The influence of seascape spatial features on the fish and macroinvertebrates in seagrass beds
Jane E. Jelbart
Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy at the University of Western Sydney, Hawkesbury
August 2004
© J.E. Jelbart 2004
Acknowledgements
I would like to thank my supervisor Dr Pauline Ross for her support and guidance. Her diligence and experimental design and research skills were a great contribution to the thesis. She always encouraged me to aim high with this project and ensured I had the support to do so.
I am also very grateful to Rod Connolly from Griffith University for his ideas and guidance on experimental design and for discussions about the meaning of results. His expertise on fish ecology, seagrass habitat and estuarine research was invaluable.
I am indebted to the team of people that assisted me with the fieldwork component of the project including Robin Janus, Caroline Herlihy, Karen Stephenson, Marcus Scnell, Chris Baker, Sonia Claus, Andrew Hayes, Brendan Haine, Paul Thomas, Vimla Rao, Nicholas Gay, Karen O'Neill and Bethany Alexander.
Financial support made this project possible from the University of Western Sydney (UWS) in the form of a Hawkesbury Postgraduate Award (3 year stipend); The NSW Royal Zoological Society (Ethel Mary Read student research grant); Gosford City Local Council (student ecological research grant); the Centre for Landscape Ecology Management UWS (9 week stipend); and the Research Office UWS, which financed my attendance to two international conferences.
Special mention must be made of the support, camaraderie and great assistance provided by Sonia Claus without whom this project would have been so much harder.
Finally special thanks go to my parents (Elizabeth and Ray), Robin Janus and Nicholas Gay.
Statement of Authentication
The work presented in this thesis is original except as acknowledged in the text. I hereby declare that I have not submitted this material, either in whole or in part, for a degree at this or any other institution.
Jane E. Jelbart
Table of Contents
Abstract 1 1 General Introduction 3 1.1 Landscape / seascape ecology 3 1.2 Seagrass habitat in estuaries 4 1.3 Fragmentation of seagrass habitat 5 1.4 Importance of seagrass for estuarine fauna 6 1.4.1 Importance of seagrass for fisheries 7 1.5 Structural features of seagrass habitat that promote the abundance and diversity of estuarine fauna 8 1.5.1 Small scale structural features 8 1.5.2 Influence of size and shape of seagrass bed 9 1.5.3 Are there edge effects in beds of seagrass for fish? 11 1.5.4 Habitat heterogeneity of a seagrass bed 12 1.5.5 Position of seagrass bed within the estuary 14 1.5.6 Proximity to other habitats, e.g. mangroves 16 1.6 Marine and Estuarine Protected Areas 17 1.7 This study 18 2 General Methods 19 2.1 Study locations 19 2.2 Identification and mapping of the estuarine habitats 19 2.2.1 Description of Zostera capricorni 23 2.2.2 The shoot density and blade length of Z. capricorni 23 2.3 Fish survey techniques 26 2.3.1 Identification of fish and macroinvertebrates 27 2.4 Pilot study: Influence of tidal states 27 2.4.1 Methods 28 2.4.2 Results and Discussion 30 3 A test of the species area relationship and landscape ecology theories 39 3.1 Introduction 39 3.1.1 The species-area relationship 39 3.1.2 Location of the seagrass bed within the seascape 41
3.1.3 Study Aims 41 3.2 Methods 42 3.2.1 Study area and description 42 3.2.2 Seagrass bed categories 42 3.2.3 Estimation of habitat characters 43 3.2.4 Fish Survey 46 3.2.5 Data analysis 47 3.3 Results 52 3.3.1 Habitat heterogeneity 52 3.3.2 Size of the seagrass bed 56 3.3.3 The perimeter to area ratio of a seagrass bed 56 3.3.4 Testing for the passive sampling hypothesis 56 3.3.5 The distance of the seagrass bed from the estuary mouth 57 3.4 Discussion 64 3.4.1 Comparison of day & night sampling 64 3.4.2 Heterogeneity of seagrass and the assemblage of fish 64 3.4.3 Small beds have greater fish densities and species richness 65 3.4.4 Passive sampling hypothesis 67 3.4.5 Location of the bed influences the fish assemblage 68 3.4.6 Implications for the species-area relationship and marine conservation 69 4 Do the edges of seagrass beds influence small fish and macroinvertebrates? 70 4.1 Introduction 70 4.2 Methods 72 4.2.1 Selection of the seagrass beds 72 4.2.2 Defining the edge, inner and central regions of the beds 74 4.2.3 Fish and macroinvertebrate survey 74 4.2.4 Data analysis 76 4.3 Results 77 4.4 Discussion 89 4.5 Conclusion 92 5 The influences of seagrass patch area, perimeter length and perimeter to area ratio on small fish and macroinvertebrates 93
