Large-Scale Spatial Dynamics of Intertidal Mussel (Mytilus edulis L.) Bed Coverage in the German and Dutch Wadden Sea

Reef-building bivalves such as intertidal blue mussel (Mytilus edulis) are important for the community composition and functioning of coastal ecosystems for various reasons. Intertidal mussel beds provide structural heterogeneity on otherwise homogeneous mudflats and thus provide habitat for other species (Gutiérrez and others 2003; Buschbaum and others 2009; van der Zee and others 2012). Mussels may compete with other suspension feeders by consuming large amounts of phytoplankton (Kamermans 1994; Cadée and Hegeman 2002). Consequently, secondary production and biomass of intertidal mussel beds is high (Asmus 1987) making them an important food resource for, amongst others, birds and crabs (Nehls and others 1997). Mussels may also promote planktonic primary production by accelerating nutrient cycling (Asmus and Asmus 1991) and local benthic primary production by deposition of nutrient-rich feces and pseudofeces. Lastly, mussel beds may locally reduce hydrodynamic forces which influences local sediment properties (Widdows and Brinsley 2002). To develop a better understanding of the roles that intertidal mussel beds play in coastal ecosystems, it is important to identify the main determinants of mussel bed distributions and dynamics. Insight into spatial distributions, habitat suitability, reproduction, dispersal, settlement, and survival are prerequisites to understand and model large-scale spatio-temporal variability of marine bivalve populations. The purpose of the present paper is to foster our understanding of long-term spatio-temporal development of adult intertidal mussel bed distributions in the Wadden Sea as a step towards the development of spatial metapopulation models of blue mussels.

Spatial distributions and dynamics of intertidal mussel beds are determined by the combination of larval supply, settlement and various post-settlement loss processes which in their turn depend on the interplay of large-scale hydrodynamic forces and small-scale facilitation (van de Koppel and others 2005, 2012). The blue mussel is a gonochoric bivalve species with a flexible reproductive strategy that adjusts spawning according to environmental conditions (Gosling 1992b). Larvae and settling individuals are observed in spring, summer, and autumn (Bayne 1976; Pulfrich 1996; de Vooys 1999; Philippart and others 2012). Dispersal of the pelagic planktotrophic mussel larvae is strongly determined by water currents (Levin 2006; Metaxas and Saunders 2009; Carson and others 2010; White and others 2010) and the duration of the planktonic larval stage which depends on environmental factors such as food availability, temperature, and salinity (Gosling 1992b). Blue mussel larvae have been estimated to reside in the water column for approximately 3–5 weeks (Gosling 1992b; de Vooys 1999). When larvae reach a size of approximately 0.2 mm they settle and produce byssus threads to attach to substrata and to each other which protects them from dislodgement due to currents and waves (Bayne 1964, 1976; De Blok and Tan-Maas 1977; McGrath and others 1988). Important substrata on the intertidal soft sediments in the Wadden Sea are adult mussels, Pacific oysters (Crassostrea gigas), cockles (Cerastoderma edule), beds of dead shells, protruding tubes of Lanice conchilega (Callaway 2003), macroalgae and seagrass (Pulfrich 1996; wa Kangeri and others 2013). Bed formation results in increased bottom surface roughness which reduces hydrodynamic force (Widdows and Brinsley 2002; Donker and others 2012). Intertidal mussel beds that have established on soft sediments provide substratum for attachment and shelter of postlarvae which results in positive feedback between mussel bed presence and local recruitment allowing a bed to persist over long time spans (Bayne 1964; McGrath and others 1988; Gutiérrez and others 2003; Herlyn and others 2008). Foundations of dead shells which are built up underneath mussel beds may also provide substratum for postlarvae thereby upholding the positive feedback even when living mussels have disappeared (Herlyn and others 2008). Settlement may, however, also occur outside mussel beds or shell foundations in years when there are large quantities of settlers. Predation by shrimp and crabs (van der Veer and others 1998; Andresen and van der Meer 2010) is an important loss factor shortly after settlement while waves and currents (Nehls and Thiel 1993; Brinkman and others 2002; Hammond and Griffiths 2004; Herlyn and others 2008; Donker and others 2012) and ice-scouring during cold winters (Strasser and others 2001) may be important during all life stages of mussels and mussel beds.

