Production and Mortality of Anchoveta (Engraulis ringens) Eggs off Peru* DANIEL PAULY MINA SORIANO International Center for Living Aquatic Resources Management MC P.O. Box 1501, Makati Metro Mani/a, Philippines

PAULY, D. and M. SORIANO. 1989. Production and O1.\lltality of anchoveta (Engrau/is ringens) eggs off Peru, p. 155-167. In D. Pauly, P. Muck, J. Mendo and I. Tsukayama (eds.) The Peruvian upwelling ecosystem: dynamics and interactions. ICLARM Conference Proceedings 18,438 p. Instituto del Mar del Peru (IMARPE), Callao, Pcru; Deutsche Gesellschaft fUr Technische Zusammenarbeit (G12) GmbH, Eschbom, Federal Republic of Germany; and International Center for Living Aquatic Resources Management (ICLARM), Manila, Philippines.

Abstract The egg production of anchoveta was estimated by month from 1953 to 1985 based on estimates of mature female biomass batch fecundity and related factors for the period 1953-1985 off Peru (4-14°S). These theoretical cstimates of production were related to empirical estimates derived from egg surveys conducted from 1964 to 1985. This yielded estimates of egg mortality; a multiple regression model including parent stock size, sardine biomass, SST and SST anomaly explained a large fraction of the variance of these egg mortality estimates. However, how results suggest parent concentration (rather than parent stock size) to be the key parameter affecting anchoveta egg mortality, a hypotheses which will have to be tested in a future contribution.

Resumen Sc cstim61a producci6n de huevos mensual de la anchoveta peruana (4-l4°S) de 1953 a 1985 basado en estimaciones de la fecundidad biomasa de hembras maduras y factores relacionados para el perfodo 1953-1985 frente a Peru (4-l4°S). Estas estimaciones de producci6n te6ricas fueron relacionadas con estimaciones empfricas obtenidas de cruceros de evaluaci6n de huevos realizados de 1964 a 1985. Esto proporcion6 estimaciones de mortalidad de huevos; un modelo de regresi6n multiple que incluye el tamano del stock de padres, biomasa de sardina, TSM y anomalfas de TSM, explic6 una gran parte de la varianza de estos estimados de mortalidad de huevos. Sin embargo, como los resultados 10 sugieren, la concentraci6n de padres (mas que el tamano del stock de padres) es el parametro clave que afecta la mortalidad de huevos, una hip6tesis que debera ser probada en una contribuci6n futura.

IntrOduction Recruitment to a stock depends on (i) the size of the spawning stock and (ii) the survival of the eggs and larvae. Within fishery biology, an immense literature exists on how to forecast the recruitment of fish stock given a knowledge of spawning stock size, i.e., from the "top-down" (reviews in Ricker 1954; Cushing 1988). There have also been numerous attempts to approach the recruitment problem from the "bottom-up", i.e., through detailed analyses of the factors controlling the survival of eggs and larvae (see, e.g., papers in Sharp 1980 and Rothschild 1986). While generally, "bottom-up" approaches have been more costly, more data intensive, yet less successful than "top-down" approaches in providing management advice, it is nevertheless clear that management of important fish resource species should be based on an understanding of all aspects of their life history, including the early stages. In the case of the Peruvian anchoveta, sufficient data are available on the dynamics of the . parent stock, and on the distribution and density of the eggs to justify an attempt to identify the, major causes of anchoveta egg mortality. *ICLARM Contribution No. 507.

155

156

Materials and Methods

Two sources of data were tapped for this contribution: i) length-structured biomass estimates obtained by VPA (Pauly and Palomares, this voL) and which were used to estimate parent biomass, mature female biomass and egg production, by month, for the period January 1953 to July 1985, using a model to be briefly described further below; and ii) Ninety maps of anchoveta egg distribution based on egg surveys conducted from 1964 to 1985, and which were published and subjected to preliminary analyses by Santander (1987) and Senocak et aL (this vol.). The model used here to estimate the monthly reproductive output (RO), Le., egg production, of anchoveta has the form: n ... 1)

ROi = !-I Bij ' Pij' Aj' Sk' PFj' RF J=

where j is the number of length classes for which, in a given month i, anchoveta biomass estimates are available (see Table 1 for the definition of the other terms). Fig. 1 gives an example of the type of maps published by Santander (1987) and the first five columns of Table 2 summarize the key information pertaining to or extracted from these maps. Most important here is the fifth column, Le., the estimates of egg standing stock, derived by planimetry (Santander 1987). [Some of these maps represented surveys which had covered less than the 4-140S stretch of the Peruvian coast used here as reference. The estimated egg standing stocks for these maps have been, in these cases, extrapolated to the whole reference area using a procedure documented in Santander (1987). The detailed planimetry of Senocak et al. (this vol.), which did not involve any large-scale extrapolation, shows that no detectable bias was introduced by Santander's extrapolation procedure.] Table I. Specifications of the multiplicative model used by Pauly and Soriano (1987) to estimate the egg production of Peruvian anchoveta (equation I), by month (i), based on biomass data by length class (j)a.

Tabla 1. EspecljicaclOnes del modelo mulliplicalivo usado por Pauly y Soriano (1987) para eslimar la producc/on de huevos de anchovela (ecuaci6n 1) por meses (i) basado en dalos de biomasa por clase de longilud (i).a Variable or constant

Definition

Remarks and/or source of estimate

(units)

Bij

Biomass of male and female anchoveta

(tonnes)

From VPA ill output of Pauly and Palomares (this vol.)