5.1 Introduction 93 5.2 Methods 96 5.2.1 Construction of the artificial seagrass units. 96 5.2.2 Size, shape and arrangement of the artificial seagrass units. 96 5.2.3 Sampling of fauna 97 5.2.4 Data analysis 99 5.3 Results 102 5.4 Discussion 103 5.5 Conclusion 117 6 Fish assemblages in seagrass beds can be correlated with proximity of mangrove 118 6.1 Introduction 118 6.2 Methods 120 6.2.1 Study areas and descriptions 120 6.2.2 Experimental design 120 6.2.3 Fish Survey 123 6.2.4 Univariate data analysis 124 6.2.5 Multivariate data analysis 125 6.3 Results 126 6.3.1 Univariate analysis: abundance and density of fishes 126 6.3.2 Multivariate analysis: composition and abundance of fish assemblages 130 6.4 Discussion 138 6.5 Conclusion 141 7 General Discussion 142 7.1 Seagrass beds contain small fish and macroinvertebrates142 7.2 Fish assemblages in seagrass beds support landscape ecology theories 142 7.2.1 The spatial structure (i.e. size and shape) of seagrass beds 143 7.2.2 The position of a seagrass bed within the estuary 144 7.2.3 The proximity of seagrass beds to mangrove forests 145 7.2.4 The patchiness or heterogeneity of Z. capricorni beds 146 7.2.5 Edge effects in seagrass beds 147 7.2.6 Self-similarity across multiple scales 148 7.2.7 Seascape approach required 148
7.3 Patch size and the species-area relationship 149 7.4 Models to explain the greater densities of small fish in small compared to larger seagrass beds 151 7.4.1 The "settle and stay" model 151 7.4.2 The 'grain' or scale of response of small fish 152 7.4.3 The absence of large predators in small seagrass beds 152 7.4.4 Life-history strategies (r and K selection) 153 7.4.5 The source or sink model 154 7.5 Reasons for correlations of fish and macroinvertebrate in the inners of the small beds 154 7.6 No change in the abundances of macroinvertebrates between small and large seagrass beds 157 7.7 Implications for seagrass regeneration projects 157 7.8 Implications for fragmentation of seagrass beds 158 7.9 Conservation and protection of seagrass beds for estuarine fauna 159 7.10 Final Comment 161 8 References 162
List of Figures
Figure 2.1. A map of Broken Bay (-33.6(S, 151.3(E) showing the Pittwater estuary and the Brisbane Water estuary and its location within Australia. 20 Photo 2.1. The southern end of the Pittwater showing the extensive recreational use of the estuary (Geoscience Australia 2001). 21 Photo 2.2. The western side of the Pittwater estuary showing a Z. capricorni seagrass bed in the foreground and A. marina mangroves. 21 Photo 2.3. A picture of a small and patchy seagrass bed in the Pittwater estuary. 22 Photo 2.4. A photo of a Z. capricorni bed and A. Marina mangroves in the Brisbane Waters estuary. 22 Figure 2.2. The number of seagrass shoots measured in accumulated quadrats in two types of seagrass beds (patchy or uniform) in the Pittwater estuary. 24 Photo 2.5. The small seine net (8 x 2m, 1mm mesh) used for the surveys in Chapters 2,3,4 and 6. 25 Figure 2.3. The length (cm) of seagrass blades measured from sampling seagrass shoots in one quadrat in the Pittwater estuary, January 2000. 25 Figure 2.4. The average number of fish species and fish individuals collected during the day and night sampling from each haul of the net in the pilot study (Pittwater 2000). 32 Figure 2.5. The average number of fish species and fish individuals collected during the three tidal states (low, mid & high) from day and night sampling in the pilot study (Pittwater 2000). 33 Figure 2.6. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the day. The results are shown for all water depths sampled (from 22 to 140 cm) 34 Figure 2.7. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the day. The results are shown for water depths from 30 to 100 cm in the pilot study (Pittwater 2000). 35 Figure 2.8. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the night. The results are shown for water depths from 30 to 100 cm in the pilot study 36
