sign in sign up
<<
Home , Atlantic bluefin tuna, Atlantic cod, Atlantic salmon, Cod, Eel, European perch

SCRS/2006/088 Col. Vol. Sci. Pap. ICCAT, 60(4): 1141-1153 (2007)

DO SOME ATLANTIC BLUEFIN SKIP SPAWNING?

David H. Secor1

SUMMARY

During the spawning season for , some adults occur outside known spawning centers, suggesting either unknown spawning regions, or fundamental errors in our current understanding of bluefin tuna reproductive schedules. Based upon recent scientific perspectives, skipped spawning (delayed maturation and non-annual spawning) is possibly prevalent in moderately long-lived marine like bluefin tuna. In principle, skipped spawning represents a trade-off between current and future reproduction. By foregoing reproduction, an individual can incur survival and growth benefits that accrue in deferred reproduction. Across a range of species, skipped reproduction was positively correlated with longevity, but for non- species, adults spawned at intervals at least once every two years. A range of types of skipped spawning (constant, younger, older, event skipping; and delays in first maturation) was modeled for the western Atlantic bluefin tuna population to test for their effects on the -production-per-recruit biological reference point (stipulated at 20% and 40%). With the exception of extreme delays in maturation, skipped spawning had relatively small effect in depressing mortality (F) threshold values. This was particularly true in comparison to scenarios of a juvenile (ages 4-7), which substantially depressed threshold F values. Indeed, recent F estimates for 1990-2002 western Atlantic bluefin tuna stock assessments were in excess of threshold F values when juvenile size classes were exploited. If western bluefin tuna are currently maturing at an older age than is currently assessed (i.e., 10 v. 8 years), then F thresholds for a juvenile-adult fishery decrease to very low levels (F20=0.13; F40=0.07). Because juveniles of either population are known to mix widely, recovery of western Atlantic bluefin tuna should depend upon sufficiently protective minimum size limits that are commonly applied throughout North Atlantic and Mediterranean .

RÉSUMÉ

Au cours de la saison de frai du thon rouge de l’Atlantique, la présence de certains adultes a été constatée à l’extérieur des zones de frai connues, ce qui donne à penser qu’il existe des zones de frai inconnues ou que nos connaissances actuelles sur l’évolution de la reproduction du thon rouge sont fondamentalement erronées. Sur la base de perspectives scientifiques récentes, l’omission de la fraye (retard de la maturation et fraye non annuelle) est possiblement un phénomène courant chez les espèces marines ayant une longévité modérée comme le thon rouge. En principe, l’omission de la fraye représente une compensation entre la reproduction actuelle et future. En ne réalisant pas la reproduction, un individu peut obtenir des bénéfices en terme de survie et de croissance provenant de la reproduction différée. Dans une gamme d’espèces, l’omission de la fraye a été corrélée positivement avec la longévité, mais pour les espèces autres que l’esturgeon, les adultes frayent à des intervalles d’au moins une fois tous les deux ans. Une gamme de divers types d’omission de la fraye (constante, à un âge plus précoce, à un âge plus avancé, omise ponctuellement et retards de la première maturation) a été modélisée pour la population du thon rouge de l’Atlantique ouest afin de tester ses effets sur les points de référence biologiques de la production d’œufs par recrue (EPR) (stipulés à 20% et 40%). A l’exception de retards extrêmes de la maturation, l’omission de la fraye n’avait un effet que relativement faible sur la réduction des valeurs seuil de la mortalité par pêche (F). Cela s’est avéré particulièrement vrai par rapport aux scénarios de pêcheries de juvéniles (âges 4-7), où les valeurs seuil de F ont été considérablement réduites. En réalité, les récentes estimations de F pour les évaluations de stock de thon rouge de l’Atlantique ouest de 1990-2002 dépassaient les valeurs seuil de F lorsque l’on exploitait les classes de taille de juvéniles. Si la maturation actuelle du thon rouge de l’ouest se produit à un âge plus avancé que ce qui est actuellement évalué (c’est-à-dire 10 v. 8 ans), alors les seuils de F pour une pêcherie de

1 Chesapeake Biological Laboratory, University of Maryland Center for Environmental Science, Solomons, Maryland 20688 , e- mail [email protected] 1141

juvéniles-adultes diminuent jusqu’à des niveaux très faibles (F20=0,13; F40=0,07). Etant donné que l’on connaît le mélange important des juvéniles des deux populations, le rétablissement du stock de thon rouge de l’Atlantique ouest devrait dépendre de limites de tailles minimales suffisamment protectrices qui sont généralement appliquées dans les pêcheries de l’Atlantique Nord et de la Méditerranée.

RESUMEN

Durante la temporada de puesta del atún rojo del Atlántico, algunos adultos desovan fuera de las zonas de desove conocidas, lo que sugiere o bien zonas de desove desconocidas o errores fundamentales en nuestros conocimientos actuales sobre la temporada reproductiva del atún rojo. Basándose en recientes perspectivas científicas, la omisión de la reproducción (maduración retardada y reproducción no anual) es posiblemente dominante en especies marinas de moderadamente larga vida como el atún rojo. En principio, la omisión de la reproducción es una compensación entre la reproducción actual y futura. Al no realizar la reproducción, un ejemplar puede obtener beneficios de crecimiento y supervivencia que se incrementan con el aplazamiento de la reproducción. En una gama de especies, se ha hallado una correlación positiva entre la omisión de la reproducción y la longevidad, pero para las especies distintas a los esturiones, los adultos desovan en intervalos de al menos una vez cada dos años. Se modeló una gama de tipos de omisión de la reproducción (constante, en etapas más jóvenes, en etapas adultas, omisión ocasional, y retrasos en la primera madurez) en la población de atún rojo del Atlántico occidental para probar sus efectos en los puntos de referencia biológicos de la producción de huevos por recluta (estipulados en 20% y 40%). Con la excepción de retrasos extremos en la madurez, la omisión de la reproducción tuvo un efecto relativamente escaso en reducir los valores umbral de la mortalidad por pesca (F). Esto era especialmente cierto en comparación con los escenarios de pesquería juvenil (edades 4-7) en la que los valores umbral de F descendieron significativamente. De hecho, las estimaciones recientes de F para las evaluaciones del stock de atún rojo del Atlántico occidental de 1990- 2002 superaban los valores umbral de F cuando se explotaban las clases de talla juvenil. Si actualmente el atún rojo occidental está madurando a una edad más avanzada que la estimada actualmente (a saber, 10 v. 8 años), entonces los umbrales F para una pesquería de adultos- juveniles descenderían hasta niveles muy bajos (F20 = 0,13; F40 = 0,07). Dado que se sabe que los juveniles de cada una de las poblaciones presentan un amplio nivel de mezcla, la recuperación del atún rojo del Atlántico occidental dependería de tallas mínimas suficientemente protectoras, que se aplican comúnmente en todas las pesquerías del Atlántico norte y Mediterráneo.

