BULLETIN OF MARINE SCIENCE, 48(1): 137-149, 1991

BIOLOGY AND PHENOLOGY OF SIM () IN NEW CALEDONIA, A SARDINE OF THE CORAL ENVIRONMENT François Conand

ABSTRACT The biology of the sardine, Anzblygaster sirin, is studied from fish caught in the lagoons around New Caledonia. About 500 fishing hauls were carried out between 1980 and 1983. This sardine is a fast-growing, medium sized fish (Lmax:24 cm) with a short lifespan. Repro- duction takes place from October to December. First maturity is attained when fish reach 1 year of age and at this time their size is about 16-1 7 cm. Mortality is high and most fish die before the age of 2 years. This phenology compared with other seasonal spawners indicates a tendency towards semelparity.

A study of the resources of small pelagic fishes from the lagoon of New Caledonia was conducted by the Centre ORSTOM of Noumea from 1980 to 1983 (Conand, 1988). About 20 species of small pelagics occur regularly in the catches. They are mostly Engraulids, Clupeids, Atherinids and Carangids. Five or six species do not exceed 10 cm (TL) and have no commercial interest apart from their use as bait for pole and line fishing. Juveniles of the larger species can be used as bait .. and adults for human consumption. One of these species Amblygaster sirm (Wal- baum, 1792) is a sardine which can reach 25 cm. It is common and is often the dominant species of the catch. This tropical species inhabits coastal waters in the Indian Ocean and in the Pacific Ocean eastward of the Samoa Islands. Some studies have been conducted on its biology: in the Red Sea (Rafail, 1970; 1972; Sanders and Kedidi, 1984), on Sri Lankan coasts (Dayartne and Gjosaeter, 1986), in Indonesia (Sadhotomo and Atmadja, 1985) and in the Philippines (Ingles and Pauly, 1984). I present results obtained in the lagoons of New Caledonia on growth, reproduction and mortality of the species. The life cycle and phenology are then considered.

MATERIALAND METHODS From March 1980 till June 1983, 18 fishing cruises, each lasting 2 weeks were made every other month, in the lagoons around New Caledonia (Fig. 1). Fish were attracted with a light, and caught with a "boke ami" (a Japanese lift-net). The most commonly used net was 11 m wide and 13 m deep and the mesh size was 5 mm. Usually two sets were made: the first in the middle of the night and the second before sunrise. In addition, Dumbea Bay near Noumea, was visited monthly from July 198 1. The whole catch was weighed, a sample was taken to determine the species and weight composition of the catch and length frequency distributions were computed for each species (total length to the lower half centimeter). Then, samples of each species were frozen and brought back to the laboratory for reproductive studies. Biometric relationships were adjusted with a balanced number of individuals per size class. Linear relations describe the relationships among standard length, fork length and total length. The length- weight relationship W = aLn has been fixed according to the principal axis of the log,, weight-log,, length relationship. For the study ofgrowth, the mathematical expressions used are, the von BertalanQ growth function (VBGF) and a modified version of this function with a sinusoidal oscillation to simulate the seasonal growth variation (Pauly and Gaschütz, 1979). Its form is: K 1 - exp[-K(t - h) + a - sin w(t -t 91 W where L,, K, t, parameters from standard VBGF, a = intensity 137 .. \\ i\ I 165"E 167"E Figure 1. XIap uf New Caledonia showing the main fishing stations.