Figure 2.9. The two-dimensional configurations for MDS ordinations of the fish assemblages collected in the pilot study (Pittwater 2000). 37 Figure 2.10. The accumulation of fish species during the day and night sampling in the Careel bay seagrass bed and Barrenjoey beach seagrass bed in the pilot study (Pittwater 2000). 38 Figure 3.1. A map of Australia showing the location of the Pittwater estuary. The Z. capricorni beds are shown in their relative size and numbered according to the scheme in table 1. 44 Figure 3.2a. The design for the ANOVA comparing; day and night sampling (level 1, fixed orthogonal), seagrass beds close and far from the estuary mouth (level 2, fixed and orthogonal), and seagrass beds in three different size categories (level 3, fixed orthogonal) with seagrass beds a random factor nested in levels 2 and 3. 49 Figure 3.2b. The design for the ANOVA comparing; day and night sampling (level 1, fixed orthogonal), seagrass beds in three different size categories (level 2, fixed, orthogonal), heterogeneity of seagrass beds (level 3, fixed orthogonal) with seagrass beds a random factor nested in levels 2 and 3. 49 Figure 3.3. The number of fish species and fish individuals per net in the small, medium and large seagrass beds during the day and night sampling in Pittwater, spring 2000. 53 Figure 3.4. Two-dimensional configurations for MDS ordinations (square root transformation) of the composition and abundance of fish in seagrass beds in Pittwater, spring 2000. 60 Figure 3.5. Plot of the species richness and total number of fish individuals against the area of each seagrass bed from the day and night sampling in Pittwater, spring 2000. 61 Figure 3.6. Plots of the mean number of fish species and fish individuals per net against the perimeter area ratio of each seagrass bed during the day and night sampling, in Pittwater, spring 2000. 62 Figure 3.7 The species accumulation curves for day and night sampling in small, medium and large seagrass beds in Pittwater, spring 2000. 63 Figure 4.1. A map of the Pittwater in Broken Bay NSW and the Zostera capricorni beds surveyed (blue = small, green = large) in autumn and spring 2001. 73 Figure 4.2. A diagram representing the regions (edge, inner and central) of the seagrass beds sampled in autumn and spring in the Pittwater, 2001. 75
Figure 4.3. A diagram representing the experimental design of the study 76 Figure 4.4. The mean abundance (±S.E.) of seagrass fauna in small and large seagrass beds during autumn and spring 2001. 79 Figure 4.5. The mean abundance (±S.E.) of seagrass fauna in the different regions of small and large seagrass beds during autumn and spring in Pittwater 2001. 80 Figure 4.6. The number per net of the eastern-striped trumpeter Pelates sexlineatus in the edge, inner and central regions of all seagrass beds (large and small) in Pittwater, autumn and spring 2001. 84 Figure 4.7. Regression plots between A) the number of fish species and macroinvertebrate individuals in each net taken from large beds in autumn; B) the number of fish and macroinvertebrate individuals in each net taken from the small seagrass beds in both autumn and spring; C) the number of fish and macroinvertebrate individuals in each net taken from the inner regions of the small seagrass beds in both autumn and spring. 87 Figure 4.8. The two-dimensional configurations for MDS ordinations of the composition and abundance of fish and mobile invertebrates in seagrass beds. The replicates are categorized into samples taken; A) during autumn and spring, B) from small and large beds and C) from different regions (edge inner and central regions) of the seagrass beds in the Pittwater 2001. 88 Figure 5.1. The area, perimeter length and shape of the three artificial seagrass patch units. 95 Figure 5.2. Locations of the study sites 1 & 2 in the Pittwater 2002. 100 Figure 5.3. The arrangement of the artificial seagrass units at location one and location two in the Pittwater, 2002. 101 Figure 5.4. The mean number or densities of fish (± S.E.) collected from the three artificial patch designs (n = 4) in the Pittwater, 2002. 109 Figure 5.5. The mean number or densities of macroinvertebrates (± S.E.) collected from the three patch unit designs in the Pittwater, 2002. 110 Figure 5.6. The plots of the significant regressions comparing: a) The abundance of fish species and macroinvertebrate individuals in the units with a shorter perimeter length (small rectangles and large squares). b) The density of fish species and macroinvertebrate individuals in all units. c) The density of fish species and macroinvertebrate individuals in the units with a shorter perimeter length. 111
