recruitment generalization

7/30/2019 Recruitment Generalization

1/26

BAKUN: COMPARATIVE STUDIES AND THE RECRUITMENT PROBLEMCalCOFI Rep., Vol. XXVI, 1985

COMPARATIVE STUDIES AND THE RECRUITMENT PROBLEM:SEARCHING FOR GENERALIZATIONS

ANDREW BAKUNPacific Fisheries Environmental GroupSouthwest Fisheries CenterNational Marine Fisheries ServicePO Box 831Monterey, California 93942

ABSTRACT

Empirical attempts to relate recruitment variation tovariability in the environment or in the biologicalcommunity have not been notably successful. This isdue in part to the very short data series generallyavailable for empirical analysis. Obviously, we mustdevelop generalizations to provide a basis for unifyingthe available fragments of experience in a coherentframework.

The ecosystems of the four major subtropical east-ern ocean boundary regions, Le., the California, Peru,Canary, and Benguela current regions, appear to becontrolled by similar environmental dynamics (char-acteristic late spring distributions of coastal tempera-ture anomaly and Ekman transport for the four region-al systems, supplementing previously reported winterand summer distributions, are presented). These re-gions contain very similar assemblages of exploitablepelagic fishes, and exhibit comparable histories offishery growth and abrupt decline. The similaritiessuggest that fish communities in these different sys-

tems may function alike with respect to their environ-ments and may have reproductive strategies for similarenvironmental problems. Because natural selectionimplies that reproductive strategies are responses tothe most crucial factors regulating reproductive suc-cess, compelling patterns of correspondence amongreproductive habits and environmental characteristicsare likely to reflect important causal mechanisms.Such findings, which are independent of time seriesresults, furnish a guide for selecting variables for timeseries modeling in a way that makes improved use ofthe minimal degrees of freedom that are available.

Beyond comparisons of spawning habitat climatolo-gy, it may be fruitful to compare actual time seriesrelationships among similar regional ecosystems (anumber of recent studies showing environmental/biological relationships for eastern boundary pelagicfishery stocks are cited). Under an assumption ofanalogy, weak empirical relationships having a similargeneral form in several systems could be assignedgreater confidence than otherwise warranted. Con-versely, a high correlation found in one system might


2/26

be questioned if no suggestion of a similar linkagewere found in other systems appearing to be analo-gous. Arraying identically formulated empirical mod-els from different regional systems might yield pat-terns among model parameters that could provideinsights for predicting effects of human actions ornatural perturbations on a given system from the rec-ord of previous similar events in other systems.

RESUMEN

Los intentos empiricos de relacionar las variacionesen el reclutamiento con variaciones ambientales o enlas comunidades biologicas no han sido demasiadoexitosos. Ello se debe, en parte, a las series demasiadobreves de datos que generalmente se dispone para 10sanalisis empiricos. Es obvio que debemos desarrollargeneralizaciones para proveer una base con el fin deunificar 10s fragmentos de experiencia disponibles enun marc0 de trabajo coherente.

Los ecosistemas de las cuatro principales regionesde corrientes de margen oriental, i.e., California,

Peru, Canarias, y Benguela, parecen estar controladospor esquemas diniimicos semejantes (se presentan lasdistribuciones primaveral tardia caracteristicas deanomalias tCrmicas costeras y transporte de Ekman en10s cuatro sistemas regionales, complementando lainformacion previa acerca de las distribuciones inver-nal y estival). Estas regiones contienen conjuntos depeces explotables muy semejantes e historias similaresen lo que se refiere a1 crecimiento y abrupta declina-cion de las pesquerias. Estas semejanzas sugieren quelas comunidades de peces en estos sistemas diferentespueden funcionar de manera semejante con respecto asus ambientes, y pueden tener estrategias reproduc-

tivas para problemas ambientales similares. Dado quela seleccion natural implica que las estrategias repro-ductivas consistan en respuestas a 10s factores miisimportantes que regulan el Cxito reproductivo, 10sobligatoriamente analogos modelos de habitos repro-ductivos y de caracteristicas ambientales pueden refle-jar mecanismos causales importantes. Estos hallazgos,que son independientes de 10s resultados de las seriestemporales, proveen una guia para seleccionar vari-


3/26


ables para modelos de series de tiempo de manera talque se optimiza el aprovechamiento de la cantidadminima de grados de libertad disponibles.