0 and I ). ù = period (here XI) and + = phase (the beginning of the growth oscillations bvith respect to T = O). hlultit'an software (Fournier et al.. 1989) has hecn used for estimating growth parameters. hlultifan utili7t's a likelihood method hased on the approach of Schnute and Fournier (1980) and Fournier and Hreen (I083) to simultaneously anal>-rescvcral length frequency samples. Several csti- mates were made zither cumulating the size distribution frequency on a monthly haais for the whole of New Caledonia. or with ohsenations of one night per month in one hay. The growth of Soung fish hetween 30 and 80 mm has been studied hy counting increments on otoliths ohsend with an optical microscope. Althougli the readings have not heen validated by rearing or tagging experiments. it was assumed that for young fish each increment corresponds to one day. Specimens brought to the laborator, for reproductive lvorh \vere measured (TL. mml and the body (g)and the gonad neighed (cg). Tht' \ex and maturity stage of the gonads wre noted using the Fontana I 1969) scale which comprises 7 stages ( 1 :immature: Z: resting: 3: matunng: 4: prespa\vning: 5: spaxvning: 6: post-spawning: 7: inxolutinn). The gonad indes tG11 given b) the relation: GI = gonad weight 100 total weight has heen caliulated and the \-ariation of its monthl) mcan. analyzed. Natural mortalit> can be inferred from the catch cune (Ricker. 197.51. To delelop this curw. the monthl) mean of the numher ot'fish caught per iishing haul has first heen calculated for cach size class. Then. considering the consistency of the variations of the 1-ield and the si7es. during the 3 years. ohservarions were grouped per month nn a yearly basis. Finally. \\ith thc hach calculatzd liges the catch cune has heen estahlished. L'sin& the natural mortalit!-. the length-weight relationship :md the gro\\th function. it is possiblr to estimate the theoritical et olution ot' the instantaneous biomass ot'a cohort. from the relation: B, = N-e-"\V! \\here ß, = biomass at time t. N = starting ahundance in numbers. 2 = total mortalit). and \V, = mean weight of a fish at time t.

RESULTS

-4t1~11dI irriariom qf .-16irrir/anct7.-The variation during the >ear of the mean monthly catch of this sardine. in weight and in number of indixiduals. (Fig. 2) CONAND BIOLOGY OF A SARDINE OF NEW CALEDONIA 139

MONTH Figure 2. Mean monthly catch per fishing set. clearly shows that recruitment starts in December at the end of southern hemi- sphere spring. Catch is highest in January then gradually decreases until Novem- ber. Biometric Relationships. -Relations amongst fork length (FL), total (TL) and standard length (SL), calculated from 66 individuals with a TL between 59 and 201 mm are: FL = 0'86 TL + 3'92 (r = 0'998) and SL = 0'81 TL + 2'74 (r = 0'998) The length-weight functions: W = aLn, given in Table 1 were established for males, females, and a balanced sample of juveniles and adults of both sexes. The function adjusted for males or females only takes into account fishes larger than 80 mm. The slope is higher for the function adjusted from the sample including juveniles.

Table 1. Length weight relationship; weight in grams, total length in millimeters; N, number; I,,,,,, and L,,,,,,,observation interval

N L," Lox r ax 10-6 N Males 192 91 218 0.994 2.62 3.225 Females 248 80 229 0.998 2.47 3.237 Males 1 Females 397 42 229 0.999 3.49 3.123 Juveniles 1 140 HULLETIN OF hl4KrNE SCIENCE. VOL. 48. NO I. IQ41

8i

6 -1

4 1

I I I) 30 60 Increments Figure 3. Relationship between otolith increnirnt count and total length fnr yxing fish. tDotc for right and left otolith are linked with a line.)

Groir~h.--Increment counts for 30 to 80 nim fish are presented in Figure 3. They show that I-month-old sardines measure betLseeen 30 and 40 min and those 1 months-old between 70 and SO mm. VBGF adjustments were made. on month]) samples from Dumbea Bah in 1983. on Saint Vincent ßal samples from 1980 to 19S1 and the monthly grouping of all the obsenations made in New Caledonia. Histograms and back calculated modal length are given in Figure 4. Parameters of the functions and back calculated age-length kalues are given in Table 2. With slight variations according to places and )ears. .4. sirrt^ in New Caledonia reaches about 19 cm in 6 months. I7 cm in 1 year and 20-2 1 cm in 1 years. The modified VßGF was used for Dumbea Bay station \\hich has been sampled bery regularly and map be said to have it5 o\\n population. Adjustment is more precise and reflects a fast growth near March-April at the end of the hot rain) season and a slow growth between September and Noxeniber at the end of the maturation and during spa-ning. CONAND BIOLOGY OF A SARDINE OF NEW CALEDONIA 141

