Short-term fluctuations in bivalve larvae compared with some environmental factors in a coastal lagoon ( South Portugal ) *

The Ria Formosa has a long tradition of bivalve harvesting, especially Ruditapes decussatus and Cerastoderma edule, contributing 90% of Portuguese production. The high productivity of shallow coastal lagoons, such as Ria Formosa, is the result of the contact of the water column and the sediment with good light conditions (Sprung, 1994a). Despite their ecological importance, bivalve larvae have not received much attention in this area or in other locations. Seasonal and annual observations of the occurrence and abundance of bivalve larvae in the plankton have been made by a few authors, such as Lebour (1938), Sullivan (1948), Rees (1954), Hrs-Brenko (1971), Alcaraz and Domínguez (1985), Mann (1988), Kalyanasundaram (1987), Leal (1989) and Chícharo (1993, 1996). General knowledge of bivalve larvae ecology will allow a better understanding of the environmental processes that determine the magnitude of bivalve recruitment. The dynamics of larvae from planktonic populations have important implications for the life histories of marine benthic invertebrates (Vance, 1973a,b; Strathmann, 1985; Hines, 1986). Thorson (1950, 1966) highlighted the impact of larval mortality in the recruitment of benthic populations. In fact, rates of growth and natural mortality can vary among populations of larvae (Chícharo and Chícharo, 2000), and the availability of settling larvae may influence patterns of recruitment. From the perspecSCI. MAR., 64 (4): 413-420 SCIENTIA MARINA 2000


INTRODUCTION
The Ria Formosa has a long tradition of bivalve harvesting, especially Ruditapes decussatus and Cerastoderma edule, contributing 90% of Portuguese production.The high productivity of shallow coastal lagoons, such as Ria Formosa, is the result of the contact of the water column and the sediment with good light conditions (Sprung, 1994a).
Despite their ecological importance, bivalve larvae have not received much attention in this area or in other locations.Seasonal and annual observations of the occurrence and abundance of bivalve larvae in the plankton have been made by a few authors, such as Lebour (1938), Sullivan (1948), Rees (1954), Hrs-Brenko (1971), Alcaraz and Domínguez (1985), Mann (1988), Kalyanasundaram (1987), Leal (1989) and Chícharo (1993Chícharo ( , 1996)).General knowledge of bivalve larvae ecology will allow a better understanding of the environmental processes that determine the magnitude of bivalve recruitment.The dynamics of larvae from planktonic populations have important implications for the life histories of marine benthic invertebrates (Vance, 1973a,b;Strathmann, 1985;Hines, 1986).Thorson (1950Thorson ( , 1966) ) highlighted the impact of larval mortality in the recruitment of benthic populations.In fact, rates of growth and natural mortality can vary among populations of larvae (Chícharo and Chícharo, 2000), and the availability of settling larvae may influence patterns of recruitment.From the perspec-tive of supply-side ecology, the importance of larval abundance for the recruitment of benthic marine invertebrates is a determining factor in structuring benthic marine communities (Gaines and Roughgarden, 1985;Minchinton and Scheibling, 1991;Miron et al., 1995).However, this approach has been widely overlooked by benthic and planktonic ecologists.Most of these investigations monitored larval patterns with a coarse temporal resolution (often months, according to Grassle andGrassle, 1974 andMaurer et al., 1979), and therefore failed to identify periods when structuring phenomena occurred (Roegner and Mann, 1995).Difficulties in the identification of bivalve larvae also hamper the calculation of overall larval abundance in the plankton.Only a very intense sampling regime, involving laborious and time consuming laboratory work with short-term intervals and simultaneous temporal sampling of planktonic larvae allows data series to be obtained in realistic time intervals.
The principal aim of this study was to analyse short-term fluctuations in bivalve larvae abundance in a coastal lagoon (Ria Formosa, South-Portugal), for a period of sixteen months, and compare them with some environmental factors.Water temperature, wind speed, tidal amplitude and a food availability indicator (chlorophyll a) were analysed in order to determine the triggering factors of bivalve larvae abundance fluctuations.

Study site
Bivalve larvae and post-larvae were collected at the Ria Formosa lagoon, in the south of Portugal.This is a tidal lagoon, consisting of a system of salt marshes, creeks and tidal flats extending for about 55 km in length and up to 6 km in width (Fig. 1).Connection with the Atlantic Ocean is established through two major openings (Faro-Olhão bar and Culatra bar).The exchange of water between the Ria and the adjacent ocean is about 50-75% in each tidal cycle (Águas, 1986).Water temperature varies from about 12-13ºC in winter to 27-28ºC in summer.However, in summer and during ebb tides, water temperature can rise to ≈ 30ºC in more interior and shallower regions of the Ria.Only one river, the Rio Gilão, flows into the lagoon.This is an unimportant water source with little impact on variation in salinity, which ranges between 25.5 and 36.9PSU (Falcão et al., 1985).

