Previous Page Table of Contents Next Page

Studies on the population structure of skipjack tuna, Katsuwonus pelamis, in the central and eastern Pacific Ocean: Signs of interaction potential using environmental data

James N. Ianelli
National Marine Fisheries Service, Building 4
7600 Sand Point Way, NE
Seattle, Washington 98115 USA


In the past decade, research on the population dynamics of skipjack tuna (Katsuwonus pelamis) has shifted from questions on stock productivity to concerns about interaction among fisheries, including the intense and growing fisheries in the eastern Pacific Ocean (EPO), where few if any skipjack spawn. In this region market pressure will limit the catch of yellowfin tuna (Thunnus albacares) associated with dolphins, hence, fishing effort on the skipjack resources is likely to grow. The impact of increased harvest is difficult to assess because the mixing of fish between areas is poorly understood.

In this paper, the population structure of skipjack and the potential for fishery interaction between the EPO and Hawaii, were evaluated. The main objective of this study was to present an alternative method of studying the stock structure and interaction potential of skipjack tuna in the Pacific Ocean. Previously, the chemical composition of skipjack otoliths was analysed using electron-beam microprobe technology. Chemical analyses of different growth zones on the otolith structure contain information on the potential origins of skipjack caught in different regions. However, the usefulness of the results for fisheries management purposes remains open to interpretation. Analyses of skipjack population structure using fishery-generated length-frequency data allowed tests of hypotheses on skipjack population structure through models of seasonal recruitment patterns. The recruitment patterns indicated, however, that the majority of the catch in the central and eastern tropical Pacific originated from broods spawned at different times of the year. The majority of skipjack production for the respective fisheries is based on a single stock that spawns over an extended period in the central Pacific. The seasonal recruitment pattern observed may be due to an extended spawning season with eastward migration of the larval and juvenile stages following seasonal extensions of the equatorial water masses. Alternatively, relatively distinct stocks with different peak spawning periods may explain the observed seasonality of brood events. To examine these recruitment hypotheses, oceanographic data were compiled to show seasonal habitat shifts. I believe that the underlying environmental variability both within and between years plays an important role in the movement of skipjack, and consequently, interaction potential between fisheries in different regions.


The world-wide landings of tuna have doubled in each decade since 1950 and exceeded 2 million mt per year in the 1980s. Skipjack tuna (Katsuwonus pelamis, hereinafter referred to as skipjack) have made up a significant and growing share of the total tuna harvest. In the 1960s, skipjack represented about 25% of the world tuna landings; in recent years skipjack comprise over 40% of the total (FAO, 1975; 1991). The skipjack tuna fisheries operating in the Pacific Ocean are primarily responsible for these increases, taking more than 80% of the skipjack harvest.

Early stock assessments of skipjack dealt with issues of maximum sustainable yield and optimal fishing strategies on a yield-per-recruit basis (e.g., Schaefer, 1954; Joseph and Calkins, 1969). As more information was collected and analysed it became apparent that skipjack were widely distributed and that skipjack fisheries in many areas depended to a large extent upon immigration of fish from elsewhere. This realisation has drawn attention to the problem of understanding the skipjack population structure. For example, regarding skipjack assessments the IATTC (1991) stated that:

"... [stock production methods] are not applicable unless the fish in question belongs to a discrete population, i.e. there is relatively little interchange between the fish in the area under consideration, in this case the eastern Pacific Ocean, and those in other areas where the species in question is exploited, or the exchange rates among subpopulations are known.... In the absence of definition of the population being exploited, production modeling is of little or no value for skipjack in the eastern Pacific fishery."
Reliable fisheries management may not be possible without understanding the geographic range of the reproducing segment of the stock. Once the geographic distribution of the stock(s) and movements are better understood, a set of hypotheses on biological productivity can be evaluated.

In addition to biological stock assessment issues related to overall productivity and resilience to fishing, questions regarding the effect of fisheries within national jurisdictions or areas on fishing conditions in other areas are currently being asked. This problem is commonly referred to as a form of fishery interaction. Because much of the skipjack now come under jurisdiction of coastal states declaring 200 mile economic zones, the potential for interactions among different fisheries has increased. Assessment of interaction requires quantification of the amount of exchange and/or interdependence of fish between areas. Knowledge of the spatio-temporal aspects of skipjack life history stages for the stock or stocks in question is also required.