KEYWORDS

Bluefin tuna, reproductive behavior, spawning seasons

1. Bluefin tuna: Spawning in the wrong place at the wrong time?

During the spawning season (May-July), Atlantic bluefin tuna sometimes occur outside spawning centers (the [GOM] or ) in parts of the Central Atlantic and elsewhere. Electronic tagging has provided definitive evidence of this behavior. Lutcavage et al. (1999) relocated twelve large bluefin tuna (190-263 cm SFL) over a broad region of the Central-North Atlantic during May-July. Previous conventional tagging studies and subsequent intensive archival tagging efforts supported this mode of behavior in large bluefin tuna. Based upon continuous tracking of individuals, Block et al. (2005) observed that many adults (> 200 cm CFL) persisted in North American and central Atlantic waters during spring and summer months. This behavior prompts three hypotheses: (1) Spawning occurs in regions outside the GOM and Mediterranean (Lutcavage et al. 1999); (2) Bluefin tuna are becoming mature at larger sizes and older ages than previously observed or expected (Block et al. 2005); or (3) Bluefin tuna are not obligate annual spawners: they sometimes “skip” spawning. Several years of directed sampling in the central Atlantic have failed to find spawning condition bluefin tuna (ICCAT, 2007). In addition, collections in the central Atlantic have failed to recover larval bluefin tuna (Fromentin and Powers, 2005). Certainly, scientific sampling of unknown spawning regions in the Central Atlantic and elsewhere is a daunting task. Therefore it will be difficult to reject the hypothesis of alternate spawning areas for bluefin tuna. On the other hand, the second and third 1142

hypotheses have not received due treatment and may be the more logical inferences given the variability of reproductive schedules in marine . The purpose of this essay is to provide a prospectus for skipped spawning (SS), with delayed maturation as a special “case” of skipped spawning in Atlantic bluefin tuna.

This review was prompted by increased empirical evidence for SS and modeling exercises showing its consequence to population dynamics (Rideout et al. 2005; Jørgensen et al. 2006). Evidence supports the generality of SS, particularly for long-lived migratory fishes with large energetic demands for reproduction (Figure 1; Table 1). Here, SS refers narrowly to the coupled migration and spawning by bluefin tuna and other fishes, also known as non-annual spawning cycles. More broadly, SS also includes failed oogenesis due to retention and/or resorption of oocytes (Rideout et al. 2005). The focus here is on Atlantic bluefin tuna and other species that as adults sometime occur in alternate during spawning periods.

2. Which fishes skip spawning?

Skipped spawning implies an iteroparous life history, common among migratory marine fishes. Iteroparity is favored when larval and juvenile mortality is high and unpredictable (Charnov and Schaeffer, 1973; Schaefer and Elson, 1975). This reproductive strategy contributes resiliency through the storage effect (Warner and Chesson, 1985), where year-after-year recruitments are stored in the spawning stock (i.e., a seed bank) (Secor in press). SS is believed to result from a trade-off between current and future reproduction (Bull and Shine, 1979; Rideout et al. 2005). By foregoing spawning, an individual may incur survival and growth advantages in the current year to maximize its lifetime reproductive success. Therefore, SS is expected to be positively correlated with reproductive life span (Rideout et al. 2005).

Typically, the prevalence of SS is determined on the spawning ground based upon histological analysis of gonads, but this approach relates more specifically to failed egg ripening following a spawning migration rather than SS as it is defined here. Therefore, an attempt was made to represent taxa where SS was associated with a missed spawning migration or behavior. Approaches to identify missed spawning migrations include mark- recapture in sequential years (Schaefer and Elson, 1975; Quinn and Ross, 1985), comparison of abundances of adults in foraging and spawning habitats (Bell et al. 1992; Livingston et al. 1997), electronic tagging (Block et al. 2005), and microstructural or microchemical analysis of hardparts (Roussow, 1957; Zlokovitz et al. 2003; Engelhard and Heino 2005). In some instances, gross histological approaches were accepted because these were applied to determine non-annual spawning and related behaviors rather than failed egg ripening on the spawning grounds (Kennedy, 1953, 1954; Schwalme and Chouinard, 1999).

Building on Rideout’s et al.’s (2005) review, 21 marine and freshwater species were documented to exhibit SS, with species exhibiting a range of 9% to 86% SS on an annual basis (Figure 1). These species represent a diverse set of fishes; most of which are important to commercial and recreational fisheries or conservation goals. After log10 transformation of maximum ages (obtained from FishBase http://www.fishbase.net/ and Carey and Judge, 2002), a positive linear relationship was detected between %SS and longevity:

2 %SS = -25.76 + 46.97 log10 Longevity; n=21; P=0.001; r =0.45

Freshwater and marine taxa did not deviate in the association between %SS and longevity, but (Acipenseridae), well known for SS, represented extreme values for both longevity and %SS. Indeed, removal of sturgeon species resulted in a non-significant correlation (P=0.14). Note that for non-sturgeon species, SS was less than or equal to 50%, indicating that most fishes probably at intervals once every two years or less. According to the regression above, 40% SS is predicted for Atlantic bluefin tuna.

