Bivalve grazing, nutrient cycling and phytoplankton dynamics in an estuarine ecosystem.

This thesis has considered the impact of the suspension feeding bivalve Mytilusedulis on nutrient cycling and phytoplankton in an estuarine ecosystem. The research was started within the framework of an extensive research project with the objective to evaluate the changes in the Oosterschelde ecosystem as an effect of a coastal engineering project (Nienhuis & Smaal, 1994). The Oosterschelde estuary is a system where mussels are dominant consumers, which is at least partly due to the strong regulation of mussel biomass by fisheries. As a consequence, the Oosterschelde estuary is a typical example of an ecosystem where herbivores have a strong impact on the entire pelagic system.The main emphasis in this study has been on the impact of the mussels on the exchange of material between the water column and the benthic system. This exchange has been studied with an in situ method designed to measure exchange of particulate and dissolved material between an undisturbed bivalve community and the water column, under natural conditions with respect to food supply, current speed, temperature etc. The method was evaluated in chapter 2. It was concluded that the method was suitable for in situ measurements of the fluxes of particulate and dissolved matter between the mussel bed and the water column.Mussel beds filter considerable amounts of material. The amount of material filtered is determined by mussel biomass, mussel activity and the supply of particulate material to the mussel bed. The in situ experiments, carried out in the years 1987-1989, showed that the quantity and quality of the suspended particulate matter varied considerably at a short time scale. A tidal variation was observed, with the supply of relatively phytoplankton-rich water to the intertidal flats during flood tide. During ebb, phytoplankton concentrations in the water were generally reduced, probably as a consequence of depletion by suspension feeders. Superimposed on this tidal variation changes in seston quantity and quality at the time scale of hours were observed. This short-term variation was related to wind-induced resuspension of bottom material. The observed increases in SPM, POC and phaeophytin- a levels during these resuspension events indicated that resuspension of algal detrital material, probably biodeposits, occurred.It was shown in Chapter 3 that under calm weather conditions a good correlation between the concentrations of SPM, POC and chlorophyll- a in the water column and the size of the respective fluxes to the mussel bed was observed. From a comparison of the composition of the fluxes with the composition of the seston it became clear that the fluxes contained a relatively high proportion of phytoplankton. After filtration, mussels are able to selectively ingest phytoplankton, while other particles are rejected through the pseudofaeces. As a result, the pseudofaeces have a reduced chlorophyll- a content (Kiørboe & Møhlenberg, 1981; Prins et al., 1991). Mussels eject the pseudofaeces into the water column. Pseudofaeces have a low settling velocity and are easily resuspended (Risk & Moffat, 1977; Nowell et al., 1981). This makes it probable that in the sequence of filtration, selection, ingestion and pseudofaeces formation, a fraction of the non-algal material that was filtered by the mussels and then rejected as pseudofaeces, was exported from the mussel bed. Consequently, the net flux of material to the mussel bed had a high proportion of chlorophyll- a , whereas part of the POC and of the particulate inorganic matter, filtered by the mussels, was exported immediately. Under rougher weather conditions sometimes a significant net export of SPM and POC from the mussel bed was observed as a result of wind-induced resuspension. The results indicated that only a part of the material that was filtered by the mussels, was stored in the mussel bed, due to immediate export of pseudofaeces by tidal currents or to more occasional wind-wave resuspension.Previous estimates indicated that the mussel population in the Oosterschelde may filter the entire volume of the Oosterschelde in approximately 10 days (Smaal et al., 1986; Van Stralen & Dijkema, 1994). The generally low phytoplankton concentrations in the Oosterschelde have been attributed to the severe 'top-down' control exerted by the mussel, together with the other dominant bivalve suspension feeder Cerastoderma edule (Smaal et al., 1986; Herman & Scholten, 1990). The observed in situ filtration activity of the mussel beds, presented in chapter 3, agrees with the values used in earlier published estimates of mussel grazing pressure.Recent ecophysiological studies have shown that bivalves respond to short-term changes in the quantity and quality of the seston by regulating ingestion rates. This regulation may be achieved by changes in clearance rates, or by rejection of variable amounts of filtered material as pseudofaeces in combination with pre-ingestive selection. The nature of this physiological response depends on both quality and quantity of the seston, and on acclimation (Bayne, 1993; Navarro & Iglesias, 1993). When exposed to diets with a low organic content, mussels mainly respond to these short-term changes by modifying the rate of pseudofaeces production (Bayne, 1993; Navarro & Iglesias, 1993). Our in situ experiments carried out in the Oosterschelde, showed no effect of the changes in seston concentration and composition during a tidal cycle on clearance rates (Chapter 3); this is consistent with the hypothesis that ingestion is regulated by pseudofaeces production when seston