One approach to evaluate the spatio-temporal patterns of skipjack life history stages is to examine age or size-specific recruitment to different areas and months. Ianelli (1993) examined this in detail and found that skipjack recruitment in the eastern Pacific Ocean (EPO) and Hawaii were variable yet revealed an underlying seasonal pattern (Figure 1). The observed recruitment patterns in these areas may be due to seasonal spawning that occurs locally or in adjacent waters or to seasonal oceanographic events superimposed on a stock that spawns in the central Pacific with relatively uniform intensity during all months of the year.

An analysis of the physical environment of skipjack was needed to examine these possibilities. The ideal way to describe the spatial dynamics of skipjack population behaviour would be to have direct observations on egg, larvae, and juvenile abundance over a broad geographic region and long time period. Such data do not exist and we rely on synoptic studies such as Nishikawa et al. (1985) and Argue et al. (1983) for information on the general distribution of skipjack in early life stages. It is reasonable to propose a relationship between skipjack larval distribution and oceanographic conditions based on these and other studies. As water temperature is the most widely available type of oceanographic data, I used it as a proxy for examining the dynamics of skipjack spawning and larval-development habitat.

In this paper, I examine oceanographic data to evaluate the relationship between the physical environment and observed skipjack recruitment patterns. The analysis is organised in two parts. First, oceanographic data on mean monthly conditions in an average year are compiled. This type of information is commonly referred to as a "climatology." Second, a time series of oceanographic data was used to examine interannual variation in sea surface temperature. The time series data included monthly mean sea surface temperatures for the period 1955-89 while the climatology data contained both sea surface temperature and temperature-at-depth data. In both cases, the physical variables were qualitatively compared with the recruitment patterns from the fishery data.

Figure 1. Lack of fit profile to fishery size composition data for three areas showing differences in patterns of cohorts between the central and eastern Pacific. This figure indicates that the model fit the data best with recruitment of the first age class (defined to have a mean length of 40 cm) occurring from March to June in Hawaii and from November to December in the EPO. For further details see Ianelli (1993).


This section examines oceanographic conditions on a seasonal basis and makes inferences on the spawning/larval habitat of skipjack. The goal is to determine if the dynamics of the physical habitat are consistent with the recruitment patterns observed and to gain insight on the role these dynamics may have on the distribution and movement of skipjack. Given the seasonal recruitment patterns in the EPO and Hawaii presented above, several propositions can be constructed about the relationship between regional spawning activity and movement potential between the regions (Table 1).

Table 1. Alternative hypotheses on spawning and movement conditions that support the observed seasonal fishery recruitment patterns.

Hypothesis (case)

Central Equatorial Pacific Spawning

Recruitment Mechanism

Spawning in the EPO or Hawaii














Seasonal or Uniform



The recruitment mechanism column of Table 1 refers to a physical process that provides an increased opportunity for skipjack to move between regions. Case 1 states that the cause of the recruitment pattern observed in the fishery data is due to spawning in the central Pacific. With a uniform intra-annual recruitment pattern, these skipjack are equally likely to move into the EPO or Hawaii at all times of the year but would pulse in because of the seasonal nature of central Pacific spawning. Cases 2 and 3 state that the majority of skipjack spawning takes place in the central Pacific, but that the opportunity to move between regions is not equal at all times of year. Under these hypotheses, seasonal oceanographic events create a "window" of recruitment opportunity for skipjack to move into the EPO and Hawaii from the central equatorial Pacific. Under Case 4, movement between regions is negligible and the majority of harvested skipjack represent products of local, seasonal spawning. An examination of the oceanographic dynamics can help evaluate each of these scenarios.

Temperature information was selected as the primary habitat indicator for skipjack spawning and/or recruitment mechanism. The abundance of skipjack larvae appears to be related to ambient water temperature (Matsumoto et al., 1984). The highest incidence of skipjack larvae occurs in surface waters from 23-28°. The presence of larvae indicates spawning in the vicinity (skipjack complete all larval stages in less than two weeks), thus examination of seasonal changes in sea surface temperature might suggest a seasonal recruitment mechanism. Alternatively, because skipjack larval stages are short and skipjack are found throughout the water column after metamorphosis, an investigation of mean temperature from the surface to 100 m was also pursued.

The sea surface monthly average temperature was based on the COADS (Woodruff et al., 1987) data set averaged over 1955-89. The monthly average temperature at depth was based on Levitus' (1982) data and was collated from 1946-81. The geographic resolution of the data were averaged over 2° latitude by 2° longitude grids. The software package FERRET1 specialises in manipulating gridded data sets and was used throughout this analysis.

