Optimization of off-bottom spat collectors for restoration and production of the European flat oyster ( Ostrea edulis ) in Dutch coastal waters

1. Efficient spat collection is essential to both the successful restoration and cultivation of the European flat oyster ( Ostrea edulis ). 2. The results of four different experiments investigating larval abundance, and the use of seven different types of suspended spat collectors in seven locations in the SW Netherlands were compared. These different types included Chinese hats, Vexar mesh, PVC tubes, and four types of bivalve shells. 3. Larval abundance, and timing of peaks in abundance for both O. edulis and the non-indigenous Pacific oyster ( Crassostrea gigas ) varied between locations and years. 4. The most successful spat collectors were sacks of bivalve shells. 5. In general, most spat were observed on collectors deployed 1 – 2 weeks after the peak in larval abundance was detected. 6. Fouling was heavy and may have prevented larval settlement on collectors deployed too early. Suspended sacks of bivalve shells show great promise as O. edulis spat collectors intended for reef restoration purposes but may need further development for cultivation application.


| INTRODUCTION
The European flat oyster (Ostrea edulis), was once widespread in the southern North Sea (Olsen, 1883). The flat oyster not only acted as a keystone species for a rich and productive ecosystem, but also served as an economically profitable fisheries product (Gercken & Schmidt, 2014). Over recent decades, O. edulis has declined considerably due to increased mortality as a result of overfishing around the end of the 19 th and beginning of the 20 th centuries (Gercken & Schmidt, 2014) and more recently due to the parasite Bonamia ostreae (Engelsma et al., 2010). However, the flat oyster population in the Dutch Delta area is showing signs of recovery, particularly in Lake Grevelingen (Smaal, Kamermans, Van der Have, Engelsma, & Sas, 2015). This may therefore be the ideal time to attempt to restore flat oyster reefs. Recently, several O. edulis restoration projects have been set up in the North Sea, off the coast of the Netherlands (e.g. Pogoda et al., 2019). Efforts to restore O. edulis in the North Sea and Dutch delta are motivated by: 1) restoring the local ecosystem and increasing the biodiversity in and around the reefs; 2) utilizing the possible ability of the reefs to reduce erosion around the base of planned wind turbines; and 3) the economic possibility of restoring the flat oyster market to its former glory.
Restoring O. edulis in the North Sea is not a simple task. The oyster experiences environmental pressures and consequent population bottlenecks at every life stage. The spawning success of adults can vary between locations and seasons (Korringa, 1940). In addition, flat oysters are a brooding species, i.e. the eggs are fertilized within the mantle cavity where they develop into larvae over 6-10 days (Korringa, 1940). Subsequently, O. edulis larvae are released into the water column and after a pelagic stage of about 10 days the larvae search for suitable hard substrate on which to settle and develop into oyster spat, this last stage is called spatfall. There is generally a considerable loss of individuals between the larval and spat phases, potentially due to unsuccessful metamorphosis, predation of larvae in the water column, and the lack of appropriate substrate on which to settle (Filgueira, Brown, Comeau, & Grant, 2014;Korringa, 1940). Settled spat is also under pressure from predation, and from competition for space with conspecifics or other settling organisms (Engelsma et al., 2010). Since its introduction in the Netherlands, the Pacific oyster (Crassostrea gigas) has become widely established in the same waterbodies that O. edulis occupies (Troost, 2010). Larvae of both oyster species are often found at similar times in the water column, and as spat, the two oyster species compete for space on appropriate substrate (Smaal & Lucas, 2003). However, there is evidence suggesting that adult C. gigas may facilitate settlement of larvae and survival of juvenile O. edulis by providing substrate and refuge from predation (Christianen et al., 2018).