Mas all5 de la comparacih de la climatologia de 10shabitat de desove, puede resultar provechoso compa-rar las relaciones entre series de tiempo reales entreecosistemas regionales similares (se mencionan variosestudios recientes que muestran las relaciones ambien-tales/biol6gicas en stocks pesqueros pelagicos en areasde margen oriental). Asumiendo la analogia, podriaasignarse mayor confianza a las relaciones empiricasdCbiles que poseen un formato general semejante envarios sistemas. Reciprocamente, una alta correlacihen un sistema podria ser cuestionada si no se encuentraevidencia de una vinculacih semejante en otros siste-mas que parecen ser analogos al primero. El ordena-miento de modelos empiricos de formulacih idCnticaprovenientes de sistemas regionales diferentes puedebrindar patrones entre 10s parametros de 10s models;

estos patrones podrian suministrar informaci6n uti1para predecir 10s efectos del accionar humano o de lasperturbaciones naturales sobre un sistema dado, sobrela base del registro de eventos anteriores semejantes.

INTRODUCTION

The recruitment problem, which has been called thecentral problem of fish population dynamics (Beyer1981), remains unsolved after decades of scientificinterest and effort. The term recruitment refers tothe quantity of youngerfish surviving the various egg,larval, juvenile, etc., stages to reach a size at which

they become susceptible to fishing gear and thus beginto be sampled by thefishery. The term therefore inte-grates the entire period during which the populationsare invisible to humans, between the formation ofreproductive products in the parental fishes and thelater remanifestation of these products as progeny ofharvestable size. It thus spans a lengthy continuum ofsusceptibilities to mortality involving a great range ofprocesses by which the thousands of eggs spawned inthe lifetime of a given female must, on average, bereduced to approximately two surviving adult fish.

As might be expected from such a tenuous chain of

circumstances, recruitment is notoriously variable. Inpopulations of fishes like anchovies, sardines, andmackerels, the ratio between parental stock size andresulting recruitment typically varies from year to yearby factors up to several hundred. Such extreme vari-ability, since it cannot be accounted for mechanis-tically, introduces a very large noise component intofishery population time series, and largely obscuresthe signals essential for managing human impacts onthe biological system. These signals include evidences


4/26

of stock-recruitment relationships, multispecies in-

teractions, etc.-knowledge of which could governfishery development and management strategies. In-deed, the need to account for quasi-random, large-amplitude variations in order to reveal essentialmechanisms responding to human alterations ofmarine biological population structures probably pre-sents, in many cases, a stronger argument for scien-tific efforts on the recruitment question than does themore often cited need for early prediction of a follow-ing years recruitment. In recognition of the need forprogress in this area, an International RecruitmentProject (IREP) is the primary focus of the major newinternational program Ocean Science in Relation toLiving Resources, cosponsored by the Intergovern-mental Oceanographic Commission and the Food andAgriculture Organization of the United Nations (Anon1985). For CalCOFI, in particular, the recruitmentquestion could be considered the core scientific issue(Marine Research Committee 1950).

There seems to have been a remarkable lack ofprogress since the importance of the recruitment prob-

lem was recognized early in this century. However,when one considers the nearly total lack of directinformation about the fairly complex series of eventsinvolving nutrition and associated growth or starva-tion; predation (which is size-dependent and thereforegrowth-dependent); transport processes (perturbationsof which may disrupt well-tuned adaptations for plac-ing large numbers of individuals in proper habitats forrequired life-cycle transitions); physiological stresses;etc., the problems resistance to solution is not sosurprising (Table 1). In addition, even where somedata are obtainable, difficulties in coping with theinterference of strong intrinsic annual periodicities has

led to a conventional practice of pooling shorter-scalevariations into annual composites; this has further ex-acerbated the difficulties of empirical analysis by (1)lowering the available signal-to-noise ratios in bothenvironmental and biological time series, (2) severelylimiting the number of data points available in rea-sonably stationary time series, and (3) scramblingtogether different causal mechanisms controllingdifferent shorter-period survival variations within thesame annual composite data point.

Because of the wide variety of possible linkages ofenvironmental fluctuations to recruitment variations

(Table l), and because a variety of oceanographic, me-teorological, or other proxy time series may be avail-able to somehow be associated with one or more ofthese linkages, correlations with short available seriesof annually composited survival estimates can gener-ally be found and readily justified according to someecological mechanism. When such a correlation hap-pens to be high enough to meet standard significance


5/26

BAKUN: COMPARATIVE STUDIES AND THE RECRUITMENT PROBLEMCalCOFl Rep., Vol. XXVI, 1985

TABLE 1Some Hypothetical Controls on Survival ofEarly Life Stages of Fishes

I. Starvation hypothesesTurbulent mixing of fine-scale food particle strataLow productivity of systemWrong type of potential food organismsDispersion of food due to divergent flow patternMismatch with seasonal food succession caused byanomalies in growth rate

11. Predation hypothesesIncidence of small planktonic predatorsIncidence of large planktonic predators(coelenterates, etc.)Incidence of predatory adult fishIncidence of predatory larval fishVariations in growth with size-dependent predation

111. Advection hypothesesOffshore transport (removal of drifting larvae fromfavorable habitat)Onshore transport (exposure of larvae to damage insurf zone, etc.)Disruption of normal current patterns to whichreproductive habits are tuned

IV.Physiological stress hypothesesT, S, or [02]conditions not within physiological rangeEffects of environmental pollution

V. Disease hypothesesInfectious outbreaks, etc.criteria, it becomes publishable as an advance in sci-entific understanding, whereas slightly weaker corre-lations do not. This happens even though, given thecomplexity of the ecological systems involved, it maybe quite unreasonable to expect such overwhelmingcontrol of survival by any single environmental vari-able or process as to yield significant (e.g., P


6/26

such relationships, with the result that they are notoften used in serious decision-making procedures andthus presently are of little practical value.