1 Provided by the Pacific Marine Environment Laboratory, NOAA, 7600 Sand Point Way, Seattle WA 98115.
The maps of average sea surface temperature by month showed distinct north-south shifts in addition to changes in the shape and size of the tropical warm-water mass (Figure 2). From February to May a warm water band (>24°C) between the EPO and central Pacific is broadest. From August to November this band becomes quite narrow. This corresponds to periods of relative weak and strong flows of the equatorial counter-current and seasonal variability in wind-driven equatorial upwelling. In Hawaii the sea surface temperature is usually warmer than 24° C and the warmest temperatures typically occurring in August - October. The general pattern is thus one of seasonal north-south shifts in warm surface waters, along with expansion and contraction of the warm water band between the EPO and central Pacific.

The mean water temperature from the surface to 100 m showed a different seasonal pattern than the sea surface temperature plots (Figure 3). Rather than a shift in the area of the warm-water band expanding into the EPO, there is a discontinuity of the central Pacific water mass extending into the EPO. This reflects the fact that the thermocline is much shallower in the eastern Pacific (Fiedler, 1992). Seasonally, there is an east-west shift in the temperature gradient due to local changes in thermocline depth. From February to June a pocket of "central Pacific water mass" develops and extends well into the EPO between 100-120°W longitude and 10-20°N latitude. This pocket subsequently recedes to the west and a band of relatively cold water persists east of 110°W from September to January. A similar pattern occurs south of the equator. In this region, however, the pocket of water is much smaller and less persistent. In Hawaii, the mean 0-100 m temperature plots show seasonal north-south shifts similar to the sea surface temperature figures.

To examine the magnitude of average seasonal differences between these two physical variables, areas representing the EPO, central equatorial Pacific (CEP), and Hawaii were selected for closer analysis (Figure 4). Within these areas, average temperatures by month indicated different patterns between regions (Figure 5). The sea surface temperature for the CEP region was always warmer (being closest to the equator), followed by the EPO and the Hawaiian region. In the Hawaiian region, as expected, the seasonal variability in sea surface temperature was greater than in the EPO region (CV=4% vs. 2% in the EPO). An opposite pattern was apparent for the average 0-100 m temperature data: the EPO region was always colder than the Hawaiian region, on average. Also the seasonal variability in the Hawaiian region temperature-at-depth data was lower than in the EPO (CV=3% vs. 5% in the EPO). The seasonal variability was lowest in the CEP for both measures of temperature (CVs<1%). The mean temperature over 0-100m depth in the EPO region exhibited more seasonal variability than did the average surface temperature in the Hawaiian region (even though the EPO region was closer to the equator).

If temperature is a valid proxy for skipjack spawning and larval habitat and assuming the single-stock hypothesis is true, then the oceanographic data suggested mechanisms for the observed seasonal pattern of skipjack recruitment to the EPO and Hawaii. The opportunity for juveniles and skipjack in spawning condition to move to the EPO would be highest during periods when the warm-water band was broadest. If habitat is adequately indexed using temperature data, then having a larger "preferred" habitat during certain times of year may increase the chance for skipjack to move between regions. The mean temperature-at-depth data revealed more dynamic seasonal east-west processes than did the average sea surface temperature data. The mean temperature-at-depth figures characterised the habitat size along 3 dimensions, hence, it may have more accurately reflected habitat volume. As with the sea surface temperature data, an argument could be made that seasonal shifts in the physical environment may influence the incursion of skipjack from the central Pacific to the EPO and Hawaii.

With respect to recruitment timing estimated by Ianelli (1993), true age of the first age group was not assumed, consequently back calculation to month of spawning requires an added assumption. Assuming that the first age group represented 12-month old fish, then the peak spawning period of skipjack occurs about 3-4 months prior to the warmest times for both the EPO and Hawaii. Perhaps the most significant result from this is that the difference in the seasonal peak of skipjack recruitment in the EPO and Hawaii is roughly the same as the difference between peak periods of warm water for these regions. The peak temperatures are 6 months out of phase between regions as are the peak estimated recruitment periods. This suggests a link between the physical environment and the observed recruitment pattern to the fishery.


The analysis presented in the previous section involved comparisons of monthly averages combined over several years with seasonal recruitment patterns observed from fishery data. Here, results from Ianelli (1993) are used to examine interannual variability of skipjack recruitment to the EPO. The estimated proportion at age of skipjack harvested in the northern portion of the EPO by month was compared with a time series of sea surface temperature anomalies. The purpose of this exercise was to determine if the added information of sea surface temperature anomalies helped explain the observed variability in fishery recruitment data.

