Temporal and Spatial Variations in the Spawning Activity of the Micronektonic Shrimp, Sergia lucens (Hansen) in Suruga Bay, Japan*

Transcription

1 Journal of the Oceanographical Society of Japan Vol.45, pp.243 to 250, 1989 Temporal and Spatial Variations in the Spawning Activity of the Micronektonic Shrimp, Sergia lucens (Hansen) in Suruga Bay, Japan* Gretchen H. Bishopt, Makoto Omorit and Fumio MuranakaT Abstract: Temporal and spatial variations in the spawning activity of Sergia lucens were investigated in relation to some environmental conditions of Suruga Bay. The daily egg abundance varied considerably with the coefficient of variation from 81% to 269% in July. The spawning activity was most clearly affected by temperature, but the relationship to lunar period- and content of chlorophyll a were not obvious. Timing ofthe July spawning is predictable with increase of the surface temperature to 24 Ž and strorig vertical movements of the 18 Ž isotherm depth; it is also related to modal length of the shrimp in June. It is suggested that intrusion of cold water at m affects reproduction of the shrimps and vertical distribution of eggs and larvae. The shrimp population seemed to relate principally to two spawning grounds, i.e. the head part and the western part of the bay. The timing of spawning is not always synchronous throughout the bay. The spawning is sporadic and the distribution of eggs is patchy. This may reflect a recent decrease in the population of the shrimp due to increased fishing pressure. 1. Introduction The"Sakura-ebi"Sergia lucens (Hansen) is a micronektonic sergestid shrimp. As it has provided the base of a historic fishery in Suruga Bay, on the south coast of Honsyu, since.1894, research on the shrimp is well advanced. Details of the life history, the ecology, and the fishery have been described (e.g. Omori, 1969, 1974; Omori et al., 1973; Nakamura and Tsukui, 1982). The shrimp migrates up to about 50 m depth at night from a depth of m. There it * Received 13 September 1988; in revised form 25 April 1989; accepted 30 May A part of the present paper is based on the thesis submitted by G. H. B. in partial fulfilment of the requirements for her degree of Master of Science at the Tokyo University of. Fisheries in forms dense swarms which are easily captured by the local trawl fishery at night. Spawning occurs at 1 year of age; from June to October. The larvae attain a length of 20-25mm and recruit to the fishery at an age of 3-4 months old. There are two fishing periods: one in the fall from October to December, and the other in the spring from March to June. The fall catch is composed of the pre-spawning 0-year group and the spawning/post-spawning 1-year group, whereas the subsequent spring catch is composed of the 0-year group; this is an almost unavoidable situation for a fishery based on a species with a lifespan of only 1.5 years, but means that recruitment is instrumental in determining year-class strength. In considering recruitment of a species, such as S. lucens, with planktonic larval development, one may expect that the hydrography of the Tokyo University of Fisheries, 4-5-7, Konan, spawning and nursery areas is of vital importance. Minato-ku, Tokyo 108, Japan. Omori (1971) determined, by laboratory rearing Present address of G.H.B.: Mendenhall of the species, an optimum temperature range of Loop Rd., Juneau, AK 99801," U. S. A Ž for larval development and survival. Station, 3690 Kogawashioiri, Yaizu, Shizuoka 425, Japan. Omori et al. (1973) reported a positive correlation between average 50m depth water tem-