quality is low. The seasonal variation in clearance rates of a mussel bed was studied by monthly measurements on a semi- natural mussel bed (Chapter 6). Clearance rates showed a positive response to chlorophyll- a concentrations and were negatively affected by SPM concentrations. The relations observed in the latter experiment were probably the result of adaptation on a longer time scale (Bayne, 1993). In the mesocosm experiments presented in Chapter 7, mussel clearance rates were generally lower than rates observed in experiments with natural seawater, and showed considerable temporal variation. In the mesocosms the mussels were exposed to seston with a much higher proportion of algae than is commonly observed in situ . The reduced and variable clearance rates in the mesocosm experiment agree with the current idea that it is more optimal for bivalves to regulate ingestion by changes in clearance rates, when the animals are exposed to diets with a high organic content (Iglesias et al., 1992; Bayne, 1993; Navarro & Iglesias, 1993).In addition to the effects of seston quantity and quality on the seasonal variation of clearance rates, a strong inhibitory effect of the presence of the alga Phaeocystis sp . on mussel clearance rates was observed. This inhibition of mussel clearance rates may lead to a breakdown of the top-down control of phytoplankton biomass and may increase the risk of massive algal blooms.In general, a release by the mussel bed of ammonium, phosphate and silicate was observed during the in situ measurements. In June 1987 an experiment was carried A during a long period of high phytoplankton concentrations, resulting in a large flux of organic matter from the water column to the mussel bed. A high release of inorganic nutrients was observed during that experiment. In June 1988 chlorophyll- a concentrations in this part of the Oosterschelde estuary were low for a number of consecutive weeks. In that period a number of in situ measurements were carried out, all showing low uptake rates of chlorophyll- a , and a small release of inorganic nutrients. It was inferred from these results, that the difference in organic matter supply to the mussel bed was the cause for the difference in nutrient release between the observations from June 1987 and the results from June 1988. Moreover, it was shown that considerable day/night differences in fluxes occurred in June 1987. During daytime, much lower release of inorganic nutrients by the mussel bed was measured, which was attributed to immediate uptake of the nutrients by algae.Budgets presented in Chapter 5, suggested that only a small proportion of the nutrient fluxes to the mussel bed were stored in mussel biomass. The remainder was stored in the sediment as biodeposits. Mineralization of the biodeposition and excretion by the mussels resulted in a release of inorganic nutrients. In the case of nitrogen, there was an approximate equilibrium between the net uptake of particulate N by the mussel bed and the release of dissolved inorganic N by the mussel bed, indicating that storage of N by the mussel bed was of minor importance. The contribution of direct excretion by the mussels to the DIN flux from the mussel bed was relatively small. Silicon fluxes showed some retention in spring, and predominantly a release in autumn. The uptake of particulate P by the mussel bed was somewhat higher than the release of phosphate. On average approximately 35% of the P uptake was retained by the mussel bed. In addition to storage in mussel biomass, which was estimated to be small in relation to the total flux of particulate P, sorption of phosphate on sediment particles might be responsible for some of the P retention.The relatively low retention of nutrients by the mussel bed may seem surprising, and contradictory to the observed growth of mussels and accumulation of biodeposits. However, the observed mussel production at the mussel lots in the Oosterschelde is close to or lower than the biomass seeded at the lots on an annual basis (Van Stralen & Dijkema, 1994), as the growth of individual mussels is balanced by mortality. This means that net storage of nutrients in mussel biomass is of minor importance. The accumulation of nutrients in biodeposits was more or less balanced by mineralization and resuspension. It should be realized, however, that as a consequence of the large fluxes of material towards and from the mussel bed it was not possible to detect small differences between uptake and release. Retention by the mussel bed of a relatively small fraction of the filtered material could still result in the accumulation of a considerable amount of organic matter in the sediment of the mussel bed.The net result of uptake and release processes is a rapid cycling of organic matter, with a conversion of particulate organic matter into dissolved inorganic nutrients. It was estimated in Chapter 4 that the amount of nitrogen, regenerated by the mussel beds in the central part of the Oosterschelde, was of the same order of magnitude as regeneration due to mineralization in all other sediments, that cover a much larger area. A comparison of the rates of nitrogen regeneration by mussel beds to estimates of total mineralization (benthic + pelagic) in the central part of the Oosterschelde was presented in Chapter 5, and indicated that the mussels contributed significantly to the nitrogen mineralization. The rapid recycling of nutrients by the mussels may stimulate phytoplankton growth rates in summer, when primary production is limited by low levels of N and Si (Wetsteyn & Kromkamp, 1994).A mesocosm experiment was carried out to explore the relations between mussel