The following sections of this paper attempt to

illustrate a rationale for interregional comparativestudies directed toward (1) providing a framework forsystematizing fragmentary information and insights inorder to foster useful generalities concerning physical,biological, and (most important for management ac-tivities) human impacts on recruitment, and (2) devel-oping a basis for detecting spurious relationships andfor increasing confidence in certain weak relationshipsby comparisons with similar situations in other region-al ecosystems. Objections regarding the possibility ofdiscrediting certain valid relationships in situationswhere the assumption of analogy may be faulty can beanswered with the observation that such relationshipsare essentially useless anyway in a situation where thevalid cannot be distinguished from the spurious.

The comparative method and the experimentalmethod are the two great methods of science (Mayr

1982). The comparative method is particularly indi-cated in situations not amenable to controlled experi-ments, and it has underlain nearly all the revolutionaryadvances in evolutionary biology, to cite one exam-ple. It seems surprising that the method has not foundwider use in the fishery-environmental field, whichappears to have been preoccupied with diversity ratherthan synthesis, and prone to view each local situationas unique.

C0MPA RAT IV E SPAWNIN G HABITAT

CLIMATOLOGY

Parrish et al. (1983) presented maps of two-month(Jan.-Feb. and July-Aug.) long-term mean distribu-tions of various environmental characters for the fourmajor eastern boundary regions of the world: the Cali-fornia, Peru, Canary, and Benguela current systems.These maps generally represented seasonal extremesin each system; however, peak upwelling in thetemperate areas is generally earlier toward the equator.The summer and winter maps of Parrish et al. aresupplemented here by characteristic late spring (May-June, N. Hemisphere; Nov. -Dec. , s. Hemisphere)distributions of coastal temperature anomaly (Figure

1) and Ekman transport (Figure 2).

Negative coastal temperature anomalies greaterthan 2C (i.e., 2 degrees cooler than at the samelatitude some 1,000 km offshore; stippled in Figure 1)delimit the major upwelling centers. Note the definiteseparation of the two distinct major upwelling regionsoff South America. Warm anomalies (areas of positiveanomaly greater than 1C are diagonally hatched inFigure l), or areas where cold anomalies are less


7/26

intense, often correspond to spawning regions ofcoastal pelagic fishes (Parrish et al. 1983). Ekmantransport vectors illustrate the strong offshore surfacetransport that must be accommodated in reproductive


8/26

BAKUN: COMPARATIVE STUDIES AND THE RECRUITMENT PROBLEM

CalCOFI Rep., Vol. XXVI, 1985

58 N

45 N

48 N

35 N

30 N

25 N

28 Nx

m

f

5s

18 s

15 5

29 s

25 s

31s

355

48 S

45 5

58 s=

m

m

4

45 N

+

48 N

i

35N 0u


9/26

30 N

fiB0 SDJROJA

25 H

2P s

MAY -JUN

e-I

15 N .

5s

10 S NOV -DEC

15 s

28 S

25 s

I 38 sSOUTH 35s

48 S

W W W W w w

m m

m VI B Lo N N*m

Figure 1. Coastal temperature anomaly (difference from characteristic temperature at similar latitude some ten degrees longitude offshore) distributions

in the four major eastern boundary current systems, characterizing late spring conditions in the respective hemispheres. Units are degrees Celsius.Corresponding winter and summer maps, and details of their construction, are presented by Parrish et al. (1983).


10/26


30 NJUN25 Na \/'symbol5sd10 s15 s10 s NOV-DEC25 S1(38 s35s -25S30 S45 S35 s P I50 Sx 40 SW w w W W

Figure 2. Surface Ekman transport distributions in the four major eastern boundary current systems, characterizing late spring conditions in therespective hemispheres. Transport is proportional to vector symbol length. A symbol length scale is provided; units are metric tons per second acrosseach horizontal meter width. Corresponding winter and summer maps, and details of their construction, are presented by Parrish et al. (1983).