The pressure on the survival of the oysters is further confounded by the presence of the parasite B. ostreae, which affects the health of the oyster and increases mortality at a later age (Engelsma et al., 2010). The cumulative effect of environmental pressures at each life stage results in a paucity of healthy, surviving adults. Providing more appropriate substrate for larvae to settle onto will relieve some of this pressure at a particularly vulnerable life stage and may enhance the amount of spat available to grow into adults. More surviving adults may increase the success of restoring oyster reefs and compensate for the lower survival rate of farmed oysters caused by B. ostreae.
Spat collectors are structures designed to provide artificial surfaces for spat to settle onto and grow to a size where they can be collected. The amount of O. edulis spat collected is influenced by the availability, type, and placement of the substrate provided. In general, oysters from the family Ostreidae typically prefer calcium carbonate rich substrate such as the shells of other bivalves (Smyth, Mahon, Roberts, & Kregting, 2018). Dutch oyster farmers generally use loose mussel shells scattered on the sea floor as spat collectors.
These are later collected and transported to culture plots. This method of sowing loose shells minimizes fouling compared with off-bottom collectors (P. Kamermans, personal observation).
However, the efficiency of this method is dependent on collecting as many of the scattered shells as possible after the spatfall. In addition, spat growing on sowed shells probably risk being smothered by bottom sediment, and are possibly more vulnerable to predation by crabs, fish, sea stars and oyster drills compared with those growing on off-bottom structures. Off-bottom spat collectors may therefore provide an alternative method of spat collection with lower mortality rates.
Various off-bottom spat collector types are currently in use in Europe to collect the larvae of O. edulis including, trestles (tables with bags of shells), and stacked PVC cones ('Chinese hats') and tubes sometimes with a calcium based coating (Freeman & Denny, 2003;Van den Brink, 2012). The objective of these spat collectors is to concentrate the number of oyster spat onto a relatively small surface area so that the spat can be efficiently gathered, removed from the collector, transplanted, and reared. The efficiency of a collector is influenced by several factors including geographic location, depth of the collectors in the water column, temperature, and competition for space with other fouling species such as the introduced Pacific oyster C. gigas (Bataller, Burke, Ouellette, & Maillet, 2006). For the commercial cultivation of oysters as well as for restoration efforts, spat collectors must be cost efficient. For commercial cultivation the product should also have a nice shape (Nalesso et al., 2008), but for the ecological restoration of flat oyster reefs the aesthetics of the oysters are not a priority. An efficient spat collector may therefore collect a large amount of spat over the spawning season without excess investment.
Various research projects (from 2003, 2011, 2017, and 2018) investigated the efficiency of different types of off-bottom spat collectors for O. edulis. Each year larval abundances in the water column were measured and attempts were made to deploy spat collectors throughout the period of larvae settlement to study optimal timing of collector deployment. The insights presented here may prove useful for the purposes of either improving the success of oyster reef restoration or increasing commercial production of O. edulis.
The selection of collectors tested in the different studies was based on existing methods of spat collection used in commercial production, suggestions from the literature, or recommendations by oyster farmers (Smyth et al., 2018;Van den Brink, 2012).