grazing, nutrient cycling and phytoplankton development under more controlled conditions. Four mesocosms were used with different densities of mussels to establish a gradient in grazing pressure. The development of phytoplankton biomass was inversely related to mussel biomass, showing the strong effect of mussel grazing on phytoplankton standing stock. In all mesocosms phytoplankton growth was P-limited.In spite of a significant contribution to P-regeneration by the bivalves in the mesocosms with high mussel density, estimates indicated that nutrient regeneration by the mussels was not the most important source of regenerated P. A tentative P balance suggested that external loading and pelagic mineralization contributed significantly to the regeneration of nutrients, and overall nutrient regeneration was higher in the low mussel biomass mesocosms. Still, the availability of phosphate was highest in the mesocosms with the highest mussel density. The experiment demonstrated the major impact that grazing may have on the various nutrient pools. In the mesocosms with low mussel density a major fraction of P was stored in phytoplankton biomass. In the mesocosms; with high mussel density, grazing resulted in a reduction of phytoplankton biomass, a consequently lower storage of P in phytoplankton biomass and an increase of the dissolved inorganic nutrient pool. As a consequence of the increased availability of phosphate in the latter mesocosms the phytoplankton community showed increased growth rates. As was shown by Sterner (1989), even without regeneration of nutrients by the grazers, grazing will increase nutrient availability, simply by preventing the monopolization of this resource by the algal community. The increased phytoplankton growth rates in the high mussel biomass mesocosms coincided with a shift in the composition of the phytoplankton community towards a dominance of diatoms. The results showed that grazing resulted in a transfer of nutrients from the phytoplankton pool to the pools of dissolved inorganic nutrients and grazer biomass, a change in phytoplankton composition, and a change in phytoplankton growth rates.Many temperate coastal ecosystems have large populations of bivalve suspension feeders. Densities of bivalve suspension feeders, typical for bivalve dominated systems, are in the range of 2-8 g ADW m <sup><font size="-2">-3</font></SUP>. This includes systems like San Francisco Bay, Bay of Marennes-Oléron, Western Wadden Sea and Oosterschelde estuary (Smaal & Prins, 1993). The initial mussel biomass in the mesocosm experiment ranged from 0.5 to 3.8 g ADW m <sup><font size="-2">-3</font></SUP>, and the treatments with the highest mussel biomass were comparable to the above-mentioned systems with respect to bivalve density.Results presented in this thesis demonstrated that bivalves affect the pelagic system in various ways. Our mesocosm experiment showed that grazing may induce shifts in phytoplankton species composition towards faster growing species. Moreover, grazing has an effect on the relation between nutrient supply and phytoplankton production. Under nutrient- limiting conditions, bivalve grazing has been shown to have a positive effect on phytoplankton growth rates. This is caused by an increase in nutrient availability. As our observations in the Oosterschelde showed, the mussel population in that estuary filters particulate nutrients, and recycles dissolved inorganic nutrients. The contribution of nitrogen mineralization on mussel beds to total mineralization may be significant, even on the scale of an estuary like the Oosterschelde. The mesocosm experiment demonstrated that regeneration of nutrients by the grazers is not the only factor leading to an increase of the dissolved inorganic nutrient pool. Cropping of the algal community by grazing reduces the accumulation of nutrients in the phytoplankton, and this alone may be sufficient to enlarge the pool of dissolved nutrients (cf. Sterner, 1989).Our in situ observations of grazing rates of the mussels confirmed the hypothesis on top- down control of phytoplankton by bivalve grazing in the Oosterschelde (e.g. Smaal et al., 1986; Herman & Scholten, 1990). Our mesocosm experiment showed a strong regulation of phytoplankton biomass by bivalve grazing. From this it can be inferred that in bivalve dominated systems phytoplankton biomass is determined by grazing, even when nutrients are not limiting. This leads to the conclusion that the response of phytoplankton to changes in external nutrient load will be limited. Similar phenomena have been observed in freshwater systems with dominating large herbivores (e.g. Mazumder, 1994; Mazumder & Lean, 1994). As was argued by Herman & Scholten (1990), top-down control of phytoplankton biomass by bivalve grazing makes a system more resilient to increases in the external nutrient loading, and in this sense the bivalve population acts as a eutrophication control. Eutrophication control by using bivalve suspension feeders has been suggested as a means to combat algal blooms, both in marine and freshwater systems (Takeda & Kurihara, 1994; Ogilvie & Mitchell, 1995). However, it should be realized that grazing will also enlarge the pool of inorganic nutrients. As was pointed out by Herman & Scholten (1990), this large pool of unused nutrients may be profitable to any primary producer that is less susceptible to grazing by the bivalve, for example macro-algae like Ulva sp . or the colony-forming Phaeocystis sp.. Eutrophication control by bivalves involves the risk of a sudden shift in an ecosystem towards another, equally undesirable state, and should therefore be accompanied by nutrient input reduction