11/26


TABLE 2Dominant Anchovy, Sardine, Jack (Horse) Mackerel,Hake, Mackerel, and Bonito in the Four Major EasternBoundary Currents

California Current Canarv Current

Engraulis mordax Engraulis encrasicholusSardinops sagax Sardina pilchardusTrachurus symmetricus Trachurus trachurusMerluccius productus Merluccius merlucciusScomber japonicus Scomber japonicusSarda chiliensis Sarda sarda

Peru Current Benguela Current

Engraulis ringens Engraulis capensisSardinops sagax Sardinops ocellatusTrachurus symmetricus Trachurus trachurus

Merluccius merluccius Merluccius capensisScomber japonicus Scomber japonicusSarda sarda Sarda sarda

After Bakun and Parrish (1980)

strategies of marine organisms inhabiting these re-gions (Figure 2). In addition, McLain et al. (1985) andShelton et al. (1985) show that biological communi-ties, not only in the Pacific systems but apparentlyalso in the Atlantic, must cope with intermittent ex-treme warm episodes of El Niiio-type intensity.

Very similar assemblages of dominant coastal pe-lagic fish species inhabit the four regions (Table 2). Ineach region there is a dominant anchovy, sardine, jackmackerel, hake, mackerel, and bonito. In severalcases, the same nominal species inhabits more thanone of the regions. In other cases, the species are sosimilar that their difference may be more a matter ofassumption than of demonstration (R.H. Parrish, pers.comm.). There have also been similar regional experi-ences of rapid fishery expansions and precipitous de-clines (Parrish et al. 1983).

The various environmental and biological similar-

ities suggest that fish communities in these differentregional systems may function similarly with re-spect to their environments and may be solving similarproblems in their reproductive strategies. Survival ofoffspring to the age of reproduction, which oftenapproximates the age of fishery recruitment, is thedirect causative factor in the selective process. Thusnatural selection demands response to the most crucialmechanisms regulating recruitment. This implies thatwe can use the observed results of natural selection in


12/26

terms of geographical and seasonal aspects of re-productive strategies to understand the causes control-ling reproductive success. In other words, pervasivepatterns of reproduction. with respect to environmen-tal characteristics, may indicate the actual environ-mental mechanisms regulating recruitment success.

For example, Parrish et al. (1983) found a generalpattern in the spawning habits of anchovies and sar-dines that seemed to point to a simultaneous mini-mization of both wind-induced turbulent mixing andoffshore-directed transport. Thus the starvation hy-pothesis, as expressed in Laskers (1981) scenario ofturbulent dispersion of fine-scale food strata, and thetransport hypothesis, involving detrimental transportof larvae away from favorable proximity to the coast,are implicated as important mechanisms to be consid-ered in modeling and exploratory data analysis.Temperature of spawning showed much less coherentpattern, suggesting that selection of spawning habitatfor any particular optimum temperature is less impor-tant than minimizing turbulent mixing or offshoretransport.

Certain initially puzzling discrepancies from thegeneral patterns have proved enlightening. For exam-ple, Bakun and Parrish (1982) reported that the Peru-vian anchoveta, by far the largest fish stock on recordbefore its collapse, stood out as an anomaly in that itsspawning peak during austral winter occurred at thevery season when offshore Ekman transport was mostintense (Figure 3; Table 3), rather than when it wasrelatively weak, as is the general pattern. However,when mixed layer depth climatologies were producedfor the various regions (Parrish et al. 1983), the dis-crepancy was apparently resolved. Seasonal variationof mixed layer depth off Peru proceeds in phase with

that of transport but has greater amplitude (Figure 4),with the result that the thinner mixed layer of australsummer is apparently carried offshore some four timesas fast as is the deeper winter mixed layer (Figure 5),even though the winter transport (by volume) is nearlytwice as large. Thus the spawning season does indeedappear to be tuned to minimize the rate of offshoreloss of eggs and larvae within the oceans upper mixedlayer. This resolution of what initially seemed a dis-crepancy of a single situation from the general patternhas thus pointed out the generality that, in treating themechanism of offshore loss of reproductive productsdistributed in the mixed layer, estimates of wind-

driven surface (Ekman) transport should ideally bedivided by the effective mixed layer depth to yield anEkman velocity of the mixed layer. In this way we areable to combine two environmental variables to yield a

~

In the statement that these two variables should ideally be combined, the qualifica-tion ideally is important. Ekman transport is estimated from relatively abundant


13/26

surface wind reports, which reflect the fairly large spatial scales of atmosphericvariation. Mixed layer depth varies on much shorter oceanic length scales and must bedetermined from more sparsely distributed subsurface measurements. Thus there maybe cases where the effective mixed layer depth is so imprecisely determined thatthecombined variable yields a less reliable indicator of interyear vanation of offshorevelocity than would be provided by the interyear vanation in the Ekman transportestimate itself.