In an earlier literature review, Bull and Shine (1979) proposed that skipped reproduction in resulted from migration and other “accessory” costs of reproduction (brood care and live bearing). The latter type of reproductive costs was principally attributed to , reptiles and elasmobranchs. Birds are also known to skip reproduction (Cam and Monnat, 2000). Although parental care systems are diverse and frequent among fishes, migration is emphasized here as it is the principal accessory cost associated with reproduction by bluefin tuna and other migratory taxa. Indeed, bluefin tuna are unusual among oceanic pelagics due to their homing to specific spawning habitats. The spawning migrations to the Mediterranean and GOM are well known as seasonally directed behaviors (Fromentin and Powers, 2005). In the Mediterranean these spawning runs have been intercepted by fishermen for many centuries (Ravier and Fromentin, 2001). Similarly, spawning migrations by ( orientalis) to specific spawning centers in the western Pacific and Japan Sea are also well documented (Okiyama, 1974; Rooker et al. 2001). While other are known to range widely and 1143

follow predictable seasonal migration pathways, their spawning habitats are not as restricted in time and space as those of Atlantic and Pacific bluefin tunas. The migrations of these two species of temperate tunas surpass those exhibited by other homing (i.e., philopatric) species, with the exception of semelparous American and European (Table 1). We should therefore expect that migration entails a significant accessory cost for bluefin tuna, which could lead to SS or other changes to their reproductive schedule.

3. Types of skipping

The pattern of non-annual spawning has consequences to lifetime reproductive success and the effect of exploitation on population sustainability and recovery. Behaviors can range from a single missed spawning event early in life to extremely delayed maturity, emulating semelparity. Here, five classes of SS have been introduced that have relevance to stock assessment and exploitation thresholds.

1) Constant skipping: Constant skipping has been proposed for Atlantic and sturgeons. Underlying constant skipping is an energetic threshold, which remains approximately constant with size and environmental variations, but is not met by annual energetic supply. Thus, Jørgensen et al. (2006) modeled a fairly constant two-year cycle in spawning by Arctic , controlled by annual allocations of energy consumption to survival, migration, and reproduction. This same idea has been applied to lake sturgeon (Roussow, 1957), where a pattern of narrow annuli in sectioned fin spines was interpreted as a several-year period of energetic recovery following a spawning event. Constant spawning intervals have been applied to life table models approaches for estimating effects of exploitation on lifetime reproductive success (Beamesderfer et al. 1995; Secor et al. 2000; Gross et al. 2002) by decrementing annual fecundity by some fixed fraction (100-%SS). These models are insensitive to SS, because the unexploited population is affected by SS in the same way as is the exploited population. Still, if SS is occurring but not recognized, higher than expected recruitment at a given stock could be occurring (i.e., in observed stock recruitment curves, spawning stock represents total adult biomass unadjusted for %SS).

2) Younger skipping: The initial spawning migration and reproduction is hypothesized to require a disproportionate amount of stored energy in smaller first maturing individual than in older adults. Here, spawning might incur a fixed energetic cost so that a smaller sized individual is disproportionately burdened. Overlaying Jørgensen et al.’s (2006) approximate two-year cycle of spawning for Arctic cod, was an increased probability of occasional back-to-back annual spawning in older individuals. Maturation may also engender large physiological and behavioral transitions, which incur their own costs. In this instance, we should observe an abrupt change in %SS during the second year of the adult period. Evidence for the latter was observed based upon microchemical analysis of striped bass (Zlokovitz et al. 2003). A third explanation for younger skipping is related to patterns of Fisher’s fitness (Rideout et al. 2005). For younger ages, lifetime reproductive success can be increased through SS if survival and growth advantages translate into proportionate increased egg production later in life. It follows that for older ages there is less future reproductive potential to trade off against current reproduction. The effect of younger skipping could be important in assessing reproductive potential and exploitation thresholds. Age structure of the population will come into play, where %SS will be higher for a growing nascent population than for a population that is comprised of older adults. Because reproductive rates and values are proportionately higher at young adult ages, younger skipping will shift the distribution of egg production to older ages, requiring increased protection of age structure.

3) Older skipping: In opposition to younger skipping, this effect relates to spawning and accessory costs that scale proportionately to size. Such an effect could occur as large and old individuals approach their asymptotic size and can no longer generate sufficient surplus production due to increased basal metabolic demands (Pauly, 1981). This pattern has been observed in freshwater species (Trippel and Harvey, 1989; Holmgren, 2003) and has been suggested for sturgeons. In the instance of white and Atlantic sturgeons, large female carcasses have been observed during spawning season with no apparent agent of mortality (Blankenship, 1997; Veinott et al. 1999). One possible explanation is the inability to recoup anaerobic debt incurred by increased metabolism associated with spawning. For instance, for a 250 cm TL white sturgeon, full recovery of muscle glycogen stores is predicted to require over five days (Goolish, 1991). Increased mortality risk with size due to increased metabolic expenditures could result in conservative spawning energetic thresholds at older ages. The influence of older skipping will have much to do with the fecundity-fish size relationship. If egg production geometrically increases with fish size, older SS could diminish the effective spawning stock biomass below expected levels 1144

(again resulting in a higher recruitment level per spawner). Older skipping will shift reproductive rates to younger ages, leading to decreased generation times. Increased reproductive rates of younger fish would translate into increased effectiveness of minimum size limits in managing for recovery.