Pij

Fraction of mature fish

(dimensionless)

Derived from logistic curve with shape and position varying with SST (see Pauly and Soriano 1987)

A"J

Size-specific factor relating anchoveta fecundity per unit weight to length

(dimensionless)

See Tables 2 and 3 in Pauly and Soriano (1987)

Sk

No. of spawnings per month, k=1 (January); k=12 (December)

(lit)

From Table 3 in Pauly and Soriano (1987), based on Jordan (1980)

PFj

Fraction of females in parent stock

(dimensionless)

From Fig. 4 and Table 3 in Pauly and Soriano (1987), based on Clark (1954), Miiiano (1958) and Jordan (1959)

RF

Relative batch fecundity

(eggs/g)

Value of 596/g female taken from Santander et al. (1984)

109

Factor for adjusting gram to tonne

(dimensionless)

Used, but not given in Equation (I)

n

Number of length classes used in month (i)

(dimensionless)

Variable between months

a See Pauly and Soriano (1987) for further details on model derivation and specification and on data sources.

157 4'5

r------,,-------r---;r-----r---....,------,-----, Crulu

Operacion Eureka XLVI 4-70clober 1981 AnchovalQ egQI/m 2

6'5

•

I-~OO

lliillillI

501-1000

[[]

1001- 4000

•

>4000

8'5

Fig. 1. Anchoveta egg distribution off Peru (4-14OS), in early October 1981. Details on this map (No. 79 in Santander 1987). given here as an example. are provided in Table 2. Fig. 1. Dislribuci6n de huevos de anchoveta /rente al Peru (4·14°5) a comienzos de OClubre 1981. Detalles de esle mapa (No. 79 en Santander 1987), presenlado aqu£ como un ejemplo, se dan en la Tabla 2.

12"$

Mall no. 79

(0

Azul

Tombo de Mora

14'S L -_ _-----' 86"W

--'-

84W

82"W

-----L

_ _- - - - '

--'-_~___"_-'--

SooW

78"W

76"W

74"W

Pauly (1987) had attempted to estimate anchoveta egg mortality by relating theoretical egg production (as estimated via equation 1), to the egg standing stock in Santander's maps. However, the model he used to calculate egg mortality (his equation 4) produced biased estimates when egg mortality was low (I.A. Gulland, pers. comm. to D. Pauly, December 1987), and his whole section on "the cannibalization of anchoveta eggs" (inclusive of his Table 2 and Fig. 4) is thus erroneous. We used instead an approach suggested by lA. Gulland (pers. cornm.). Defining Nd as the egg standing stock at the end of period D, and R as the initial size of a "cohort" of anchoveta eggs, it follows for that cohort, that: Nd

= Re- ZD

... 2)

where Z is the egg mortality from spawning/fertilization to hatching and D (days) is the egg development time. The parameter D can be estimated from: logleP = 6.953 - 4.09IoglO(T+26)

... 3)

where T is the sea surface temperature (SST, in °C) and which was derived by Pauly (1987) based on data in Santander and Sandoval de Castillo (1973) and equation (5) in Pauly and Pullin (1988). Under steady-state condition (assumed here to prevail shortly before, during and after a given egg survey was conducted and/or during a period of one month), equation (2) implies a mean standing stock (N) whose value can be estimated from:

1 D N = _jRe-ZD

D

...

4)

0

or R N =' _ _ (1 - e-ZD) ZD

... 5)

158 Table 2. Summary of data on anchoveta egg surveys conducted from 1964 to 1985 off Pern (4-14os)a. with added ancillary infonnation and some .derived statistics. Tabla 2. Resumen de los datos de los cruceros de evaluaci6n de huevos de anchoveta /levados a cabo de 1964 a 1985 frente al Peru (4-14°S) con alguna informaci6n adicional y estadfsticas derivadas. M.p Rwming no.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 2S 26 27 28 29 308 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

73 74 75 76 77 78 79 80 81 82 83 84

M.p no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

standing stock

Momh

(eggx 1012).

1964

M" Apr lun Sop

1%5

Mar Apr lui Nov

27.4 17.0 12.9 137.0 60.3 51.4 8.03 53.3 44.0 16.2 4.94 418.0 97.4 6.36 292.0 116.0 50.2 169.0 '15.5 22.0 9.42 82.3 22.1 82.0 22.6 9.05 107.0 12.7 29.3 2.08 13.2 23.8 7.19 5.47 62.3 35.2 41.3 19.8 18.9 109.0 60.9 70.7 86.4 3.44 85.2 64.7 29.6 40.7 56.0 55.6 67.4 17.0 53.8 53.6 19.2 45.9 .313 28.0 25.4 17.1 2.67 64.7 6.44 25.5 21.6 30.1 66.2 69.8 21.8 68.2 27.3 10.3 18.1 74.5 86.4 31.6 9.79 49.6 85.6 96.2 7.76 47.4 38.0 10.8