| Chinese hats
Chinese hats are an industry standard in oyster spat collection, particularly in France (Freeman & Denny, 2003). Chinese hats are shallow plastic cones with a diameter of about 15.5 cm with a surface area calculated as 377 cm 2 (Kamermans, Brummelhuis, Poelman, Van Gool, & Troost, 2004). The hats can be stacked together into towers and deployed vertically into the water column. Despite reports of extremely high fouling rates on Chinese hats (Van den Brink, 2012), the structures were developed specifically for the purpose of efficient spat collection and have proven successful in doing so (Anonyme, 2014). Furthermore, because these structures are widely used commercially, testing Chinese hats may also act as a type of control with which the other tested collectors can be compared; if another collector is shown to be more efficient than Chinese hats, it may be worth replacing them on a commercial scale. Communication with oyster farmers indicated that a calcium coating is necessary for the efficient removal of spat from the collectors. In addition, Smyth et al. (2018) indicate that O. edulis has a preference for calcium rich substrate.

| PVC tubes
PVC tubes are used on a large scale in oyster production in France and Canada (Poiriera et al., 2019). They are 1.2 m long with a 2.2 cm diameter. Like Chinese hats, the collected spat can be removed mechanically from the tubes. This is useful for oyster farmers as their final product is individual oysters. For restoration purposes removing the oysters means an extra step, because you do not want to leave the collectors at the location indefinitely.

| Vexar mesh
Vexar mesh is made of sheets of hard, black plastic strips in a crosshatched formation. The mesh was reported to be not only successful in collecting a mid-range to high number of spat (of Crassostrea virginica), but it was cheap, reusable and user-friendly; easy to prepare, deploy, remove the spat and to store (Freeman & Denny, 2003).
Furthermore, spat on the mesh also yielded the highest spat growth of all the collectors tested by Freeman and Denny (2003). Vexar mesh bags were also successful in collecting O. edulis spat on the Canadian Atlantic coast (Bataller et al., 2006).

| Bivalve shells
Bivalve shells are widely used in shellfish aquaculture and generally readily available and successful spat collectors (FAO, 2019;Lok & Acarli, 2006;Smyth et al., 2018). Using net stocking or bags of shells allow oyster spat to recruit onto a large surface area with minimal wasted volume.

| Larvae sampling
For each year a sample of water from the same area as the spat collectors was analysed once per week between weeks 24 and 35 for the presence of oyster larvae. A 100-μm plankton net was used to filter 100 L of water. The size of the net was based on the minimum size of around 160 μm that the flat oyster larvae have during the pelagic phase (Walne, 1974). The samples were fixed in 4% buffered formalin (diluted with filtered sea water) for later microscopic examination in the laboratory. The samples of larvae were filtered using a plankton-gauze (30 μm). The volume of the samples was reduced to 20-60 ml, depending on the amount of suspended matter. From the concentrated samples subsamples were taken for counting numbers of larvae. A Hensen plunger-sampling pipette was used to take subsamples. Bivalve larvae of O. edulis and C. gigas were identified and counted using a universal camera microscope (Reichert Me-F2, 52.6×). Three subsamples of each sample were analysed. Depending on the density of the samples, subsamples of 1-2.5 ml were counted. Larvae were identified according to Hendriks, van Duren, and Herman (2005) combined with data obtained from cultured larvae.

| Experiments with collectors
While there was some variation in the testing of the different spat collectors in the different experiments, all followed a similar protocol.
Each experiment involved deploying spat collectors by attaching them onto a mounting structure that kept them constantly submerged in the water (Table 1). Collectors were deployed each week in June-August, corresponding with the period when high numbers of O. edulis larvae were measured in the water column ( Figure 1a). Once all the collectors had been deployed, they were left until October/November to allow the spat to grow to an easily visible size at which time all collectors were retrieved. The retrieved spat collectors were transported to a laboratory where they were inspected for oyster spat and other fouling organisms. The mussel, oyster, and cockle shells used in the studies presented here were obtained from local farmers. The mussel shells were cooked and the cockle and oyster shells were weathered for a least 3 months. The shells were deployed inside 5-L mesh sacks.

| 2003
Spat collectors were deployed in three areas of Lake Grevelingen

| 2011
Spat collectors were deployed near Stampersplaat (2)  The Chinese hats were identical to those used in 2003 but were coated in a calcium solution made of 50% calcium (calcium dihydroxide and magnesium dihydroxide) and 50% water. The collectors were left to dry until the coating was completely hardened prior to deployment.