4) Event skipping: An episodic environmental event, such as an anomalously cold spawning season can cause an entire population to undergo SS for a single calendar year. Such events might be expected to occur more frequently in temperate and boreal freshwater systems that are less well buffered against weather and alterations. For instance, environmental hypoxia precluded spawning by white suckers in a Canadian Lake (Trippel and Harvey, 1989). Similarly, mud or rockslides could temporarily obstruct spawning runs in stream fishes. In marine fishes such as bluefin tuna, it seems more likely that event skipping would be associated with large scale weather/climate events. The influence of event skipping on lifetime reproductive success will be dependent upon lifespan and is probably minimal in moderately long lived species such as Atlantic bluefin tuna.

5) Delayed maturation: Age and size at maturity show high phenotypic plasticity and are known to be responsive to density and changing environments (Rjinsdorp, 1989; Roff, 1991). Yet, even within a population’s generation time, large variation is observed in age and size at maturity. For instance, although eastern Atlantic bluefin tuna mature by age 3 (Corriero et al. 2005), recent electronic tagging data by Block et al. (2005) suggests that some groups (contingents) of eastern Atlantic bluefin tuna may delay initial spawning until much later (age >8 years). Delaying maturation would seem related to the same trade-offs between current and future reproduction that are expected for skipped spawning thresholds. Alternatively, delayed maturation may indicate that a particular cohort or individual has never attained sufficient surplus production to enable a spawning migration and reproduction. Such “failed reproduction” has been suggested for European eels, which incur extreme accessory costs associated with their spawning migration (Tsukamoto, 1998; Secor, 2004). In freshwater systems, Trippel and Harvey (1989) observed an extremely old white sucker (21 years) that had not yet spawned, and attributed this extreme immaturity to a limiting energetic threshold. Lifetime reproductive success is extremely sensitive to age at maturation (Roff, 1995). Delaying maturation even a year or two is expected to substantially depress lifetime reproductive rate.

4. What are the consequences of skipped spawning?

To evaluate the effects of different types of SS on bluefin tuna sustainability, various scenarios of fecundity schedule, exploitation schedule and SS were compared based upon the Egg Production per Recruit (EPR) biological reference point. This rather elemental life-table approach is used to investigate the sensitivity of reproductive rates to changes in demographic schedules. EPR thresholds are typically included in the suite of biological reference points against which stock assessments are compared (Goodyear, 1989; Myers et al. 1994; Murawski et al. 2001). First, a value of lifetime egg production under no exploitation (E) is computed:

a E = ∑ N t ⋅ Lt ⋅ St t=1 where N is cohort abundance, L is fecundity, S is proportion spawning, t designates age class, and a is maximum age. Initial cohort size (t=1) was arbitrarily set to 100,000. Lifetime egg production under varying exploitation rates and SS scenarios are then compared to E as a fraction, termed % of maximum per recruit or %EPR (Boreman, 1997). Western Atlantic bluefin tuna was chosen as the focal population because it is expected to show greater sensitivity to SS than the eastern population and is under higher priority for recovery. Parameters and scenarios for the EPR analysis (Table 2) were simplified (i.e., assumptions of constant M, knife edge maturity and recruitment to fishery) to sharpen contrasts. %EPR was examined across a range of exploitation rates (F=0.1 to 1.1, or 10-67% annual fishing mortality rate).

Thresholds for %EPR are not stipulated but precedence exists for a convention of 50% EPR for fishes that have low life time reproductive rates such as sturgeons, whereas 20% may be a protective threshold for a moderately long-lived fecund marine species such as Atlantic cod or striped bass (Boreman, 1997). Here, 20% and 40%EPR are used to represent a range of appropriate biological thresholds for management related to either sustainability (F20) or recovery (F40) (Mace, 1994). The most recent fecundity schedule data for western Atlantic bluefin is that of Baglin and Rivas (1977) for samples collected in 1967. Their published regression predicted that a 25 year old bluefin tuna should be capable of spawning 75 million eggs. On the other hand the maximum observed fecundity 1145

was reported to be ~45 million eggs. Therefore, two fecundity schedules were examined (Figure 2). Because very little influence was detected between the two maximum egg production EPR curves, a maximal egg production of 75 million eggs was accepted as the baseline.

Most scenarios of SS had minimal effects on %EPR. As stated above, constant skipping produced the same EPR curve as the no skipping baseline condition and is therefore not shown. For 20% EPR, younger skipping resulted in a reduction in F20 of 20% below the baseline (Table 3; no juvenile exploitation). Later skipping resulted in an EPR curve that was indistinguishable from the baseline condition (Figure 2). Delaying reproduction for two years resulted in reductions in threshold F values similar to those observed for early skipping. Delaying reproduction for five years showed a much more substantial reduction in threshold F values (Table 3). Importantly, the effect of a juvenile fishery on F threshold values was substantially greater than all but the most extreme scenario of SS (delay 5 years). Further, the effect of SS was substantially diminished under scenarios of juvenile fishing, indicating the dominant effect of juvenile fishing. F threshold values across modeled scenarios were predominately lower than F estimates (ages 8+) from 1990 to 2002 western bluefin tuna stock assessments (ICCAT, 2003).

An important assumption that could bias results is that of constant mortality across all ages. Natural mortality is expected to be inversely related to size and age and this change in demographic schedule should affect threshold F values. To demonstrate this idea, I assumed that 50% of total lifetime mortality (Mlife=3.0) occurred during the first three years of life and assigned different constant mortalities for fish ages <4 (M=0.5) and those ages 4-25 (M=0.068). Somewhat surprisingly, this rather extreme (and unrealistic) scenario of varying survival schedules had little influence on threshold F values. This result suggests that the western Atlantic bluefin tuna’s long life span exhibits a buffering (aka “storage”) effect against changes to early juvenile natural mortality. Other more realistic scenarios of changing mortality rates may have differing effects, but under the principal assumption of a moderately long life span (25 years or greater), these effects on threshold F values are expected to be minor.