Figure 2. Mean monthly sea surface temperature based on COADS data (Woodruff et al., 1987) data. The solid line represents the 24° isotherm.

Figure 3. Mean monthly temperature from 0-100 m depth based on Levitus' (1982) data. The solid line represents the 24° isotherm.

Figure 4. Locations selected for closer examination of seasonal and inter-annual patterns of physical variables and fishery catch data.

Figure 5. Seasonal temperature profiles using surface temperatures and average temperature to 100m depth by area. Refer to Figure 4. for the areas presented. ETP=EPO.

Mean Sea Surface Temperature

Mean Temperature 0-100m Depth

A monthly time series of sea surface temperature data was compiled from the COADS database from 1955 to 1990. First, the mean monthly sea surface temperature by 2° geographic grid was calculated over all years of observations. Then the anomalies were computed by subtracting the mean monthly values from each month-year observation. To capture the impact sea surface temperature anomalies may have on recruitment patterns to the EPO, the data were condensed as average anomalies for a geographic grid between 10-30°N and 120-130°W. This region was selected because it is longitudinally adjacent (primarily to the west of) the area where the fishery occurs. Also, this region is an area where the recruitment to the EPO is likely to take place based on the temperature pattern. The resulting time series (Figure 6) indicated a significant positive autocorrelation at about 50 months or slightly more than 4 years (Figure 7). This roughly coincides with the average period between El Niño events of 3.8 years as determined by Quinn et al. (1987) over the period 1803-1987. Also evident in this time series is a shift to warmer conditions in the mid-1970s. A similar shift has been observed by various authors (e.g., Hare and Francis, 1995; Royer, 1993; Hollowed and Wooster, 1992) in different regions and in the Pacific.

Catch and catch rate index by year class were computed using monthly estimates of proportion at age (pijk) from Ianelli (1993). Catch (cik) and standardised effort (eik) data were supplied by the IATTC in 5° latitude x 5° longitude strata. The quantities were computed as:

where Cij is the catch of age j individuals in year i, and for year class t,

Similarly, a CPUE index was calculated as:


The time series of harvests and year-class index (CPUE) is presented in Figure 8. The harvest by year-class and CPUE index are closely related. Comparing these indices with the temperature anomalies is suggestive of a positive relationship, however the correlation is not significant (Figures 9 and 10). Part of the reason for the poor relationship may be that during the 1980s, there were fishery-driven factors not accounted for in the CPUE index. During this period, the fleet was changing rapidly and many high-performance vessels were leaving the EPO to fish in the western Pacific. Also, the change may have been due to the multispecies nature of the fishery. During this period, the catch rates for yellowfin tuna, the target species for this region, reached record highs levels.


Analyses of the seasonal aspects of the physical environment provided insights into the potential spawning patterns and recruitment mechanisms of skipjack. In the CEP the environmental conditions are seasonally stable, hence the spawning is not likely to be limited by water temperature. Skipjack larvae and juveniles from this region have been sampled in abundance during all quarters of the year (Nishikawa et al., 1985). Spawning in this region is therefore not likely to be influenced by seasons and Case 1 (Table 1) is also unlikely.

Figure 6. Monthly time series of data on the average sea surface temperature (SST) anomalies between 10-30°N and 120-130°W. (Data source: COADS, Woodruff et al., 1987).

Figure 7. Autocorrelation function for the sea surface temperature anomaly time series.

Figure 8. Plots of the annual time series of data on the month-adjusted year-class harvest of skipjack in the northern EPO.

Figure 9. Estimated index of annual skipjack cohorts and mean annual sea surface temperature anomalies plotted over time for the EPO north of 10°N. The open squares correspond to points in which catch may have been influenced by changes in the fishery.

Figure 10. Estimated cumulative catch of skipjack cohorts versus mean annual sea surface temperature anomalies in the EPO north of 10°N. The open squares correspond to points in which catch may have been influenced by changes in the fishery.

The sea surface temperature and mean 0-100 m temperature data provided evidence of possible seasonal recruitment mechanisms between the CEP and the fishery regions. Of the hypotheses presented in Table 1, the argument for Case 3 is the strongest, considering the available information. If skipjack in the EPO and Hawaii came from seasonal spawning events in the CEP, then one would expect to see a similar recruitment pattern in these regions. The observed recruitment pattern between the EPO and Hawaii and the seasonal temperature peaks in these regions were both estimated to be six months out of phase. The underlying spawning pattern may not be strictly uniform, on average, over all months of the year. There may be two peak spawning periods operating on geographically distinct segments of the same stock. This gives conditional support for Case 2.