14/26


OFFSHORE TRANSPORT

1111 S CALIF BlGHT10090807060504030201on1-2 3-4 5-6 7-8 9-10 10-11

CALIFORNIA M=(SPAWNING PEAK] SPAWNING PEAK I

Figure 3. Offshore-directed Ekman transport within the spawning

grounds of the Peruvian anchoveta (Chimbote) and within the spawn-ing grounds of the central and southern subpopulations of (CaliforniaCurrent) northern anchovy, by 2-month segments of the annual cycle.Units are scaled relative to the seasonal maximum in each area, whichis assigned the value 1.0 in all three locations in order to scale themsimilarly (and thereby to emphasize seasonal timing rather than abso-lute magnitude). Timing of seasonal peak spawning in the two systemsis indicated by the wide arrows. (See Table 3 for unscaled numericalvalues.)

single more meaningful variable, thereby conservingdegrees of freedom for improving the power ofhypothesis tests.

The major point to be made is that such findings,which are based on patterns in the long-term compos-ite averages, are independent of any results based onexamination of the available time series data. Thus thefindings provide a guide to selecting variables for time

AT CHIMBOTE MIXED LAYERDEPTH10 OFFSHORE09TRANSPOR T-

0807060504030201001-2 3-4 5-6 7-8 9-10 11-12 1-2


15/26

Figure 4. Offshore-directed Ekman transport and mixed layer depth with-in the spawning grounds of the Peruvian anchoveta, by 2-month seg-ments of the annual cycle. Units are scaled relative to the respectiveseasonal maxima, which are assigned the value 1.0 in both cases inorder to place both on a common scale (and to thereby emphasizeseasonal timing rather than absolute magnitude). (See Table 3 forunscaled numerical values).

series analysis so that the selection process itself doesnot deplete the scarce degrees of freedom. One couldsay that the observed characteristic seasonality andgeography of reproductive strategies results from alarge number of trial-and-error experiments per-formed by the population over the period of develop-ment. Moreover, natural selection ensures that thisobserved result is the most correct one for the fish,at least in the long term (providing that we are observ-ing truly long-term adaptations).

TABLE 3Actual (Unscaled) Values Used to Construct Figures 3,4, and 5

Transport (T sec m I)

~

Months 1-2 3-4 5-6 7-8 9-10 11-12

Chimbote 1.1 1.7 1.8 1.8 1.7 I .4

S. Calif. Bight 0.2 0.4 0.3 0.3 0.2 0.2S. Baja Calif. 0.7 1.1 1.2 0.8 0.7 0.7Mixed Layer Depth (m)Months 1-2 3-4 5-6 7-8 9-10

11-12

Chimbote 7 12 29 40 23 12

S. Calif. Bight 37 25 11 10 14 20S. Baja Calif. 45 23 18 9 20 34Net Offshore Velocity of Mixed Layer (km day-)Months 1-2 3-4 5-6 7-8 9-10

11-12

Chimbote 13.6 12.2 5.4 3.9 6.4 10.1

S. Calif. Bight 0.5 1.4 2.4 2.6 1.2 0.9S. Baja Calif. 1.3 4.1 5.8 7.7 3.0 1.8Transport and mixed layer depth climatologies taken from Parrish et al. (1983);offshore velocities computed as the quotient of the two.


16/26

BAKON: COMPARATIVE STUDIES AND THE RECRUITMENT PROBLEMCalCOFl Rep., Vol. XXVI, 1985

OFFSHORE VELOCITY of the more tropical eastern boundary areas (the sar-

Fl

dinellas, etc.) may be playing the same basic survival

10 S CALIF BIGHT0.9 S BAJACALIF0807060504030201001-2 3-4 5-6 7-8 9-10 11-12

CALIFORNIA

LSFAWNINGFWKI ISPAWNING

PEAK

I

Figure 5. Offshore-directed velocity of the surface mixed layer (Ekmantransport divided by mixed layer depth) within the spawning grounds ofthe Peruvian anchoveta (Chimbote) and of the central and southernsubpopulations of (California Current) northern anchovy, by 2-monthsegments of the annual cycle. Units are scaled relative to the seasonalmaximum in each area, which is assigned the value 1.0 in all three

locations in order to scale them similarly (and thereby to emphasizeseasonal timing rather than absolute magnitude). Timing of seasonalpeak spawining in the two systems is indicated by the wide arrows. (SeeTable 3 for unscaled numerical values.)

Thus any empirical relationship based on short timeseries of recent data on reproductive success and onvariations in some environmental characteristic , notcorroborated by any apparent reproductive accom-modation to that characteristic, might be of question-able validity. Conversely, an empirical relationshipreflected in what seems to be an obvious adaptation incharacteristic reproductive habits could be assigned

increased confidence. This line of reasoning seemsapplicable even to a unique situation where noassumed analogues are available for comparison, butthe comparative aspect introduces a more powerfultest. For example, an apparent adaptation may bemere happenstance in a single situation. However,given the regional diversity introduced by varyingcoastline orientations, atmospheric regimes, advectedcharacteristics, etc., it is unlikely that strong interre-gional correspondence of reproductive habits with en-


17/26

vironmental characteristics is due merely to happen-stance. Therefore, recognition of such an interregionalpattern provides much stronger confirmation of realmechanistic linkage. And, as noted in the Introduc-tion, the present need in recruitment research is forany substantial error to be on the side of undue conser-vatism and rigor, if results are to be applicable to realproblems.