Year

Dec

1%6

1%7

1%8

1%9

1970

reb Sop Deo l"" lui Slop M.y Sop

0" 1971

1972

31m 33 34/35 36 37 38 39 40/41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62/63 64/65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82/83 84 85 86 87 88 89 90

reb M.y Sop Nov M.y Sop Nov

Nov M.y Aug Nov

Feb lui Aug Sop

Dc' Dec 1973

1974

l"" Mar luo lui Aug Sop Nov

Feb May Aug Sop Nov

1975

Feb Aug Sop

Dec 1976

1977

1977 1978

l.n lol Aug Nov M" Apr lui Aug

D" Apr lui

0" Dec 1979

1980 1981

Feb lui Sop Nov lm Sop

FeI>

Apr Sop 1982

D" Feb Sop

Dec 1984

Sop

Dec 1985

Feb

1986

M" Aug Scp May

Daily reproductive output (egg A 1012 )1> 16.8 3.35 4.71 68.8 ( 1.89) 35.8 7.20 20.7 60.3 19.7 1.19 74.4 123.0 0.704 49.1 52.4 32.0 185.0 127.0 16.5 19.5 69.4 53.8 (259.0 ) (253.0 ) (135.0 ) 1.24 95.4 178.0 ( 44.3 ) ( 33.1 ) ( 79.8 ) (120.0 ) (106.0 ) 9.70 ( 5.06) (10.7 ) ( 2.75) ( 20.4 ) ( 23.5 ) (52.4 ) (170.0 ) (69.5 ) ( 2.79) (173.0 ) (116.0 ) (93.7 ) (29.3 ) (105.0 ) (196.0 ) 40.5 (50.8 ) 22.8 57.3 81.2 ( 9.30) ( 2.70) (23.9 ) (49.0 ) 42.5 ( 0.901) ( 4.83) (22.2 ) ( 5.54) 13.5 10.2 26.7 15.8 3.72 28.9 21.2 4.39 111.0 158.0 69.7 127.0 9.96 2.71 1.06 3.66 4.33

30.1 (12.9 ) ( 0.163)

Allchovcta

parent stock (IA 1(6)1> 5.01 3.19 1.04 1.54 (0.532) 7.36 8.26 3.30 3.07 6.55 3.10 1.44 2.50 1.33 0.998 1.09 2.70 2.74 2.16 2.35 1.72 1.77 14.10 (3.60) (3.60) (3.60) 3.02 3.83 3.72 (2.59) (1.77 ) (1.38 ) (1.12) (0.946) 2.34 (3.05 ) (3.07 ) (4.71 ) (9.53 ) (4.44 ) (6.70) (2.28 ) (2.27 ) (1.55 ) (1.58 ) (0.864) (0.941) (1.73 ) (1.37 ) (1.78 ) 1.71 (2.22 ) 1.93 2.06 2.16 (1.68 ) (1.53 ) (0.821) (0.632) 3.13 (1.19 ) (0.394) (0.377) (0.621) 1.14 0.492 0.385 0.283 0.461 0.360 1.63 1.80 1.16 1.59 2.36 0.782 0.412 0.lD94 0.114 0.269 0603 o.W2 (0.187) (0.327)

II Adllpled from Santander (1987, Table 3). b Fran (part of) equation (1) and biomasscs estimated by Pauly and Palomares (litis vaLl. e From Table I in Muck (this vol.), wilh monlhly values interpolated between annual means (used from JUllclJuly); lhcse estimates arc d From equation (3) and SST in Pauly and Tsukuyama (1987) for 1953lo 1982, and ffIXn Scnocak ct al. (lhis vol.) from 1983 to 1985. e From equation (6). f From equation (9). g Outlier, not used {or estimates of parameters of equations (6) and (9).

Sardine biomass c (, x 1(6) 0.32 0.31 0.30 0.31 0.34 0.36 0.37 0.38 0.35 0.33 0.31 0.30 0.30 0.30 0.30 0.30 0.30 0.30 0.31 0.31 0.32 0.32 0.30 0.30 0.30 0.30 0.30 0.47 0.98 1.50 2.18 2.18 2.18 2.18 2.18 2.18 2.18 2.18 2.18 2.18 2.17 2.17 2.15 2.13 2.14 2.15 2.17 2.20 2.20 2.19 2.12 2.10 1.98 2.03 2.18 2.37 2.42 2.52 2.48 2.39 2.42 2.52 2.39 2.24 2.32 2.49 2.50 2.52 2.53 2.57 2.57 2.57 2.57 2.57 2.57 2.98 3.60 3.76 3.21 2.84 2.65 2.12 2.15 2.34

VCIY

Egg development

Z

Eslim.:ucd Z

time (day)d

(d.y·IJe

(d.y·l)r

1.53 1.74 2.02 2.10 2.02 1.37 1.27 1.55 1.72 1.37 1.78 2.14 1.96 1.89 2.19 2.25 1.77 1.98 1.85 1.60 1.80 1.84 1.61 1.89 1.82 1.87 1.60 1.70 1.85 1.21 1.29 1.42 1.57 1.55 1.26 1.08 1.27 1.94 2.06 2.16 2.12 1.85 1.67 1.61 1.91 2.04 1.85 1.69 2.04 2.06 1.98 1.84 1.50 1.54 1.64 1.35 1.35 1.77 1.87 1.94 1.54 1.94 1.94 1.82 1.63 1.80 1.87 1.66 1.58 1.94 1.58 1.66 2.12 1.87 1.58 1.78 1.03 2.08 1.96 1.93 1.78 2.10 2.08 1.91

crude lind need refinemcnt, based 011

0.049

0.206 1.212 0.822

1.139

0.018 0.134 0.921 7.722 0.225 2.002 0.527 2.330 3.104 10.989 13.369 7.353 6.006 20.661g 2.385 3.324 15.924 21.505