Adult bivalve population sizes may fluctuate vigorously as a consequence of high temporal variation in larval supply, early stage survival and variation in adult mortality rate (Thorson 1950; Gosling 1992a; Beukema and others 1993, 2001, 2010; Caley and others 1996; van der Meer and others 2001; Levin 2006). In the Wadden Sea, Macoma balthica, Cerastoderma edule, Mya arenaria, and Mytilus edulis often show recruitment peaks after cold winters (Beukema and others 2001; Strasser and others 2003). Bivalve recruitment peaks following cold winters have also been observed in other northwestern European seas and estuaries such as the Swedish west coast (Möller 1986) and the Wash, UK (Young and others 1996). Various explanations have been suggested for this. First, adults produce more or larger eggs under cold conditions because lower maintenance metabolism is required, thus leaving more energy for gonad production (Honkoop and van der Meer 1998). It should be noted, however, that the experiment that Honkoop and van der Meer (1998) performed with Macoma balthica shows that the temperature effect on gonad production is rather small. Secondly, bivalve spawning may occur earlier in the season after warm winters which may lead to a temporal mismatch with the spring bloom of phytoplankton (which, inter alia, depends on light conditions) which decreases recruitment (Philippart and others 2003). The third line of explanations concerns the temperature-related phenology of mussels and their main crustacean predators, shore crab Carcinus maenas and brown shrimp Crangon crangon (van der Veer and others 1998; Andresen and van der Meer 2010). Specifically, post-settlement survival on intertidal flats is relatively high after cold winters because of lower densities or later arrival of shrimps and crabs (Beukema 1991, 1992; Young and others 1996; Strasser and Günther 2001). Reduced densities or later arrival of predators leads to a larger proportion of young bivalves to grow too large to be handled and consumed by predators (Philippart and others 2003; Beukema and Dekker 2005). Mortality of adult bivalves may also peak during cold winters because frost is lethal for cockles (Cerastoderma edule) and ice-scouring may eradicate intertidal mussel beds (Strasser and others 2001; Büttger and others 2011).

The observation that recruitment peaks often follow cold winters suggests that winter temperature drives large-scale spatial population synchronization because winter temperatures are correlated over large geographic ranges. Particularly, Beukema and others (2001) show that recruitment of the bivalve species Macoma balthica, Cerastoderma edule, and Mya arenaria is spatially synchronized between three distant locations in the western Dutch and western German Wadden Sea between 1973 and 1999 (Mytilus was not included in the monitoring in two of the three sites).

Synchronous dynamics on smaller spatial scales can result from spatial autocorrelation in more regional growth determinants. For instance, diverging regional trends in food availability, predation, and hydrodynamic impact could lead to distinct synchronization patterns between regions. Synchronous dynamics among nearby populations can also result from distance-limited dispersal of larvae between populations (Bjørnstad and others 1999; Ranta and others 2006). The supply of planktonic larvae in a particular region depends on the proximity of larvae sources and hydrodynamic transport toward the region (Levin 2006) and, to a certain extent, also on larval swimming behavior (Knights and others 2006). As the swimming capacity of larvae is limited relative to water currents, the spatial connectivity of the mussel populations, and therefore the degree of population synchrony, will largely depend on regional geomorphology and currents during the larval stage (Young and others 1996; McQuaid and Phillips 2000; Gilg and Hilbish 2003; Watson and others 2010).

Temperature, the spatially diverging growth determinants and dispersal will lead to characteristic signatures in spatial population covariance. That is, the spatial range over which the determinants operate, will show up in the spatial range of population synchronization (Bjørnstad and others 1999). If temperature is the dominant determinant, it is expected that synchronization will be observed in the entire population. If regionally distinct determinants are important then synchronization patterns will show autocorrelation on a regional scale. And more specifically, if geomorphology and hydrodynamic transport are important determinants of population synchronization, physical barriers such as rivers will show up as boundaries of synchronizing populations. Because temperature and the regional mechanisms operate simultaneously, the temperature effect must dominate the regional mechanisms to achieve visible synchronization at a large spatial scale. Similarly, regional effects must be pronounced in order to show up if there are strong temperature effects.

The objective of this study is to obtain insight into the distribution and dynamics of intertidal mussel beds in the Wadden Sea by exploring large scale spatio-temporal trends and scale-dependent population synchronizations. First, we compare the mean spatial distribution of intertidal mussel beds for the period 1999–2010 with the spatial distribution of the period 1968–1976 to get insight into long-term spatial variability and into the determinants of habitat suitability. Secondly, we analyze spatial patterns in the synchronization of intertidal mussel bed dynamics for the period 1999–2010. Thirdly, we use a stochastic growth model to analyze the impacts of winter temperature and storminess on intertidal mussel bed dynamics for the period 1999–2010 while controlling for density dependence. The analyses presented here aim to stimulate the development of spatially explicit metapopulation models including habitat suitability, resource dynamics and hydrodynamic connectivity. Insight into these factors is relevant from a management and conservation point of view where insight into long-term variability and natural establishment of intertidal mussel beds is a basic requirement.

