1 Introduction

Seagrass meadows provide important habitats for a wide range of fish species. The often higher densities and diversity of fish in seagrass meadows compared to other coastal habitats, particularly sand and mud flats, reflects elevated structural complexity, primary and secondary production, and availability of vegetal detritus (Nagelkerken 2009 and references therein; Nanjo et al. 2014). These characteristics, in turn, are likely to provide increased protection from predation and food availability for both adult and juvenile fish compared to less complex habitats (Horinouchi 2007). For these reasons, seagrass meadows are perceived as important “nursery habitats” for fisheries production, and have received considerable attention, leading to several reviews over the last few decades (see Bell and Pollard 1989; Connolly et al. 1999a, b; Gillanders 2006; Heck and Orth 1980; Heck and Valentine 2006; Ogden 1980; Jackson et al. 2001; Whitfield and Pattrick 2015; Ogden 1977).

Many of the early generalisations of seagrass ecology emerged from studies carried out in the 1970s and 1980s, which led to an increasing effort in seagrass research in the 1990s and 2000s. In a search on ISI Web of Science using the keywords “fish*” and “seagrass*”, a total of 2,114 papers referred to fish and seagrass between 1992 and 2015. A confirmation process for those papers revealed that a total of 601 papers (not including review papers) specifically examined at least one aspect of fish ecology in seagrass meadows. The number of papers increased from 59 in 1992–95 to 166 in 2004–07, before declining to 84 in 2012–2015 (Fig. 18.1). The majority of papers originated from Australia and North America in the 1990s, but subsequently, studies from the Caribbean and Europe in the 2000s contributed to the literature (Fig. 18.1). There was also an increase in the number of papers coming from Africa and Asia during the 2000s. Perhaps reflecting the focus of research in narrow geographic ranges, a large number of studies has examined fish assemblages in meadows dominated by Zostera and Thalassia (200 and 137, respectively) (Fig. 18.1b). Fish in Posidonia and Halodule meadows received less attention (74 and 66 papers, respectively) (Fig. 18.1b). However, Posidonia received far greater attention in the 2000s, mainly associated with increased work in the Mediterranean Sea. Fish assemblages in other seagrasses, including Amphibolis, Cymodocea, Enhalus, Halophila, Syringodium and Thalassodendron, have received little attention.

Fig. 18.1
figure 1

The number of primary publications examining fish in seagrass meadows in a different years and regions, and b different seagrass genera

Full size image

In this chapter, we examine the importance of seagrass meadows to fish and fisheries, either directly as a habitat and food source, or indirectly through the provision of seagrass-associated organic matter to other habitats. We generally focus on finfish, but broaden the scope to include invertebrates when discussing fisheries. A recent increase in the number of published studies examining fish in seagrasses in a range of seagrass genera and regions around the world (Fig. 18.1), allows us to examine whether early generalizations regarding the use of seagrass meadows by fish hold for a broad suite of seagrass genera and geographic regions, and discuss these in the context of Australian seagrasses where possible. We firstly consider the spatial and seascape patterns in diversity of fishes by examining fish assemblages in seagrass meadows compared to other habitats, including comparisons across different seagrass habitats and depths, and exploring the current knowledge of structure, patch size and edge effects. We then focus on assessing the generalisations regarding the nursery function of seagrass meadows to fish and feeding ecology within those systems. We next examine the mechanisms of connectivity between seagrass meadows and coastal and offshore fisheries, focusing particularly on Australian fisheries. Finally, we discuss management issues related to seagrass and the sustainability of fisheries, followed by the main research gaps the literature that we consider need attention.

1.1 Characteristics of Seagrass Habitats

The physical characteristics of seagrasses differ markedly among genera. In a classification proposed by Walker et al. (1999), Zostera and Halodule have low biomass, are characterized by rapid turnover, and are ephemeral (Fig. 18.2). In comparison, Thalassia, Enhalus and Posidonia have relatively high biomass, slow turnover of biomass, and are persistent (Fig. 18.2, Hemminga and Duarte 2000). The different growth forms respond differently to disturbance , and are likely to interact differently with higher trophic levels (Walker et al. 1999). Furthermore, different genera have different distribution ranges, with Zostera exhibiting the greatest latitudinal range from the tropics to cold temperate regions (Moore and Short 2006), while Thalassia is restricted mainly to the tropics (Van Tussenbroek et al. 2007) and Posidonia to temperate regions of Australia and the Mediterranean Sea (Green and Short 2003). Also, the different forms of seagrasses occur in different coastal environments, and therefore form different associations with neighbouring coastal habitats. For example, Thalassia is often associated with mangroves and coral reefs (van Tussenbroek et al. 2007), while Zostera is often found in estuaries and Posidonia in marine waters (Guidetti et al. 2002; Gobert et al. 2006).

Fig. 18.2
figure 2

Conceptual model displaying the life history and functional characteristics of seagrass genera that are relevant for their roles as fish habitats (adapted from Walker et al. 1999)

Full size image

The differences in form and environment affect the interactions of seagrasses with the associated epiphytes and invertebrate fauna, and the combination of all these attributes will affect the associated fish assemblages. Seagrass species with relatively slow turnover such as Posidonia spp., and those with extensive and persistent stems such as Amphibolis spp., have high levels of epiphytic algae and sessile epifauna (Borowitzka et al. 1990; Jernakoff and Nielsen 1998). These attributes would increase food availability and habitat structure beyond the seagrass itself, and influence the densities and diversity of fish assemblages in seagrass meadows. In addition, Heck and Orth (1980) suggested that there is a latitudinal shift in habitat complexity and heterogeneity, with temperate seagrass meadows being the most homogeneous and tropical meadows being the most heterogeneous through the occurrence of associated organisms such as sponges, bryozoans, corals, rhodophytes, and calcareous algae such as Halimeda spp. The addition of these other habitats could lead to increased diversity of fishes in seagrass meadows (Nagelkerken et al. 2015).

2 Spatial and Seascape Patterns in Diversity

In early studies, the value of seagrass as a fish habitat was often based on comparisons between seagrass meadows and adjacent unvegetated sediments. The greater diversity and densities of fish, particularly juveniles, in seagrass meadows led to the general acceptance that seagrasses provide critical nursery habitats for fish, and this was often attributed to their greater structural complexity that enhances growth and survival of juvenile fish. More recently, fish assemblages of seagrass meadows have been examined in a broader spatial and seascape context, comparing them to other structured habitats or among meadows comprising different seagrass species or different spatial configurations representing fragmentation. Also, a greater focus has been placed on the influence of connectivity between seagrass meadows and other habitats at the seascape scale through the movement of fish and organic matter between habitats. Below, we discuss the role of spatial and seascape factors in influencing the value of seagrass meadows as fish habitats.