0.915 0.634 2.43 0.381 0.34 1.967 1.766 3.183 0.243 1.841 3.465 0.182 3.007 0.705 4.352

0.507 1.85 2.433

3.432 0.025

0.277 6.188 2.097 0.329 4.01 0.10

0.289

(UIlUC

0.238 0.251 0.076 0.555 0.138 0.248 0.373 3.457 3.396 0.249 U,428 0.407 1.001 0.162 0.250 0.108 0.100 1.524 0.723 0.462 1.155 2.422 1.741 2.866 3.581 1.956 0.697 5.493 3.385 0.233 2.755 6.717 8.129 6.810 0.347 0.094 0.181 1.244 1.880 1.139 0.616 3.395 0.257 0.954 2.540 1.155 1.169 0.210 1.167 1.610 0.694 0.447 4.768 8.885 6.748 0.191 0.402 1.832 1.823 0.997 0.448 0.538 1.116 0.833 0.193 1.175 1.786 1.256 0.372 1.219 0.268 0.494 1.651 3.915 0.336 4.220 0.060 0.174 0.173 0.029 0.076 0.491 0.361 0.174

VPA cstimalCS of sardine biomasscs. .

159

Both Nand R can be turned into egg production rates through division by D, Le., 1- e -ZD

N/D

... 6)

=-----

RID

ZD

N/D then corresponds to the above-mentioned map standing stock estimates divided by egg development time (see also Senocak, this vol.), while RID corresponds to the output of equation (1), reexpressed on a daily basis. Various plausible multiple regression models were then applied to the data of Table 2, our aim being to identify biologically acceptable factors explaining the observed variability in egg survival. The Rojas/Alamo database on anchoveta food and feeding habits (see Rojas de Mendiola, this vol.; Alamo, this vol.; and Pauly et al., this vol.) was also tapped for data on anchoveta egg cannibalism and the results were used to help interpret the output of our egg mortality models.

Results and Discussion Parental Biomass and Egg Production, 1953 to 1985 Fig. 2 and Tables 3, 4 and 5 present the results obtained by applying equation (1) to the biomass data generated by Pauly and Palomares (this vol.). The estimates of parental biomass in Fig. 2 resemble those published earlier by Pauly and Soriano (1987). As was the case earlier, the lone independent estimate of parental biomass obtained for August/September 1981 using the egg production method (Santander et al. 1984) is very close to our estimates of parental biomass for that period (see also Table 3, footnote).

II

24 22

I

20 18

l.\J

Mature

99

Parent stock

( 99+ 00 > Parent and juvenile stock

16

o 14

CIl CIl

10

iii

8

o

Spawning stock estimate based on the egg production method

12

o E

6 4

2

o 1955

1960

1965

1970

1975

1980

1985

Year Fig. 2. Total biomass, parent (=mature) stock and biomass of mature female anchoveta off Peru (4-14°S). January 1953 to July 1985. NQte strong, regular seasonal oscillations and very good match with independent spawning stock estimate in August/ September 1981 by Santander et al. (984). Fig. 2. Biomasa lola I, slack de padres (=maduros) y biomasa de anchovelas hembras madurasfrente al Peru (4-14°S), Enero 1953 a Julio 1985. Nolar la fuerle y regular oscilacion eSlacional y una muy buena concordancia can la estimacit5n independienle del slock desovanle en AgoslolSepliembre1981 de Sanlander el al. (1984).

160 Table 3. Estimated biomass of mature female and male anchovela (Engraulis estimates for the early/mid-1970s tend to be biased upward (see text)~

ringens) off Peru (4-14°5), January 1953 to July 1985 (in tormes). Note that

Tab/a 3. Biomasa estimada de anchovetas madwas (Engraulis ringens) hembras y machos frente a/ Peru (4-14°S). Enero 1953 a Julio 1985 (en tone/adas). Notar que los va/ores para comienzos y mediados de 1970 tienden a estar sobreestimados (ver texto). Year

Jan

1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

2,324,629 778,346 4,628,339 2,511,945 1,407,115 3,280,197 606,482 698,823 4,050,912 4,776,121 2,136,645 3,845,888 4,154,754 4,440,739 4,057,425 3,197,645 2,345,165 6,310,369 1,895,066 5,886,226 3,768,454 2,597,888 1,010,533 2,127,963 1,918,159 245,908 1,002,829 461,452 929,657 2,031,170 120,318 9,966 197,855

reb 3,687,629 1,190,877 3,299,281 4,049,643 7,022,473 3,526,073 2,214,714 1,402,145 6,233,816 4,918,827 3,503,178 4,974,556 6,281,776 6,549,658 4,122,205 2,702,713 2,769,932 10,787,530 3,194,843 6,889,323 3,882,752 3,543,061 2,187,394 4,190,732 2,460,301 806,537 1,138,779 523,174 1,630,146 2,363,049 23,904 14,545 268,764

Mar 4,255,910 958,377 1,448,142 4,450,385 6,651,020 3,371,224 1,807,238 1,602,622 4,684,939 3,830,377 4,254,088 5,426,498 7,360,888 4,728,831 4,176,276 4,174,045 6,324,308 14,137,810 5,098,249 7,445,824 3,840,114 4,410,033 5,053,889 5,050,645 2,663,668 1,063,143 1,388,163 705,087 1,708,486 2,525,817 5,454 29,025 603,246

Apr

May

Jun

Jul

Aug

Sep

0

. ·.1f1I;...