| 2017
Spat collectors were deployed at Blokkendam (7; in the Voordelta, North Sea coastal area) and Outlet (6; where Lake Grevelingen feeds into the Voordelta; Figure 1b). The collectors tested were separate similar-sized sacks with 500 g mussel shells (average of 340 shells), or 1 kg Pacific oyster shells (average of 95 shells). One sack of each spat collector was attached to a float that was suspended in the water 1 m from the bottom, tethered to a heavy tile. Another sack of each spat collector was strapped directly onto the tile. A set of collectors was deployed each week from week 24 to week 35 and retrieved for analysis in October 2017 (week 42; Figure 1a).
The spat collectors tested were separate similar sized sacks with 300 g mussel shells, 1,000 g cockle shells (Cerastoderma edulis), 500 g

| 2017
The largest number of flat oyster spat collected at Outlet occurred in week 23, the first week of collector deployment, suggesting that the peak in larval abundance had occurred prior to the beginning of the experiment. There was a secondary peak in larval abundance in week 33, but no increase in spat collection was observed in the following week.
At Blokkendam, a total of 155 flat oyster spat was collected. The highest number of spat was found on collectors deployed in week 23.

| 2018
At Blokkendam a peak in flat oyster spat collection was observed in weeks 28 and 29, one and two weeks after the peak in larval

| 2003
Fouling was heavy on the PVC tubes, consisting mainly of Pacific oyster spat (Figure 5a), and on the mussel sacks, dominated predominantly by barnacles and tunicates (Ciona intestinalis; Figure 5b). Comparatively less fouling was observed on Chinese hats, consisting mainly of macroalgae (Figure 5c).

| 2011
While all three collector types were fouled on retrieval (Figure 6), the organisms could easily be removed from the Chinese hats due to the calcium coating, and from the mussel stockings when the mussel shells were removed from the mesh. Fouling was, however, particularly difficult to remove from the Vexar mesh, which resulted in the necessity to abandon removal attempts and manually search through the fouling for oyster spat.
There was a high degree of fouling on almost all collectors deployed during the experiment. However, the dominant species of fouling organism depended on which week the collector was deployed. Collectors from the first deployment week were heavily fouled with blue mussels, and mussel presence decreased with collectors deployed in later weeks. Other dominant fouling organisms included macroalgae, barnacles, bryozoans, tunicates (C. intestinalis and Didemnum sp.) and Pacific oyster spat.

| 2017
There was limited fouling on the shells at both experimental locations in the 2017 study. The fouling present was generally attached to the sack and consisted mostly of barnacles, bryozoans, tunicates (C. intestinalis and Didemnum sp.), macroalgae, and some Pacific oyster spat (Figure 7). There appeared to be little difference in fouling amount or species between the mussel shell and oyster shell collectors, but this was not explicitly analysed in this study.

| 2018
All collectors from both locations were heavily fouled, although there was some variation in type of fouling organisms depending on the substrate. In general, cockle shells trapped the most sediment, but were the least overgrown with other organisms, and these were almost exclusively barnacles. Flat oyster shells, Pacific oyster shells and mussel shells showed more variation in fouling coverage, with F I G U R E 5 Example of epibiota fouling on the PVC tubes (a), mussel sacks (b), and Chinese hats (c) after retrieval from the water at the end of the experiment in Lake Grevelingen in 2003 some shells being completely covered in fouling organisms, while others remained almost completely bare.