Figure 5.7. The plots of the significant regressions comparing the abundance of some individual fish species with the amount of epiphytic algae that was removed from each artificial seagrass unit 112 Figure 6.1. A map of Broken Bay with the estuaries Brisbane Water and the Pittwater. 122 Figure 6.2. The mean number per net of fish species, fish individuals, juvenile fish species and juvenile fish individuals in seagrass beds close and far from mangroves during the day and night in the preliminary study 127 Figure 6.3. The regressions of; the densities of mangrove fish; the species richness of mangrove fish and; the species richness of non-mangrove fish, with the distance of each seagrass bed in study 1 (Pittwater 2002) from the mangrove forests. 136 Figure 6.4. Two-dimensional configurations for MDS ordinations of the composition and abundance of fish in seagrass beds in Pittwater 2000, Pittwater 2002, and Brisbane Waters 2003. 137
List of Tables
Table 2.1. The mean number of fish per net haul collected from the pilot study (Pittwater 2000) during the day and night sampling over three tidal states (low, mid and high). 29 Table 2.2. The results of the analysis of similarity (two way crossed) comparing the fish assemblages; at day and night, during three tidal states and in two seagrass beds for the pilot study (Pittwater 2000). 30 Table 3.1. A description of each Z.capricorni bed surveyed in Pittwater spring, 2000 and the number allocation for figure 3.1. 45 Table 3.2. The regression results comparing the total numbers of fish species or individuals with the measures of seagrass (shoot density, blade length percentage cover) and sand in each bed from day and night sampling in Pittwater 2000. 45 Table 3.3. The fish collected from beds of Z. capricorni in Pittwater, spring 2000 using four drags of a seine net during the day and night. 54 Table 3.4. The analysis of variance comparing the number of fish species and the number of fish individuals per net collected from seagrass beds during different times (day and night), from beds of different heterogeneity (patchy or uniform) and size (small, medium or large). 55 Table 3.5. The summary of analyses of variance results run separately on the most abundant fish at different times (day and night), in seagrass beds of different heterogeneity (patchy and uniform) and size (small, medium and large). 55 Table 3.6. A two way crossed analysis of similarities comparing fish assemblages in seagrass beds with time of sampling as the first factor, and the second factor being either; heterogeneity, size of the seagrass bed, or distance from the estuary mouth 58 Table 3.7. The analysis of variance comparing the number of fish species and fish individuals per net from seagrass at different times (day and night), from beds of varying distance from the estuary mouth (close or far) and size (small, medium or large). 59 Table 3.8. The summary of the ANOVA comparing the most abundant fish in beds close to and far from the estuary mouth in Pittwater, 2000. 59
Table 4.1. The total abundance of fish species in the edges, inner and central regions of small (n=3) and large (n=3) seagrass beds in Pittwater, autumn and spring 2001. 81 Table 4.2. The total abundance of macroinvertebrates groups in the edges, inner and central regions of small (n=3) and large (n=3) seagrass beds in Pittwater, autumn and spring 2001. 82 Table 4.3. An analysis of variance (ANOVA) comparing the number per net of fish species and fish individuals in seagrass beds during different sampling times (autumn and spring), from beds of different size (small or large) and from different regions within the beds (edge or inner) in Pittwater 2001. 82 Table 4.4. An analysis of variance (ANOVA) comparing the number per net of fish species, fish individuals and invertebrate individuals collected from the large seagrass beds during different sampling times (autumn and spring), and the different regions in Pittwater 2001. 