-5

Q)

u;

7

Y = -3.52 - 3.9·10- X r = 0.590; d.t. = 82

••

-4 ........""'-

o:Z: :.c a"O

~ .

A

.

... t

2

4 6 8 W Parent stock ( t )( 106 )

12

14

-8

•

l-L~l-L..L.....J-L.-'-J.-,-.L...L-'-l-L-I-l-L..L.....Ji-L-'-J.-'-J...J.--'---

4.5

5.0

5.5

6.0

6.5

7.0

6 LoCJ 10 parent stock (t )( 10 )

Fig. 3. Evidence of density-dependence in anchoveta egg survival. A. Bivariate plot, as used to derive Ricker-type stock-recruitment relationships. B. Plot of the residuals of equation (I) vs. the line linking the index of egg survival (y-axis) and parental stock. Note markedly improved fit over Plot A, resulting from the inclusion of SST and ssTl as variables (see text). Fig. 3. Evidencia de la densodependencia en la sobrevivencia de huevos de anchove/a. A. Plo/eo bivariado, /al como es usado para derivar la relaci6n s/ock-reclwamien/o de Ricker. B. Plo/eode residuales de la ecuaci6n (7) vs la /{nea que relaciona el indice de sobrevivencia de huevos (eje Y) y el stock de padres. No/ar el mejor ajus/e en relaci6n al plo/eo A, a consecuencia de la inclusi6n de TSM y TSM 2 como variables (ver /ex/o).

Table 6. Statistics of multiple regression linking anchoveta egg survival index and its predictor variables (equation 7; degrees of freedom = 80, R = 0.771). Tabla 6. Da/os es/adis/icos de la regresi6n mUltiple que relaciona el indice de sobrevivencia de huevos de anchove/a y sus variables predic/oras (ecuaci6n 7; grados de liber/ad 80; R 0.771).

=

Variables

(Units)

Estimates

=

Standard errorsa

In (N/(D·P»

(eggs 109/ days. tonnes)

loglO parent stock

(tonnes)

-2.147

0.257

Sea surface temp (SST)

(OC)

..{).250

0.942

(SST)2

(oC)2

0.0627

0.0245

(independent variable)

0.952

a Standard error of Y-estimate

Taking the partial derivative of equation (7) with respect to temperature: 8ln (N/(D·P)) = 2 (0.0663)T - 2.50

... 8)

8T and setting it equal to zero allows estimation of the temperature (Tmin ) at which In (N/(D.P)) is. minimum, i.e., Tmin = 19.8°C. Thus, our analysis of Santander's maps provides evidence for: i) parental cannibalism on anchoveta eggs; and ii) a parabolic relationship between egg mortality and SST, with maximum close to the upper limit of the optimal temperature range of anchoveta (about 15-20°C). One multiple regression model we derived, using the data in Table 2, to explain variability of our estimates of egg mortality, is: 'loglOZegg = -23.7 + 0.60810glOP + 0.37910g1OS + 0.505AT + 2.4T - 0.0751T2

... 9)

163 >


...::J Ul

CI CI

Q)

0.5

-0.5

0

Ul 0

-I

::J -1.5

:9 Ul

Q)

a::

-2

.,

-.

0

15

16

17

18

19

20

Q)

B • • •• • • • • • • • •• •• • • • • .~ • •• • • • • .,. •• • • • ••• •• • • •• • • •

2

... ,

1.5

c>

•

;: ::J 05 . Ul

CI CI

•

..

-.

>< -C

c

•• •• • •• ••• • • • • • •• ..: • • • •• • •• •~. ••• • • • • • • • • • •

0

.:;

,

A

•

•• •

•• •

•

Q).

-C

0

Q)

_-0.5 0 ~

-I

.. - ... ..

0

.g-1.5 Ul Q)

21

22

a::

23

-2

250

300

-.

350

>
> ::J

...

C •••• ••

• •

0.5

Ul

., ..., •

5.5

6.0

••

CI CI

-

•

500

550

Ul>
0 -c> -5 Q)'> u ... ._ ::J -6

•

6.5

-2

0

::J-c _c

.,"

• "{[j]

- 8 IC-_-'--_-'-_-'--_--'-_----''--_-'--_-'-_-' -8 -6 -4 -2 0

7.0

Log lo parental biomass (t x 10 6

0

-I

Q)

•• •

•.,• .•••

• 5.0

•

, "•• .-• ...• • •

,•

450

0

.. •• •

•

A

• •

400

SST 2 (OC)

SST (OC)

Observed values of egg survival index

)

Fig. 4. Selected features of model represented by equation (1). A-C: residuals of equation (1) linking egg survival index and its predictor variables. Note absence of structure. D: plot of observed vs. predicted values of the egg survival index. Fig. 4. Caracterlsticas seleccionadas del madelo represenJado por la ecuaci6n (7). A-C: residuales de ecuaci6n (7) relacionando el {ndice de sobrevivencia de huevos y sus variables predictoras. Notar la ausencia de estructura. D: ploteo de valores observados vs valores predictivos del (ndice de sobrevivencia de larvas.

•

-2

•• • • •• • ••

• )(

IU "0 C

0

-3

-4

> >

II)

..

~5

Cl Cl

W

•• •

•

• • ••

•

•

• --.!. • • • • '• • •• • • • • • • •• • • •• • • •

... :::J

,

-6

..