Recently, a concentration of skipjack larvae was found in the northern EPO (IATTC, pers. comm.). Based on the variables examined, there appears to be favourable spawning conditions in the EPO during the period February-May. This gives support to Case 4 (Table 1). The size of this region is restricted, however, and the extent to which spawning in the EPO contributes to the adult population may thus be relatively small. Based on the otolith analyses presented in Ianelli (1993), the skipjack from the EPO were similar to those from Hawaii during the early-life stages. If these samples represented typical harvests, then it is unlikely that skipjack spawning in local regions contributes substantially to subsequent recruitment. This conclusion is further corroborated by the general increase in abundance of skipjack larvae found toward the western Pacific (Matsumoto et al., 1984) and by a variety of earlier studies (e.g., Klawe, 1963; Rothschild, 1965; Joseph and Calkins, 1969; Williams, 1972).

Understanding processes that affect interannual recruitment variability is one of the more active areas of fisheries research. The role of the environment in recruitment variability has been the subject of debate since the early days of fishery science (Smith, 1988). The physical habitat of skipjack shows significant intra-annual spatial variability, however, the variability between years is greater. For example, the coefficient of variation for sea surface temperature between years was 5% versus the 2% observed for seasonal or intra-annual variability measure. The same can be said about recruitment of skipjack into the EPO and Hawaii. Based on the catch rate data presented, there are large interannual variations in skipjack abundance. Several oceanographic processes may contribute to the interannual variability not accounted for in this analysis. Forsbergh (1989) examined many physical variables and found an index of wind speed from the presumed spawning area to have the strongest relationship. In the present study, fewer physical variables were compared, however, the observations covered a broad area and allowed for a clear descriptive view on some aspects of the physical dynamics of skipjack habitat.

The degree to which the environment affects movement of skipjack between major oceanic regions is central to understanding their population structure. Quinn and Brodeur (1991) reviewed the variability of movement patterns for several different species of marine animals. They concluded that for many species environmental conditions influenced movement, but that responses to oceanographic conditions alone could not explain a large part of movement pattern variability. Based on analyses presented in this study, there appears to be a close link between skipjack spatial distribution and the dynamics of their habitat.

This study has focused on ways of analysing the stock structure of skipjack harvested in Hawaii and the eastern tropical Pacific. These areas were selected because of potential interaction between the fisheries in these regions and because of the effect of skipjack stock structure uncertainties on assessments. In addition to extensive oceanographic data, a relatively long time series of fishery data was available from both areas. This provided a unique opportunity to examine alternative hypotheses on skipjack stock structure in an area where the environmental dynamics are likely to play an important role in their distribution.

Different species of tuna have many life history characteristics in common. Tunas generally spawn in warm tropical water and have similar developmental stage durations (Bayliff, 1980). They are pelagic, broadcast spawners with positively buoyant eggs. Fecundity in tunas is high, with little variability in egg size between or within different species. Virtually all species of tunas have the ability to thermoregulate, although the vascular system can be different between genera. Similarities in other life history aspects become more variable between species. For example, Kearney (1991) remarked that:

"... from a population dynamics perspective, southern bluefin tuna are more like lobsters than they are like skipjack."
From an ecological perspective, a natural ranking of tuna species can be made along the classic r-selected to K-selected spectrum (Table 2). An r-selected species is generally one that exhibits a high rate of intrinsic growth, reaches sexual maturity at a young age, has a small body size, and displays a relatively rapid population turnover. A K-selected species, on the other hand, typically grows slowly, reproduces at older ages, lives longer and can be characterised by having a relatively large number of age groups present in the population (Ricklefs, 1979). For both species of bluefin, the geographic area of spawning is restricted and spawning occurs over a relatively short period. Pacific albacore are divided into a northern and southern stock and have larger but well defined spawning areas. The yellowfin tuna spawning area is larger and overlaps considerably with the skipjack spawning area. Skipjack have the largest spawning area of the species listed which is consistent with their rank in absolute abundance (highest). The comparisons of different species of tunas provide insight on stock-structure characteristics. Weighing the evidence of stock-structure types, K-selected species are more likely to have well-defined stocks. For skipjack, being at the r-selected end of the spectrum is consistent with the lack of evidence for discrete populations. Arguably, the lack of evidence for discrete stocks of skipjack may be due to the complexities that are beyond our current powers of observation.