83 Table 4.5. An ANOVA comparing the densities of macroinvertebrate individuals collected from seagrass beds during different times (autumn or spring), of different size (small or large) and from different regions within the beds (edge or inner) in Pittwater 2001. 83 Table 4.6. Summary of the ANOVA comparing the abundances of single species of fish or macroinvertebrate groups from seagrass beds during different times (autumn and spring), from beds of different size and from different regions within the beds in Pittwater. 85 Table 4.7. The regression values (R2) calculated from comparing the abundance of fish individuals or species with the abundance of macroinvertebrates in each haul of the net in Pittwater, 2001. 86 Table 5.1. The fauna collected from the artificial seagrass units from four replicates of each design (large rectangle, large square and small rectangle) in the Pittwater, January 2002. 104 Table 5.2. An analysis of variance comparing the abundance and density of fish species and fish individuals from two locations in the Pittwater estuary, Jan. 2002 and from three different artificial patch unit designs. 105 Table 5.3. An analysis of variance comparing the abundance and density of macroinvertebrates and decapods from two locations in the Pittwater estuary, 2002 and from three different artificial patch unit designs. 106
Table 5.4. The regression values for comparing the abundances or densities of the fish and mobile macroinvertebrates collected from the artificial seagrass patch units in Pittwater, Jan. 2002. 107 Table 5.5. Regression values for comparing the abundance of fauna with the amount of epiphytic algae that was removed from each artificial seagrass patch unit in Pittwater, Jan. 2002. 108 Table 6.1. The densities of fish per 10m2 in seagrass beds close and far from mangrove forests in the preliminary study (Pittwater 2000), study one (Pittwater 2002) and two (Brisbane Water 2003). 128 Table 6.2. The total abundance of fish collected from four mangrove forests in study one (Pittwater 2002) using a seine net and fish traps. 129 Table 6.3. The analysis of variance comparing the densities (numbers per net) of fish species and individuals during different sampling times (day and night); and in seagrass beds adjacent to mangrove with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). 129 Table 6.4. The analysis of variance comparing the densities of juvenile fish individuals and juvenile fish species during different sampling times (day and night); and in seagrass beds adjacent to mangrove with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). 131 Table 6.5. The summary of an analysis of variance comparing individual fish species collected during different times (day and night) and from seagrass beds adjacent to mangrove forest with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). 132 Table 6.6. Regressions of the abundance of different groups of fish and the dominant single species of fish with the distance of the seagrass bed from the mangrove forests for study one (Pittwater 2002). 133 Table 6.7. Regressions of the abundance of fish species, individual fish and the dominant single species with the distance of the seagrass bed from the mangrove forests for study two (Brisbane Waters 2003). 134 Table 6.8. The results of the analysis of similarity (two way crossed) for all studies. 134 Table 6.9. Simper (similarity percentages) results for the three studies selecting the most discriminating species. 135
Abstract
Abstract
Seagrass beds of Zostera capricorni are an integral part of the estuarine landscape along the east coast of Australia forming important habitats for juvenile fish and macroinvertebrates. Seagrass beds can vary in their spatial structural such as their size, shape and patchiness of seagrass cover. They can also be located within the estuarine landscape context such as their proximity to other habitats or their location within the estuary. The influence or correlation of these landscape or seascape spatial features of seagrass beds on the assemblages of seagrass fauna (fish and macroinvertebrates) was tested in this thesis.