At Pacific Environmental Group we are presentlyentertaining the notion that the coastal pelagic fishes

game as those (sardines, anchovies, etc.) in thetemperate zones, but at a warmer temperature range.If this is even partially true, the Angola, Guinea, andCosta Rica Dome coastal regions, the Ivory Coast-Ghana upwelling region, the Malabar Coast of India,etc. , would provide another set of spawning habitatseasonalities and geographies with which to sharpenthese interpretations. Other possibilities for instructiveanalogues include such western ocean boundarystocks as the Brazilian sardinella and the Japanesesardine; in these cases, basic differences in the en-vironmental dynamics of the systems should be care-

fully evaluated.

COMPARATIVE TIME SERIES MODELING

Beyond comparisons of long-term mean character-istics of spawning habitats and biological conse-quences, it may be fruitful to compare empirical (timeseries) models among similar regional systems. Therehave been a number of published suggestions of en-vironmental/biological relationships. (Some of themore recent examples for eastern boundary coastalpelagic fishery stocks are cited in Table 4.) Formula-tions and points of view have differed greatly among

researchers concerned with the different regional sys-tems. For example, I believe that Boyds (1979) for-mulation in terms of sea-temperature variance, theresults of which suggest Laskers (1981) mechanism ofturbulent dispersion of fine-scale food concentrations,has not been tried in the California Current or the othersystems.

Systematic comparisons of identical empirical mod-el formulations for all available analogous situa-tions might yield suggestive patterns among modelparameters to provide new insights into the processescontrolling the biological systems. The comparisons

could also yield information about the validity ofattempts to transfer experience among systems(e.g., predicting consequences of human actions ornatural perturbations in one system from the record ofsimilar events in other systems, etc.). Under anassumption of analogy, empirical relationships havinga similar general form in more than one system mightbe correctly afforded greater interest and confidencethan otherwise warranted. Conversely, a correlationfound in one system might be considered questionable


18/26

if no suggestion of similar linkage were found in oth-er systems appearing, in other respects, to functionanalogously.

For example, consider the following set of formula-tions (the variables illustrated in this example areperhaps most suitable to anchovies)


19/26


.....................

.....................

.....................

....................

.....................

log @,ISi) = up,i +B,,iSF +B,.iui +B,,iii

.....................

.....................

................. etc.

where Ri is recruitment to a given fish stock i,Si is adult stock size,

TABLE4

uiis upwelling index coincident in time andspace with spawning (we might generallyexpect strong upwelling to be a detrimentalfactor, involving loss of eggs and larvae viaoffshore transport),Li is a properly space- and time-lagged upwell-ing index that would reflect longer time-scalenutrient inputs to the system at upstream up-welling centers,

w.is wind-generated turbulence input to thelarval habitat (motivation for the choice ofvariables ui, iii, and wi can be found in Bakunand Parrish 1980),Ti is an index of characteristic temperature,T~ is temperature variance as used by Boyd(1979), etc.The dots indicate that other hypothesized fac-tors and combinations of factors would also beconsidered.The log (RIS) formulation simply connotes amultiplicative effect of the explanatory vari-Examples of Recent Studies Indicating Environmental/Biological Relationships for

Eastern Boundary Current Pelagic Fishery Stocks

California Current

Lasker (1981)-northern anchovy (Engraulis mordax)-negativeeffect of wind-generated turbulence (wind cubed index) duringspawning season.


20/26

Parrish and MacCall (1978)-Pacific mackerel (Scomberjaponicus)-positive effect of upwelling (index) at spawning

ground previous to spawning (one-month lag).

Bakun and Parrish (1980)-Pacific sardine (Sardinopssagax)-positive effect of upwelling (index) in upwellingmaximum region upstream of spawning grounds (800-km spatiallag) integrated over previous calendar year (2- to 14-month

spatial lag).

Bailey (1981)-Pacific hake (Merluccius producrus)-negative

effect of offshore transport (upwelling index) during planktonic

larval period.

Collins and MacCall (1977)-Pacific bonito (Sardachiliensis)-positive effect of upwelling (index) during spawningseason.

Peru Current

Various authors-anchoveta (Engraulisringens)-negative

effect of El Nifio (often attributed to lowered primary

productivity).

Benguela Current

Boyd (1979)-Southwest African anchovy (Engraulis capensis)

-negative effect of sea-surface temperature variance during

spawning season.