Table 2. Biological characteristics of selected species of Pacific tuna (from Bayliff 1980).

Tuna Species

Age at Maturity

Approx. Longevity

Maximum Length

Estimated Natural Mortality

Size of Spawning Area

Evidence of Discrete Stocks

Skipjack (Katsuwonus pelamis)

1 yr

3-6 yr

90 cm




Yellowfin (Thunnus albacares)

2-3 yr

10 yr

200 cm




Bigeye (T. obesus)

3 yr

10 yr

215 cm




Albacore (T. alalunga)

4-5 yr

10 yr

145 cm




Northern Bluefin (T. thynnus)

3-4 yr

10 yr

250 cm




Southern Bluefin (T. maccoyii)

5-6 yr

20 yr

220 cm





Studies using genetics, tagging, growth, distributional analyses, parasites, and catch rates have been conducted to elucidate skipjack stock structure in various regions of the Pacific (Table 3). Each method has weaknesses. Synthesising all information, the ability to reject or accept particular stock structure hypotheses is not great.

The physical habitat of skipjack (water temperature) was qualitatively compared with results from analyses of the fishery data. Sea surface temperatures were examined for intra-annual and interannual variability and compared with fishery data on the same scale. Additionally, mean 0-100 m temperature was used to examine the three-dimensional dynamics of skipjack habitat. Results from this study indicated that the mean temperature-at-depth data reveals more dynamic seasonal east-west processes than the average sea surface temperature data. These processes are more likely to play an important role in the recruitment of skipjack into the EPO. The interannual variability was greater than the intra-annual (seasonal) variability in sea surface temperature anomalies. The relationship between the mean anomalies and fishery year-class index suggested a weak but insignificant positive correlation.

Table 3. Summary of results of methods to investigate skipjack population structure and the implications for management (modified from Argue et al. (1986) and added to from Ianelli (1993)).


Results Pertaining to Population Structure

Management Utility

Blood genetics

SPC analyses support a clinal hypothesis, but it is unclear how the cline is maintained. Other investigators using different blood genetics data have inferred genetically isolated subpopulations.

Subjective and conflicting evidence of potential for fishery interactions.


No oceanic barriers to migration. Conditional support for clinal hypothesis.

Quantification of population parameters and fishery interactions.


Estimates of growth parameters vary in space and time suggesting a strong environmental rather than genetic effect.

Estimation of yield.

Juvenile distribution/maturity

Juvenile and spawning skipjack appear concentrated at longitudinal extremes of the study area. Conditional support for isolated subpopulation hypotheses.

Delineation of areas where spawning may be concentrated.


No evidence of discrete subpopulations.

Evidence of movement between tropical and subtropical waters.

Catch Rate Analyses

Suggests several discrete units related to apparent abundance concentrations.

Patchy distribution may relate to environmental conditions useful in forecasting.

Microconstituent analyses

Similarities between eastern Pacific and Hawaiian skipjack during early life stage but differences detected with ontogeny.

Evidence of potential for fishery interaction.

Length frequency analyses

Similar recruitment patterns between skipjack in the north and south EPO. These patterns differ from Hawaiian skipjack.

Recruitment of skipjack to the EPO and Hawaii differs in its seasonal periods suggesting that brood conditions tend to be independent.

Physical habitat

Descriptive overview of habitat dynamics indicating seasonal shift to both EPO and Hawaiian archipelago.

Evidence of potential for fishery interaction, useful for forecasting

Comparisons with other species

Suggestion that skipjack has a low potential for maintaining discrete stocks

Evidence of potential for fishery interaction.

From these and other analyses, two competing hypotheses of the skipjack stock structure in the eastern and central Pacific can be proposed. First, the majority of skipjack production for the respective fisheries is based on a single stock that spawns over an extended period in the central Pacific. The pattern of different cohort seasonality between the areas may be due to an extended spawning season where eastward migration of the larval and juvenile stages is based on seasonal extensions of the equatorial water masses. Second, relatively distinct stocks with different peak spawning periods in the same region may explain the observed seasonality of brood events. Arguments in favour of the latter hypothesis, however, require a high degree of spatial and temporal assortative behaviour between the stocks. This is not well supported by studies on the mixing rates between schools. Bayliff (1988) and Hilborn (1991) estimated that the individual turnover rate within schools is high. Thus, if schools from different stocks occur in the same region and mixed at the rates estimated, then they would have to somehow have to regroup themselves into respective stocks for reproduction. While such a feat would be remarkable, it seems unlikely to provide substantial gains in terms of overall reproductive fitness.