5. Skipping and mixing

Mixing rates and their influence on recovery of western Atlantic bluefin tuna is a fundamental challenge facing SCRS in their 2006 integrated bluefin tuna assessment (ICCAT, 2007). Recent evidence by Block et al. (2005) provides compelling support for the premise of two centers of spawning for Atlantic bluefin tuna with no straying of spawning adults between them. Clearly, the differences in reproductive schedules between eastern and western Atlantic bluefin tuna, coupled with large differences in abundance indicate that the eastern Atlantic bluefin tuna population is much more productive. We should expect therefore that eastern bluefin tuna population dynamics play a major role on the effects of mixed fisheries on the western population. Indeed, recent microconstituent research indicated approximately equivalent contributions of eastern and western bluefin tuna to U.S recreational fisheries (Rooker et al. 2006).

Skipped spawning was shown here to have somewhat minor effects on F thresholds, with the exception of delays in first maturation. Some evidence exists for this mode of behavior in the late maturing contingent of eastern Atlantic spawning fish observed by Block et al. (2005). Among SS scenarios, those emulating changes in first maturation would also warrant further examination in terms of the effect of mixing on threshold F values. In comparison to most SS scenarios, exploitation on juvenile bluefin tuna (ages 4-7) had a much greater influence on F thresholds. Thus, it would seem productive for SCRS to examine various scenarios of juvenile exploitation and mixing on EPR and other biological reference points. Here, because juveniles of either population mix widely, it seems inevitable that a single minimum size limit that is protective for recovery of the western Atlantic population must be applied to all mixed stock fisheries (western, central, eastern Atlantic, and Mediterranean Sea).

6. Preview of skipped spawning in the information age

Although skipped spawning and large variation in age at first spawning has yet to be ascribed to Atlantic bluefin tuna in the scientific literature, archival data records provide compelling evidence for their occurrence. By providing a complete record of geolocations for individuals, Block et al. (2005) observed that scores of fish occurred outside known spawning centers during spawning periods. This pattern was in part explained by inclusion of sub-adults (<200 cm CFL for western Atlantic bluefin tuna) in the data set. Several such individuals showed patterns of >1 year residency in the western Atlantic followed by presumed spawning forays to either the GOM or Mediterranean. Interestingly, at tagging many individuals destined to spawn after a >1 year interval in 1146

the Mediterranean were already of a size considered mature for the eastern Atlantic bluefin population (50% maturity=3 years; c. 105 cm CFL; Corriero et al. 2005). These individuals could represent a unique contingent of eastern Atlantic bluefin tuna and/or an example of SS (i.e., extreme immaturity). For larger adults, Block et al. (2005) observed geolocation records consistent with occurrence in either spawning center, but also observed a large fraction of adults persisted outside the spawning centers during spring and summer spawning months. Based upon a length distribution recently observed for spawners in the Gulf of Mexico, Block et al. (2005) suggested that many western Atlantic bluefin may initially spawn at a larger size and older age (>10 years) than previously observed or expected. Thus, they argued that presumed adults in their sample (>200 cm CFL) may in fact have been immature fish. If maturation were now 10 years or older, recent estimated fishing mortality rates (F=0.16 to 0.40) are likely substantially in excess of those expected to support sustainability at EPR=20% or 40% (F=0.07 to 0.21, depending upon juvenile mortality and fishing scenario; Table 3).

Arguably, electronic tagging studies for bluefin tunas (Lutcavage et al. 1999; Itoh et al. 2003; Block et al. 2005) are challenging previous held concepts on bluefin tuna biology and ecology. A principal challenge will be to summarize the rich datasets on individual behaviors, generated from such studies, into meaningful population- level modalities and parameters for improved conservation and management. For instance, electronic tagging studies may represent our best opportunity to quantify SS and other variations in reproduction schedule, but accumulating enough observations over multiple years currently limits population-level inferences. Advances in technology now permit more reliable tracking of geolocations over multi-year periods (B. Block, pers. comm.). Depending upon how electronic tags are deployed in current and future studies, important opportunities exist to better estimate reproductive schedules and model their relationship with population structure, migration patterns, climate, and exploitation.

Acknowledgements

Support for preparation of this paper came from the National Science Foundation (OCE-0324850). Drs. B. Block, R. Kraus, and J. Rooker provided valuable comments and feedback on an earlier version of this manuscript.

References

BAGLIN, R.E. and L.R. Rivas. 1977. Population fecundity of western and eastern North Atlantic bluefin tuna (Thunnus thynnus). Col. Vo.Sci. Pap. ICCAT, 6(2): 361-365. BEAMESDERFER, R.C.P., T.A. Rein and A.A. Nigro. 1995. Differences in the dynamics and potential production of impounded and unimpounded white sturgeon populations in the Lower Columbia River. Transactions of the American Fisheries Society 124: 857-872. BELL, J.D., J.M. Lyle, C.M. Bulman, K.J. Graham, G.M. Newton, D.C. Smith. 1992. Spatial variation in reproduction, and occurrence of non-reproductive adults, in , Hoplostethus atlanticus Collett (Trachthyidae), from South-Eastern Australia. Journal of Fish Biology 40 (1): 107-122. BLANKENSHIP, K. 1997. Ghosts from the past may still haunt bay waters. Bay Journal. Chesapeake Bay Program. November. 3 p. BLOCK, B.A., L.L.H. Teo, A. Walli, A. Boustany, M.J.W. Sokesbury, C.J. Farwell, K.C. Weng, H. Dewar and T.D. Williams. 2005. Electronic tagging and population structure of Atlantic bluefin tuna. Nature 434: 1121- 1127. BOREMAN, J. 1997. Sensitivity of North American sturgeons and paddlefish to fishing mortality. Environmental Biology of Fishes 48: 399-405. BULL, J.J., R. Shine. 1979. Iteroparous that skip opportunities for reproduction. American Naturalist 114 (2): 296-303. CAM, E., J.Y. Monnat. 2000. Apparent inferiority of first-time breeders in the kittiwake: the role of heterogeneity among age classes. Journal of Ecology 69 (3): 380-394. CAREY, J. R., D.S. Judge. 2002. Longevity records: Life Spans of Mammals, Birds, Amphibians, Reptiles, and Fish Monographs on Population Aging, 8 Odense University Press.