•

15

16

17

18

I

19

20

21

22

23

SST (OC) . Fig. 5. Relationship of egg survival index and SST; note curvature. suggesting a minimum at 19.8°C. Fig. 5. Relaci6n entre el (ndice de sobrevivencia de huevos y TSM; notar curvatura, sug iriendo un minima a 19.B oC.

164

which has an R2 = 0.562 and whose statistics are given in Table 7. Fig. 6 shows that this model gives a particularly good fit to SST and SST anomaly, and a less.er fit to the biological variables (parent stock and sardine stock). Again, SST has a parabolic relationship with egg mortality. Taking the partial derivative of equation (9) with respect to T gives:

a10glO L.egg

= 2.4 - 2(0.0751) T

aT

... 10)

which set equal to zero and solved for T gives T = 16.0°C. Thus, we find again that anchoveta egg momiity has a peak within the range of temperature that is optimal for anchoveta. ;;.,

.g

>.

o.a

:

Q

0.2

01

o

..-l

••

•••

~ ~.

~

-0.4 -0.6

•

•• __

~ -o.~~

QI

B •

..~....

~t: ~:: o

A

• •

I. 2 1.0

• • •,

-o.a

I

•

,

!

•

,. ,

,, 6.0

I

o ...

0.4

~

0

...

~

•

•

.t :

01

.-

I

5.5

•

:

•

>.

0.4

,!j -0.8 I

!

,

, ,

6.5

o

..-l

-1.2

I

7.0

.

••

2

•• • •••

~

>.

0

o... ...o

-I

:t:

E 01 01 QI

o

cil

C

..

".4 ...•

.. ~

-3 -4

.g

2.5

.r; _

2.0 1.5

~

1.0

01

0

E 0.5 ~ -0.5 o -I -1.5

g>

o

..-l

L-......L_...L._!----l._-'-_!----I._......._..J

15

16

17

18

19

20

21

22

23

24

SST 1°C) I

..-:.._--------~ ..I:. . ·

~

••

•

• 2.5

0.6 0.4 0.2

o

-0.2 -0.4 -0.6 -I -1.2

• , 3.0 3.5 6 Log10sardine stock It x 10 )

•

• • •• . :~•.... •

•• ~

•

..-l

-3

-2

•

-I

Monthly

. 0

I

2

3

4

5

6

SST anomaly 1°C)

..E~ .. · · :y- .. .: - .. • •

o.a

-ua

.

T:., 3.0

.~

-2

~

-r..

Log lo anchoveta parent stock It x 106 )

'>. o

• •• •

•

•

•

•

•

•

-1.6 -1.4 - 1 -0.6 -0.2 0.2 0.6 I Log observed egg mortality (da{l)

1.4

lo

Fig. 6. Relationship between anchoveta egg mortality and its prediclor variables. The solid lines of Plots A-D represent the axes of equation (9), with intercepts set at the mean of all other variables; the dots represent the residuals, plotted from the solid lines. Plot E shows the predicted vs. the observed mortalities. Note overall good fit of equation (9) lo the data, the absence of structure in the residuals, and the strong relationship between mortality, SST and SST anomaly. Fig. 6. Relaei6n entre la mor/alidad de huevos de anehove/a y sus variables predie/oras. LAs Uneas s6lidas de los plo/eos A-D representan el eje de la eeuaei6n (9), eon in/ereep/os fijadas al promedio de lodas las o/ras variables; los puntas represen/an los residuos, ploleados de las Uneas s6lidas. EI Plo/eo E mues/ra la mor/alidad es/imada vs la observada. No/ar el buen ajusle general de la eeuaei6n (9) a los dalas, /a auseneia de es/rue/UTa en los residues y la fuer/e relaci6n entre mor/alidad, TSM y la anomalia de TSM.

165 Table 7. Statistics of equation (9), relating anchoveta egg monality and its predictor variables (degrees of freedom 43; R 0.751). Tabla 7. Datos estadlsticos de la ecuacion (9), que relaciona la mortalidad de huevos de anchoveta y sus variables predictoras (grados de /ibertad 43; R 0.751).

=

=

=

Variables log 10 egg mortality log 10 parent stock log 10 sardine stock SST anomaly (AT) Sea surface temp. (SST) (SST)2

=

(Units)

Estimates

(day-I) (tonnes) (tonnes) (0C) (OC)

(indep. variable) 0.6078 0.3788 0.5049 2.4007 -0.0751

(Oq2

Standard errors a 0.4937 0.2851 0.2160 0.0758 0.6881 0.0185

a Standard error of Y-estimate.

Fig. 6D, which shows the positive partial correlation between anchoveta egg mortality and SST anomaly, suggests, however, that tbe observed parabolic relationship between SST and egg mortality do not reflect a causal linkage. Rather, SST, which fluctuates seasonally, is lowest in September-October (see Fig. 8 and Bakun 1987), during a period when anchoveta concentrate under the coast for spawning (Jordan 1971; FAa 1981; Csirke, this vol.). Anchoveta feed during the spawning season and hence, the peak consumption.of anchoveta eggs by adult anchoveta occurs in September-October (Fig. 7; Table 8). This could explain the apparent relationships between SST and egg mortality. As for SST anomaly, we assume that it is linked with anchoveta egg mortality via two mechanisms: (1) increased predation on anchoveta eggs by zooplankton and other predators not included in our models, and whose metabolic rate and hence, food consumption, can be expected to increase when the SST anomaly increases and (2) increased concentration of anchoveta within the small inshore patches of low SST waters that remain during high SST anomalies (Muck et al., this vol.). 1.5

6

._go

CII

-CI)

.t::

0.>

0

0

Eo

0

c c:::

E 5

-

:J.t::

CII 0

CII

00

C

...

