent dissi-pation in shelf seas, J. Phys. Oceanogr. 26 (1996) 15791590.

[139] Simpson J.H., Hunter J.R., Fronts in the Irish Sea, Nature250 (1974) 404406.

[140] Sinclair M., Marine populations: an essay on populationregulation and speciation. Washington Sea Grant Program,Univ. Washington Press, Seattle, 1988, 252 p.

[141] St. John M.A., Lund T., Lipid biomarkers: linking theutilisation of frontal plankton biomass to enhanced condi-tion of juvenile North Sea cod, Mar. Ecol. Prog. Ser. 131(1996) 7585.

[142] Stavn R.H., The horizontal-vertical distribution hypothesis:Langmuir circulation and Daphnia distributions, Limnol.Oceanogr. 16 (1971) 453466.

[143] Sundby, S., Wind climate and foraging of larval and juvenileArcto-Norwegian cod (Gadus morhua), in: Beamish R.J.(Eds.), Climate Change and Northern Fish Populations,1995, pp. 405415.

[144] Sundby S., Ellertsen B., Fossum P., Encounter rates betweenfirst-feeding cod larvae and their prey during moderate tostrong turbulent mixing, ICES Mar. Sci. Symp. 198 (1994)393405.

[145] Sundby S., Fossum P., Feeding conditions of Arcto-Norwe-gian cod larvae compared with the Rothschild-Osborn the-ory on small-scale turbulence and plankton contact rates, J.Plankton Res. 12 (1990) 11531162.

[146] Suthers I.M., Fraser A., Frank K.T., Comparison of lipid,otolith and morphometric condition indices of pelagic juve-nile cod Gadus morhua from the Canadian Atlantic, Mar.Ecol. Prog. Ser. 84 (1992) 3140.

[147] Svendsen E., Aglen A., Iversen S.A., Skagen D.W., SmestadO., Influence of climate on recruitment and migration of fishstocks in the North Sea, Can. Sp. Publ. Fish. Aquat. Sci. 121(1995) 641653.

374


[148] Taggart, C.T., Thompson, K.R., Maillet, G.L., Lochmann,S.E., Griffin, D.A., Abundance distribution of larval cod(Gadus morhua) and zooplankton in a gyre-like water masson the Scotian Shelf, in: Watanabe Y., Yamashita Y., Yosh-ioki O. (Eds.), Survival Strategies in Early Life Stages ofMarine Resources, Proc. Int. Workshop, Yokohoma, Japan,11.14 October 1994, A.A. Balkema Publishers, Rotterdam,1996, pp. 155173.

[149] Tennekes H.H., Lumley J.L., A first course in turbulence,MIT Press, Cambridge, MA, 1972, 293 p.

[150] Theilacker G.H., Bailey K.M., Canino M.F., Porter S.M.,Variations in larval walleye pollock feeding and condition: asynthesis, Fish. Oceanogr. 5 (1996) 112123.

[151] Thetmeyer H., Kils U., To see and not be seen: the visibilityof predator and prey with respect to feeding behaviour, Mar.Ecol. Prog. Ser. 126 (1995) 18.

[152] Tiselius P., Nielsen G., Nielsen T.G., Microscale patchinessof plankton within a sharp pycnocline, J. Plankton Res. 16(1994) 543554.

[153] Viitasalo M., Kirboe T., Flinkman J., Pedersen L.W., Vis-ser A.W., Predation vulnerability of planktonic copepods:consequences of predator foraging strategies and prey sen-sory abilities, Mar. Ecol. Prog. Ser. 175 (1998) 129142.

[154] Visser A.W., MacKenzie B.R., Turbulence-induced contact

rates of plankton: the question of scale, Mar. Ecol. Prog. Ser.166 (1998) 307310.

[155] Wada T., Jacobson L.D., Regimes and stock-recruitmentrelationships in Japanese sardine (Sardinops melanostictus),1951-1995, Can. J. Fish. Aquat. Sci. 55 (1998) 24552463.

[156] Ware D.M., Thomson R.E., Link between long-term vari-ability in upwelling and fish production in the NortheastPacific Ocean, Can. J. Fish. Aquat. Sci. 48 (1991) 22962306.

[157] Watanabe Y., Zenitani H., Kimura R., Population decline ofthe Japanese sardine Sardinops melanostictus owing to re-cruitment failures, Can. J. Fish. Aquat. Sci. 52 (1995) 16091616.

[158] Wiebe P.H., Mountain D.G., Stanton T.K., Greene C.H.,Lough R.G., Kaartved S., Dawson J., Copley N., Acousticalstudy of the spatial distribution of plankton on GeorgesBank and the relationship between volume backscatteringstrength and the taxonomic composition of the plankton,Deep-Sea Res. II 43 (1996) 19712000.

[159] Wright P.J., Metcalfe N.B., Thorpe J.E., Otolith and somaticgrowth rates in Atlantic salmon parr, Salmo salar L.: evi-dence against coupling, J. Fish Biol. 36 (1990) 241249.

[160] Wroblewski J.S., Richman J.G., The nonlinear response ofplankton to wind mixing events - implications for survival oflarval northern anchovy, J. Plankton Res. 9 (1987) 103123.

.

375

Documents

MacKenzie 2000 Ocean-Acta