1147

CHARNOV, E.L., W.M. Schaffer. 1973. Life history consequences of natural selection: Cole’s result revisited. Am. Naturalist 107, 791-793. CORRIERO, A., S. Karakulak, N. Santamaria, M. Deflorio, D. Spedicato, P. Addis, S. Desantis, F. Cirillo, A. Fenech-Farrugia, R. Vassallo-Agius, J.M. de la Serna, Y. Oray, A. Cau, P. Megalofonou, G. De Metrio. 2005. Size and age at sexual maturity of female bluefin tuna (Thunnus thynnus L. 1758) from the Mediterranean Sea. Journal of Applied , 21 (6): 483-486. ENGELHARD, G.H., M. Heino. 2005. Scale analysis suggests frequent skipping of the second reproductive season in Atlantic . Biology Letters, 1 (2): 172-175. FREDRICH, F., S. Ohmann, B. Curio, F. Kirschbaum. 2003. Spawning migrations of the chub in the River Spree, . Journal of Fish Biology, 63 (3): 710-723. FROMENTIN, J.M., J.E. Powers. 2005. Atlantic bluefin tuna: population dynamics, ecology, fisheries and management. Fish and Fisheries, 6 (4): 281-306. GOODYEAR, C.P. 1990. Spawning stock biomass per recruit: The biological basis for a tool. Col. Vol. Sci. Pap. ICCAT, 32(2): 487-497. GOOLISH, E.M. 1991. Aerobic and anaerobic scaling in fish. Biological Review, 66: 33-56. GROSS, M.R., J. Repka, C.T. Robertson, D.H. Secor and W. Van Winkle. 2002. Sturgeon conservation: insights from elasticity analysis. Am. Fish. Soc. Sympos, 28: 3-10. HARDEN JONES, F.R. 1968. . St. Martin's Press, New York. 325 pp. HENSEL, K., J. Hol…ik. 1997. Past and current status of sturgeons in the upper and middle Danube River. Environmental Biology of Fishes, 48 (1-4): 185-200. HOLMGREN, K. 2003. Omitted spawning in compensatory-growing . Journal of Fish Biology, 62 (4): 918-927. HURLEY, P.C.F., T.D. Iles. 1987. Age and growth estimation of Atlantic bluefin tuna, Thunnus thynnus, using otoliths. In NOAA Tech. Report NMFS 8. Proceedings of the International Workshop on Age Determination of Ocean Pelagic Fishes: Tunas, , and . ICCAT. 2003. Report of the 2002 Atlantic Bluefin Tuna Stock Assessment Session. Col. Vol. Sci. Pap. 55(3): 710-937. ICCAT. 2007. Report of the 4th Meeting of the Working Group to Develop Integrated and Coordinated Atlantic Bluefin Tuna Management Strategies. In Report for Biennial Period, 2006-07, Part I. Vol. 1 (in press). ITOH, T., S. Tjsuji, A. Nitta. 2003. Migration of young Pacific bluefin tuna (Thunnus orientalis) determined with archival tags. Fish. Bull. 101: 514-534. JØRGENSEN, C., B. Ernande, O. Fiksen, U. Dieckmann. 2006. The logic of skipped spawning in fish. Canadian Journal of Fisheries and Aquatic Sciences. 63 (1): 200-211. KENNEDY, W.A. 1954. Growth, maturity, fecundity and mortality in the relatively unexploited lake , Cristivomer namaycush, of Great Slave Lake. J. Fish. Res. Brd. Can. 11: 827-852. KENNEDY, W.A. 1953. Growth, maturity, fecundity and mortality in the relatively unexploited whitefish, Coregonus clupeaformis, of Great Slave Lake. J. Fish. Res. Brd. Can. 10: 413-441. LIVINGSTON, M.E., M. Vignaux, K.A. Schofield. 1997. Estimating the annual proportion of nonspawning adults in hoki, Macruronus novaezelandiae. Fish. Bull. 95 (1): 99-113. LUTCAVAGE, M.E., R.W. Brill, G.B. Skomal, B.C. Chase, P.W. Howey. 1999. Results of pop-up tagging of spawning size class fish in the Gulf of : Do North Atlantic bluefin tuna spawn in the Mid-Atlantic? Canadian Journal of Fisheries and Aquatic Science. 56: 173-177. MACE, P.M. 1994. Relationships between common biological reference points used as thresholds and targets of fisheries management strategies. Canadian Journal of Fisheries and Aquatic Sciences. 51 (1): 110-122.