We analyzed the spatial variability of the coverage of mussel beds in the international Wadden Sea on the basis of data collected in the periods 1968–1976 and 1999–2010. Within each period, notable regional differences in mussel bed coverages were observed. Between the periods, however, these distribution patterns were remarkably stable in that the regions that were densely populated in the period 1968–1976 were also densely populated in the period 1999–2010. The highest coverages occur in the relatively small and sheltered tidal basins in the eastern Dutch and in the western Lower Saxon Wadden Sea whereas mussel beds were virtually absent in tidal basins that are not sheltered by barrier islands (TB 11–18). The strikingly stable mussel bed distributions may be explained by the regionally variable levels of exposure to hydrodynamic forces which are relatively constant on the scales of decades.

Although analysis of long-term regional differences in mussel bed coverage provides insight into the relatively invariable determinants of habitat suitability, analysis of year-to-year variability including spatial synchrony, sheds light on the determinants of short-term dynamics. We expected that yearly increases in coverage would be explained by episodic recruitment events following cold winters and that decreases would occur during stormy winter seasons. In particular, previous studies performed at various locations have shown that bivalve recruitment events are often preceded by cold winters (Möller 1986; Young and others 1996; Beukema and others 2001; Strasser and others 2003) which has led to the commonly held view that winter severity is the main driver of bivalve recruitment events and population synchronization in North West Europe. Because winter temperatures are strongly correlated in the entire study region (Figure 5) we hypothesized that growth peaks would show up at the same time throughout the entire study area. If regional growth determinants or larval dispersal would dominate, then population synchronization would show up at smaller spatial scales. Our analyses did not show that cold winters systematically induced increases in mussel bed coverages. Nor did we find strong evidence for negative effects of storminess. We did find, however, that synchronic growth was constricted to confined blocks of tidal basins bordered by hydrodynamic barriers such as rivers and large extents of unsuitable habitats. These results point toward the importance of considering larval transport and early stage survival to further develop understanding of the spatial dynamics of mussel beds in the Wadden Sea. The importance of larval transport has also been demonstrated for coastal mussel populations in southwest England (Gilg and Hilbish 2003) and the southeast South Africa (McQuaid and Phillips 2000) where coastal circulation is an important determinant of larval dispersal. Although our hypothesis and results are consistent with findings from other systems, we do not rule out alternative explanations. For instance, regional differences in the change of eutrophication (van Beusekom and others 2012) and carrying capacity or predation rates could lead to regional differences in vital rates which could cause the observed synchronization patterns. Another possible cause of diverging dynamics in the separate blocks may be related to local hydrodynamic disturbances which were not detected by regressing adult mussel bed dynamics on storminess. To distinguish between the above-mentioned hypotheses it is important to obtain detailed long-term spatial data on competition and food limitation, reproductive capacity and output, early stage predation, recruitment, hydrodynamic disturbance and transport. In this context it is also important to have better understanding of the positive and negative interactions between Pacific oysters and mussels. Particularly, mussels and oysters benefit from each other in that they provide suitable substrate but they suffer from each other due to competition for food resources. The long-term consequences of the establishment of Pacific oyster populations for mussel populations will depend on the responses of both populations to environmental conditions and on the interactions between the populations.

Open image in new window

The fact that we did not find a temperature effect on growth does not imply that temperature may not have an effect on the long-term dynamics and sustainment of intertidal mussel beds at all. There are several reasons why we might have missed to detect a winter temperature effect. Firstly, our time series covered the period 1999–2010 which was relatively warm. Beukema and others (2001), who did find a temperature effect, used data for the period 1969–1999 which contained several cold winters. In particular, the severe winters of 1978/1979 and 1995/1996 were followed by extraordinary recruitment events for different species of bivalves including Mytilus edulis. The study of Möller (1986) also included the year 1979 after which an eightfold increase in the numerical density of Mya arenaria was observed. The study of Strasser and others (2003) focused on recruitment of several species of bivalves in the whole Wadden Sea following the cold winter of 1995/1996. They found higher than average densities of recruits of Cerastoderma edule in 1996 in the entire Wadden Sea. For the species Macoma balthica and Mya arenaria higher than average recruitment was only observed in the southern Wadden Sea. It is therefore possible that we would have found statistically significant temperature effects if our time series had included more and colder winters. It is important to note, however, that cold winters are not a prerequisite for high recruitment events as relatively mild winters may also be followed by high recruitment events. For instance, after the relatively mild winters of 1974/1975 and 1990/1991, Beukema and others (2001) found high densities of recruits for different species of bivalves. And in our data set the mild winters of 2000/2001 and 2004/2005 were followed by substantial increases of adult mussel beds in the Dutch Wadden Sea (observed as peaks in 2002 and 2006). Because cold winters are expected to occur less frequently in the future due to climate change, the factors underlying recruitment after relatively warm winters will become increasingly important determinants of bivalve metapopulations. Another potential cause of the absence of a winter temperature effect is that loss processes operate between the moments of recruitment and the surveys of adult mussel beds which may distort the correlation between winter temperature and mussel bed growth to some extent. This implies that winter temperature effects might be more easily detected for endobenthic species such as Macoma balthica than for epibenthic species that are more vulnerable to hydrodynamic disturbance. Note, however, that we did observe bivalve-characteristic episodic growth peaks by considering yearly changes in the coverage of adult mussel beds (Figure 3).

