AQUACULTURE Archimer http://www.ifremer.fr/docelec/ JUNE 2000; 186(1-2) : 117-144 Archive Institutionnelle de l’Ifremer http://dx.doi.org/10.1016/S0044-8486(99)00373-7 Copyright © 2000 Elsevier Science B.V. All rights reserved Ecophysiological model of growth and reproduction of the black pearl oyster, Pinctada margaritifera: potential applications for pearl farming in French Polynesia Stéphane Pouvreaua, b, Cédric Bachera and Maurice Hérala a : CREMA (IFREMER-CNRS), BP 5, 17137, L'Houmeau, France b : IFREMER COP, BP 7004, Taravao, Tahiti, French Polynesia *: Corresponding author : Tel. : 33 5 46 50 94 40; Fax: 33 5 46 50 06 00; E-mail : [email protected] Abstract: A model of bioenergetics of the black pearl oyster (Pinctada margaritifera) was built to simulate growth, reproduction and spawning in suspended culture at field sites in Takapoto lagoon (French Polynesia). This model was based on allometric scaling of physiological functions and scope for growth (SFG) calculations. The input functions were clearance rate (CR, l day−1), retention efficiency (RE, %) for each kind of particle encountered in suspended matter, pseudofaeces and faeces productions (PF and F, mg C day−1), excretion and respiration rate (U and R, mg C day−1). The assimilated carbon (i.e., SFG, mg C day−1) was partitioned to the three internal state variables (somatic tissue, shell and gonad) according to the asymptotic increase of the reproductive effort (ER, %) with the age. Given organic and mineral particulate matter in suspension in lagoon water (POM and PIM, mg L−1) and assuming that the taxonomic composition of POM was fairly constant throughout the year, the model predicted annual evolution of total tissue weight (W , g dry weight), Tissue shell weight (W , g DW) and gonad weight (W , g DW) of pearl oysters at various ages. Data on Shell Gonad tissue and shell growth, but also on gonad development of cultivated pearl oysters, acquired in 1997– 1998 in Takapoto lagoon, were used to validate the model outputs. Results of the simulations indicated that the P. margaritifera growth model provided realistic growth trajectories for shell, somatic tissue and gonad, for pearl oysters aged from 1 to 4 years. For scientists, this validated model is very useful to understand more extensively feeding processes and reproduction features of P. margaritifera in Polynesian lagoons. It provides also valuable information for pearl farmers and for management of cultured populations, especially concerning the time to produce a pearl, spat collection, density in system culture, choice of farming site and carrying capacity of the whole lagoon. In this respect, the model demonstrates that local overdensity must be avoided especially when water current is low (<1 cm s−1) and that global consumption of the cultivated pearl oyster stocks plays a insignificant role in comparison with the consumption of all the natural bivalves occurring in the lagoon. Nevertheless, this first growth model developed for P. margaritifera was only validated for a Takapoto atoll and presents several limitations concerning especially the temperature effect and the range of food concentration. Consequently, some improvements will be necessary to apply this work to other pearl farming sites in French Polynesia. 1 Abstract A model of bioenergetics of the black pearl oyster (Pinctada margaritifera) was built to simulate growth, reproduction and spawning in suspended culture at field sites in Takapoto lagoon (French Polynesia). This model was based on allometric scaling of physiological functions and Scope for Growth (SFG) calculations. The input functions were clearance rate (CR, l d-1), retention efficiency (RE, %) for each kind of particle encountered in suspended matter, pseudofaeces and faeces productions (PF and F, mg C d-1), excretion and respiration rate (U and R, mg C d-1). The assimilated carbon (i.e. SFG, mg C d-1) was partitioned to the three internal state variables (somatic tissue, shell and gonad) according to the asymptotic increase of the reproductive effort (ER, %) with the age. Given organic and mineral particulate matter in suspension in lagoon water (POM and PIM, mg l-1) and assuming that the taxonomic composition of POM was fairly constant throughout the year, the model