7 10 13 16 19 LT 7 10 13 16 19 LT 10 15 20 LT A B C Figure 4. Length frequency histograms and back calculated size from von Bertalanffy growth func- tions; A) Dumbea Bay in 1982; B) Dumbea Bay in 1982 with the modified function (note the im- provement in the adjustment); C) Saint Vincent Bay from 1980 to 1982, with the modified function. Repr,ducri~~t?.-Studies of sex-ratio were limited to a few observations (Table 3). The) show that one sex is often dominant in a fishing haul. which could result from the dominance of one ses in a school. Obserb ations also show that females are less abundant amongst small fish. but more abundant than males amongst large ones. The variations of the GI from I980 to 1981 (Fig. 5) clearl) C~OLTSthat the sexual cpcle of .4. siritz in New Caledonia is regular and annual \vith a peak occurring in the fourth quarter. This allons the calculation of a inem on a month11 basis of the 3 years of obsenations. Figure 6 gives the annual cycle of: (a) the monthly proportion of males. females and individuals undetermined by eye ob- servation: (b) the monthly proportion of the maturity stages for males and females: (c)the female gonad index. In Januarq and February. almost all sardines are small juveniles. and gonads develop gradually from March till September. Betneen October and December. a11 the fish are mature and spawning. CONAND: BIOLOGY OF A SARDINE OF NEW CALEDONIA 143

c 100- s n Males

50- Females t A mn] Undeterminated 3 I. n- number 174 170 383 139 371 124 341 88 14 187 59 97

c 100- ou. MALES o

50 - Sexual stages

4,%6

R O- Y J FMAM J JA SOND REI3 5 loo- n Em* I

50 -

FEMALES 0-

JFMAMJJASOND Figure 6. Reproduction of4. sinn. Mean of 3 years of observations: A) monthly proportion of males, females and undetermined; B) monthly proportion of maturity stages for males and females; C) GI as a function of month for females.

All fish reach maturity in October and at that time they are 10 to 12 months old. Smallest specimens are 12 cm but the mode is 16 cm and this value represents the mean size at first maturation. To estimate batch fecundity, the ovocytes of the last mode of their size distri- bution were counted for 24 females measuring between 168 and 215 mm caught in October and November. Fish were at stage 4. The relation between batch fecundity and weight (in grams) is: Table 4. Catch per unit effort (number of'indi\.iduals per fishing set). hlontlily \-alue and total. per 9ize class

35 78 n o n i 5n "8 45 14s 43 o O 818 I .DO6 55 814 I03 n n 2.513 3-13 il5 2.286 466 31 -7 1.429 42IV -5 I .wn 856 153 0 Wil 3.903 85 1 .;Y 450 7; 1 7 2x8 3.73 95 415 393 310 II Ihl l.S(lIl 1 o5 1.433 306 '65 16.5 O -.-7 7ILj 115 515 473 33 1 1 h3 n 1.510 175 --77 213 5Oh 191 o 1.158 I35 0 13 570 216 O 1.140 145 il n 420 64 n 1 .il98 155 n il ---171 17 -7 977 1 hS n o I I3 3 3 OX5 175 o Ci 33 n 1 0 364 185 o n '8 o 11 11 il 105 n 0 7 n h 178 705 o n - n I 54 71 5 o o 4 n n 31 7 375 CI n I o 1 10 135 n 11 1) n n 1 245 n O n n (ï 1

F = 300w - 132 with a standared variation of 83 and r = 0.97. It has not been possible to establish how man? batches of eggs are laid in a spawning season. so. the total fecunditg is still unkno-n.

.\Jorta/itj~, T/i.c~ireii~*~iIEi~diiti~~ii ,$ti Cdiort. -From month13 ohsen ations of size frequent) and CPUE. expressed in number of fish per fishing haul. a CPUE per size class for each month and for the )ear is inferred (Table 3). From this salue and the growth function it is possible to calculate the catch per age (,Annex 11. to draw the catch curie (Fig. 7)and to estimate the total mortalit?: z = 7.9 kO.1 As there is almost no fishery fbr sardine in New Caledonia the total mortalit) is also the natural mortality. The theoretical e\ olution of the bioniass of a cohort of 100.000 fish at 1 month. (age of the recruitment) is shown in Figure 8. The maximum biomass occurs at 7 months and this critical size is attained slightl? before the first niaturit>. Lift Cjde. -In Nen Caledonia the sardine --litih/j~,qu.Uw~irm reaches maturit) as it approaches 1 gear of age. The spa~ningseason occurs from October to Decem- ber. before the hot. rain) wason. There are probab13 selrra1 spa~ning5(most Clupeids are serial spmners) [Hunter and Goldberg. 19SO: Alheit. 1988). The lifespan is short and the fish uwally die after spanning or at least if it sunives. before reaching the next spawning season.