2.1 Seagrass Versus Unvegetated Sediments

Species richness and densities of fish are most often higher in seagrass meadows than unvegetated sediment throughout the world, including Australia (Ferrell and Bell 1991; Franco et al. 2006; Connolly 1994b), but there are exceptions (Heck and Thoman 1984). While those general conclusions have often been based on meadows comprising smaller species, such as Zostera spp, they have also been observed for meadows of larger seagrasses, such as Posidonia or Amphibolis. In expansive studies across southern Australia, Edgar and Shaw (1993, 1995) showed that fish production was usually, but not always, greater in shallow seagrass meadows comprising either Posidonia or Zostera than adjacent unvegetated sediment. At some locations, fish production and abundances were far higher over sand than in Posidonia meadows (Edgar and Shaw 1993, 1995), while other studies have shown similar abundances between Posidonia or Zostera meadows and adjacent sandy areas (Hyndes et al. 2003; Gray et al. 2011). Furthermore, abundances of fish are not always greater in seagrass compared to sandy areas in Thalassia meadows of the Caribbean Sea/Central America region (Adams and Ebersole 2004; Arrivillaga and Baltz 1999; Sheridan et al. 1997) and Enhalus meadows in Japan (Nakamura and Sano 2004). These inconsistent patterns likely relate to the landscape structure or density of the seagrasses (Hyndes et al. 2003), or other factors such as sampling bias and diel differences in fish assemblages.

Many studies have compared fish assemblages across habitats during the day, but fish assemblages can shift between day and night (Harmelin-Vivien 1982; Gray et al. 1998; Johnson et al. 2008). Such diel shifts can reflect the movement of certain species in and out of seagrass during the diel cycle. For example, a comparison of fish in Australian Posidonia meadows and adjacent bare sediment by Hyndes (unpublished data) showed far higher densities of fish in the former habitat during the day, but no differences at night. However, these results reflected the pronounced diel differences in densities of schooling species within the family Atherinidae, whose members appear to move into shallow sandy areas adjacent to seagrass at night (Humphries and Potter 1992). This has also been observed for labrid and sparid species in P. oceanica meadows in the Mediterranean Sea (Guidetti 2000). However, such conclusions need to consider the ability of some species to avoid capture during the day, as highlighted by Gray et al. (1998) for the schooling Liza argentea. From the above, any assessment of seagrass meadows as fish habitat against unvegetated sediments needs to consider location of the habitat, species of seagrass, and potential sampling biases (e.g. time of day and type of sampling gear).

2.2 Differences Among Seagrass Species

Coastal environments can contain a mosaic of seagrass meadows comprising a number of species, and it is overly simplistic to assume that all seagrass meadows provide the same functions. While there has been considerable attention paid to the role of meadow structure, few studies have compared fish assemblages across different seagrass habitats that vary markedly in their species composition and physical structure (Fig. 18.2). For example, species diversity and densities of fishes differ substantially between meadows of Enhalus acroides and Thalassia testudinum in Zanzibar (Gullström et al. 2008) and Halodule wrightii, Thalassia hemprichii and Syringodium filiforme in the Gulf of Mexico (Ray et al. 2014). In Australia, several studies have shown that fish assemblages differ between meadows of P. australis and Z. capricorni in eastern Australia (Middleton et al. 1984; Rotherham and West 2002; Young 1981), although these differences may reflect depth gradients across different seagrass species. Similarly, fish assemblages in P. oceanica beds differ from those in beds of the smaller Zostera and Cymodocea species (Bussotti and Guidetti 1999), which may reflect the seasonal dynamics of the seagrasses. That is, the canopy of P. oceanica is present year round (Guidetti et al. 2002), while that of small-sized seagrasses almost disappears in the winter (Guidetti and Bussotti 2000). Also, on the west coast of Australia, fish assemblages in Posidonia sinuosa and Amphibolis griffithii differ (Fig. 18.3, Travers and Potter 2002; Hyndes et al. 2003). Hyndes et al. (2003) suggested that the greater biomass and size of fish in Amphibolis meadows is related to the space among the long, permanent stems below the canopy of Amphibolis, along with the high load of epiphytes on those stems , allowing larger fish to occupy and forage in those gaps. The influence of seagrass structure on fish assemblages has been tested using simpler structural characteristics such as leaf density and height (see later section), but more complex measures of structural complexity as well as food availability need to be examined.

Fig. 18.3
figure 3

From Hyndes et al. (2003). nMDS ordination plots of the denisties and biomass of fish assemblages recorded in seagrass meadows comprising different seagrass species and unvegetated areas in south-western Australia using two different trawling methods. A = Amphibolis griffithii, C = Posidonia coriacea, S = Posidonia sinuosa, U = unvegetated areas

Full size image

Even for seagrass species of the same genus, fish assemblages can differ substantially. This is highlighted through comparisons of Australian Posidonia species, in which species richness and densities can be greater in P. sinuosa than over sand, while such differences are not apparent for P. coriacea (Hyndes et al. 2003). Posidonia coriacea belongs to the Posidonia ostenfeldii group of species that form patchy clumps of seagrass in more dynamic sandy environments compared to P. sinuosa in the P. australis group that form continuous meadows in more stable environments (Gobert et al. 2006). The lack of differences between fish assemblages in P. coriacea and sand (Fig. 18.3) likely reflects the presence of large sand patches within this habitat. At the finer patch scale , fish assemblages are likely to differ between seagrass and sand, but the degree of such differences will depend on patch size and edge effects (see below). Thus, the landscape structure of seagrass species, even within a genus, could have a profound influence on fish assemblages occupying seagrass habitats.

The different forms of seagrasses and their meadows, and their associated organisms, such as macroalgae , mangroves, corals and sponges, introduce additional layers of complexity when attempting to generalise about the role of seagrass for fish communities. Clearly, the mosaic of different seagrass species in the coastal region contributes to its overall biodiversity of fish, and the threat of replacement of one seagrass species with another (Pogoreutz et al. 2012) is likely to have significant ramifications for the biodiversity and ecosystem services these coastal ecosystems provide.

2.3 Seagrass Versus Other Structured Habitats

Much of the discussion on the value of seagrass as a habitat has focused on seagrass versus unvegetated sediment (see above), while few studies have compared fish assemblages between seagrass and other forms of structured habitats such as reefs. Jenkins and Wheatley (1998) showed that fish assemblages were more similar between structured habitats (Zostera and algal reef) than unvegetated sediment, while densities and species richness did not differ between the two structured habitats. In comparison, Sogard and Able (1991) showed a hierarchy in densities of fish in habitats, firstly those in Zostera, then the filamentous green alga Ulva, and then unvegetated mud. Guidetti (2000) showed that species richness was greater in P. oceanica than on nearby rocky reefs with macroalgal cover, while densities of fish were similar in the two habitats. While the species composition differed between the two habitats, they were more similar to each other than when compared to unvegetated sediment (Guidetti 2000). Species that make part of artisanal fishing catches, like the sparids Diplodus annularis and S. cantharus and the labrid S. ocellatus occurred predominantly in P. oceanica, while the economically and ecologically relevant sparid D. sargus and a range of labrids occurred mainly over rocky reef. However, D. annularis also settles into macroalgae of the northern Adriatic where seagrasses are not present (Guidetti and Bussotti 1997). In another study, densities of larger S. ocellatus were higher over rocky reef than P. oceanica, but medium-sized fish were more abundant in the seagrass (Mouillot et al. 1999). In comparison to those studies, Francour (1994) showed that, in Corsica, the density and biomass of fish were lower in P. oceanica than in rocky areas, particularly inside a marine reserve. However, outside the no-take area, densities were similar in the two habitats. Thus, there appears to be no consistent pattern among regions, but results may be strongly dependent on local fishing pressure (Minello et al. 2003; Sheridan and Hays 2003). In addition, the methods used to assess fish density or biomass could have influenced these trends. For example, large species mainly inhabit reefs at adult stages, while juveniles occupy seagrass meadows (Francour 1994; Dorenbosch et al. 2005; Campbell et al. 2011). If the sampling technique does not effectively record these large-sized species, the calculated density or biomass will be underestimated.