2 Bishop, Omori and Muranaka perature of the spawning grounds from June- August and the resulting year-class strength. Nakamura (1982) observed a positive relation between year-class strength and yearly average width of the Ž temperature zone from June through August at the head of the bay. In order to correlate the hydrography and abundance of the eggs and to assess year-class strength of S. lucens, Shizuoka Prefectural Fisheries Experimental Station (S. P. F. E. S.) continues sampling of the eggs and larvae once a month at a number of stations covering the entire Suruga Bay. However, short-term variability of release of the eggs has not been well documented. Thus, for, precise estimates of egg production, it is necessary to know how spawning activity varies by day and by place and whether the spawning is synchronous in the whole shrimp population of the bay. The present paper deals with the variation of daily production of eggs at two selected locations in the spawning ground. We attempted to determine the variability in egg abundance and the possible causes affecting the spawning activity, and to improve sampling strategy accordingly. 2. Materials and methods The eggs of Sergia lucens are primarily spawned at the head of Suruga Bay where the pre-spawning aggregation is found near the mouth of the rivers. The eggs are planktonic, and are distributed mostly at depths ranging from 20 to 50m. They hatch into nauplii after hrs under normal environmental temperature conditions (Omori, 1969, 1974). In 1985 and 1986, a programe of daily sampling of the 0-50m water column by vertical tow of a NORPAC net with 100 ttm mesh size was planned for the months of July at two locations, Stations A and B, situated offshore at Yui and Kanbara respectively (Fig.1); samples were actually obtained on 26 and 25 occasions at two locations respectively in 1985, and 23 and 21 occasions respectively in Sampling was conducted once every 5 or 7 days during August and September. Ocassionally, supplementally samplings were made east of Station B, near the mouth of the Fuji River. Filtered volume of water was determined using a flowmeter (Rigosha Co. Ltd.), and samples were fixed with buffered formalin to 5%, split twice, and the resulting 1/4 counted Fig.1.Location of daily sampling (A and B) and monthly survey (18-29) stations in Suruga Bay. to determine occurrence of eggs and larval stages of S. lucens. Water temperature measurements from the surface to 100m depth by a digital display bathythermograph (Kankyo Keisoku System Co. Ltd., model DBT-D1) were conducted simultaneously with the net sampling. In July 1986, waters from 0, 5, 10, and 15 m depth were sampled for determination of particulate chlorophyll. The sample water was filtered using Whatman GF/C filters and spectrophotometrically analysed for chlorophyll a content. Abundance and distribution of eggs and larvae were analysed in relation to possible causes of the temporal and spatial variations. In order to know abundance and distribution of eggs and larvae in the entire bay, samples were obtained by S.P.F.E.S. once a month from June to October with vertical hauls using the same NORPAC net from 150m to the surface at 8 stations (Fig.1). Vertical profiles of water temperature and salinity were determined at each station. Egg production was estimated using an equation (Omori, 1983): where C is the total number of eggs per square meter in the sampling area during the period t;pni is the estimated number of eggs per square meter; D is the time elapsing between the egg being laid and the nauplius hatching out; Wi is

3 Spawning Behavior of Sergia lucens the area represented by station i; ti is the timeweighting given to station i, i.e., one-half the time elapsing since the preceding sampling plus one-half the time elapsing before the succeeding sampling. The values for monthly egg production were then calculated. Duration of egg to hatching was calculated for average monthly temperatures of 20-50m depths at the end of the month. The start of heavy spawning was roughly associated with warming of the surface waters to 24 Ž and preceded vertical oscillation of the 18 Ž isotherm depth, which is the lower limit of the optimum temperature range for larval development and survival. The temperature around 40 m rapidly increased to days before the spawning peaked. In August, the surface temperature continued to increase, reaching maxima of 28 Ž and 26 Ž in 1985 and 1986, respectively. The shallowest depths of the 18 Ž isotherm for the two years were 31 and 39 m. The 18 Ž isotherm depth at all stations where the eggs occurred. An equation, Y=-2.69X , where Y is the duration between the egg being laid and the nauplius hatching out (hr) and X is temperature ( Ž) (S.P.F.E.S., 1984), was used to determine the duration at the ambient temperature. 3. Results Monthly average egg abundance was always greater at Station B than at Station A (Table 1), and often many eggs were also collected near the mouth of the Fuji River. Figure 2 shows egg abundance at Stations A and B in relation to the m temperature profile at Station B. Daily fluctuation in egg abundance was considerable. The coefficient of variation of egg abundance in a month (standard deviation/sample mean ~100%) varied from 81% to 269% in July (Table 1). There was no significant difference between temperatures at the two locations. Egg abundance at the two locations was positively correlated at the 95% level (r=0.74 in 1985, and 0.92 in 1986) indicating similar temporal trends in the spawning activity. In 1985, the spawning occurred from the beginning of July; numerous eggs were found until the end of the month. However, in 1986, eggs became numerous after the middle of July and heavy spawning was seen for only 3 days Fig.2.Sergia lucens. Variations of egg abundance at Stations A and B and m temperature profile at Station B for July-September Shadow indicates zones where temperatures are favorable for larval development and survival (18-25 Ž). Table 1.Sergia lucens. Monthly average egg abundance (7, no./m3), coefficient of variation (cv, %) and variance (s2) at Stations A and B. N means number of samples.

