Growth and mortality rates of yellowfin tuna , Thunnus albacares ( Perciformes : Scombridae ) , in the eastern and central Pacific Ocean

Age and growth parameters were estimated for the yellowfin tuna Thunnus albacares (Bonnaterre, 1788). Atotal of 443 individuals were sampled from China longline fisheries in the eastern and central Pacific Ocean from February to November 2006. The von Bertalanffy growth parameters were estimated at L = 175.9 cm fork length, k = 0.52 year-1, and t0 = 0.19 year. The total mortality rate (Z) was estimated to be from 1.19 to 1.93 year -1, the fishing mortality (F) and the natural mortality (M) were calculated to be 0.91 year-1 and 0.65 year-1, respectively. The rate of exploitation (U) was estimated to be 0.46. This study provides estimates of growth and mortality rate for yellowfin tuna in the eastern and central Pacific Ocean, which may be used as biological input parameters in future stock assessments for the oceanic region. However, age analysis with other techniques, additional validation of the size composition and stock structure are also needed.

The yellowfin tuna Thunnus albacares (Bonnaterre, 1788) is a commercially important species of tuna inhabiting tropical and subtropical seas worldwide, except the Mediterranean Sea (MARGULIES et al. 2001).The species supports major fisheries throughout its range (COLLETTE & NAUEN 1983, NISHIKAWA et al. 1985) and represents near half of the catch, both at the industrial and artisanal levels, in the equatorial Brazilian EEZ (TRAVASSOS 1999, ZAGAGLIA et al. 2004).Little is known about key biological parameters of the yellowfin tuna, such as age and growth within the eastern and central Pacific region.
Despite these possibilities, very few validated age studies have been conducted for large tropical pelagic species like the yellowfin tuna in the eastern and central Pacific Ocean.The last study on yellowfin tuna's age and growth in the area was conducted in the late 80's (Eastern Pacific Ocean, WILD 1986), except for HAMPTON's (2000) (Western Tropical Pacific) and SU et al. (2003) (Western Pacific) studies.
In the present study, data on the growth and mortality of yellowfin tuna in the eastern and central Pacific Ocean is presented.The objective of the present study is to provide the first detailed information on the growth and mortality of yellowfin tuna collected in this area, by using fork length data and the ELEFAN I technique (PAULY 1987).This may be useful in managing the rapidly developing fishery of yellowfin tuna in the eastern and central Pacific Ocean.February to November 2006.The von Bertalanffy growth parameters were estimated at L ϱ = 175.9cm fork length, k = 0.52 year -1 , and t 0 = 0.19 year.The total mortality rate (Z) was estimated to be from 1.19 to 1.93 year -1 , the fishing mortality (F) and the natural mortality (M) were calculated to be 0.91 year -1 and 0.65 year -1 , respectively.The rate of exploitation (U) was estimated to be 0.46.This study provides estimates of growth and mortality rate for yellowfin tuna in the eastern and central Pacific Ocean, which may be used as biological input parameters in future stock assessments for the oceanic region.However, age analysis with other techniques, additional validation of the size composition and stock structure are also needed.KEY WORDS.Biological parameters; size frequency analysis; yellowfin tuna.G. Zhu et al. ZOOLOGIA 28 (2): 199-206, April, 2011

Study area and data collection
Samples were collected from the eastern and central Pacific Ocean (11°00'S-05°00'N, 134°00'-153°00'W) (Fig. 1) using Chinese longline vessels from February to November 2006.Fork length (FL) measured to the nearest 1 cm (and pooled into 5 cm length classes), round weight (RW) and dressed weight (DW) to the nearest 1 kg were obtained.The specimens were sexed by inspecting gonad morphology.To examine the temperature experienced by yellowfin in the eastern and central Pacific Ocean, daily sea surface temperature (SST) data and vertical profile were obtained from the same Chinese longline vessels using a CTD (Conductivity Temperature and Depth sensors, Sea-Bird 37, Sea-Bird Electronics, Inc.), within the same time period.log-transformed data, and the association degree between variables (log W and log L) was calculated by the determination coefficient (r 2 ).Regression analysis was employed on log-transformed data for males and females separately and the slopes were tested for significant differences between sexes through analysis of covariance (ANCOVA).The hypothesis of isometric growth (RICKER 1975) was tested using the t-test (p < 0.05).
The present study simulated the relationship between dressed weight and round weight using a linear regression analysis.The ANOVA analysis was used to verify whether the linear model is appropriate for describing the relationship.