2.4 Influence of Depth

Amongst seagrasses that form beds of dense and tall plants, the species with the widest depth distribution is P. oceanica (from the surface to 40 m; den Hartog 1970). Several species of Halophila occur from the surface to depths of 30 m or more (Den Hartog 1970), but the diminutive Halophila species modify the environment much less than P. oceanica, which forms a thick mat of dead and living rhizomes and roots. There have been limited comparisons between fish faunas in shallow versus deeper meadows, but a greater biomass of herbivorous fish has been observed in shallow (1–5 m) versus deep (15–20 m) P. oceanica meadows of the Mediterranean (Francour 1997), and juvenile sparids (D. annularis and S. cantharus), and the labrid (Symphodus spp.), were observed mainly in shallower waters (Francour 1997; Bussotti and Guidetti 1999; Francour and Le Direach 1994, 1998). Similarly, Travers and Potter (2002) showed greater species richness and densities of fish in shallower meadows of both Posidonia australis and Amphibolis antarctica across depths in Shark Bay on the west coast of Australia. Based on this, differences in the fish assemblages in Posidonia australis and Zostera capricorni in Australia (Middleton et al. 1984) and P. oceanica and a complex of Cymodocea nodosa and Zostera noltii in the Mediterranean (Guidetti 2000) could be confounded by the smaller seagrass species being located in shallower water. Middleton et al. (1984) showed that smaller species and smaller individuals of some species occupied Z. capricorni while larger individuals occupied P. australis. This could result from larvae of some species settling preferentially into Zostera and subsequently migrating to Posidonia meadows (Middleton et al. 1984) or preferentially settling into shallow habitats, regardless of structure, where predation levels may be lower (Ruiz et al. 1993) and growth rates higher (Sogard 1992).

2.5 Structural Complexity

There is considerable longstanding evidence of a significant positive relationship between seagrass density and the abundance of small resident and transient fishes, and this type of relationship has been reported from individual seagrass meadows across the globe (see Gillanders 2006 for a review). However, this significant relationship between seagrass density and fish abundance does not always exist at larger spatial scales (that is, among, as opposed to within, seagrass meadows), and sparse meadows may sometimes harbor higher fish densities than dense meadows, a point made and discussed in some detail by Bell and Westoby (1986) in south-eastern Australia 30 years ago. The usual explanations for the large number of juvenile fishes in seagrass meadows is that they survive and grow at higher rates in structurally complex habitats like seagrass meadows , and data generally, but not always, support these explanations (Fig. 18.4, Heck et al. 2003). Bell and Westoby (1986) and Bell et al. (1987) tried to understand the reasons for such differences and proposed that stochastic settlement from the plankton could explain the inconsistencies at larger scales, with young-of-the-year fish choosing to settle in the first, but not necessarily the most dense, meadows they encountered when recruiting from offshore waters, and later redistributing themselves into preferred parts of the meadows in which they settled.

Fig. 18.4
figure 4

a Models describing the relationship between habitat complexity and predator success: (dotted line) Canion and Heck (2009), (solid line) Nelson (1979) and (dashed line) Crowder and Cooper (1982). b Predation rate (mean ± S.E.) of Lagodon rhomboides on Palaemonetes pugio in field experiments in Perdido Bay, FL by Canion and Heck (2009). Letters indicate significant differences (One-way ANOVA, p < 0.005)

Full size image

Recent studies, using different experimental procedures, help understand the variance existing between local and regional seagrass density and juvenile finfish and shellfish abundance of the type discussed by Bell and Westoby (1986). Earlier tank experiments increased only prey, not predator densities, with increasing seagrass density to examine the relationship between seagrass density and predation rate. However, recent studies of predator-prey relationships in mesocosms (e.g. Canion and Heck 2009, Mattila et al. 2008; Scheinin et al. 2012) showed that when densities of both prey and predators were increased with seagrass density (as they typically do in nature), the survival rate of individual prey was no different among a wide range of seagrass densities (Fig. 18.4). Survival in any density of seagrass was, however, clearly higher than on unvegetated substrates, thereby demonstrating the survival benefits of recruiting into seagrass habitats regardless of shoot density. Thus, it is likely that the first meadow encountered by a post-larval fish is a good choice for settlement.

2.6 Edge Effects and Fragmentation

Seagrass meadows are often spatially heterogeneous (patchy) habitats from fine spatial scales to seascape scales, resulting from variation in a range of environmental and biological factors that affect their distribution (Bostrom et al. 2006). These heterogeneous characteristics are often expressed as seagrass patch sizes , edge characteristics and fragmentation patterns. Understanding the effect of fragmentation of seagrass meadows on fish assemblages is crucial due to the ongoing worldwide loss of seagrass (Waycott et al. 2009) and the associated fragmentation of seagrass meadows. The resulting smaller, fragmented patches of seagrass increase edge to area ratios, which can either negatively or positively affect fish depending on their preference for the interior or edge of the meadow. Fragmentation is a process, but many studies that have examined its effect on fish assemblages have used a static state as a representation of the process of fragmentation (Bostrom et al. 2006; Macreadie et al. 2009). Studies in south-eastern Australia that simulated fragmentation using artificial seagrass units (ASUs) showed that loss of seagrass area during fragmentation may not lead to reduced fish abundances if fish are more abundant at the edge of seagrass meadows (due to increased perimeter/area ratio), as occurred in this case with pipefish (Macreadie et al. 2009). Moreover, these studies showed that treatments that were actively fragmented had higher species richness than those already fragmented at the start of the study, illustrating the difference in the effects of fragmentation compared to static patchiness (Fig. 18.5) (Macreadie et al. 2009). In the Mediterranean Sea, Vega Fernandez et al. (2005) reported interesting results based on a P. oceanica meadow that was partially destroyed by excavation, resulting in areas of seagrass habitat of equal shoot density but different degree of fragmentation. The fish assemblages associated with differently fragmented beds showed fairly variable patterns: (i) for some species, the abundance was positively related to the degree of fragmentation, (ii) some species were more abundant in fragmented beds, without any difference related to with the degree of fragmentation, (iii) some species were more abundant in large seagrass patches or in the continuous meadow.