The absence of evidence of a storminess effect in the growth model does not imply that storminess does not influence mussel bed dynamics. The absence of a storminess effect may be explained by the fact that we considered dynamics of adult mussel beds while especially settlement and newly settled mussel beds are sensitive to hydrodynamic disturbance and storminess because individual mussels have not yet attached themselves firmly to the sediment and each other. As a result establishment and development of adult mussel beds tends to occur at relatively sheltered locations where the impact of storms on adult beds is relatively low. For example, Nehls and Thiel (1993)—who studied survival of mussel beds including newly recruited mussel beds—found that only 49 out of 94 beds survived their first winter and that all of the surviving beds were located in sheltered areas. These areas were also found to be populated with mussel beds in several decades before and after their investigation. Our paper concerns the dynamics of adult mussel beds which implies that we are analyzing the dynamics of beds that already survived the initial young stages and thus occur in relatively sheltered locations only.

The limitation of our empirical approach to detect storminess effects could be cleared up by integrating observations of young and adult intertidal mussel beds with hydrodynamic models (Donker and others 2012). Although such integrated models provide insight into the mechanic interactions between hydrodynamic forces and intertidal mussel beds, the approach will be difficult to apply to large temporal and spatial scales for the following two main reasons. Firstly, the hydrodynamic conditions of all the relevant locations—which include tides, wind and fetch—need to be known which requires very detailed data sets, extensive monitoring and complex and computer intensive modeling. Secondly, because waves and currents are attenuated when they travel over the rough mussel bed surfaces, they cannot be considered independent variables. Yet, it would be useful to have more insight into the long-term impacts of waves and currents on the tidal flats in order to identify suitable mussel bed habitat in terms of hydrodynamics independently of the observations presented in this paper.

In this paper, we have suggested that synchronic spatial population dynamics within blocks of neighboring tidal basins may result from distance-dependent dispersal of larvae. In particular, larvae that are produced in one tidal basin may be transported to hydrodynamically connected tidal basins. This suggestion is supported by the fact that the borders of the synchronic blocks are located at tidal basins with estuaries and rivers (in which water currents form a barrier toward the next tidal basins) and at regions that lack suitable habitat. As mentioned above, the turnover time for the Ems estuary (Figure 1, TB 30) is substantially longer than the turnover times for the other Dutch tidal basins (Ridderinkhof and others 1990) so that the chance for a larva to survive the time it takes to pass the estuary and arrive at suitable habitat is relatively low. The chance to survive the passing of the Elbe estuary (TB 18) and arrive at suitable habitat would conceivably be even lower. Our results, that is, that the population dynamics within the blocks separated by the Ems and Elbe estuaries are distinct, are in line with this observation. Hydrodynamic transport modeling, on the basis of tides and wind, could elucidate possible fates of mussel larvae and provide insight into the connectivity of suitable habitat patches within tidal basins. This type of modeling would help to provide insight into the structure of the metapopulation network and would be able to show which subpopulations produce larvae with high survival and settling probability. In addition to water movement, the characteristics of the water affecting larval growth and development (such as food availability, salinity, and temperature) should be considered (Metaxas and Saunders 2009). Particularly, a lower growth- and development rate during the larval phase increases the time a larva needs to complete development which increases the risk of being predated by ctenophores and benthic feeders or of being transported offshore (Young and others 1998; Philippart and others 2003; Metaxas and Saunders 2009). Hydrodynamic connectivity modeling and water quality measurements in combination with larval requirements could provide a framework for the development of a dispersal model describing the corridors between suitable habitats from the perspective of larvae.

In the Wadden Sea there is high quality, comparable long-term distribution data available for intertidal mussel beds but not for subtidal mussel beds. Because our quantitative analyses required homogeneous data we have focused on intertidal mussel beds only. Subtidal mussel populations may, however, influence the dynamics of intertidal populations by producing larvae that end up in the intertidal and vice versa. Hence, a complete model of spatial population dynamics would include survival, production, dispersal, and settlement of both intertidal and subtidal populations.

The spatial configuration of suitable habitat and population growth dynamics presented in this paper is relevant from a natural resource management perspective. If it is combined with the outcomes of further research suggested above, it will be possible to pinpoint those locations and sub-populations that are of pivotal importance for the resilience of the metapopulation of mussel beds and hence require special protection.