| DISCUSSION
Enhancement of natural production of flat oysters, whether for ecological restoration or cultivation purposes, is a challenging objective.
The environmental requirements of the oysters need to be met while other factors such as predation, competition, diseases, and parasites should be actively mitigated to give the oysters their best chance for survival. The early life stages show high mortality rates, and are a significant bottleneck in the population due to high levels of predation of both larvae and young spat and the requirement that the larvae have for available, appropriate substrate on which to settle (Korringa, 1940;Rodriguez-Perez et al., 2019;Smyth et al., 2018). By providing more appropriate substrate on which to settle, at the right time in a way that minimizes the competition pressure exerted by other species, more flat oyster spat will survive this population bottleneck. In this way, the potential for successful restoration and cultivation of the oyster will increase.
The different experiments presented here explored possible methods to actively promote flat oyster populations during the larval settlement stage by using off-bottom spat collectors in the Dutch delta and coastal waters. Lake Grevelingen is a relatively sheltered oyster production area, and the Voordelta is open to the influence of the North Sea. Despite the environmental differences between these two areas, flat oysters are known to grow and reproduce in both areas (Pogoda et al., 2019;Smaal et al., 2015).
The comparative abundance of flat oyster and Pacific oyster larvae varied among the experiments. Due to a high abundance of Pacific oyster larvae in 2003, they may have outcompeted the flat oyster spat for space on the collectors, resulting in the low numbers F I G U R E 6 Example of epibiota fouling the spat collectors (Chinese hats, a; Vexar mesh, b; and mussel sacks, c) deployed in the first week (a1, b1, c1) and ninth week (a2, b2, c2) of the experiment in Lake Grevelingen in 2011 of flat oyster spat observed following the peaks in larval abundance.
In 2011 flat oyster larvae greatly outnumbered Pacific oyster larvae in the water column, and much higher numbers of spat were observed on the collectors compared with 2003. The reason for this is unclear; perhaps peak larval abundance for flat and Pacific oysters occurred at different times at Stampersplaat in 2011, and the experiment coincided with the peak in flat oyster larval abundance and not that of the Pacific oyster. In general, differences in flat oyster larval abundance between years has been observed in the past in the nearby Oosterschelde estuary (Korringa, 1947).
Pacific oysters were also one of the most common fouling organisms on the collectors. Pacific oysters are a non-indigenous species in the Netherlands, introduced in the 1960s to supplement the flat oyster stocks that were failing due to the severe winter of 1963/64 (Engelsma et al., 2010). The Pacific oyster is a successful species with a wide environmental tolerance range and high reproductive rate (Troost, 2010). The introduction of this species therefore may have added yet another pressure to the already stressed population of flat oysters in the Netherlands. However, Christianen et al. (2018) suggested that the presence of the Pacific oyster may facilitate the survival of the flat oysters by providing refuge against predation. In any case, developing a method of spat collection that distinguishes between these two oyster species is unlikely to be successful, but understanding the specific characteristics of the flat oyster related to spatfall and settlement may help increase the amount of surviving flat oyster spat.

| Timing of deployment of collector
To study the timing of deployment of collectors, testing it in different years and in different locations increases the strength of the conclusions. The results indicate that timing the collector deployment with the peak in larval abundance is important to the success of a spat collector. The pelagic phase of O. edulis larvae lasts only a few weeks in summertime (Korringa, 1947). In two of the four experiments (2011 and 2018), the peak in spat settlement occurred on collectors deployed 1-2 weeks after the peak in larval abundance. It would therefore be most efficient to concentrate the deployment of collectors shortly after the peak in larval abundance is detected.
If collectors are deployed too early, they may become fouled with other organisms that will prevent the spat from settling, or smother or overgrow spat that does manage to settle (Mackenzie, 1970). This competition for space was observed on the first collectors placed in 2011 when mussels settled before the flat oyster spatfall so that much of surface area on the collectors deployed was already occupied before the flat oyster spat settled. Other dominant fouling organisms observed on the collectors in all studies included algae, bryozoans, and colonial tunicates. These organisms can completely cover a surface, and essentially replace the hard, suitable substrate of the collector with a soft, gelatinous surface on which the spat will not settle.
If deployed too late, the collectors will miss the peak in spatfall, rendering them essentially obsolete for that season. In 2017, collectors were probably deployed after the peak in larval abundance, which was absent in our data but probably occurred earlier in the season.
Consequently, collectors probably missed the peak spatfall as spat collection was highest on collectors deployed in the first week of the experiment. It would therefore be useful to regularly sample the larvae in the water and deploy the collectors only after a high number of larvae is detected.
Oyster breeding success varied from year to year and between locations. It is likely that high variation in timing of settlement exists amongst different geographical regions within the oysters' natural range. Therefore, it is difficult to predict when a peak in larval abundance occurs, or even what a 'high number' of larvae for that season will be. What causes this variation in reproductive success is not well understood but may be linked to water temperature. Temperature sum (accumulated temperature) is an important variable explaining larval abundance in modelled scenarios for bivalves (Broell, Mccain, & Taggart, 2017). Maathuis, Coolen, Van der Have, and Kamermans (2020) developed a model where the seawater F I G U R E 7 Example of epibiota fouling on the mussel (left) and Pacific oyster (right) sacks in the experiment in 2017 (top); and cockle, mussel, flat oyster, and Pacific oyster (top to bottom) shell sacks in 2018 (bottom), after retrieval from the water at the end of the experiments in the Voordelta temperature sum in early summer could be applied as a crude predictor of peak flat oyster larval abundance. In this study, larval counts of Korringa (1947) and the data presented here for 2003 and 2011 were included. Therefore, it is advisable to monitor the seawater temperature sum along with the larval abundance in the water column and deploy collectors 1-2 weeks after maximum larvae numbers are detected.  (Tamburri, Luckenbach, Breitburg, & Bonniwell, 2008). Our study shows that collectors placed 1-2 weeks after the larval peak yielded most spat. This probably allowed a biofilm to develop on the collectors during the pelagic phase of the larvae. The artificial spat collectors such as Chinese hats and PVC tubes are designed to simplify and mechanize the retrieval process by making individual spat easy to remove from the collector (FAO, 2019). However, the results suggest that the efficiency of these collectors was limited as they tended to collect fewer spat than the bivalve shells. The calcium coating on these artificial spat collectors is a practical solution to allow for easy detachment of spat during retrieval.