Fig. 18.5
figure 5

Modified from Macreadie et al. (2009). Mean species richness (±standard error) in artificial seagrass units . Treatments for the fragmentation experiment were constructed from artificial seagrass units (1 m2) and included: Control, a continuous 9-m2 patch; Fragmented, a 9-m2 patch fragmented to 4 single 1-m2 patches (resulting in a 56% loss of seagrass habitat); Pre-fragmented (PF), 4 single 1-m2 patches; and Disturbance control, fragmented and then immediately restored to a continuous 9-m2 patch

Full size image

Earlier studies did not show strong evidence for seagrass edge effects on fish distributions (Connolly and Hindell 2006). Most studies showed no edge effect, and for those that did show an effect, fish abundances either increased (positive effect) or decreased (negative affect) near the edge. Many studies inferred edge effects from studies of different patch sizes (smaller patches have a larger perimeter/area ratios (Bell et al. 2001), but assumptions that changes in faunal abundance in differing patch sizes are due to edge effects is not well founded (Connolly and Hindell 2006). However, recent studies have made more direct measurements of the distribution of fish in seagrass meadows, or have used an experimental approach with artificial seagrass units (ASUs). Direct measurements of fish distribution in Zostera beds in south-eastern Australia showed strong edge effects for some species: pipefish (Syngnathidae), weedfish (Clinidae) and gobies (Gobiidae) showed positive effects, while garfish (Hemirhamphidae) and weed whiting (Labridae) showed negative effects (Smith et al. 2008, 2012). These effects were, however, not uniform in space or time. For example, pipefish were influenced by edge effects on the seaward but not landward edge of the beds, while weedfish only showed an effect at night (Smith et al. 2008). Edge effects for both these taxa varied with patch size, indicating that patch size cannot be used as a simple proxy for edge effects, and that both patch size and edge effects must be considered (Smith et al. 2010). Furthermore, edge effects may vary over the life cycle of fish. For example, King George whiting (Sillaginidae) showed a negative edge effect for small recruits (Smith et al. 2011) but a positive edge effect for older juveniles (Smith et al. 2012).

Heterogeneous characteristics of seagrass meadows, such as patch size and edge effects, can influence fish assemblages by altering: (1) water flow, physical disturbance and sediment structure; (2) predation pressure; (3) movement and behaviour of fish; and (4) reproductive strategies of fish (Bostrom et al. 2006; Macreadie et al. 2009). In support of some of these mechanisms, recent studies in south-eastern Australia showed that fish displaying edge effects were influenced by variation in the physical structure of seagrasses (Smith et al. 2008), the distribution of food resources (Macreadie et al. 2010), and the influence of predators (Smith et al. 2011). Experiments using ASUs showed that the positive edge effect for pipefish in Zostera was related to food resources (Macreadie et al. 2010), in particular the supply of copepod prey in the water column for these ambush predators (Macreadie et al. 2010; Warry et al. 2009). In contrast, negative edge effects for King George whiting recruits are likely the result of predation effects, with tethering studies showing that predation by piscivorous fish was much higher near the edge of the bed than its interior (Smith et al. 2011). In summary, recent mensurative and manipulative experiments in temperate Australia show that many fish species exhibit strong, but variable, responses to seagrass edges and fragmentation.

3 Role of Seagrass as a Habitat

The functional roles of seagrasses as habitat varies among species, which use seagrass beds as permanent and seasonal residents, transients and occasional migrants (Kikuchi 1974). Much research has focused on seasonal residents, particularly those recruiting into meadows as larvae or juveniles before migrating to other habitats, which has led to seagrasses being considered critical “nursery” habitats for many economically important species.

3.1 Nursery Role

For decades, seagrass meadows have been considered to provide critical nursery habitats, principally due to many earlier studies demonstrating high abundances of juveniles in seagrass meadows compared to unvegetated substrates (see Beck et al. 2001), and the large number of economically important species using seagrass meadows as juveniles. The underlying premise of the nursery role is that juveniles in a putative nursery habitat should disproportionately contribute to the adults in other habitats, through some combination of greater densities, growth rates and survival of juveniles in the nursery habitat, followed by their successful movement to adult habitats (Fig. 18.6, Beck et al. 2001). In support of the nursery role of seagrass meadows, a meta-analysis by Heck et al. (2003) showed that density, survival and growth of fish were generally greater in seagrass meadows, although this generality did not hold as well between seagrass and unvegetated areas in the Southern Hemisphere, or between seagrass and other structured habitats, such as saltmarshes (Heck et al. 2003). Heck et al. (2003) showed that growth rates of fish are generally greater than in surrounding unvegetated areas, but not necessarily other structured habitats. This suggests that structure and presumably increased food availability associated with habitat structure, play major roles in supporting increased growth rates. Similarly, structurally complex habitats increase survival by reducing the foraging success of predators (Heck et al. 2003). Nevertheless, there is still limited evidence that seagrass meadows disproportionately support juveniles that contribute to the adult stages of fish populations, due to the difficulty of assessing the movement patterns of juveniles from seagrass meadows into adult habitats. A recent, update meta-analysis by McDevitt et al. (2016) confirmed most of the generalisations reported by Heck et al. (2003), thereby strengthening the support for the seagrass nursery hypothesis .

Fig. 18.6
figure 6

Conceptual diagram of the nursery role concept (adapted from Beck et al. 2001). Size of symbols represents the relative magnitude

Full size image

Dahlgren et al. (2006) have further developed the discourse around the nursery role hypothesis by proposing the new term “Effective Juvenile Habitat” or “EJH”, in which an EJH should contribute more juveniles on average to adult habitats than the average of all juvenile habitats. However, this can only be evaluated by determining the proportion of fish in adult habitats that originate from all possible juvenile habitats. Recent advances in microtagging, otolith microchemistry and stable isotopes are allowing this to be tested (e.g. Gillanders and Kingsford 1996), but progress has been slow due to the complexity and cost of such studies.

Extensive meadows of many seagrass species are restricted to the protected waters of estuaries (see Green and Short 2003), which can be extensively used as juvenile habitats (Sogard 1992), regardless of whether seagrass is present. The concept of the nursery habitat was built around the focus of many studies on estuaries (Beck et al. 2001), leading to a bias towards estuarine systems (e.g. Chesapeake Bay, Jones 2014). Indeed, the meta-analysis by Heck et al. (2003) indicated that warm temperate and tropical seagrasses in the Northern Hemisphere appeared to play a more important nursery role than those in cool temperate and boreal waters (Heck et al. 1989; Heck and Coen 1995). In Australia, seagrass meadows are extensive in sheltered estuarine and marine waters as well as more exposed and open marine waters (Abrantes et al. 2015). In this context, in expansive studies across southern Australia, Edgar and Shaw (1993) and Edgar and Shaw (1995) demonstrated the highly variable use of seagrass as a juvenile habitat for fish. Edgar and Shaw (1995) considered the lower abundances of juvenile fish in seagrass meadows along southern, as compared to eastern Australia, to be related to the focus of studies on estuarine versus marine environments in the respective regions. This was supported by studies of more exposed waters along the west coast of Australia, where seagrasses played a minimal role as a juvenile habitat (Travers and Potter 2002; Hyndes et al. 2003).

Although not mutually exclusive, the dynamics of seagrass as a habitat provides another explanation of the observed variation in the role of seagrasses as juvenile habitats. Many species of seagrass are highly seasonal in above-ground biomass, e.g. Halodule, Zostera and Thalassia (Hemminga and Duarte 2000). The highly ephemeral nature of these seagrasses means that habitat availability is not consistent throughout the year, and therefore, it is perhaps not surprising that fish do not use these habitats throughout their life cycles. For example, the seasonal variability in leaf density of Cymodocea nodosa and Zostera noltii compared to P. oceanica in the Mediterranean Sea appears to result in different fish assemblages in those habitats (Guidetti 2000). However, persistent (or perennial) species such as those of Posidonia provide above-ground biomass throughout the year, allowing fish species to occupy these productive habitats throughout their lifecycles, and therefore resulting in a smaller proportion of fish utilizing seagrasses as juvenile habitats alone (see section below). In southern Australia, the seagrasses Posidonia australis , P. sinuosa and Amphibolis antarctica and A. griffithii form extensive meadows, and studies in these meadows have formed the majority of those concerning seagrass-associated fish in the region (Travers and Potter 2002; Hyndes et al. 2003; Edgar and Shaw 1993, 1995). The perennial nature of these seagrasses would allow resident species to occupy structured habitats throughout their life cycles, and these species characterize those meadows in southern Australia (see below). Conversely, the ephemeral nature of other seagrass species would prevent species that require structure to remain in those habitats when the above-ground biomass is absent in the winter months. Supporting this hypothesis, meadows of Zostera capricorni in estuaries on Australia’s east coast provide a juvenile habitat for a range of species (Smith and Sinerchiab 2004). The seasonality of seagrass meadows would alter the availability of this habitat through the year, which would have presumably had a profound effect on the evolution of life history strategies of fish in coastal waters.