DISCU'SSION Results on growth. from selera1 studies. are summarized in Table 5. Earl) studies (Rafail. 1972 Burhanuddin et al.. 1 974: Sanders and kdidi. 1'383) prob- CONAND BIOLOGY OF A SARDINE OF NEW CALEDONIA 145

Annex I. Showing the steps for the construction of the catch curve ofArnblygaster sirni with L, = 221.5 (TL in mm), K = 1.58 (yr-I), to = -0.003 (yr), Dt, time spend in the size class; N, number of fish

Class limits Age Lower Upper In out Dt N NiDt Lo&NlDt Mean age 30 40 0.089 0.123 0.034 228 6,706 8.81 0.106 40 50 0.123 0.159 0.036 1,006 27,944 10.24 0.141 50 60 0.159 0.197 0.038 3,432 90,316 11.41 0.178 60 70 0.197 0.237 0.040 4,219 105,475 11.57 0.218 70 80 0.237 0.281 0.044 3,903 88,704 11.39 0.259 80 90 0.281 0.327 0.046 2,725 59,239 10.99 0.303 90 100 0.327 0.377 0.050 1,800 36,000 10.49 0.352 100 110 0.377 0.431 0.054 2,219 41,093 10.62 0.404 110 120 0.431 0.491 0.060 1,510 25,167 10.13 0.461 120 130 0.491 0.557 0.066 1,158 17,545 9.77 0.523 130 140 0.557 0.630 0.073 1,290 17,671 9.78 0.592 140 150 0.630 0.713 0.083 1,098 13,229 9.49 0.670 150 160 0.713 0.808 0.095 927 9,758 9.18 0.759 160 170 0.808 0.921 0.113 685 6,062 8.71 0.862 170 180 0.921 1.057 0.136 364 2,676 7.89 0.985 180 190 1.057 1.232 0.175 210 1,200 7.09 1.138 190 200 1.232 1.473 0.241 89 369 5.91 1.341 200 210 1.473 1.870 0.397 54 136 4.91 1.641 210 220 1.870 3.159 1.289 31 24 3.18 2.230 220 230 3.159 10

ably underestimated sardine growth when they adjusted a VBGF giving sizes of 10 to 1 1 cm at 1 year and 15 to 17 cm at 2 years. On the other hand, estimates of growth based on otolith readings with the assumption of the formation of one increment per day (Gjosaeter et al., 1984; Dayartne and Gjosaeter, 1986) give a size of 16 to 17 cm at 6 months and 20-2 1 cm at 1 year and might be overestimated. Several studies of otolith increments were made with checks on free or cultured tagged fish. They show that on some species, on some days, no increment is formed (Jones, 1986), or the increment is very thin and below the detection power of the light microscope (Campana and Neilson, 1985). VBGF adjustments made by Ingles and Pauly (1 984) and Sadhotomo and Atmadja (1 985) correspond, up to the second year, to those observed in New Caledonia and are within the interval of the variation observed according to the locations and the years. In , Dayartne and Gjosaetér (1986) have observed A. sirm sardines, maturing in February and they found them to be totally mature from April to June. These observations from the northern hemisphere agree with those of New Caledonia with a 6 months shift. Knowledge of batch fecundity has limited interest as this fish is probably a serial spawner and spawns a number of batches during the season. Studies carried on several species of Clupeids and Engraulids, using Hunter and Goldberg (1980) method, show a spawning frequency varying usually between 2 and 8 days. By another way, Thorrold (1989) has shown, with otolith daily increments analyzed for Herklotsichtys castelnaui, a sardine from Queens- land estuaries, that spawning occurred mostly twice per lunar month, at neap tides. These two techniques could give interesting information on the real fecundity of A. sinn. The mortality rate estimated in New Caledonia is higher than those given by Ingles and Pauly (1984) and Sadhotomo and Atmadja (1985) who observe growth rates similar to ours for the first 2 years. For Ingles and Pauly (1984) 4- or 5-year-old fish would not be exceptions. From our study it seems that in New 146

Figure 7. Catch curve used to infer mortalit?: points are means for all arras.