Growth
The growth process can be described by growth velocity and growth acceleration.Length frequency data was also used to calculate the von Bertalanffy growth rate (k) and the asymptotic length (L ϱ ) by model progression analysis using the program ELEFAN I (PAULY 1987) within the FiSAT program (GAYANILO et al. 1994).
The growth performance index ' (phi-prime) was calculated based on the growth parameter estimates following the equation of PAULY & MUNRO (1984).The index was compared with estimates obtained by other authors to facilitate the intra and interspecific comparison of the growth performance (PAULY & MUNRO 1984).

Mortality
For the calculation of the instantaneous annual mortality rate (Z) the length-converted catch curve (PAULY 1983, MUNRO 1984) was applied to the pooled length frequency data using the estimated growth parameter.The calculation was done with the FiSAT program (GAYANILO & PAULY 1997).
The natural mortality was calculated by Pauly's empirical equation: logM = 0.1228 -0.192 log L' ϱ + 0.7485 log k + 0.2391 log T, where: T = the mean annual temperature (in °C), which is assumed to reflect the locally sea surface temperature in the survey area (PAULY 1980) (in the present study, T = 26.4°C);M = natural mortality.
In order to obtain the L' ϱ value, the present study used the relationship between fork length and total length (TL) on yellowfin tuna from the Fishbase database (TL = 1.108L, www.fishbase.org)and combined the L ϱ value estimated from the equation.
For the calculation of the fishery mortality (F), the M value was subtracted from the Z value in order to get the fishing mortality (F = Z -M) (SAINSBURY 1982, APPELDOOM 1984, 1988).
With the estimated values of F and Z the rate of exploitation (U) was calculated according to LANDAU (1979) andGULLAND (1985).

Dressed weight -round weight relationship
The relationship between dressed weight and round weight estimated using a linear model led to the following equa-

Length-weight relationship
Confidence interval for mean FL was assessed by bootstrapping 1000 bootstrap pseudo-samples.We used Kolmogorov-Smirnov test (K-S test) to examine the difference in FL distributions between sexes.
The length-weight relationship was calculated by applying a power regression W = aL b e ⑀ , ⑀~N (0, 2 ), where W is the round weight (kg), L is the fork length (cm), and a (initial growth coefficient or condition factor) and b (growth coefficient, i.e., relative growth rate of fish) are both constants.
This equation may also be expressed in its logarithmic form: log W = log a + b log L. The parameters log a, b and the standard errors of b value (S.E.b ) of W-L relationships were estimated by linear regression analysis (least-squares method) on tion where RW = 2.0961DW+1.0906(r 2 = 0.9897, n = 88, p < 0.001) (Fig. 2).F value from ANOVA test indicated that the linear model is appropriate for describing the relationship (F = 8228.9,df = 87, p < 0.001).

Length-weight relationship
Among the 443 specimens collected, 387 (90/215 females/ males) were used for growth analysis because the fork lengths of some specimens were not measured.The length of yellowfin tuna ranged between 93 and 167 cm FL and mean FL was 130.2 cm (128.7-131.7 cm for the bootstrapped 95% confidence interval).The fork length of about 80% individuals ranged from 110 cm to150 cm (Fig. 3).The maximum size of males (93-167 cm FL) slightly exceeded that of females (99-162 cm FL).The mean FL of females and males were 128.4 cm (125.5-130.0cm) and 131.9 cm (129.8-133.9cm) respectively.K-S test found no significant differences between FL distributions of females and males (H = 1.2040, p > 0.05).

Age and growth
The growth parameters estimated by ELEFAN I routine and the performance index (') were as follow: L ϱ = 175.9cm, k = 0.52 year -1 , t 0 = 0.19, ' = 4.21.The value of L ϱ was higher than the maximum observed fork length of 167 cm (Fig. 5).The goodness-of-fit index Rn was 0.332.Z = 1.56 (with 95% confidence interval of 1.19-1.93)year -1 , F = 0.91 year -1 .The instantaneous natural mortality rate (M) obtained using the equation of PAULY (1980) was 0.65 year -1 .The reliability of the estimated M was ascertained using the M/K ratio because this ratio has been reported to be within the Comparing the present findings with previous works, we find a close similarity between studies conducted in similar waters (Tab.I, Fig. 6).The ' value of yellowfin tuna in the eastern ) is lower than that in the western Pacific Ocean (4.08-4.62)(Tab.I).Generally, our results fall within a reasonable range (Fig. 7).

Mortality
The length-converted catch curve is shown in figure 8.The estimated instantaneous rates of mortality for all fish were   ⌬ range 1.12-2.50for most species (BEVERTON & HOLT 1957).The value of M/K ratio was 1.25.The rate of exploitation was U = 0.46.