Canary Current

Belveze and Erzini (1983)-sardine (Moroccan central zone;Sardina pilchardus)-positive relation of catches to upwelling(index) averaged over the three previous spawning seasons (catchis composed of one- two-, and three-year-olds).-mixed smallpelagics (Moroccan northern zone; mostly S. pilchardus)-highcatches in low-rainfall years.

Freon (1983)-sardinella (Senegal)-positive relation of CPUEto wind speed during upwelling season (previous winter); authorsuggests autocorrelation in wind series as explaining a

recruitment effect.

Guinea Current

Bakun (1978)-Ghanain herring (Sardinella aurira)-negativerelation of catch to rainfall and sea-surface temperature (relatedto upwelling),

Binet (1982)-sardinella (Ivory Coast and Ghana; S. maderensisand S. aurira)-positive effect of upwelling (temperature index)


21/26

and negative effect of coastal runoff on catch.

Cury and Roy*-sardinella (Ivory Coast and Ghana; Smaderensis and S. aurita)-positive effect on catch and onreproductive success of local upwelling (temperature index),which is related to remote forcing by wind stress variations in thewestern equatorial Atlantic (i.e., off Brazil).

*Cury, P., and C. Roy. MS. Modelisation de Iabondance des especes pelagiques cotieres de Cote dIvoire integrant leffort de peche et unindice dupwelling ORSTOM, 24 Rue Bayard, 75008 Pans, France.


22/26


ables; i.e., if an increase of a given variablewere to double predicted survival per unitspawning biomass, then an equal decreasewould halve predicted survival rather than re-duce it to zero, as would be required by anadditive (Le., nonlogarithmic) formulation.The exponent c represents MacCalls (1980)suggestion for the incorporation of variablehabitat dimension in a density-dependentmechanism such as adult cannibalism (where ifc = 1, there is no habitat size variation andthe result is a normal Ricker curve; if c = 0,the habitat size is proportional to stock size andtherefore density does not increase with abun-dance, so that there is no stock-size depen-dence through the density-dependent mecha-nism; O


23/26

situations, even though the relationships are too weakindividually to meet ordinary significance criteria,valuable insight may have been gained. If nothing isfound but chaos suggesting unrelated random com-binations, one might conclude that little confidenceshould be placed in any such models based on thereally inadequate degrees of freedom represented byshort autocorrelated recruitment time series.

CONCLUDING REMARKS

This paper is intended to serve as the convenersintroduction to the invited papers for the Symposiumon Comparative Studies of Eastern Ocean BoundarySystems at the 1984 CalCOFI Conference. Certain ofthese introductory remarks may seem negative; theirpurpose is to indicate that at this stage in the develop-ment of fisheries/environmental science, the likeli-hood that there is no such thing as an exact analogyamong large marine ecosystems does not rule outinterregional comparison as a useful deductive tool. Inusing interregional comparison as a gauge for validityof findings, inconsistency with an interregional pat-tern is of course no proof of spuriousness; however,

interregional consistency may provide otherwise un-available support for a statistically weak but potential-ly useful result.

In view of the great resources required for experi-ments covering the range of time and space in recruit-ment variability of a fish stock, the cost-effectivenessof the comparative approach deserves mention. Inmany cases the observational investment will alreadyhave been made. Maritime reports (such as those usedin the construction of all the figures in this paper) areavailable, at varying density, for all the oceans of theworld. Some fishery research studies will have been

made in any area containing a substantial exploitedfish stock. Thus the costs of a rudimentary compara-tive study may be little more than those involved inacquiring existing data and arraying them in similarformats.

Fishery/environmental science clearly needs gener-alities and unifying principles with broad applicabil-ity. Consider, for example, the situation at the FA0Technical Consultation to Examine Changes inAbundance and Species Composition of Neritic Re-sources (held in Costa Rica in 1983) where, at a majorgathering of fishery scientists from all over the world,

no consensus could be found (e.g., Bakun 1983) as tothe probable consequences of continuing to take rec-ord sardine catches off northern Chile in the presenceof the 1982-83 El Nifio, even then recognized as thestrongest environmental anomaly in the eastern Pacificfor at least fifty years. We certainly expect that validgeneralities do exist in the fishery/environmental


24/26


field, as in other branches of science. Identifying themand making them available for fishery managementand industrial decision making should be given highpriority.

The following symposium papers present selectedcomparative aspects of the four major eastern oceanboundary ecosystems of the world and their associatedfisheries. No attempt has been made to assemble acomprehensive and coherent comparative treatment.Our purpose is to provide some examples of ongoingcomparative research, some useful insights into thedegree of analogy presented by these four similarsystems, and some basic information to aid compara-tive interpretation and deduction. Most important, wehope to stimulate fisheries scientists to maintain acomparative focus in their research efforts.