CI)

1.0

4

-.-

o~

E :J

0' 0'

CI)

.t::>. 0"CI

c:::

0 0..Q

c CI)

CI)

~ ~

3

CI)

0

0' .... 0'0

CI)_

..Q

c

°

0.5

2

:E

>.01

='o

CI)

00.

a

J

FMAMJJ ASOND Month

a

Fig. 7. Mean number of anchoveta eggs in the stomach of anchoveta sampled from 1953 to 1982 by Rojas de Mendiola (this vol.) and Alamo (this vol.), as extracted from their database (see Table 8 for details). Note strong seasonality, with a major peak during th.e major spawning season (September/October) and a minor peak during the minor spawning season (February/March). The scale on the right side provides an approximate conversion to egg consumption rate (see Table 8). . Fig. 7. Nl1mero promedio de huevos de anchoveta en el est6mago de anchoveta muestreada de 1953 a 1982 por Rojas de Mendiola (este vol.) y Alamo (este vol.J, extraido de su base de datos (ver tabla 8 para detalles). Notar lafuerte estacionalidad, con un pica mayor durante Ia estacion de desove prj~cipal (SeptiembreIOctubre) y un pica menor durante la estacion de menor desove (FebreroIMarzo). La escala del lado derecho proporciona una conversion aproximoda de la tasa de consumo de huevos (ver Tabla 8).

166 Equation (~) predicts, for anchoveta eggs in August/September 1987, a value of Z = 2.13 day-l;.this is much higher than the value of 0.91 day-I presented by Santander et al. (1984). . Data are available which might allow resolving this discrepancy, and testing of the hypotheses presented above. Notably, a detailed analysis of the Rojas/Alamo database on anchoveta stomach contents emphasizing seasonal and interyear variations of egg cannibalism (i.e., going beyond the mean seasonal cycle in Fig. 8), and combined with corresponding estimates of anchoveta concentration, could contribute toward elucidating the role of egg cannibalism in anchoveta population dynamics. Table 8. Data for quantification of egg cannibalism in anchoveta. as extracted from the Rojas/Alamo database (see Rojas de Mendiola, this vol.; Alamo, this vol.; Pauly et al., this vol.). Tabla 8. Datos para la cuanJijicaci6n del canibalismo de hlU!vos de anchoveta, extraidos de la base de datos de Rojas/Alamo (ver Rojas de Mendiola, esle vol.; Alamo, esle vol.; Pauly et al., este vol.).

Month

No. of anchoveta with sampling month and record for eggs

No. of anchoveta with eggs in the stomach

616 769 852 633 409

38 77 22 17 7 6 23 204 520 99 112 53

January February March April May June July August September October November December

266 589 996 1698 869 1057 971

Total no. of eggs in stomach

Mean eggs per anchoveta stomach

Daily Anchoveta egg consumptiona

0.205 1.300 0.445 0.095 0.042 0.128 0.681 1.639 6.440 1.915 1.390 0.194

0.051 0.325 0.111 0.024 0.010 0.032 1.170 0.410 1.610 0.479 0.348 0.048

126 1000 379

60 17 34 401 1632 10935 1664 1469 188

a Approximate daily anchoveta egg consumption per g body weight of anchoveta (see also Fig. 7), as computed from the daily ration of 0.448 g estimated by Paulyet al. (this vol.) for'anchoveta of about 20 g live weight, and a mean stomach content of 0.122 g. The quotient of mean stomach content over ration is '" 3.7, but this estimate of turnover rate was increased to 5 to account for the facts that (I) most stomachs with eggs were sampled early mornings, when stomach contents are less than average and (2) that eggs are more rapidly digested than other food items.

19.5

~

T :>. c

p---o, I

4

't:J

/

/

I

o

''

'0 \ \

\ \ \

N :>.

0

E

18.5

\

3

\

U

\

:: '0

-.

19

0

\

~

\ \

2

'0\

01 01 CI)

I

I

18

I

\

\

\

\

C

,

C CI)

'0_

~

0

I

I

I

I

I

I

I

I

I

I

I-

en en

I

17.5

17

--0

I

J

F

M

A

M

J

J

A

S

0

N

0

Month Fig. 8. Solid line: annual cycle of mean monthly anchoveta egg mortality, 1964 to 1985, based on map-specific Z values in Table 2, or on estimates from equation (9) when these could not be computed via equation (6); monthly means were smoothed over 3 months. Note strong summer peak, coincident with peak' occurrence of eggs in anchoveta stomachs (see Fig. 7). Dotted line: mean seasonal cycle of sea surface temperature, as derived using only SST values for the months represented in Table 2, and smoothing over 3 months. . Fig. 8. Lfnea s6lida: ciclo anual del promedio mensual de la mortalidad de hlU!vos de anchoveta, 1964 a 1985, basado en valores de Z obtenidos de mapas espec(ficos de la Tabla 2, 0 en eSlimaciones de la ecuaci6n (9j cuando islos no pudieron ser COmpUlados por la ecuaci6n (6);10$ promedios mensuales fueron suavizados mediante promedio corrido de 3 meses. Notar elfuerte pica de verano, coincidente con la ocurrencia del pico de' huevos en los est6magos de anchoveta (ver Fig. 7). Lfnea interrumpida: ciclo estacional promedio de la temperatura superficial derivado usando solo valores de TSM para los meses presenJados en la Tabla 2, y suavizados con promedios corridos de 3 meses.