The economically important Sillaginodes punctatus provides a useful cautionary example of generalizations regarding the role of seagrasses as a nursery habitat. Young-of-the-year of this species recruit into Zostera meadows in south-eastern Australia (Robertson 1977; Connolly 1994a, c; Jenkins et al. 1997; 1998), but can also recruit to reef algae (Jenkins and Wheatley 1998) and sandy areas in some locations (Jenkins and Hamer 2001). In comparison, the species appears to recruit only into sheltered sandy areas rather than Posidonia australis meadows in south-western Australia (Hyndes et al. 1997; 1998). The presence of seagrass, therefore, does not appear to be the primary requirement for successful recruitment of this species. Indeed, the higher recruitment of S. punctatus into more sheltered unvegetated habitats where food availability is higher (Jenkins and Hamer 2001), suggests that food availability drives habitat choice of the recruiting juveniles of this species. In support of this conclusion, experiments by Connolly (1994b, c) showed that juvenile abundances were not affected by the removal of Zostera, and the species’ choice for Zostera as a habitat was over-ridden by food availability. However, higher predation on juvenile S. punctatus in unvegetated areas (Hindell et al. 2002), suggests that there is a trade-off between high food availability and predation in the sheltered unvegetated habitats. In addition, proximity of seagrass meadows to the currents and wind conditions that transport larvae into coastal habitats appears to play a major role in the recruitment of this species (Jenkins et al. 2000) and others (Ford et al. 2010) into juvenile habitats in the region.

3.2 Seagrass for Residents

The ability of fish species to complete their life cycles in seagrasses would require the persistence of the habitat in suitable conditions for reproduction to occur. Firstly, as stated above, many studies on seagrass-associated fish have been carried out in estuaries, which are used for extensive periods by the juveniles of many species before returning to the marine environment to spawn (Potter et al. 2015). While not mutually exclusive, the presence of above-ground biomass of particularly persistent species of seagrasses, such as Posidonia spp. in coastal marine waters of southern Australia and the Mediterranean Sea, is likely to provide habitat for juveniles and adults of many seagrass-associated species throughout the year (Bell and Harmelin-Vivien 1982; Edgar and Shaw 1993; Harmelin-Viven 1984; Hyndes et al. 2003). Resident species, therefore, appear to be common in meadows in coastal marine waters, but the proportion of life-long residents may also be related to depth, since P. oceanica meadows in shallow (3 m depth) waters of the Mediterranean Sea contain more juveniles (Francour 1997; Guidetti and Bussotti 1997; 1998) than those in relatively deep (12–18 m) waters (Bell and Harmelin-Vivien 1982; Harmelin-Viven 1984).

Fish assemblages in persistent, marine seagrass meadows are likely to be influenced by the reproductive and early life history strategies of resident species. For example, several resident fish families common to Posidonia meadows, e.g. Syngnathidae, Apogonidae and Monacanthidae (Kendrick and Hyndes 2003; Steffe et al. 1989) and Labridae (Bell and Harmelin-Vivien 1982; Francour 1997) are characterized by parental care such as mouth brooding, giving birth to live young, or depositing demersal eggs in nests (Patzer 2008). The larvae from species with these reproductive strategies may therefore recruit directly and earlier into the natal or nearby seagrass meadows of their parents, and bypass the high potential for mortality in the pelagic phase and other recruitment processes and potential bottlenecks that influence the settlement of pelagic larvae into seagrass meadows.

4 Feeding Ecology of Fish Assemblages

A range of feeding guilds occurs in the fish assemblages in seagrass meadows worldwide: (1) predators of fish and other larger nekton; (2) meso-carnivores feeding mainly on small crustaceans, such as copepods and amphipods, shrimps or molluscs; (3) omnivores that feed on mobile or sessile epifauna, as well as epiphytic algae and/or seagrass; (4) herbivores that feed on epiphytic algae and/or seagrass; and (5) detritivores that consume detritus within the meadow. Fish assemblages in southern Australian seagrass meadows are dominated by meso-carnivores feeding mainly on small crustaceans or molluscs (Edgar and Shaw 1993, 1995; Connolly et al. 2004), which in turn, often feed on epiphytes and detritus (Cook et al. 2011; Ebrahim et al. 2014). These meso-grazers play critical roles in controlling epiphytic algae on seagrass leaves in Australia (Cook et al. 2011; Ebrahim et al. 2014) and elsewhere (Myers and Heck 2013; Whalen et al. 2013), and contribute to secondary production (Lepoint et al. 2000; Smit et al. 2005, 2006; Vizzini and Mazzola 2003; Vizzini et al. 2002).

Evidence for direct grazing on seagrasses comes mainly from tropical systems (Valentine and Duffy 2006). However, it must be recognized that the trophic structure we see in seagrasses today may not reflect those of the past due to the direct and indirect effects of historical over-harvesting of large predators and grazers (Jackson et al. 2001). Apart from supporting dugongs, manatees and green turtles, tropical seagrass meadows can support a diversity of herbivorous fish, with species typically belonging to the families Labridae (Sparisoma clade), Acanthuridea and Siganidae (see Valentine and Duffy 2006 and references therein). Species within these families often move from coral reefs to adjacent seagrass meadows, where they can create “haloes” of unvegetated areas due to high levels of grazing (Randall 1965; Ogden 1977). Indeed, Sparisoma radians has been shown to consume nearly all daily seagrass production in some locations in the Gulf of Mexico (Kirsch et al. 2002). However, species richness and abundance of herbivores decrease with increasing latitude and decreasing water temperature (Floeter et al. 2005, González-Bergonzoni et al. 2012). While the processes leading to this pattern are not well understood (Clements et al. 2009), grazing on temperate seagrasses by fish is considered to be limited.

Sarpa salpa (Sparidae) in the Mediterranean Sea is an exception to the above generalisation. Early studies suggested a mixed diet for this species, comprising seagrass, epiphytes and reef algae (Havelange et al. 1997; Lepoint et al. 2000), with an ontogenetic shift towards seagrass (Verlaque 1990). Early studies also suggested that S. salpa consumed 4–15% of daily production of P. oceanica (Havelange et al. 1997; Velimirov 1984), but more recently, Tomas et al. (2005) found that consumption rates can exceed local production rates of temperate seagrass meadows. In some locations, seagrass biomass was reduced by as much as 50%, while epiphytic load was reduced by about 30% (Tomas et al. 2005). As a consequence, Valls et al. (2012) regarded this species as a keystone modifier (sensu Mills et al. 1993) in P. oceanica meadows.