DISCUSSION
Yellowfin tuna growth has been studied by various methods in the Pacific Ocean (Tab.I).The study of growth using length-frequency analysis has long been the most frequently used method, even in other oceans, such as the Atlantic Ocean (LE GUEN & SAKAGAWA 1973, FONTENEAU 1980, GAERTNER & PAGAVINO 1991) and the Indian Ocean (MARSAC & LABLACHE 1985, MARSAC 1991).
The growth curves estimated from this study agree well with the growth curves estimated earlier by other authors (such as MULTIFAN technique, SU et al. 2003) and even scales (such as LE GUEN & SAKAGAWA 1973)  Differences in growth patterns may be the result of differences in genetic structure and/or differences in temperature, density of food and diseases (PAULY 1994, WOOTTON 1998).The estimation of ' value was 4.21 in the present study and is  1975).However, for short-lived fish species, a length-converted catch curve is more suitable (GAYANILO & PAULY 1997), even though the results from the latter estimation are more precise (PAULY et al. 1995).The Z value in the present study is 1.56 year -1 using the length-converted catch curve and is closer to the Z value of SCHAEFER (1967) (1.72 year -1 ) and SU et al. (2003) (1.71 year -1 ).WISE (1972) concluded three different Z values from three different tuna fisheries: Z = 1.52 year -1 for the pole and line fishery with live baits, Z = 2.32 year -1 for the purse seine fishery and Z = 1.88 year -1 for the tuna longline fishery.WISE (1972) also showed that the reasonable Z value ranged from 1.4 to 2.4 year -1 and the Z value in the present study is similar to other results (SCHAEFER 1967).
The estimation of natural mortality poses some difficulty because it may be affected by the selection of the estimation method and the study area (SU et al. 2003).The natural mortality in the present study was 0.65 year -1 using PAULY's (1980) empirical equation and is similar to the M values of HENNEMUTH (1961) (M = 0.64-0.90year -1 ) and HAMPTON (2000) (M = 0.7-1.2year -1 for the individuals with fork length over 100 cm) and SU et al. (2003) (M = 0.56 year -1 ).The records on yellowfin's natural mortality ranged between 0.6-1.2year -1 (MURPHY & SAKAGAWA 1977) and M in the present study was within a similar range.
Mortality and exploitation rate estimates should be treated carefully since they were estimated from two discontinuous sampling periods and may be biased by annual differences in year class strength.It is rather difficult, and probably unwise, to describe the current state of the stock because of the lack of information on the effect of fishing on the recruitment, behavior and migration pattern of yellowfin tuna in the eastern and central Pacific Ocean.

Figure 1 .
Figure 1.Map of the sampling area and sampling sites for yellowfin tuna in the eastern and central Pacific Ocean from February to November 2006.

Figure 2 .
Figure 2. Linear relationship between dressed weight and round weight for yellowfin tuna in the eastern and central Pacific Ocean from February to November 2006.

Figure 3 .
Figure 3. Length frequency distribution of yellowfin tuna in the eastern and central Pacific Ocean from February to November 2006.

Figure 4 .
Figure 4. Relationship between dressed weight (DW) and fork length (FL) of yellowfin tuna in the eastern and central Pacific Ocean from February to November 2006.

Figure 5 .Figure6.
Figure 5.The von Bertalanffy growth curves of yellowfin tuna in the eastern and central Pacific Ocean as superimposed on the length-frequency histograms.

Figure 8 .
Figure 8. Length-converted catch curve for all yellowfin tuna specimens collected from the eastern and central Pacific catch samples from February to November 2006."Not used" indicate the data refer to length classes not fully recruit to the fishery.

Figure 7 .
Figure 7.Comparison on the relationship between estimate of growth coefficient and asymptotic length for yellowfin tuna in the eastern and central Pacific Ocean in the present study with the relationship estimated by other authors.n = data source.The data are partly sourced from Table I of the present study and Fishbase's data about yellowfin tuna.(Source: http://www.fishbase.org/PopDyn/PopGrowthList.cfm?ID = 143&GenusName = Thunnus &SpeciesName = albacares&fc = 416).
using other length-frequency analyses.Estimations on L ϱ , k and ' in the present study is close to the results from a similar area in the central Pacific concluded by MOORE (1951) (L ϱ = 172.7 cm, k = 0.86 year -1 and ' = 4.41) and SU et al. (2003) (North-Western Pacific, L ϱ = 175 cm, k = 0.39 year -1 and ' = 4.08).