1148

MERRIMAN, D. 1941. Studies of the striped bass (Roccus saxatilis) of the Atlantic coast. USFWS Fish. Bull. 50: 1-77. MURAWSKI, S.A., P.J. Rago, E.A. Trippel. 2001. Impacts of demographic variation in spawning characteristics on reference points for fishery management. ICES Journal of Marine Science. 58 (5): 1002-1014. MYERS, R.A., A.A. Rosenberg, P.M. Mace, N. Barrowman, V.R. Restrepo. 1994. In search of thresholds for recruitment . ICES Journal of Marine Science. 51 (2): 191-205. OKIYAMA, M. 1974. Occurrence of the postlarvae of bluefin tuna, Thunnus thynnus, in the Japan Sea. Japan Sea Reg. Fish. Res. Lab. Bull. 25: 89-97. PAULY, D. 1981. The relationship between surface area and growth performance in fish: a generalization of von Bertanffy's theory of growth. Meereforsch 28: 251-282. PEARCY, W.G. 1992. Ocean Ecology of North Pacific Salmonids. Washington Sea Grant, Univ. WA Press, Seattle. , B.R. 1984. Relations Between migration, reproduction and nutrition in yellowfin bream Acanthopagrus australis. Marine Ecology-Progress Series. 19 (1-2): 17-23. PULLIAINEN, E., K. Korhonen. 1990. Seasonal-changes in condition indexes in adult mature and non-maturing burbot, Lota lota (L), in the North-Eastern , Northern . Journal of the Fish Biology. 36 (2): 251-259. QUINN, S.P., M.R. Ross. 1985. Non-annual spawning in the white sucker, Catostomus commersoni. Copeia. (3) 613-618. RAVIER, C., J.M. Fromentin. 2001. Long-term fluctuations in the eastern Atlantic and Meditteranean bluefin tuna population. ICES Journal of Marine Science. 8: 1299-1317. RIDEOUT, G.A., G.A. Rose, M.P.M. Burton. 2005. Skipped spawning in female iteroparous fishes. Fish and Fisheries. 6: 50-72. RIVAS, L.R. 1977. Age composition, seasonal distribution of age groups, and longevity of the western North Atlantic bluefin tuna (Thunnus thynnus). Col. Vo.Sci. Pap. ICCAT, 6 (2): 287-304. RJINSDORP, A.D. 1989. Maturation of male and female (Pleuronectes platessa L.). J. Cons. int. Explor. Mer 46: 35-51. ROFF, D.A. 1991. The evolution of life-history variation in fishes, with particular reference to . Journal of Sea Research 27: 197-207. ROFF, D.A. 1995. Life History Theory: A Framework. The Evolution of Life Histories, Chapman and Hall, NY. pp. 35-61. ROOKER, J.R., D.H. Secor, G. DeMetrio, E. Rodríquez-Marín. 2006. Evaluation of population structure and mixing rates of Atlantic bluefin tuna from chemical signatures in otoliths. Col. Vol. Sci. Pap. ICCAT, 59(3): 813-818. ROOKER, J. R., D. H. Secor, V. S. Zdanowicz and T. Itoh. 2001. Discrimination of northern bluefin tuna from nursery areas in the Pacific Ocean using otolith chemistry. Mar. Ecol. Prog. Ser. 218:275-282. ROUSSOW, G. 1957. Some considerations concerning sturgeon spawning periodicity. Canadian Journal of Fisheries and Aquatic Science 14: 553-572. SCHAEFER, W.M., Elson, P.F. 1975. The adaptive significance of variations in life history among local populations of Atlantic in North America. Ecology 56: 577-590. SCHWALME, K., G.A. Chouinard. 1999. Seasonal dynamics in feeding, organ weights, and reproductive maturation of Atlantic cod ( morhua) in the southern Gulf of St Lawrence. ICES Journal of Marine Science. 56 (3): 303-319. SECOR, D.H. The year-class phenomenon and the storage effect in marine fishes. J. Sea Res. (in press).

1149

SECOR, D.H. 2004. Fish migration and the unit stock: three formative debates. In Steven X. Cadrin, Kevin D. Friedland, John R. Waldman (eds.). Stock Identification Methods. Elsevier Inc., Burlington. P. 17-44. SECOR, D. H., V. Arefjev, A. Nikolaev, A. Sharov. 2000. Restoration of sturgeons: lessons from the Sturgeon Ranching Programme. Fish and Fisheries. 1:215-230. STORR-Paulsen, M., K. Weiland, H. Hovgard, H.J. Ratz, H.J. 2004. Stock structure of Atlantic cod (Gadus morhua) in West waters: implications of transport and migration. ICES Journal of Marine Science. 61 (6): 972-982. TESCH, E. W. 1977. The : Biology and management of Anguillid eels. Chapman and Hall Ltd. Halsted Press, New York. TRIPPEL, E.A., H.H. Harvey. 1989. Missing opportunities to reproduce- an energy-dependent or fecundity gaining strategy in white sucker (Catostomus commersoni. Canadian Journal of 67 (9): 2180-2188. TSUKAMOTO, K., I. Nakai, W.V. Tesch. 1998. Do all freshwater eels migrate? Nature 396: 635-636. TYUS, H. 1990. Potamodromy and reproduction of Colorado squawfish in the Green River basin, Colorado, Utah. Transactions of the American Fisheries Society. 119 (6): 1035-1047. WARNER, R.R., P.L. Chesson. 1985. Coexistence mediated by recruitment fluctuations: a field guide to the storage effect. Am. Nat. 125: 769-787. VEINOTT, G., T. Northcote, M. Rosenau, D. Evans. 1999. Concentrations of strontium in the pectoral fin rays of the white sturgeon (Acipenser transmontanus) by laser ablation sampling-inductively coupled plasma- mass spectrometry as an indicator of marine migrations. Can. J. Fish. Aquat. Sci. 56: 1981-1990. ZLOKOVITZ, E.R., D.H. Secor, and P.M. Piccoli. 2003. Patterns of migration in striped bass as determined by otolith microchemistry. Fisheries Research 63: 245-259.

Table 1. Spawning migrations for highly migratory species known to show homing to specific spawning habitats. Approximate maximal distances between foraging and spawning habitats are shown. Two semelparous species are shown in bold. Spawning Species Reference Migration (km) Anguilla anguilla 7000 Tesch, 1977 Pacific bluefin tuna Thunnus orientalis 6000 Itoh et al. 2003 Atlantic bluefin tuna Thunnus thynnus 5000 Block et al. 2005 Onchorynchus nerka 2500 Pearcy, 1992 Icelandic cod Gadus morhua 2000 Storr-Paulsen et al. 2004 sturgeon Huso huso 2000 Hensel and Hol…ik, 1997 Orange roughy Hoplostethus atlanticus 1500 Bell et al. 1992 harengus 1500 Harden Jones, 1968 Striped bass Morone saxatilis 1500 Merriman, 1941

1150

Table 2. Life table construction for Egg Production per Recruit comparisons for western Atlantic bluefin tuna. Fecundity regression from Baglin and Rivas (1977). Longevity and maturity schedule from Rivas 1977, Hurley and Iles, 1987, Fromentin and Powers, 2005 and Secor, pers. obs).