Caledonia. 2-year-old fish are few and it \+auld be xery unusual for a fish to reach 3 years. Size dispersion around the mean let e\en suppose that the largest spec- imens were in fact only 7 )ears old.

CONCLUSION

. Inrhljgmtu SI~I~Iin Neu Caledonia is a fast growing fish of medium size with a short lifespan. The reproductixe season is well defined and lasts 2 or 3 months and first maturity is reached \\hen fish attain their first year. hlortalit> i5 high

Table 5. Companwn of'the results on grtwth and mortality of 4. .tirm from studies in various region of the Indo-Pacitic

LcnE!th lim agri L, h C1 Rclcrcn'c Lx GltI( Ill 6 mth I II 2 ?r5 irmi ¡ir 'I i>r '1 Rafiil (1972) Egspt 6 I I 17 72.6 0.65 1.45 Sanders and Kedidi ( 1 P84j EEYPt - IO IS 26.1 n.37 1.49 Gjosaeter et al. (19S-1) Xíozam biq tie 16 2n - 22.n 2.5 3.41 Dayartne and Gjosaetcr ( 108hj Sri Lanha I' '1 - --.'3 (4 2.38 - Burhannudin et al. (197-11 Indonesia (, 17 23 1-1.3 0.59 1.34 Sadhotomo and Xtmadja ( 1935) Indonesia 11 I7 23 '5.7 1.17 2.08 Ingles and Paul! ( 1qS-l) Philippines ICI 16 23 '7.3 ir.80 I.hh Presrnt study New Calrdonia 12 17 21 22.2 1.58 ?.'I CONAND: BIOLOGY OF .4 SARDINE OF NEW CALEDONIA 147

M Io0i 501 O O 6 12 18 24 30 AGE (months) Figure 8. Theoretical form of the instantaneous biomass of a cohort of 100,000 one month old individuals.

- serial spawning vvvv - isochronal spawning Onchorynchus gorbuscha ( 2 yr) V - spawning season V ( n yr) life span Amblygaster sirm ( 1yr )

Pleuronectes platessa (10-15 yr ) --_---_ V V V

Engraulis mordax (4-6 yr) ____--_------vvv vvvv vvvv vvvv vvvv

Figure 9. Example of four reproductive strategies of fish with seasonal spawning period. and fish usuallq die before the second >ear. however some inditiduals can have a second spawning season. This phenology is compared with other seasonal spatvn- ers (Fig. 9). Isochronal spawners like for example P1tw"re.s plarc&w in teni- perate waters (Simpson. 195 1 ) or serial spawners like Eiigruirlw í~iordíis(Hunter and Goldberg. 1980) have a limited annual spawning season which repeats for ses eral years. With adults dying after thel ha\ e spawned. Oiiclior1i~'ii~'liii'sgor- birscha, a Pacific salmon, is on the opposite typicall) semelparous (Mann and Mills. 1979). .í. ririli appears intermediate and shows a tendency touards se- melparity as it dies after a unique breeding season which lasts 2 or 3 months. duringwhich it spawns several times. "Semelparity is an option for a short lifespan wirh a high rcpruciuciilr t.Küïi aiid feciiiidiíj" {?\:aiiii aiid ?)?ills. !'??Q?). Bü: i: requires a strategy ensuring that the population maintains itself' if unfax orahle conditions. occuriiig during larval or juvenile life lead to a recruitment failure. The intermediate strategq of 4. .siri?icombines the adyantage of high reproductive effort and the increase of survi\ al probability with several cohorts.