Direct grazing by fishes on temperate Australian seagrasses appears to be very limited. Some consumption of seagrass has been observed in species of Labridae (from the Odacine clade), Hemirhamphidae, Monacanthidae, Tetraodontidae and Terapontidae, but these species are omnivorous, feeding predominantly on epiphytes and macro-invertebrates, and not seagrasses (Bell et al. 1978; Bell and Harmelin-Vivien 1983; Burchmore et al. 1984; Burkholder et al. 2012; Conacher et al. 1979; Macarthur and Hyndes 2007; Kwak et al. 2015). Despite evidence of some grazing on seagrasses, there is limited evidence that fish significantly influence seagrass production. Through exclusion experiments and observations, Burkholder et al. (2012) inferred that the terapontid Pelates octolineatus was responsible for grazing large portions of the Halodule and Halophila seagrasses in Shark Bay , Western Australia , but this species is omnivorous, feeding largely on epiphytic algae (Edgar and Shaw 1993) and therefore is likely to play a greater role in controlling epiphytes on seagrass leaves. As is true for many omnivorous species, the proportion seagrass in the diets of Monacanthus chinensis (Bell et al. 1978) and Haletta semifasciata (Macarthur and Hyndes 2007) increases with fish size. It may therefore be relevant that several omnivorous species attain relatively large sizes in perennial Posidonia or Amphibolis seagrass meadows (Hyndes et al. 2003) that form extensive meadows in southern Australia. These species are therefore likely to play a considerable role in the removal of particularly epiphytic algae in those meadows. There is limited information on grazing on seagrasses by fishes in tropical Australia, and this is clearly a research gap. Kwak et al. (2015) showed that the juveniles of a range of fish species in seagrass meadows of northern Queensland consumed very little seagrass. However, the parrotfish Leptoscarus vaigiensis is common in tropical Australia (Lim et al. 2016), and has been shown to consume up to 10 times the rate of local seagrass growth another Indo-Pacific region (Unsworth et al. 2007a, b). The species, therefore, has the capacity to remove considerable biomass of seagrass in the tropics. Also, with global ocean warming and the movement of tropical species to higher latitudes (Fig. 18.7), grazing on seagrasses in more temperate regions is predicted to increase (Hyndes et al. 2016).

Fig. 18.7
figure 7

The current (blue) and predicted end-of-century (orange) distributions of seagrasses (top) and herbivores (bottom) along the west coast of Western Australia. See the supplemental material for methods to determine distributions and their shifts. From Hyndes et al. (2016)

Full size image

Grazing on seagrass can vary among seagrass species. For example, Thalassia testudium appears to have a greater resistance to grazing than Halodule wrightii in Florida, USA (Armitage and Fourqurean 2006), and elevated nitrogen content can influence grazing (Goecker et al. 2005). In Western Australia, consumption rates were greater on the low-nutrient, slow-growing Posidonia and Amphibolis species compared to high nutrient , faster growing Halodule and Halophila species (Burkholder et al. 2012). Despite the low nutrient content of Posidonia species, approximately 80% of the organic soluble content of P. oceanica is assimilated during gut passage in S. sarpa (Velimirov 1984). This suggests that the species gains nutrients from seagrass even though seagrass fragments in its hindgut appear to be undigested. Similarly, Monacanthus chinensis and Hyporhamphus melanochir assimilate nutrients from Posidonia leaves (Conacher et al. 1979; Nichols et al. 1986), providing evidence that Posidonia supports the energetic requirements of some (albeit a small number of) fish species in both the Mediterranean Sea and Australia.

5 Connectivity Across Seascapes

A major source of seascape connectivity is the process of larval dispersal from spawning in adult habitat, settlement into different juvenile habitats, and subsequent movement back to the adult habitat (Ford et al. 2010; Fowler and Short 1996; Haywood and Kenyon 2009; Verweij et al. 2008). Indeed, seascape connectivity via propagule (eggs and larvae) dispersal (at a biogeographic scale of hundreds of kms) is considered to critically affect the dynamics of fish communities associated with P. oceanica beds in the Adriatic Sea, in both space and time (Melià et al. 2016). The larvae of fish settling in seagrass habitats are often derived from adults in different, sometimes distant, habitats (Jenkins et al. 2000; Ford et al. 2010; Hyndes et al. 1998). This pattern of connectivity occurs in both tropical and temperate systems, with the adults of tropical species utilising coral reefs or deeper sedimentary habitats as adults (Huijbers et al. 2013, 2015; Mcmahon et al. 2012; Nagelkerken 2009; Verweij et al. 2008), and temperate species occupying coastal and offshore rocky reefs and sedimentary habitats (Ford et al. 2010, Gillanders and Kingsford 1996; Hyndes et al. 1998; Gillanders 2002; Jenkins and Wheatley 1998), and the juveniles occupying seagrass habitats in estuaries and embayments.

Use of a mosaic of juvenile habitats, termed the ‘seascape nursery’ (Nagelkerken et al. 2015), may follow a sequential pattern (Jenkins and Wheatley 1998; Pardieck et al. 1999). For example, Grol et al. (2014) showed that larvae of a coral-reef fish species initially settled into rubble habitat before moving into seagrass and then mangrove habitat with increasing size. With further growth, juvenile fish may gradually move into deeper waters before making the migration offshore (Hyndes et al. 1998). These movements from juvenile to adult habitats can occur over kilometres to 100s of kilometres (Gillanders et al. 2003; Huijbers et al. 2013; Mcmahon et al. 2012). If the adult habitat is a relatively short distance from the juvenile habitat then there may be a movement back and forth between the habitats before residence is taken up in the adult habitat (Huijbers et al. 2015). The timing of the ontogenetic migration from the seagrass nursery habitat to adult habitat may be a trade-off between higher survival in the juvenile habitat versus higher growth rates in the adult habitat (Nagelkerken 2009), or coincide with reproductive maturity and the onset of spawning (Hyndes et al. 1997, 1998).

The movement and migration of fish across a range of habitats, including seagrass meadows, significantly contributes to seascape connectivity (Olds et al. 2012; Grober-Dunsmore et al. 2008, 2009). This movement and migration may be part of the daily ambit of the individual or may have a regular diel or tidal periodicity (Krumme 2009), or may occur in the process of adults forming spawning aggregations (Nemeth 2009). Many studies have focused on fish assemblages associated with coral reefs and their proximity to seagrass habitats (Berkström et al. 2013; Olds et al. 2012; Grober-Dunsmore et al. 2009), with a general consensus that proximity of seagrass meadows to coral reefs has a significant positive effect on the assemblage structure of reef fish (Berkström et al. 2013; Olds et al. 2012). Focusing more on seagrass-associated fish, Unsworth et al. (2008) found that assemblages in seagrass were influenced by proximity to mangroves and reef, with a marked increase in fish abundance and species richness in seagrass meadows near mangroves. These links can be associated with strong diurnal and tidal migrations of fish (Krumme 2009). For example, grunts (Haemulidae) rest on patch reefs during the day but forage in surrounding seagrass meadows at night (Krumme 2009; Nagelkerken et al. 2008), while parrotfish (Scarine in Labridae) forage in seagrass during the day but hide on the reef at night (Krumme 2009). Similarly, tidal migrations of fish can connect subtidal seagrass meadows with intertidal mangrove habitats (Jelbart et al. 2007), subtidal habitats with intertidal seagrass meadows (Robertson 1980), and coral reefs with seagrass meadows (Unsworth et al. 2007a, b).