| Type of collector
The potential and justification for investment into off-bottom bivalve shells as spat collectors depends on the intended purpose for the oyster spat. From a cultivation perspective, empty shells are most productive when a single oyster spat survives on each shell.
As the oyster grows, the shell will eventually break away leaving an individual, unattached oyster. If more than one oyster grows on a shell, they may crowd each other and become attached to each other or warped in shape, which is not desirable from a producer's perspective as these oysters may only be suitable for the oyster meat market (Nalesso et al., 2008). From an ecological restoration perspective, crowded shells are not a problem as the aesthetics of the oyster are not a priority. In addition, potential empty bivalve shells can simply be translocated along with the oysters to the restoration site, thereby providing more substrate for future larval settlement.
For cultivation purposes, the costs of spat collection remain a regular, necessary investment. Collectors need to be prepared, deployed, retrieved and processed every season. Shells as spat collectors cannot be reused, because they degrade as the oysters grow. In addition, the availability of shells depends on shellfish production, which shows natural fluctuations. This has prompted the development of artificial, reusable spat collectors and mechanized harvesting techniques as seen with Chinese hats and PVC tubes in France (Anonyme, 2014). If the increased spat collection on off-bottom bivalve shell collectors could offset their drawbacks, and if the investment involved with their use could be made more economically viable with mechanization, bivalve shell collectors may be a good option as commercial spat collectors.
The intention of restoration efforts for the flat oyster are to create an eventually self-sustaining, natural oyster reef (Kamermans et al., 2018). The short-term investment intended for this type of restoration, therefore requires less financial and labour investment than for cultivation. Off-bottom bivalve shells are therefore a promising method of spat collection for restoration purposes.

| CONCLUSIONS
This paper describes results of spat collection in areas where adult oysters are present and larval abundance is not limiting. Under these circumstances the use of off-bottom spat collectors for restoration and production of flat oysters can be optimized when taking into account the following points: • Monitor the abundance of larvae in the water column, along with the temperature of the water to determine larval peak.
• Deploy collectors 1-2 weeks after the peak in larval abundance to allow for maximum spat settlement, and to minimize the impact of other fouling species.
• The most successful spat collectors were suspended sacks of bivalve shells.
• For biosecurity reasons only cooked or weathered shells should be used.
• The use of off-bottom sacks of bivalve shells as spat collectors show great promise for oyster reef restoration and, if developed to be more economically viable, commercial cultivation purposes.