predicted annual evolution of total tissue weight (W , g dry weight), shell weight (W , g DW) and gonad weight Tissue Shell (W , g DW) of pearl oysters at various ages. Data on tissue and shell growth but also on gonad development Gonad of cultivated pearl oysters, acquired in 1997-98 in Takapoto lagoon, were used to ground truth the model outputs. Results of the simulations indicated that the P. margaritifera growth model provided realistic growth trajectories for shell, somatic tissue and gonad, for pearl oysters aged from 1 year to 4 years. For scientists, this validated model was very useful to understand more extensively feeding processes and reproduction features of P. margaritifera in Polynesian lagoons. It provides also valuable information for pearl farmers and for management of cultured populations, especially concerning the time to produce a pearl, spat collection, density in system culture, choice of farming site and carrying capacity of the whole lagoon. In this respect, the model demonstrates that local over-density must be avoided especially when water current is low (<1 cm s-1) and that global consumption of the cultivated pearl oyster stocks plays an insignificant role in comparison with the consumption of all the natural bivalves occurring in the lagoon. Nevertheless, this first growth model developed for P. margaritifera was only validated on Takapoto atoll and presents several limitations concerning especially the temperature effect and the range of food concentration. Consequently, some improvements would be necessary to generalise this work on other pearl farming sites of French Polynesia. Key words : Tropical lagoon - Suspension feeding - Carrying capacity - Scope for growth - Food depletion - Pearl farming. 2 1. INTRODUCTION In French Polynesia, black pearl aquaculture has played an increasing role since 1980. Today, production is approaching six metric tons of pearls (i.e. 160 millions of US dollars). After the large scale mortality which decimated cultivated pearl oysters, Pinctada margaritifera, in several atolls in 1985, French Polynesian authorities decided to set up a general research programme on the pearl oyster (PGRN). The main objective of PGRN was to assess the carrying capacity of a lagoon for pearl oyster farming, by taking Takapoto lagoon (Tuamotu archipelago, French Polynesia) as a study site. Although the black pearl oyster was naturally one of the most abundant bivalves in the benthic fauna of Polynesian atoll lagoons, the development of pearl culture moved these benthic bivalves progressively into the pelagic environment, since farmers suspend pearl oysters on long-lines in mid-water. Thus, understanding the relationships between cultivated pearl oyster and pelagic food web is essential to study the carrying capacity of a given lagoon. To reach this aim, a first step is to build a deterministic ecophysiological model for cultivated pearl oysters in lagoon environment, explaining individual growth over several years, as a function of food availability and composition. Such a model is very useful since it provides validated bivalve requirements which are necessary in carrying capacity calculations. Physiological models explaining the growth of cultivated molluscs in their environment in relation to food supplies have already been achieved on numerous bivalves species (e.g. Bacher et al., 1991; Powell et al., 1992; Schneider, 1992; Raillard et al., 1993; Van Haren and Kooijman, 1993; Barillé et al., 1997; Grant and Bacher, 1998; Scholten and Smaal, 1998). Generally, these models are based on the Scope for Growth concept and use a complete sequence of steps in nutrition and resources allocation. This approach is applied here for the first time on P. margaritifera. In this respect, all the feeding processes previously studied in Polynesia (Robert et al., 1998; Pouvreau, 1999; Pouvreau et al., 1999a,b) were integrated in the present model. Furthermore, since adult pearl oysters are able to spawn several times a year (Thielley, 1993; Buestel et al., 1995; Pouvreau, 1999), reproductive effort was also incorporated in this growth model. First, this study presents a global dynamic carbon model explaining individual growth and reproduction of P. margaritifera in a lagoon environment (Takapoto lagoon) over several years, as a function of food availability and composition. In order to validate the model, simulations of growth and reproduction were systematically tested against measurements. 