4 Bishop, Omori and Muranaka Table 2. Sergia lucens. Average temperature at 20-50m depth and egg abundance at 8 monthly survey stations in 1985 and Number of eggs/m3 in 1985 was estimated using average volume of water filtered by a net (5.5m3 for 0-50 m vertical tow). A flwmeter was used in Egg abundance above the monthly average values are indicated with black letter. was lowered at the beginning of August. During the sharp thermocline and narrow zone of optimum Table 3. Sergia lucens. Monthly egg abundance in Suruga Bay in 1985 and 1986 ( ~1012 eggs). temperature (18-25 Ž) spawning did decrease. In 1986, the thermocline was less steep and the optimum temperature zone was wider than the previous year, but egg densities were low. In 1985, sampling at Stations A and B indicated a second peak in September, although the peak was inferred from only one sample showing high egg abundance. In 1986, the spawning activity did not recover at the head of the bay. 18 Ž depth in July 1985, but not in July 1986, when there was a period with a shallow 18 Ž isotherm and low surface temperature at the beginning of the month. Correlations with surface temperature and 18 Ž isotherm depth were insignificant in September of both years. Figure 3 shows how spawning activity in July 1986 was related to the surface temperature, lunar period, relative tidal height, and content of chlorophyll a. Correlations between egg abundance and tidal height, and between egg abundance and lunar period were not significant at the 95% confidence level. There was also no consistency in the sign of the relation between spawning activity and chlorophyll a content. The surface water in September 1986 was cooler than in 1985, but the average temperature at m was warmer than that in Table 2 shows monthly abundance of eggs in the entire bay in the two years. There are two main spawning grounds, i.e. at the head and in the western part of the bay. It is noteworty that the principal spawning area in 1986 apparently shifted from the head to the western part, centering around Station 25, off Yaizu. Eggs were particularly abundant in August and September, and more than 60% of the annual egg production was made in this area (Table 3). Table 4 shows correlation coefficients of egg abundance with surface temperature, 18 Ž isotherm depth, and tidal height for July-September. The correlation coefficients were positive and significant between spawning activity and surface temperature in July of both years. They were negative and significant between spawning and 4. Discussion Many external and internal factors might be involved in the spawning activity of Sergia lucens. Among them, however, temperature

5 Spawning Behavior of Sergia lucens Table 4. Sergia lucens. Correlation coefficients of egg abundance at Stations A and B with surface temperature, 18 Ž isotherm depth and tidal height for July-September Numbers with asterisk mean significant at the 95% confidence level. Fig.3.Sergia lucens. Variations of egg abundance, surface temperature, chlorophyll a in 0-15m layer, relative tidal height and lunar period in June 1986 (average of Station A and B). The vertical profile of chlorophyll a at Station A on 16 July 1986 is presented in the upper left corner. seems to be the most important. Spawning starts after the water at 20-50m depth is warmed to 18 Ž (Omori, 1974) and reaches a peak when the surface water temperature reaches 24 Ž. A seasonal thermocline with characteristic "secondary minimum temperature" becomes evident in he main spawning period. This phenomenon, associated with the advection of subsurface cold water, cools the water at m depth in Suruga Bay, (Nakamura, 1982). The effect varies both in time and space, but increases markedly in summer. The strong intrusion of cold water at 20-50m seems to affect shrimp's reproduction and the vertical distribution of eggs and larvae. As Nakamura (1982) suggested, the hydrography of the spawning area during the early formation of the seasonal thermocline seems particularly to influence the spawning. According to Matsuyama (1988 and pers. comm.), remarkable baroclinic tides are often observed in Suruga Bay in July-September, resulting in rapid change of temperature in the subsurface layers (maximum temperature range at the semidiurnal period was about 5 Ž in Uchiura Bay at the east head of