Baseline Max. fecundity fixed Juvenile fishery Longevity=25 yrs; M1-25=0.12; Baseline conditions, except Knife-edge recruitment Fecundity=1967 estimates for W. Atl. fecundity=1967 estimate for at age 4.0 yrs population, ages 8-25 yrs according to 1967 W. Alt. population but set age-dependent linear regression model, constant to 45 million at ages allowing maximum of 75 million at age 25 > 18 years yrs; annual spawning at ages > 7 years; knife edge recruitment to fishery at age 8.0 years

Skipping Scenarios Younger Skipping 50% Annual spawning for age 8-12 yrs Older Skipping 50% Annual spawning for ages 21-25 Skip after First Spawn 0% Annual spawning at age 9 Delay 2 Years First spawning at age 10 Delay 5 Years First spawning at age 13

Table 3. Biological reference points for annual fishing mortality (F) under scenarios of juvenile fishing and skipped spawning. Skipping Juvenile F@ %EPR=20 F@ %EPR=40 scenario fishery Constant M >Juvenile M Constant M >Juvenile M None (annual) No 0.28 0.22 0.13 0.11 Yes 0.15 0.13 0.08 0.07 Younger No 0.22 0.18 0.11 0.10 Skipping Yes 0.13 0.12 0.07 0.07 Delay 2 Years No 0.21 0.19 0.11 0.10 Yes 0.13 0.12 0.07 0.07 SCRS Range of 8+ Western Stock Fishing Mortality, 1990-2002 ~0.16 to 0.40

1151

Longevity (years) 1020 30 50 90 150 Lake sturgeon 80 Atlantic sturgeon 5.0 Russian sturgeon Beluga sturgeon 4.0 70 Spawning Interval (yr) Shortnose sturgeon White sturgeon 3.0 60 European perch Yellowfin bream 50 Atl. herring Lake whitefish Stellate sturg Lake trout 2.0 Atl. Salmon (repeat Orange roughy 40 spawner) Hoki 1.5 30 Atlantic cod

Skipped Spawning (%) Skipped Spawning Burbot 20 Co pikeminnow Striped bass White sucker 10 European chub 1.0 1.0 1.5 2.0 2.5

Log10 Longevity (years) Figure 1. Skipped spawning (as percent of sample) and related spawning interval (100/%SS) versus longevity for marine (solid symbols)Figure and1. Skipped freshwater spawning (empty symbols). (as percent Sturgeons of sample) are shown and in the related grey box. spawning References interval for SS (100% are European SS) versus chub (Fredrichlongevity et al.for 03), marine Striped (solid bass (Zlokovitzsymbols) et and al. 03), freshwater White sucker (empty (Trippel symbols). and Harvey Sturgeons 89), Colorado are shown pikeminnow in the (Tyus grey 90), box. Burbot Refrences (Pullianian for andSS Korhonenare European 90), Atl. chub Cod (Schwalme(Fredrich andet al.Chouinard 2003), 99), striped hoki (Livingstonbass (Zlokovity et al. 97), et Atlantic al. 2003), salmon white (Schaefer sucker and Elson(Trippel 75), and OrangeHarvey, roughy 1989), (Bell Coloradoet al. 92), Lake pikeminnow trout and Lake (Tyus, Whitefish 1990), (Kennedy burbot 53, (Pullianian54), Atlantic herringand Korhonen, (Engelhard and1990) Heino, Atlantic 05), Stella codte, Russian,(Schwalme and Belugaand Chouinard, sturgeons (Secor 1999), et al. hoki 00), Yellowfin(Livingston bream et (Pollock al. 1997), 84), EuropeanAtlantic perchsalmon (Holmgren (Shaefer 03), andshortnose, Elson, white, 1975), and Atlantic sturgeon (Gross et al. 02), Lake sturgeon (Roussow 57). orange roughy (Bell et al. 1992), lake trout and lake whitefish (Kennedy, 1953, 1954), Atlantic herring (Engelhard and Heino, 2005), stellate, Russian and beluga sturgeons (Secor et al. 2000), yellowfin bream (Pollock, 1984), European perch (Holmgren, 2003), shortnose, white, and Atlantic sturgeon (Gross et al. 2002), lake sturgeon (Roussow, 1957).

1152

Annual Spawning Scenarios 100 Baseline: Max. Fecundity=75 million 80 Max. Fecundity=45 million 60 40 Juvenile Fishery: Ages 4-7 years 20 0 0 0.1 0.3 0.5 0.7 0.9 1.1 F Skipping Scenarios 100 No Skipping Younger Skipping 80 Older Skipping % EPR Unfished% Population 60 Skip After 1st Spawn 40 Delay-2 Years 20 Delay-5 Years 0 0 0.1 0.3 0.5 0.7 0.9 1.1 F FigureFigure 2. %2. Egg% egg Production production per recruit per inrecruit unfished in unfished population populationof western Atlantic of western bluefin Atlantictuna across bluefin a range tuna of fishing across mortali a rangety rates of for scenarios of skipped spawning, maximum fecundity, and juvenile exploitation. See text and Table 2 for description of scenarios. Notefishing that becausemortality threshold rates for values scenarios were computed of skipped at F=0. spawni 0.1, 0.3,ng, 0.5, maximum 0.7, 0.9 and fecundity, 1.1, curves andappear juvenile unrealistically exploitation. jointed. See text and Table 2 for description of scenarios. Note that because threshold values were computed at F=0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1, curves appear unrealistically jointed.

1153