3 Second, applications of the validated model for pearl farming in Polynesia are also presented. Information of potential interest for pearl farmers is given, especially concerning spat collection and rate of nacreous deposition. Carrying capacity of Polynesian pearl farming sites is also approached : (1) at local scale (i.e. the system culture), we tested potential effect of density (POM-depletion) on the growth of pearl oyster ; (2) at a global scale (i.e. the whole lagoon), we computed, on the basis of model calculations, the clearance time of the cultivated and natural bivalve populations (Dame and Prins, 1998). 2. MATERIALS AND METHODS 2.1. Model design and states variables Several simplifications were made to streamline the model for which a conceptual design is described on Fig. 1. The model uses a balanced carbon budget approach to simulate separately growth in shell and growth in total soft tissue including gonad. This model is based on (1) the widely applied (reviewed in Bayne, 1998; Hawkins et al., 1998a) Scope for Growth (SFG) concept (Bayne, 1976); (2) allometric relations between physiological functions and dry tissue weight (W ). The SFG concept, already used in many models (e.g. Bacher et al., Tissue 1991; Schneider, 1992; Raillard et al. 1993; Barillé et al., 1997), assumes that energy or matter gained by food acquisition is equal to the energy or matter lost for maintenance, growth and reproduction. In this work, SFG is calculated as the difference between carbon acquired by feeding processes and lost by respiration (supposedly a measure of maintenance) and excretion for one individual, as follows : SFG = FR – (PF+F+R+U), where SFG = scope for growth (mg C d-1 ind-1) FR = consumption (retained organic matter, mg C d-1 ind-1) PF, F = pseudofaeces and faeces production (mg C d-1 ind-1) R = respiration (mg O d-1, converted in mg C d-1 ind-1) 2 U = urea and amino acid excretion (converted in mg C d-1 ind-1) If SFG>0, carbon is partitioned into shell, soma and gonad (Resources allocation). If more carbon is required for maintenance (Respiration) than available from food (Assimilation), SFG becomes negative and carbon is mobilised firstly from gonad (utilisation of energy reserves). Therefore, states variables of the model were : (1) shell, (2) somatic tissue and (3) gonadal tissue. Shell represents the organic matrix. Somatic tissue includes gill, mantle, muscles and digestive gland (+viscera). Gonadal tissue includes developing and mature gametes. 2.2. Environmental factors Temperature in this tropical environment, and especially in Takapoto lagoon, does not present great seasonal variations (from 26 to 30°C during "normal year", i.e. without El Niño event), so that temperature effect on pearl 4 oyster physiology has not yet been demonstrated (Robert et al., 1998; Pouvreau et al., 1999b) and, consequently, was not included in the present model (this first limitation is analysed in discussion part). Alternatively, forcing functions in the model were pelagic food concentration and composition in Takapoto lagoon water. Potential food for pearl oysters was expressed as total particulate matter (TPM, mg l-1), and consisted of particulate inorganic matter (PIM, mg l-1) and particulate organic matter (POM, mg l-1). PIM and POM were monitored once a week from March 1997 to March 1998, always at the same hour (14:00, Fig. 2). This sampling hour generally corresponds to the daily maximum of seston concentration whereas minimum values were generally reached around 6:00 (Charpy, 1996; Buestel and Pouvreau, 2000). In order to take into account these night and day variations and avoid over-estimation of the daily mean, PIM and POM measured at 14:00 were corrected by 0.7. This value was obtained previously on the basis of several night and day measurements. Afterwards, POM was converted into carbon (POC, mgC l-1) by using a conversion factor (POM/POC=2, Mann, 1979 for bivalve flesh; Parsons et al., 1961 for algae composition; Quemeneur and Marty, 1992 and Galois et al., 1996 for total suspended matter). In