It was found that the spatial structure of seagrass beds (size and shape), their patchiness of the seagrass cover and location within the estuary (close or far from estuary mouth) were correlated with the assemblages of fish within seagrass beds. In particular it was demonstrated that there were greater densities of small fish species in the small compared to the large beds of Z. capricorni. This occurred regardless of the placement of the seagrass bed within the estuary context, its proximity to other habitats or patchiness of cover. Further experimentation using artificial seagrass patches demonstrated that this effect of patch size was independent of the perimeter length or perimeter to area ratio of the seagrass beds. The experimental work demonstrated that the small and large patches received similar total abundances of fish and macroinvertebrates species and individuals. In the smaller seagrass patches, however, the fish were required to concentrate in greater densities than what was found in the larger patches i.e. fish per unit area.
It was hypothesised that the greater density of small fish species in small seagrass beds could be attributed to the greater proportion of edge habitats in small beds i.e. edge-mediated effects. However, the number of fish species per net haul in edges and inner regions of small and large seagrass beds were measured and found not to be different. There was, however, a positive correlation between fish and macroinvertebrate abundances in the inner regions of the small seagrass beds. This correlation was not found, however, in the edge regions of the
1 Abstract small seagrass beds nor in any region of the large seagrass beds. A positive correlation was also found between the densities of fish species and macroinvertebrate individuals in the artificial seagrass units that contained a short perimeter length. This correlation was not evident in the artificial seagrass patch units with a long perimeter length. These results indicate that the relationship between fauna (fish and macroinvertebrates) may change around the edges of seagrass beds.
Z. capricorni beds and Avicennia marina mangroves forests are commonly found adjacent to each other in estuaries of temperate Australia. The fauna in seagrass beds that were of varying distance from mangrove forests were sampled in two estuaries, over three periods. The proximity of mangrove forests was correlated with the fish assemblages in seagrass beds. In particular, those fish species that use seagrass beds and mangrove forests were more abundant in seagrass beds closer to mangrove forests than beds further away. This confirms that the connectivity of habitats (seagrass beds and mangrove forests) within the seascape can contribute to the faunal assemblages of those habitats.
The outcomes of this research suggest that to conserve the small fish species within an estuary, it is essential to protect even the small and patchy seagrass beds. A network of seagrass beds from all regions of the estuary is also required and the adjacent mangrove forests must be included.
2 General Introduction
1 General Introduction
1.1 Landscape / seascape ecology
Landscape ecology investigates the dynamics of spatial and temporal patterns of habitats within the greater landscape and attempts to relate these patterns with ecological processes (Risser et al. 1984; Turner 1989; Robbins & Bell 1994; Mazerolle & Villard 1999). Temporal and spatial changes are considered a part of the landscape mosaic and not as aberrations or ‘noise’. Currently, most studies in landscape ecology investigate terrestrial fauna, especially birds and insects (see review by Mazerolle & Villard 1999). The body of marine and estuarine research is, however, growing (see Bell & Hicks 1991; Robbins & Bell 1994; Irlandi 1994; Irlandi et al. 1995; Irlandi & Crawford 1997; Fonseca & Bell 1998; Bell et al. 1999; Brooks & Bell 2001) and the term seascape has been put forth to represent the uniqueness of the marine/estuarine system (Bartlett & Carter 1991).
Seagrass beds are considered appropriate habitats to investigate questions based in landscape / seascape ecology incorporating spatial scale, spatial patterns and habitat size. This is because seagrass beds are distributed within an estuary over different spatial scales, while they maintain a relative structural homogeneity (Robbins & Bell 1994; Turner et al. 1999). It is important to link the seascape patterns within estuaries (such as seagrass- mangrove mosaics) with the ecological processes (such as recruitment, competition, predation) to determine the factors influencing the faunal communities of estuaries. Furthermore, it is essential for marine conservation to estimate how these processes vary over different spatial scales (Turner et al. 1999). Seagrass beds occur in meadows that can extend over kilometre-wide areas (i.e. historically defined landscape) and it is at this scale of patchiness that marine studies are relatively scarce (Robbins and Bell 1994).