Environmental parameters
These parameters were measured two to three times per week during the spring and summer months in the larval sampling station (Fig. 1).This is when the more intense bivalve spawning occurs in the area (Chícharo, 1996).During the rest of the year, measurements were made once a week.Water temperature and salinity were measured with a Kent Eil 5005 MC5 probe.Wind velocity and direction data were obtained weekly from the Meteorological Service of Faro Airport, which is directly adjacent to the study site.Chlorophyll a was determined spectrophotometrically and measurements were corrected for phaeopigments (Lorenzen and Jeffrey, 1980).

Sampling and laboratory processing
Plankton samples were collected for a period of 16 months, from April 1990 to July 1991.One hundred and thirty two plankton samples, each of 20-40 litres of water, were taken using a water sample, at the surface on the ebbing tide on the larvae sampling station (max depth 3 m) (Fig. 1).All bivalve larvae densities were converted to an equivalent volume (one cubic meter).Plankton samples were replicated and filtered through 63-µm-mesh gauze immediately after collection.The collected material was fixed with 90% alcohol.Bivalve larvae were subsequently treated with 5% sodium hypochlorite solution to digest organic matter.During periods of peak abundance, plankton samples were divided using a Stempel pipette (UNESCO, 1968).Bivalve shells were counted and identified using a Zeiss IM35 inverted microscope.Individual larvae were measured in length (anterior-posterior axis), height (dorso-ventral axis) and hinge area at 100 or 400-X using an ocular micrometer.Where possible, larvae were identified to the genus or species level using the keys of Chícharo (1996) and Le Pennec (1978).

Data analysis
The total abundance of bivalve larvae was expressed in number/m 3 .The relationship between parameters was analysed by Pearson's correlation.To avoid assuming a significant correlation due to random processes, the Bonferroni inequalities (Snedecor and Cochran, 1989) were used in the analysis.The value of t 0.05 was corrected to t 0.05/n' (where n' is the number of pairs of correlations in the matrix).Only after applying this correction did we verify if a correlation was significant.

Environmental factors
Water temperature ranged between 13.1ºC and 27.5ºC, in January 1991 and July 1990, respectively (Fig. 2A).Salinity values normally varied close to that of normal sea salinity.The lowest value occurred during the winter months (32 PSU in February and March 1991), and a maximum of 38.4 PSU was observed in July 1991 (Fig. 2B).Wind velocity ranged between 0 and 9.7 m/s.Higher values of wind velocity were registered in December 1990 and May 1991, but no clear seasonal pattern was observed (Fig. 2C).The highest value of chlorophyll a was observed during summer (July; 13.23 µg/l) and lowest during winter (December; 0.4µg/l) (Fig. 2D).During the spawning season, the tidal amplitude (High tide-Low tide), between May 4 and August 31 1990, was registered and the spring tides occurred May 23-25, June 23-25, July 23-26 and August 22-24 (Fig. 3).Planktonic bivalve larvae were present throughout the entire sampling period (Fig. 2).Bivalve larval abundance began to increase in April 1990 and 1991, and the highest abundance occurred in the beginning of May of 1990.Counts of greater than 1,000 larvae/m 3 were frequent, except during the winter months (November-January).A small increase in abundance was noted from January until April 1991.However, an important variation in the abundance was observed between successive samples.
The analysis of the temporal distribution of the more abundant species showed that Mytilus galloprovincialis larvae were caught almost in all the months, but the maximal value of 4721 individuals/m 3 in May was counted.The abundance of larvae Ruditapes decussatus was high during spring and summer months (Fig. 4).Its maximum larval abundance (3,564 larvae/m 2 ) was observed in June 1990, and the minimum larvae abundance (0 larvae/m 2 ) was caught between November 1990 and March 1991.
The larval abundance of Cerastoderma edule was highest (3,942 larvae/m 2 ) in April and they were absent between December and November 1990 (Fig. 4).