LITERATURECITED

Alheit. .I. 198% Reproductive hiology nf sprat (Spriirfrtr cpnifwl:factors determining annual egg producticm J. Cons. Int. Explnr. bier 44: 1 h2-168. ßurhanuddin, Al. Hutomo. S. hlartosewjo and D. A. Djamali. I V74. ßeherapa iispek biolog¡ ¡kan 1t.mtu-u Smjiiii~llrl,+iwi (R'albrium) di perairan Pulou Panggang. Oseannlogi di Indonesia 2: 17-25. Campana, S. E. and .I. D. Neilson. 1985. hlicrostructure of fish otoliths. Cam. J. Fish. .Aquat. Sci. -12: ini.t-103~ Conand. F. 1988. Biologir et kccilogie des poisson pe1;ìgiyues du lagon de Nau\-elle-Cal~donieuti- lisables comme appit thonier. Paris: ORSTOhl. Etudes et Thèses. 239 pp. Dayartne. P. and J. Gjnsaeter. 1986. .Age and growth of four S¿?di~d/dspecies from Sri Lanka. Fish. Res. 4: 1-33. Fontana. .A. 19hY. Etude de la maturite seanzlle des sardinelles .Ymiimsl/d thi (val) et Siir~iiiit,h tii(rirci C. et Y. de la region de Pointe-Noire. Cah. ORSTOM. ser. Oc6anogr. 7(2):101-1 14. Foumlcr. D. A. and P. A. ßrem. I qS3. Estimation ofabalone mortalit) rates gro\vth analysis. Trans. .Amer. Fish. Soc. 1 13: 40341I. -. J. K. Sihert. J. hlajkowshi and J. Hampton. 1968. hlultilhn R likelihood-hased method tor estimating grrnvth parameters and age composition from multiple length frequenc! samples nith an application to southern bluetin tuna (771rriirrir.r iw'¿qvi). Nanaimo. Canada: Otter software. 23 pp. Gjosaeter. .F.. P. Da>-artne. O. A. ßergstad. \ . GjoSwtrr. hi. I. Soura and I. Al. Beck. 1'JX-t. Ageing tropical fish by gro\\th rings in the otolith\. FAO Fi5h. Circ. 176. 54 pp. Hunter. J. R. and S. R. Cioldherg. 1 WO. Spnnningincidence and hatch ft-cundit? in northern anchov). Erz.gnrirli~i?wrdrl-y. Fi\h. Bull.. U.S. 79: 21 5-23fl. Ingles. J. and D. Paul). 19S4. -\n atlas of the groxvth mortality rind recruitment of Phil~ppinefishes. ICLARXI Tech. Rep. 13. 117 pp. Jones. C. 1986. Determining age of larval fish ivith the otolith increment technique. Fish. Bull.. U.S. s4: 9l-l(l2. hlann. R. H. K. and C. A. hlills. 187'J. Demographic aspects of tish fecundity. Foges Ihl-177 iii P. .l. Xliller. ed. Fish phenology anaholic adaptatiyness in Teleost. .Academic Press. London. Paulj. D. and G. Gaschiit;.. 1974. .A simple method for titting oscillating length gronth data. .sith program fbr pockct calculator, C'IEhl-Ch1 1970 G: 74-3. Rafml. S. 2. 1970. Studies of populations and exploitation status of Eg>ptian Red Sea ahundant sardines. Hull. ht.Oceanogr. Fish. (L1.A.R.I I: 13&1.19. -. 1973. Studies of Red Sra fisheries by light and purse-wine near Al-Ghardaqa.Hull. Inst. (Iceanogr. Fish. (!J..A.R.) 2: 35-49. 1 Ricker. \V. E. 1975. C~mputatioiiand interprctntion of biologic.al statistics id fish popuIatii:mc. Bull. Fish. Res. Hoard Can. 191. 3S2 pp. Sndhotiimo. ß. and S. B. .Atmadja. 1985. On the growh oisomc pelagic fish in the .la\-;ì Sea. J. Pcn. Perikanan Laut 33: 53-60. Sanders. hi. J. and S. hl. hedidi. I 98.1. Stoch assessment for the spotted Sardinzlla c.Smhrrc/br \irni1 caught hl purse seine adjarcnt to the border hctwern EgSpt and Sudan. IJNDP FAO. R:\B S3 1113 03. 18 pp. CONAND BIOLOGY OF A SARDINE OF NEW CALEDONIA 149

Schnute. J. and D. Fournier, D. 1980. A new approach to length frequency analysis: growth structure. J. Fish. Res. Bd. Canada 37: 1337-1351. Simpson, A. C. 1951. The fecundity of plaice (Pleuronectesplatessa L.). Fishery Invest., Lond. (2) 17: 1-27. Thorrold, S. R. 1989. Estimating some early life history parameters in a tropical clupeid, Herklot- sichthys castebzaiti, from daily growth increments in otoliths. Fish. Bull., U.S. 87: 73-83.

DATEACCEPTED: June 6, 1990.

ADDRESS: Centre ORSTOM de Brest, B.P. 70, 29280 Plouzaiw, France.