·

'

167

References Albeit, J. 1986. Egg caIUlibalism versus egg predation: their significance in anchovies. ICES C.M. 19861H:59, Pelagics Cttee, 9 p. Bakun, A. 1987. Monthly variability in the ocean habitat off Peru as deduced from maritime obselVations, 1953 to 1984, p. 46-74. In D. Pauly and I. Tsukayama (OOs.) The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Smdies and Reviews 15, 351 p. Instimto del Mar del Peru (IMARPE), Callao, Peru; Deutsche Gesellschaft fUr Teclmische Zusammenarbeit (GTZ), GmbH, Eschbom Federal Republic of Germany; and International Center for Living Aquatic Resources Management (lCLARM), Manila, Philippines. Cark, F.N. 1954. Biologla de la anchoveta. Bo1. Cient Cia. Admin. Guano 1(1): 98-132. Cushing, D.H. 1988. The study of stock and recruitment, p. 105-128. In J.A. Gulland (ed.) Fish population dynamics. 2nd ed. J. Wiley and Sons, NewYor}c. FAO. 1981. Atlas of the living resources of the sea. Prepared by the FAO Fisheries Department FAO Fish. Ser., Rome. Jordan, R. 1959. ObselVaciones sobre la biolog(a de la anchoveta (Engraulis ringens J.) de la zona pesquera de Huacho. Bo1. Cient Cia Admin. Guano 35(11): 3-22 + app. Jordan, R. 1971. Distribution of anchoveta (Engraulis ringens) in relation to the environment Inv. Pesq. 35(1): 113-126. Jordan, R. 1980. Biology of the anchoveta. I: Summary of the present knowledge, p. 249-276. In Proceedings of the Workshop on the Phenomenon known as 'El Nino', Guayaquil, Ecuador, 4-12 December 1974. United Nations Educational, Scientific and Culmral Organization, Paris, 284 p. Minano, J. 1958. Algunas apreciaciones relacionadas con la anchoveta peruana y su fecundidad. Bol. Cient. Cia. Admin. Guano 34: 11-24. Pauly, D. 1987. Managing the Peruvian upwelling ecosystem: a synthesis, p. 325-342. In D. Pauly and I. Tsukayama (eds.) The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Studies and Reviews 15, 351 p. Instituto del Mar del Peru (lMARPE), Callao, Pe'rU; Deutsche Gesellschaft fur Teclmische Zusammenarbeit (G12), GmbH, Eschbom, Federal Republic of Germany; and International Center for Living Aquatic Resources Management (ICLARM), Manila, PhilippiI)es. Pauly, D. and R.S.V. Pullin. 1988. Hatching time in spherical pelagic marine fish eggs in response to temperature and egg size. Environ. BioI. Fish 21(2): 261-271. Pauly, D. and M. Soriano. 1987. Monthly spawning stock and egg production of Peruvian anchoveta (Engraulis ringens), 1953 to 1982, p. 167178. In D. Pauly and I. Tsukayama (eds.) The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Smdies and Reviews 15, 351 p. Instituto del Mar del Peru (lMARPE), Callao, Peru; Deutsche Gesellschaft fUr Teclmische Zusanunenarbeit (G12), GmbH, Eschbom, Federal Republic of Germany; and International Center for Living Aquatic Resources Management (ICLARM), Manila, Philippines. Pauly, D. and I. Tsukayama, Editors. 1987. The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Smdies and Reviews 15,351 p. Instituto del Mar del Peni (lMARPE), Callao, Peru; Deutsche Gesellschaft fiir Teclmische Zusammenarbeit (G12), GmbH, Eschbom, Federal RepUblic of Germany; and International Center for Living Aquatic Resources Management, Manila, Philippines. Ricker, W.E. 1954. Stock and recruitment. J. Fish. Res. Board Can. 11: 559-623. Rothschild, B.J. 1986. Dynamics of marine fish populations. HalVard University Press, Cambridge, 277 p. 0 Santander, H. 1987. Relationship between anchoveta egg standing stock and parent biomass off Peru (4-l4 S), p. 179-207./n D. Pauly and I. Tsukayama (eds.) The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Studies and Reviews 15, 351 p. Instituto del Mar del Peni (lMARPE), Callao, Peru; Deutsche Gesellschaft fUr Teclmische Zusammenarbeit (G12), GmbH, Eschhom, Federal RepUblic of Germany; and International Center for Living Aquatic Resources Management (ICLARM), Manila, Philippines. Santander, H. and O. Sandoval de Castillo. 1973. Estudio sobre las primeras etapas de vida de la anchoveta. Informe Inst. Mar Peru-Callao 41(1):28. Santander, H., J. Albeit and P.E. Smith. 1984. Estimacion de la biomasa de la poblaci6n desovante de anchoveta Peruana Engraulis ringens en 1981 por aplicaci6n del "metodo de produccion de huevos". Bol. Inst. Mar. Peru-Callao 8(6):213-250. Sharp, G.O., Editor. 1980. Workshop on the effect of environmental variation on the sUlVival of lalVal pelagic fishes. Intergov. Oceanogr. Comm. Workshop Rep. No. 28. 323 p.