6 Bishop, Omori and Muranaka Table 5. Sergia lucens. Modal lengths of the commercial catch, March-June Suruga Bay.); the effect of such short, rhythmic variations on spawning activity cannot be disregarded. Reduced spawning in August may be related to unfavourable, warm temperatures. Although it has not yet been experimentally proved, it is possible that under favourable conditions S. lucens produces two broods in one season with a considerable interval between. This may result in two spawning peaks in July and, September Moreover, the shrimp may not release eggs when phytoplankton is low, as total phytoplankton abundance declines to a summer minimum in August when the seasonal thermocline is most evident (S.P.F.E.S., 1977). There are accounts of penaeid shrimps showing lunar periodicity; in spawning Panaeus duorarum, for example (Roessler et al., 1969). Ho et al.(1983) reported a relation between lunar period and the depth, density, and shape of S. lucens swarms, so lunar periodicity in the spawning activity of this species might be expected. In the present investigation, however, the first July spawning peak in each year came at the quarter moon and neap tide, but the second occurred at the new moon in 1985 and the first quarter moon in 1986, respectively, thus a consistent relation with lunar period is not indicated. In the discussion of spawning timing, it is necessary to consider not only environmental conditions, but also the biology of the shrimp population. Body length frequencies of the commercial catch from March through June, , were compared using data from S.P.F.E.S. (Table 5). Considering each calendar year as a unit, rather than examining the progression of each year-class, the spring length frequencies showed a trend of decreasing modal size from 35.5mm in 1985 to 33.5mm in A coincidence between modal length in June and data of the July maximum spawning peak suggests a possible close relation between size of the spawning animals and timing of the spawning. Omori (1969) found the smallest "post-spawned" female at 39 mm length and the smallest female with mature ovaries at 37 mm, so 37 mm is considered the smallest size at which the shrimp spawn. Using the growth curve given by Omori (1969), we find that the shrimp with modal length of 35.5 mm in 1985 would have reached sexual maturity by approximately 10 July, but in 1986, with modal length 2 mm less, the spawning may have been delayed to the end of July. This timing agrees with the observation that there was a difference of about 10 days between the start and the peaks in July, fitting in well with the approximate dates of "modal maturity". The fact that egg abundance at Station B was greater than that at Station A and the result of supplementally samplings east of Station B indicated that center of spawning ground at the head of the bay was near the mouth of the Fuji River. Therefore egg abundance at the head of the bay alone does not necessarily reflect total spawning effort. The 1986 samples at Stations A and B differ from the monthly survey, in that only one peak, in late July, was apparent, after which there was almost no spawning. Average egg densities near the head of the bay showed trends which also differed from those in the bay as a whole. Egg abundance in 1985 was high at the head of the bay, but the monthly survey suggested that the total egg production in 1985 was not much greater than in These discrepancies may not be particularly surprising in a plankton survey, since the complexity of plankton distribution with large temporal and spatial variability is extremely difficult to analyse from a limited number of samples. This study suggests that spawning of S. lucens is sporadic. With a gradual decrease in the shrimp population due to increased fishing pressure (S.P.F.E.S. 1984), the fishermen are of the opinion a) that the areas of aggregation, consisting of a number of swarms, of the adult population have become small and b) that the distance between areas of aggregations has increased. If the swarms are small and distinct, egg abundance at one location may vary considerably. As a swarm always consists of a single cohort, the duration of spawning of a swarm will decrease, and eggs will only occur for a short time. The results presented here seem to give cause for concern about the shrimp resource. Frequent sampling during the spawning season