LITERATURE CITED

Anonymous. 1985. IOC-FA0 Guiding Group of Experts on the Pro-gramme of Ocean Science in Relation to Living Resources, Reports ofmeetings of experts and equivalent bodies SC-85/WS/18. Intergovern-mental Oceanographic Commission, Unesco, Paris, 35 p.

Bailey, K.M. 1981. Larval transport and recruitment of Pacific hake

Merluccius producrus. Mar. Ecol. Prog. Ser. 6:l-9.Bakun, A. 1978. Guinea Current upwelling. Nature 271:147-150.-. 1983. Part 111-Report of the working group on environmental

studies and monitoring. In G.D. Sharp and J. Csirke (eds.), Reports ofthe expert consultation to examine changes in abundance and species

composition of neritic fish resources, San Jose, Costa Rica, April 1983.FA0 Fish. Rep. 291(1):41-54.

Bakun, A,, and R.H. Parrish. 1980. Environmental inputs to fisherypopulation models for eastern boundary current regions. In G.D. Sharp(ed.), Workshop on the effects of environmental variation on the surviv-al of larval pelagic fishes, Lima, Peru, April-May, 1980. IOC WorkshopRep. 28, Unesco, Paris, p. 67-104.

-. 1982. Turbulence, transport, and pelagic fish in the Californiaand Peru Current systems. CalCOFl Rep. 23:99-112.

Belveze, H., and K. Erzini. 1983. The influence of hydroclimatic factorsof the availability of the sardine (Sardina pilchardus Walbaum) in theMoroccan Atlantic fishery. In G.D. Sharp and I. Csirke (eds.), Proceed-ings of the expert consultation to examine changes in abundance and

species composition of neritic fish resources, San JosC, Costa Rica,April 1973. FA0 Fish. Rep. 291(2):285-327.

Beyer, J.E. 1981, Aquatic ecosystems-an operational research approach.Univ. Wash. Press, Seattle, 315 p.


25/26

Binet, D. 1983. Influence des variations climatiques sur la pkherie desSardinella aurita ivoiro-ghaneennes: relation sicheresse-surpeche,Oceanologica Acta 5(4):443-452.

Boyd, A.J. 1979. A relationship between sea surface temperature variabil-ity and anchovy Engraulis capensis recruitment of Southwest Africa,1970-1978. Fish. Bull. S. Afr. 12:80-84.

Collins, R.A., and A.D. MacCall. 1977. The California Pacific Bonitoresource, its status and management. Calif. Dept. Fish Game, Mar. Res.Tech. Rep. 35, 39 p.

Freon, P. 1983. Production models as applied to sub-stocks depending onupwelling fluctuations. In G.D. Sharp and J. Csirke (eds.), Proceedingsof the expert consultation to examine changes in abundance and speciescornposition of neritic fish resources, San Jose, Costa Rica, April 1983.FA0 Fish. Rep. 291(3):1047-1064.

Lasker, R. 1981. Factors contributing to variable recruitment of the north-ern anchovy (Engraulis mordax) in the California Current: contrastingyears, 1975 through 1978. In R. Lasker and K. Sherman (eds.), Theearly life history of fish: recent studies. The Second ICES Symposium,Woods Hole, Mass., April 1979. Rapp. P.-v. Riun. Cons. Int. Explor.

Mer 178:375-378.

MacCall, A.D. 1980. The consequences of cannibalism in the stock-recruitment relationship of planktivorous pelagic fishes such as En-graulis. In G.D. Sharp (ed.), Workshop on the effects of environmentalvariation on the survival of larval pelagic fishes, Lima, Peru, April-May, 1980. IOC Workshop Rep. 28, Unesco, Paris, p. 201-220.

Marine Research Committee. 1950. California Cooperative Sardine Re-search Program progress report 1950. California State Printing Office,54 p.

Mayr, E. 1982. The growth of biological thought. Harvard Univ. Press,

Cambridge, Mass., 974 p.

McLain, D.R., R. Brainard, and J. Norton. 1985. Anomalous warm eventsin eastern boundary current systems. CalCOFl Rep. 26:(this volume).

Parrish, R.H., and A.D. MacCall. 1978. Climatic variation and exploita-tion in the Pacific mackerel fishery. Calif. Dept. Fish Game, Fish. Bull.167, 109 p.

Parrish, R.H., A. Bakun, D.M. Husby, C.S. Nelson. 1983. Comparativeclimatology of selected environmental factors in relation to easternboundary pelagic fish reproduction. In G.D. Sharp and J. Csirke (eds.),Proceedings of the expert consultation to examine changes in abundance

and species composition of neritic fish resources, San Jose, Costa Rica,April 1983. FA0 Fish. Rep. 291(3):731-777.

Shelton, P.A., A.J. Boyd, and M.J. Armstrong. 1985. The influence oflarge-scale environmental processes on neritic fish populations in theBenguela Current system. CalCOFI Rep. 26:(this volume).


26/26

recruitment generalization

Documents