The sequential movement of fish from juvenile to adult habitats likely provides an important process for transferring nutrients across seascapes, with the series of predator-prey interactions along that path, i.e. trophic relay (sensu Kneib 1997), contributing to a net flow of nutrients to other habitats in coastal or offshore waters (Hyndes et al. 2014). This is highlighted by the estimated export of 7400 t of silver perch Bairdiella chrysoura from Zostera meadows to other coastal habitats in Chesapeake Bay, USA (Sobocinski and Latour 2015). Similarly, the ontogenetic shift of the pinfish Lagodon rhomboides from seagrass meadows to offshore Gulf of Mexico (GOM) is estimated to contribute approximately 25% of the overall potential production in the northeastern GOM (Nelson et al. 2013). Thus, the maintenance of the connectivity of fish assemblages between seagrass and other habitats has significant implications for conservation planning, including habitat restoration, the location of marine protected areas, and the management of fisheries. Planning must include consideration of the mosaic of inter-connected habitats in an area rather than just the characteristics of individual habitats (Berkström et al. 2013; Cheminée et al. 2014; Olds et al. 2012, 2014, 2016). In Australia, few studies have directly examined this process for fish (but see Jelbart et al. 2007), although our understanding of this process has been compiled from a range of studies for some economically important finfish species, e.g. the King George whiting S. punctatus (Hyndes et al. 1998, Jenkins et al. 2000), and invertebrate species, e.g. the Western Rock lobster P. cygnus (Chittleborough and Phillips 1975; Chittleborough 1970; Macarthur et al. 2008).

6 Seagrass Links to Fisheries

6.1 Direct Links

There is strong evidence around the world that numerous fish and invertebrate species that support important fisheries use resources available in seagrass habitats at some stage in their life cycles (Connolly et al. 1999a, b; Jackson et al. 2001; Verweij et al. 2008; Haywood and Kenyon 2009; Mizerek et al. 2011; Jones 2014; Lilley and Unsworth 2014; Seitz et al. 2014; Jackson et al. 2015), although the relative usage seems to vary regionally, which may in part simply reflect the amount of research in a region (Jackson et al. 2001). The usage of seagrass habitat by fisheries species is particularly well documented for Australia (Mcneill et al. 1992; Haywood et al. 1995; Jenkins et al. 1997; Connolly et al. 1999a, b; Curley et al. 2013), where species have been shown to use seagrass meadows as juvenile habitats, foraging habitats, or spawning habitats. For example, seagrass meadows in Australia are used as juvenile habitat by the King George Whiting Sillaginodes punctatus (Jenkins et al. 1996; Robertson 1977; Connolly 1994a) and prawns (Coles et al. 1987, 1993; Loneragan et al. 1998), a foraging habitat for the juveniles and sub-adults of the Western rock lobster P. cygnus (Macarthur et al. 2008, 2011), a foraging habitat for the adults of the rock flathead Platycephalus laevigatus (Klumpp and Nichols 1983b; Jenkins et al. 1997), and a spawning habitat for the Southern Calamari Sepioteuthis australis (Pecl et al. 2006). Some species, such as the grass whiting Haletta semifasciata and the six-spine leatherjacket Meuschenia freycineti are associated with seagrass throughout their entire life cycles (Jenkins et al. 1997). The southern sea garfish, Hyporhamphus melanochir, has an unusual link to seagrass in that the species feeds on seagrass tissue (either directly or as fragments in the water column) during the day and diurnally migrating invertebrates over seagrass beds at night (Robertson and Klumpp 1983; Klumpp and Nichols 1983a). Seagrass meadows are considered to enhance the value of fisheries linked to that habitat in south-eastern Australia by ~$A32,000 ha−1 year−1 (Blandon and Zu Ermgassen 2014). In the gulf waters of South Australia alone, seagrass has been estimated to contribute $A114 M per year to the economy (Blandon and Zu Ermgassen 2014; Mcarthur and Boland 2006). Similarly, the seagrass meadows of Cairns Harbour were estimated in the 1990s to contribute 178 t and $A1.2 M per year to the prawn fishery in tropical Australia (Watson et al. 1993). Based on the link between seagrass and other important fisheries around Australia, its contribution to the economy of Australia will be substantially higher. For example, in Western Australia, the fishery for the western rock lobster, whose juveniles and sub-adults in the southern part of the population forage in seagrass meadows, was valued at ~$A234 M in 2013. However, the quantification of the role of seagrasses in supporting fisheries is less clear in other regions. For example, many papers state that seagrasses, including Posidonia oceanica , have important nursery roles for many fishes, including commercial species in the Mediterranean. However, the quantification of this role is still in its infancy (Guidetti et al. 2009) and there has been no assessment of the related economic benefits for fisheries in that region.

6.2 Indirect Links Through Trophic Relays and Outflow of Organic Matter

Seagrass meadows can export, on average, 0.22 kg C m−2 year−1, mostly in the form of leaves (Heck et al. 2008; Hyndes et al. 2014). This exported detrital material can accumulate in a range of coastal and offshore habitats. For example, accumulations of seagrass and seagrass epiphyte detritus in unvegetated mudflats adjacent to seagrass meadows of subtropical Australia support the food chain to fish, including the economically important whiting, Sillago schomburgkii (Connolly et al. 2005). Similarly, the transfer of seagrass detritus to nearby coral reefs in the region appears to contribute to the food resources of the labrid Pseudolabrus guentheri (Davis et al. 2014). The export of seagrass to more distant habitats can also support fisheries species. In south-western Australia , seagrass leaves (and kelp thalli) provide an important juvenile habitat in the surf zones for a range of economically important fish species (Lenanton 1982; Crawley et al. 2006). Furthermore, prawns in offshore waters in northern Australia (Loneragan et al. 1997), and larvae of the blue grenadier Macruonus novaezelandiae in offshore waters of south-eastern Australia (Thresher et al. 1992), are likely to be supported by transported seagrass detritus. Thus, the export of seagrass detritus to near and distant habitats can, at least partly, support fisheries in other coastal habitats, but direct evidence of this link is still limited.

6.3 Evidence of Links Through Seagrass Loss and Fisheries Production

A strong dependence on seagrass by fisheries should be reflected in declining catches when seagrass is lost, however, the evidence to date is largely equivocal with seagrass loss resulting in fishery declines in some regions but not others (Connolly et al. 1999a, b; Gillanders 2005). One clear example of impacts to fisheries from seagrass loss was the marine wasting disease that caused catastrophic loss of eelgrass, Zostera marina , in the Atlantic basin in the early 1930s and subsequent collapse of the fishery for bay scallop, Argopecten irradians, on the east coast of the United States (Fonseca and Uhrin 2009). In Australia, a 70% decline in eelgrass, Zostera, over 10 years from the early 1970s in Western Port , Victoria, resulted in catch declines of about 40% for species known to be associated with seagrass, but not for other fishery species (Macdonald 1992; Jenkins et al. 1993). King George whiting provides a good example of this link. Prior to the 1970s, catches were variable but showed an increasing trend in three bays in Victoria (Fig. 18.8). However, after the seagrass loss in Western Port, catches declined, whereas they continued to increase in the two other nearby bays (Port Phillip and Corner Inlet) .