DISCUSSION
Bivalve larvae occurred throughout the year in the plankton of the study area.This seems to be common in coastal lagoons.Alcaraz and Domínguez (1985) also observed this in the Ria of Pontevedra (Spain) as did Morgado (1993) in the Ria de Aveiro (Portugal), and Sprung (1994b) in the Ria Formosa.Lebour (1938) noted the continuous presence of bivalve larvae in the plankton throughout the year in Plymouth (U.K.), as did Rees (1954) in the North Sea and Hrs-Brenko (1971) in the Adriatic Sea.
The taxonomic composition of bivalve larvae agreed with the abundance of adult bivalves in the area.In fact, Cerastoderma edule and Ruditapes decussatus are the most abundant bivalves in the lagoon (Calvário, 1995).However, Mytilus galloprovincialis adults are scarce in the Ria and yet the larvae were the most abundant in the plankton and were always present in the samples.Alcaraz and Domínguez (1985), Leal (1989) and Cáceres-Martinez and Figueras (1998b) have also noted their presence almost continuously throughout the year in similar Iberian ecosystems.The rarities of adults in the lagoon are probably because of the few hard substrata available in the Ria Formosa (Chícharo and Chícharo, 2000).
Variation in the total abundance of bivalve larvae showed a clear seasonal pattern, with highest abundance in the spring and summer months and minima in autumn and winter.This was clearly related to the seasonal fluctuation of water temperature.A similar trend between larval abundance and salinity could also be observed.However, temperature seems to be the more important factor in the timing of the reproduction of bivalves (Guillou et al., 1992) and therefore in the fluctuations of the abundance and occurrence of planktonic bivalve larvae.
SHORT-TERM FLUCTUATIONS IN BIVALVE LARVAE 417 High values of R. decussatus larvae in the plankton, usually between May and September, agree with the spawning periods of R. decussatus (Vilela, 1947;Banha, 1984).According to these authors, spawning seems to occur when the water temperature rises above 20ºC and, in general, this was confirmed in the present study.The correlation coefficient indicates that temperature explained a significant component of variation in the abundance larvae of this species.
In contrast to the situation for R. decussatus larvae, C. edule larvae were present during almost all the year in the plankton of the lagoon.However, intense short-term fluctuations occurred, as has also been observed by Alcaraz and Domínguez (1985).Abundance was higher in 1991 and maximum values occurred in spring.This is in accord with the spawning periods of C. edule (Vilela, 1947;Santos et al., 1986;Guillou et al., 1992).The relationship of this species with environmental parameters revealed that temperature was not related significantly with the abundance of larvae.This can be due to the fact that C. edule adults need water temperatures greater than 15ºC to spawn (Le Pennec, 1978), and water temperatures in the Ria Formosa lagoon were above this limit for almost all the year.
Besides water temperature, chlorophyll a levels also seem to have had some relationship with bivalve larvae.In fact, the quantity and quality of food seem to determine the bivalve larval abundance (Calabrese and Davis, 1970).Nevertheless, according to Sprung (1994b) the net contribution of planktonic forms to primary production in the Ria Formosa is low (only about 45 g C m -2 y -1 ) and is mainly restricted to the summer months.The major primary production probably arises from green algal mats that cover mud flats during winter, in the form of inter-tidal and sub-tidal Zoostera beds and in salt marshes surrounding the tidal flats.However, the suspended detritus resulting from these sources could have dimensions unsuitable for filtration by small organisms such as bivalve larvae.The high contribution of phytoplankton to primary production in summer also coincides with the major abundance of bivalve larvae.Moreover, the phytoplankton biomass of the lagoon in late spring and summer is usually dominated by small flagellates (von Brockel, personal communication), which is an appropriate food source for bivalve larvae.In fact all the species analysed displayed significant correlations with levels of chlorophyll a.
Wind velocity, seemed to have little influence on the total bivalve larval abundance.However, there are some reports giving different results.According to Pearce et al. (1998) andVerdier-Bonnet et al. (1997), high water turbulence negatively affects bivalve larvae.This can be related to the fact that low wind velocity was registered during the more intense period of spawning of this species in the present study (< 7 m/s).
Besides monthly and seasonal fluctuations of bivalve larvae, great oscillations of larvae were registered, with small periods when an absence was registered, alternating with other periods of high abundance.These oscillations seem to be frequent in bivalve larvae.Le Pennec (1978) and Sprung (1994b) have also noted this variation.This could result from the patchy distribution of zooplankton (Hunter, 1981).Nevertheless, the decline of larvae in the plankton can be also related to the settlement of young bivalves (Cáceres-Martínez and Figueras, 1998a, b).However, we believe that the offshore transport by tidal currents is determining this important variation in the larval abundance between successive samples, especially registered during the spawning period.During spring tides, when the renovation of water of the lagoon is almost total, a great reduction of abundance was observed.In fact, tidal amplitude was related negatively with the abundance of larvae.The existence of two bars allowing changes of water masses may result in important losses of larvae to the adjacent ocean.However the existence of some synchronised mechanism of spawning with neap tides may also have developed in some bivalves species, similar to the one described for crustacean and fish (Sinclair, 1988), to avoid the loss of the larvae to unfavourable habitats.
It can be concluded, despite the known limitation of correlation analysis with several environmental factors, that the seasonal fluctuations of bivalve abundance seem to be controlled by temperature, the major factor in the timing of the reproduction of bivalves.However, the short-term fluctuations are related to tidal amplitude (advection).Identifying cause-effect relationships of abundance with some environmental parameters would be a predictive tool for aquaculture and/or natural beds management.Further effort must be applied to obtain longer data series on short-term fluctuations of larval abundance and newly post-larval abundance and of environmental parameters.Moreover, accurate field estimations of mortality rates during planktonic larval life and benthic post-larval life are also necessary for the bivalve species in this coastal lagoon.
FIG. 2. -Larval abundance and environmental parameters, during the sampling period: A. water temperature, B. Salinity C. Wind velocity D. Chlorophyll a