The question remains, what role does migration play in the stock structure of skipjack in the central and eastern Pacific? Fish migration can be treated like any other evolutionary adaptation; its function is to increase reproductive fitness. For skipjack to have populations that exhibit a migratory propensity, the benefits of doing so must be significant. There must be an energetic "reward" (increase in fitness) over the long term. Other tuna species that have well-defined population structures make non-spawning migrations. For example, in Japan the IATTC tagged several 3-4 month-old northern bluefin (Thunnus thynnus). Roughly one year later fish tagged within days of each other were recaptured on both sides of the Pacific. For some reason part of the young-of-the-year cohort travelled to the eastern Pacific while others remained in the waters off Japan. This suggests that non-spawning migration tendencies may not be inherited, or if they are, the quantitative genetic traits involved are highly variable. A similar argument could be made for EPO skipjack. I believe that skipjack that migrate to the EPO are not necessarily the progeny of adult skipjack that have also migrated to the EPO. There may be phenotypic traits that influence individual propensity for long migrations, but such individuals are not likely to constitute a stock. It is possible that a stock may have evolved to spawn during specific seasons to take advantage of environmental conditions conducive to eastward movement as juveniles and larvae.

Migration, homing, or the general tendency to disperse (i.e., diffusive behaviour) can affect the potential for interaction between fisheries. Different levels of homing and dispersion have implications for regional management effectiveness and for the potential for interaction between fisheries (Table 4). If the level of dispersal is low and the migration/homing propensity is low, then there is a moderate chance for effective regional management. Under all other scenarios, the potential for interaction between fisheries is high. The potential for interaction implies that exploitation rates in different areas are significant relative to the total population turnover rate. Ultimately, under the multiple or single stock hypothesis, the effectiveness of local management regimes will suffer without cooperative, transboundary harvest strategies and assessments.

Table 4. Summary of competing mechanisms for alternative hypothesis on the stock structure of skipjack in the Pacific and the likely impact on potential local management and fishery interaction (Modified from Kleiber et al., 1987).

Stock Structure Hypothesis

Level of Dispersal w/in Habitat

Migration/Homing Propensity

Potential for Effective Local Management

Potential for Fishery Interaction
















Multiple stocks




High in mixing zone, low in breeding areas

Multiple stocks




High in mixing zone and in breeding areas


The author thanks Drs. T. Quinn, R. Francis and R. Deriso for their advice and assistance during various stages of preparing this manuscript. The Inter-American Tropical Tuna Commission (IATTC) provided the primary support for this research.


Argue, A.W., F. Conand and D. Whyman. 1983. Spatial and temporal distributions of juvenile tunas from stomachs of tunas caught by pole-and-line gear in the central and western Pacific Ocean. Tech. Rep. Tuna Billfish Assess. Programme, S. Pac. Comm. 9: 47 p.

Argue, A.W., P. Kleiber, R.E. Kearney and J.R. Sibert. 1986. Evaluation of methods used by the South Pacific Commission for identification of skipjack population structure. In: Symons, P.E.K., P.M. Miyake, and G.T. Sakagawa (eds.). Proceedings of the ICCAT Conference on the International Skipjack Year Program; 21-29 June 1983, Tenerife, Canary Islands, Spain. ICCAT: 242-251.

Bayliff, W.H. (ed.). 1980. Synopsis of biological data on eight species of scombrids. Spec. Rep. I-ATTC 2: 530 p.

Bayliff, W.H. 1988. Integrity of schools of skipjack tuna, Katsuwonus pelamis, in the eastern Pacific Ocean, as determined from tagging data. Fish. Bull. NOAA-NMFS 86(4): 631-643.

FAO. 1975. Catches and landings, 1975. FAO Yearb. Fish. Statist. 40: 417 p.

FAO. 1991. Catches and landings, 1991. FAO Yearb. Fish. Statist. 72: 653 p.

Fiedler, P.C. 1992. Seasonal climatologies and variability of eastern tropical Pacific surface waters. NOAA Tech. Rept. NMFS 109: 65 p.

Forsbergh, E.D. 1989. The influence of some environmental variables on the apparent abundance of skipjack tuna, Katsuwonus pelamis, in the eastern Pacific Ocean. Bull. I-ATTC 19(6): 433-569.