P. margaritifera, retention efficiency is nearly null for 1 µm particles and becomes maximal for > 5 µm particles (Pouvreau et al., 1999a; Yukihira et al., 1999). Therefore, a part of POM cannot be directly retained on P. margaritifera gill. In this respect, POM in Polynesian lagoon need to be reviewed (Table 1) so that our model uses POM according to its size as input for consumption. Suspended matter is generally composed of a living (bacteria, phytoplankton, protozoan and zooplankton) and detritus material. In lagoonal waters, free bacteria were the smallest living particles (0.4 µm). A part of them are fixed on larger particles but scarce data (Torréton and Dufour, 1996) are not sufficient to establish an order of magnitude. Phytoplankton is divided into two groups: (1) picophytoplankton (< 2 µm) which includes prokaryotic cyanobacteria (Synechococcus sp. and Prochlorococcus sp.) and eukaryotic algae (picoeukaryotes, Charpy and Blanchot, 1998), and (2) phytoplankton, which includes nano and micro-algae. Detritus is also divided into two groups : (1) pico-detritus (<2 µm), and (2) larger detritus called micro-detritus. Protozoa are mainly heterotrophic nano- and micro-plankton and include ciliates, heterotrophic flagellates and dinoflagellates, foraminifers and radiolarians. Zooplankton consists mainly of micro-plankton and includes especially appendiculates and larvae of various species (naupliar copepods and bivalves, crustaceans and polychaetes larvae). Particulate matter >200 µm (mesoplankton) are not taken into consideration in this work because of their probable low representation in the diet of the bivalves. The contribution of each of these previous components to the total particulate organic matter was assumed to be 5 fairly constant (trophic web in equilibrium, suggested by the low standard deviation in Table 1) and was estimated on the basis of the review described in Table 1. 2.3. Physiological mechanisms formulations Each term presented in the SFG equation (see previous paragraph) was computed as a function of the forcing variables (exogenous factors) or/and the dry tissue weight, W (endogenous factor). The physiological Tissue functions were studied extensively in previous works (Robert et al., 1998; Pouvreau, 1999; Pouvreau et al., 1999a,b) and the useful equations are summarised in Table 2. In the following paragraphs, we remind only what it is necessary to understand the model conceptualisation. 2.3.1. Food acquisition Pearl oysters obtain their food by processing large amount of waters thanks to a high clearance capacity (Pouvreau et al., 1998 and 1999a for Polynesian water, Yukihira et al., 1998a for Australian water; and Hawkins et al., 1998b for Malaysian water). Using biodeposition method (Iglesias et al., 1998), the clearance rate CR (l d- 1) was estimated in situ on 196 oysters of various sizes under several natural seston load conditions, and varied according to the PIM and POM concentration as an allometric function of dry tissue weight W (g), as Tissue follows (Pouvreau et al., 1999b): CR = 647.04 (± 60.48 SE) •PIM-0.42 (± 0.05 SE) •POM 0.96 (± 0.11 SE) •WTissue 0.61 (± 0.04 SE) ; (n=196, R²=0.75, P<0.05) The well-developed gill of pearl oysters can efficiently retain particles > 2 µm (Pouvreau et al., 1999a ; Yukihira et al., 1999). So each of the previous compartments in seston is retained differently according to its approximate mean size Sz (µm, equivalent spherical diameter) and according to the retention spectrum RE (%) given by Pouvreau et al. (1999a), which exhibits the following logistic pattern : RE = 100/(1+52.16 • 0.137Sz) ; (n=16, R² = 0.86, P<0.05) Mean size of bacteria, cyanobacteria, pico-detritus, picoeukaryotes was assumed to be approximately equal in mean to 0.4, 0.8, 1 and 1.2 µm, respectively, giving a retention efficiency equal to 2, 6, 12 and 15 %, respectively, for these pico-particles (<2 µm). Micro-particles (> 2 µm) were supposed to be retained in mean at 60 %. Then, the amount of retained organic matter, FR (mg C d-1 ind-1) was computed as follows : POM FR =CR •[ 0.02 POC +0.06 POC +0.12 POC +0.15 POC +0.6 POC ] POM bacteria cyanobacteria picodetritus picoeukaryotes microparticules 6 2.3.2. Biodeposition Prior to ingestion, particles with low organic content