Fig. 18.8
figure 8

Historical annual catches of King George whiting from Port Phillip Bay , Western Port and Corner Inlet. Arrow indicates beginning of reported seagrass decline

Full size image

In many cases, seagrass loss has not resulted in declines in fisheries linked to seagrass habitats, but this is likely to depend on the targeted species and the degree of its dependence on seagrass. For example, there was no catastrophic collapse of European fisheries associated with the eelgrass wasting disease in the early 1930s (Gillanders 2005). The equivocal link between seagrass loss and declines in fishery catches is most likely explained by facultative use of seagrass by many species, which can also use other structured habitats (Jenkins and Wheatley 1998; Heck et al. 2003). Therefore, the loss of seagrass may be ameliorated to some extent if alternative structured habitats that provide similar resources are present (Gillanders 2005; Jenkins et al. 2015). However, the level of facultative use of alternative habitats is often unknown or not quantified, making it difficult to determine the effect of seagrass loss and adaptively manage fisheries when seagrass is being lost. But if several structured habitats simultaneously decline, as is occurring in the Mediterranean Sea for shallow Posidonia oceanica meadows due to coastal building (Duarte 2002) and Cystoseira forest due to overgrazing (Gianni et al. 2013), declines of small-scale fisheries are likely.

6.4 Management of Seagrass-Associated Fisheries

Ecosystem-based management of fisheries is a goal now embraced by many countries including Australia (Fletcher et al. 2010; Hobday et al. 2011), and as such, includes consideration of both fishing and external impacts on seagrass (Hobday et al. 2011). Most fisheries management agencies, however, do not have direct responsibility for seagrass habitat, but rather take on an advocacy role to influence the policies of coastal and catchment managers that directly influence impacts on seagrass. Fisheries management agencies can, however, have a direct policy role in the case of fishery methods and practices that have a destructive impact on seagrass meadows (Short and Wyllie-Eciieverria 1996; Airoldi and Beck 2007) and also in the development of aquaculture areas and their associated impacts on seagrass (Delgado et al. 1997). Where seagrass is lost, fisheries managers may advocate for seagrass restoration to support and improve fisheries (Tanner et al. 2014). In the Mediterranean, the “Habitat Directive” developed by the EU provides the framework to member states to protect Posidonia oceanica , which is a habitat forming species of ‘priority importance’ for protection. Management occurs via the creation of Natura 2000 sites or other types of marine protected areas (MPAs) at national or trans-boundary levels. Also, the adoption of the Marine Strategy Framework Directive and Common Fishery Policy by EU countries drives an effort to adopt an ensemble of protection/management measures aimed at protecting the marine biodiversity and ecosystem functioning, while simultaneously supporting fisheries.

7 Gaps and Future Directions

Since the 1990s, studies on fish assemblages in seagrass meadows have expanded in terms of both the geographic range and the seagrass species forming the habitat. This has allowed us to broaden our understanding of the role of seagrass habitats to fish assemblages, but there are still a number of gaps that need addressing: (i) how seagrass loss will impact fish assemblages linked to seagrass, particularly as juveniles and their contribution to fisheries production; (ii) how habitat fragmentation versus habitat loss affects fish assemblages; (iii) how fish species interact with seagrass meadows in tropical Australia; and (iv) how temperate seagrass habitats respond to warming sea temperatures with global warming.

The 29% global areal loss of seagrass meadows since the late 1800s, and more rapid losses in the last two decades (Waycott et al. 2009), causes major concerns regarding the various ecosystem services they provide, including its importance as a habitat for fish assemblages and particularly fisheries species. Interpretation of the potential effects of seagrass loss requires a thorough understanding of the role of not only seagrass habitat in the life history of the fish, but also the potential role of alternate habitats. For many seagrass associated fish species, assumptions about the importance of seagrass habitat have been based only on sampling in seagrass habitat rather than a broader suite of potential habitats. While seagrass meadows have been estimated to enhance the value of fisheries in south-eastern Australia by ~$A32,000 ha−1 year−1 (Blandon and Zu Ermgassen 2014), there is limited empirical data linking seagrass-associated fish to fisheries production usually in more offshore waters, and limited data on the relative importance of different juvenile habitats in their contribution to the adult population (Beck et al. 2001; Dahlgren et al. 2006). More studies using approaches, such as otolith microchemistry and stable isotopes (e.g. (Gillanders and Kingsford 1996), are therefore needed to quantify the contribution of juveniles from seagrass meadows to adult spawning habitats for particularly fishery species. This may be more difficult to assess when fisheries in offshore regions are influenced through trophic relays of fauna originating from seagrass meadows, thereby diluting the contribution of seagrass to the offshore production. However, this export from seagrass meadows can be an important means of carbon transfer from tropical systems (Hyndes et al. 2014), and deserves attention.

While understanding fragmentation of seagrass and its effect on associated fishes is important, novel approaches are needed to separate the effects of fragmentation from seagrass loss. Seagrass loss and fragmentation tend to occur together, and therefore, the two processes tend to be confounded (Fahrig 2003). Furthermore, fragmentation is a process rather than a state, and most studies actually consider habitat configuration rather than fragmentation (Boström et al. 2006, 2011). Moreover, fragmentation effects can be further confounded by differences in within-patch structural characteristics of seagrass. Some of these issues can be addressed experimentally using ASUs (Macreadie et al. 2009), however, the scale of the experiments is inevitably small compared with seascape scales. Novel research approaches are therefore required to determine the relative importance of habitat fragmentation and habitat loss to seagrass associated fishes at seascape scales.

With less than 10% of the published studies on fishes in Australian seagrass meadows being undertaken in the tropics, there is a clear need to gain more expansive data on the role of seagrasses for fish assemblages in this climatic region. For example, our understanding of grazing by fishes on seagrasses comes mainly from tropical systems (Valentine and Duffy 2006), but predominantly from Caribbean/Gulf of Mexico region and more recently, Indonesia and Africa. Few papers (Kwak et al. 2015) have examined trophic interactions between fish and seagrasses in tropical Australia. Since species richness and abundance of herbivores is higher in lower latitudes (Floeter et al. 2005; González-Bergonzoni et al. 2012), we would expect relatively high grazing on tropical seagrasses in the region. However, this pattern needs to be confirmed, and its importance is becoming even more pronounced with global warming and the poleward shifts in the distribution of tropical species in the region.

Globally, warming waters have also increased the presence of tropical species in the harvest of finfish from temperate waters (Cheung et al. 2013), and lush kelp forests have dramatically shifted to barrens in some regions such as Australia, Mediterranean and Japan due to the poleward shift in herbivores (Verges et al. 2014). Less is known about this process known as “tropicalisation” and its impacts on temperate seagrasses. Tropical herbivores have, however, moved into seagrass meadows in temperate Gulf of Mexico, where it has been predicted that with continued immigration of those herbivores, seagrass meadows could be grazed to the “height of closely mowed lawn” (Heck et al. 2015). Hyndes et al. (2016) predicted that the influx of tropical herbivores into the temperate seagrass meadows of south-western Australia will have major consequences on the ecosystem services they provide, including a shift in food web structure from one that is detrital based to one that is mostly based on direct consumption. This would lead to reductions in the abundances of seagrass-associated fauna through reduced habitat structure and food availability (Hyndes et al. 2016). With future projections of sea temperature rises, this is clearly an issue that needs research attention in Australia and elsewhere.