Hare, S.R., and R.C. Francis. 1995. Climate change and salmon production in the Northeast Pacific Ocean. In: Beamish, R.J. (ed.). Climate change and northern fish populations. Spec. Publ. Can. J. Fish. Aquat. Sci. 121: 357-372.

Hilborn, R. 1991. Modeling the stability of fish schools: exchange of individual fish between schools of skipjack tuna (Katsuwonus pelamis). Can. J. Fish. Aquat. Sci. 48: 1081-1091.

Hollowed, A.B., and W.S. Wooster. 1992. Variability of winter ocean conditions and strong year classes of northeast Pacific groundfish. Symp. ICES J. Mar. Sci. 195: 433-444.

Ianelli, J.N. 1993. Studies on the population structure of skipjack tuna Katsuwonus pelamis in the central and eastern Pacific Ocean. Ph.D. Thesis. University of Washington, Seattle. 215 p.

IATTC. 1991. Annual report of the Inter-American Tropical Tuna Commission, 1989. Ann. Rep. I-ATTC: 270 p.

Joseph, J., and T.P. Calkins. 1969. Population dynamics of the skipjack tuna (Katsuwonus pelamis) of the eastern Pacific Ocean. Bull. I-ATTC 13(1): 1-273.

Kearney, R.E. 1991. Extremes in fish biology, population dynamics, and fisheries management: Pacific skipjack and southern bluefin tuna. Reviews Aquat. Sci. 4: 289-291.

Klawe, W. 1963. Observations on the spawning of four species of tuna (Neothunnus macropterus, Katsuwonus pelamis, Auxis thazard and Euthynnus lineatus) in the eastern Pacific Ocean, based on the distribution of their larvae and juveniles. Bull. I-ATTC 6(9): 449-540.

Kleiber, P., A.W. Argue and R.E. Kearney. 1987. Assessment of Pacific skipjack tuna (Katsuwonus pelamis) resources by estimating standing stock and components of population turnover from tagging data. Can. J. Fish. Aquat. Sci. 44(6): 1122-1134.

Levitus, S. 1982. Climatological Atlas of the World Ocean. NOAA/ERL GFDL Professional Paper 13, Princeton, New Jersey. 173 p.

Matsumoto, W.M., R.A. Skillman and A.E. Dizon. 1984. Synopsis of biological data on skipjack tuna, Katsuwonus pelamis. NOAA Tech. Rept. NMFS Circ. 451: 92 p. Also published as FAO Fish. Syn. 136: 92 p.

Nishikawa, Y., M. Honma, S. Ueyanagi and S. Kikawa. 1985. Average distribution of larvae of oceanic species of scombroid fishes, 1956-1981. Ser. Far Seas Fish. Res. Lab. 12: 99 p.

Quinn, T.P., and R.D. Brodeur. 1991. Intra-specific variations in the movement patterns of marine animals. Am. Zoo. 31: 231-241.

Quinn, W.H, V.T. Neal and S.E. Antunez de Mayolo. 1987. El Niño occurrences over the past four and a half centuries. J. Geophys. Res. 92(c13): 14,449-14,461.

Ricklefs, R.E. 1979. Ecology, 2nd Edition. Chiron Press. New York. 966 p.

Rothschild, B.J. 1965. Hypotheses on the origin of exploited skipjack tuna (Katsuwonus pelamis) in the eastern and central Pacific Ocean. Spec. Sci. Rep. U.S. Fish Wildl. Serv. (Fish.) 512: 20 p.

Royer, T.C. 1993. High latitude oceanic variability associated with the 18.6 year nodal tide. J. Geophys. Res. 98: 4639-4644.

Schaefer, M.B. 1954. Some aspects of the dynamics of populations important to the management of commercial marine fisheries. Bull. I-ATTC 1(2): 25-56.

Smith, T.D. 1988. Stock assessment methods: The first 50 years. In: Gulland, J.A. (ed.). Fish Population Dynamics: The Implications for Management, 2nd edition. John Wiley and Sons, New York. 422 p.

Williams, F. 1972. Consideration of three proposed models of the migration of young skipjack tuna (Katsuwonus pelamis) in the eastern Pacific Ocean. Fish. Bull. NOAA-NMFS 70(3): 741-762.

Woodruff, S.D., R.J. Slutz, R.L. Jenne and P.M. Steurer. 1987. A comprehensive ocean atmosphere data set. Bull. Am. Meteorol. Soc. 68(10): 1239-1250.

Previous Page Top of Page Next Page