7 Spawning behavior of Sergia lucens Table 6.Sergia lucens. Variance [V(Y)] and coefficient of variation [cv(y),%] of average egg abundance (no./m3) at Stations A and B in July 1985 and may greatly improve the precision of the estimate of egg production. This is needed to determine the timing of peaks and to plan the monthly surveys in relation to the spawning peak. For planning of future sampling strategies and/or data processing, we suggest samplings at two locations, one near the mouth of the Fuji River and the other off Yaizu in the western part of the bay. Additional samplings along the 200m isobath between the two locations would allow further determination of whether temporal trends in the entire bay are, in fact synchronous. This would avoid situations such as that of August and September 1986, when only temporal trends at Stations A and B were available and showed no spawning peaks for the two months, although more than 80% of the spawning activity occurred offshore of Yaizu. If the temporal variation of spawning in the bay could be represented by observations at one station from the head part and one from the western part of the bay, the two data sets, i.e. temporal variability data and monthly survey data, could be used to produce a time-space integrated egg production estimate. In order to determine how often it is necessary to sample at the spawning ground in order to obtain reliable data on monthly average egg abundance, the following equations may be used. V(Y)= N n' / N-1 * In principle, s2 contains daily variance and variance in a day, but the latter is ignored here due to lack of data. E s2 / n' where V(Y) is variance of average egg abundance (no./m3); N is number of days in a month; n' is number of sampling days in a month; s2 is daily variance of egg abundance. The precision of Y at different sample frequencies is estimated by the coefficient of variation of cv(y)=v(y)/y where cv(y) is the coefficient of variation of Y (%); Y is average egg abundance in a month (no./m3). Using actual data from Stations A and B in July 1985 and 1986 (Table 1), supposing Y is equal to Y and 52 means daily variance only*,v(y) and cv(y) were estimated for five levels of sample frequency (n'). The result (Table 6) indicates that, even though samples were collected every two days, values for cv(y) are comparatively large, for example 20% and 44% for Station B in 1985 and 1986 respectively. Apparently, a few random samplings in a month are not meaningful; only frequent, appropriate sampling based on hydrographic and biological information relevant to spawning activity can give precise estimates of egg production of S. lucens. Acknowledgements We would like to thank Professors M. Murano and S. Tanaka of the Tokyo University of Fisheries and Messrs. T. Ai and F. Tsukui of the Shizuoka Prefectural Fisheries Experimental Station for their advice, cooperation and encouragement during the study. The Yui-Harbour Fisheries Cooperative allowed us use of their facilities, while the "Sakura-ebi" Fisheries Cooperative generously lent the fishing vessels used for sampling. Special thanks are due to Messrs. A. Yamada and T. Hamamura of the latter Cooperative for their assistance in the sampling at sea. The present work was financially supported by the "Sakuraebi" Fisheries Cooperative and Mr. Hosaka's grant provided to Omori. Valuable advice given by Dr. J.B.L. Matthews and two anonymous reviewers is also acknowledged. Bishop is indebted to Mr. and Mrs. A. Asahi for hosting her visit to Yui for sampling and analysis. References Ho, C., H. Shimizu, and T. Aoyama (1983): The relationship between lunar phase and the fishing condition of a sergestid fishery. Bull. Jpn. Soc.

8 Bishop, Omori and Muranaka Fish. Oceanogr., 42, Matsuyama, M. (1988): Circulations and tidal currents in Suruga and Sagami Bays. Bulletin on Coastal Oceanogr., 26, (in Japanese). Nakamura, Y. (1982): Oceanographic features of Suruga Bay from view point of fisheries oceanography. Bull. Shizuoka Pref. Fish. Exp. Sta., 17, (in Japanese). Nakamura, Y. and F. Tsukui (1982): On the effect of qualitative change of fishing effort on the resources of sergestid shrimp, Sergia lucens, in Suruga Bay. Bull. Shizuoka Pref. Fish. Exp. Sta., 16, (in Japanese). Omori, M. (1969): The biology of a sergested shrimp Sergestes lucens Hansen. Bull. Ocean Res. Inst. Univ. Tokyo, 4, Omori, M. (1971): Preliminary rearing experiments on the larvae of Sergestes lucens (Penaeidea, Natantia, Decapoda). Mar. Biol., 9, Omori, M. (1974): The biology of pelagic shrimps in the ocean. Adv. Mar. Biol., 12, Omori, M. (1983): Abundance assessment of micronektonic sergestid shrimp in the ocean. Biol. Oceanogr., 2, Omori, M., T. Konagaya, and K. Noya (1973): History and present status of the fishery of Sergestes lucens (Penaeidea, Decapoda, Crustacea) in Suruga Bay, Japan. J. Cons. Int. Explor. Mer, 35, Roessler, M. A., A. C. Jones, and J. L. Munro (1969): Larval and postlarval pink shrimp Penaeus duorarum in south Florida. FAO Fish. Rept., 57(3), Shizuoka Prefectural Fisheries Experimental Station (1977): Surugawan gyojokaihatsuchosa hokokusho (Report on exploitation of the fishing grounds in Suruga Bay). 242 pp. (in Japanese). Shizuoka Prefectural Fisheries Experimental Station (1984): Sakuraebi zoshokutaisakuchosa hokokusho (Report on mariculture and resource management of Sergia lucens). 106 pp.(in Japanese).