are generally rejected in pseudofaeces. In P. margaritifera, the quantity of organic matter lost in pseudofaeces (PF , mg C d-1 ind-1) was measured in situ (Takapoto POM lagoon) and varied according to POM, PIM and W , as follows (Pouvreau et al., 1999b): Tissue PFPOM = 32.16 (± 2.16 SE) •POM• [PIM - 0.18 (± 0.02 SE)] •WTissue ; (n=196, R²=0.69, P<0.05) For PIM<0.18 mg l-1, there is no pseudofaeces production. This minimum PIM-concentration is called the pseudofaeces production threshold. The remaining organic material (FR -PF ) is ingested. A great part of the ingested organic matter is POM POM absorbed during intestinal transit. The non-absorbed organic material is ejected in faeces. Mineral matter is assumed to be not at all absorbed. In P. margaritifera, the organic faeces production (F , mg C d-1 ind-1) was POM measured in situ for various seston load and composition and followed the equation (Pouvreau et al., 1999b): F = 480.00 •W 0.49 (± 0.72 SE) • (1 - e-0.66 (± 0.06 SE) . TPM) ; (n=196, R²=0.66, P<0.05) POM 2.3.3. Oxygen consumption and excretion Routine oxygen consumption (Bayne, 1976) was monitored in situ (Takapoto lagoon) on various sizes of animal and at different seasons (Robert et al., 1998). This routine respiration rate represented in fact the metabolic requirements, i.e. routine metabolism and overheads for growth (soma and shell matrix) and also for reproduction, since a pearl oyster is more or less always in gametogenesis. These authors showed that respiration R (converted in mg C d-1) varied mainly according to W , as follows (Robert et al., 1998) : Tissue R = 7.38 (± 1.20 SE) •RQ•WTissue0.73 (± 0.004 SE) ; (n=113, R²=0.76, P<0.05) Knowing the proximal composition of the diet of pearl oysters in Takapoto lagoon (proteins=0.5, carbohydrates=0.25 and lipids=0.25, Buestel and Pouvreau, 2000), it was possible to estimate the respiratory coefficient, RQ, i.e. moles of CO liberated per mole of O consumed (%), according to Gnaiger (1983). RQ was 2 2 equal to 0.92, falling in the range given by Hawkins and Bayne (1985). Anaerobic metabolism is supposed to be insignificant since oysters are never emersed in the lagoon environment and oxygen is always in over-saturation in Takapoto lagoon (Haumani Gaby, Pers. comm.). Excretion rate, U (converted in mg C d-1), was measured under laboratory conditions (Méro, 1996; Pouvreau et al., 1998) which simulated those of the lagoon environment. Carbon losses appeared when urea and amino acids 7 were excreted. Urea and amino acids represented less than 10% of the total ammonia excreted, and increased with dry tissue weight, as follows : U amino acids, urea= 0.02 (± 0.001 SE) •WTissue0.78 (± 0.06 SE) ; (n=36, R²=0.84, P<0.05) 2.3.4. Resources allocation to growth and reproduction Rules specifying the animal priorities for resources partitioning are not well-known. Different prior assumptions have been made concerning energy allocation between shell growth, tissue growth and reproductive outputs (Ross and Nisbet, 1990; Van Haren and Kooijman, 1993). For this model, we assume the two following rules : (1) If absorbed carbon exceeds uptake for maintenance (Respiration) and losses in excreted urea and amino acids, the scope for growth (SFG, mgC h-1) is positive and this surplus is partitioned into shell matrix, somatic and gonadal compartments. Dry shell, tissue and gonad weight (W , W , and W ) are, then, calculated shell tissue gonad from their respective carbon input and ash ratio (Ash , Ash and Ash ). Allocation to gonad, i.e. Shell Tissue Gonad reproductive effort (ER), increases with age (e.g. Bayne and Worrall, 1980; Thompson, 1984; Rodhouse et al., 1984). Considering that ER was approximately equal to 0, 9, 22 and 35 % for pearl oysters of year class 0, 1, 2 and 3, respectively, (experimental data, Pouvreau, 1999), a classical logistic pattern according to the age was proposed here (see Table 2) to model ER. For the present, we make no difference between ER for males and ER for females, since during the 4 first years of life, cultivated pearl oyster are mainly males. The sex inversion causes probably physiological transformation, but we have not enough data to model this aspect. Part of carbon, which is not driven to gonad, i.e. (1-ER), is allocated to soma and shell matrix. Shell growth is proportional to somatic growth with a constant allocation to shell (Alloc =0.6, Pouvreau, 1999) and to soma (1-Alloc ). Shell shell (2) When respiratory requirements and excretory losses exceed available absorbed carbon, SFG becomes negative and carbon is mobilised from gonad tissue, playing a kind of storage role. In fact, storage is virtually non-existent in this tropical bivalve (Buestel et al., 1995). So that, as soon as the food supply is sufficient, gonad development commences. Such an opportunistic strategy reproduction differs from the conservative strategy in many temperate bivalves, for which reserves are accumulated before gametogenesis. Validity of the model (accordance between simulation and observation) will imply that assumptions (1) and (2) are reasonably reliable. 2.3.5. Gametes release Those (Thielley, 1993; Buestel et al., 1995; Pouvreau, 1999) who have worked on reproduction of this species in Polynesian lagoons have come to the same conclusion : pearl oysters are able to spawn several times and 8 throughout the year. When food is sufficient for several weeks (i.e. POM x Days), the gonad expands due to gametogenesis. Various stressful exogenous factors (e.g. change in temperature, in salinity, in atmospheric pressure) are supposed to constitute spawning cues. Nevertheless, endogenous factors can also be involved : in this first model, gamete release, which is supposed to be complete and instantaneous, occurs when gonado- somatic ratio exceed a maximal value (RGS > 0.29, Pouvreau, 1999). max 2.3.6. Shell and pearl formation Dry shell weight (W ) is used to compute the shell height (H , Dorso-ventral axis, mm) and the shell Shell Shell thickness (T , Hinge line, mm) according to the specific allometric relationships (Table 2 and Pouvreau, Shell 1999). Assuming that nacreous deposition on shell by mantle activity exhibits the same rate as nacreous deposition on the nucleus implanted in gonad diverticula, time of pearl formation, t ,can be calculated from the end integral function : dT t Shell T = , where Pearl ∫tend dt 0 T = Thickness of the nacreous layer for a marketable pearl (T = 1.2 mm) pearl pearl t = Time to obtain a marketable pearl. end 2.3.7. Depletion calculation The filtration activity of dense aggregations of bivalves could locally deplete seston in water (Fréchette et al., 1989) especially if the currents that supply seston are too low to replenish completely the ingested particles. Since (1) P. margaritifera exhibits high clearance activity, (2) lagoon currents are generally low and (3) food supply is limited, depletion conditions can occur in culture system according to density and flow velocity (Fig. 3). The seston concentration POM(x) within a culture of suspended bivalves is determined as a balance between the transport of particles from remote sources (flow velocity, U), the local production of new particles (primary production, PP), and the removal of particles due to feeding of bivalves populations (density.consumption, d.FR ) or sinking (sinking rate, SR) : POM ∂POM ∂POM = - U . + PP - d.CR.POM - SR ∂t ∂x 9 Two assumptions were made to simplify calculations : (1) the local production and the sinking rate are negligible ∂POM ∂POM in comparison with transport (PP<<U• and SR<<U• ); (2) the system is in steady state ∂x ∂x ∂POM ( = 0). With such assumptions, depletion ratio after i long line systems, D, , i.e. (POM after i long i ∂t lines)/(POM entered the system), is equal to : d . CR . i - D = e U , where i D = depletion ratio after i long lines i d = linear density on long lines (ind. m-1) CR = clearance rate of pearl oysters (converted in m3 s-1) i = number of juxtaposed long lines U = flow velocity (m s-1) 2.4. Simulations and validation The model was implemented in STELLA software (High Performance Systems, Hannover, NH, USA). List of equations, parameters and variables are presented in Table 2 and Table 3. Model results (predicted values) were compared to data on shell, soma and gonad growth of various age-groups of pearl oysters measured in 1997- 1998 (observed values, Pouvreau et al., 2000). This allowed assessment of both the shape of the trajectory of total tissue and shell weight as well as the numerous peak weights observed for gonads. 10