Research in landscape ecology measures landscape structure, function and change (Robbins & Bell 1994). The principal aim of this current study was to
3 General Introduction examine correlations between the seascape structural patterns and the faunal assemblages of a seagrass beds. This would supplement the growing body of research that has investigated the links between seascape patterns and ecological processes such as predation and competition (see review by Parrish 1989; Irlandi 1994; Micheli & Peterson 1999; Nagelkerken et al. 2001).
Seascape structural patterns of seagrass within an estuary can refer to the spatial structure (size and shape) of the seagrass beds, their position within an estuary, their interconnectedness with other habitats and their heterogeneity or patchiness.
1.2 Seagrass habitat in estuaries
Seagrasses are a prominent feature of the tropical and temperate coastline of Australia. They are specialized marine flowering plants (angiosperms) in soft sediments in estuaries and are not true grasses (Poaceae) (Walker et al. 1999).
Seagrasses are anchored to sediments by a system of rhizomes and roots that stabilize the fine sediment particles (Scoffin 1970; Short & Wyllie- Echeverria 1996) and create clear water conditions in areas that would otherwise be often turbid. They reduce the water movement (Fonseca et al. 1982) and provide a stable surface for epiphytes (Harlin 1975). Seagrasses support complex food webs by way of their physical structure and primary production. They are the basis of an important detrital food chain (Harrison & Mann 1975; Short & Wyllie-Echeverria 1996) and provide an important role in the trapping and recycling of nutrients (Hemminga et al. 1991). The leaves, roots and rhizomes provide high structural complexity (compared to bare sediments) and provide spatial niches for seagrass fauna (Heck & Wetstone 1977; Knowles and Bell 1998). In fact they rank with mangroves and coral reefs as some of the most productive coastal habitats (Short & Wyllie-Echeverria 1996) and provide the basis of a unique coastal ecosystem. Seagrasses support a dense and diverse epifaunal assemblage containing numerous fish and invertebrates (Petersen 1918; Heck & Orth 1980; Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984).
4 General Introduction
1.3 Fragmentation of seagrass habitat
Seagrasses are fragmented naturally by wave action, currents and storm events and bottom feeding animals (Townsend & Fonseca 1998) into patches or beds that range from a scale of metres to thousands of metres (Orth & Moore 1986). Coastal regions, however, are among the most rapidly urbanising places on earth (Crooks & Turner 1999; Ehrenfeld 2000) and this has resulted in unprecedented habitat loss. The destruction of seagrass habitat is a worldwide phenomenon (Walker & McComb 1992) and seagrasses are commonly targeted for preservation in marine protected areas or reserves (Hovel 2003). This has prompted a body of research investigating the influence of seagrass fragmentation on seagrass fauna; in particular the relationships between seagrass patch size, shape, and complexity with seagrass fauna (Bell & Hicks 1991; McNeill & Fairweather 1993; Irlandi 1997; Eggleston et al. 1998; 1999; Frost et al. 1999; Irlandi et al. 1999; Bologna & Heck 2000; Bowden et al. 2001; Hovel & Lipcius 2001; 2002; Hovel 2003).
Seagrass habitat is dynamic on a spatial and a temporal scale (den Hartog 1970) because they undergo seasonal fluctuations due to variable growth conditions (Ramage & Schiel 1999). They can also experience sudden declines from storm events (Clarke & Kirkman 1989), epidemic diseases (Olesen & Sand-Jensen 1994a,b), pollution (Cambridge & McComb 1984), erosion (Kirkman 1978), propeller scaring and vessel groundings (Sargent et al. 1995), sediment and nutrient loading (Kirkman 1978; Orth & Moore 1983; West 1983; King & Hodgson 1986; Dennison et al. 1993), land reclamation and changes in land use (Kemp et al. 1983; Larkum & West 1990).
The most widespread and common cause of seagrass decline is from the reduction of light availability (Walker & McComb 1992) because seagrasses have high minimum light requirements compared to other plants (Dennison et al. 1993). This reduction in light availability occurs by numerous mechanisms including increased nutrients leading to a proliferation of algae
5 General Introduction and increased turbidity from suspended sediments (Walker & McComb 1992). These events can be chronic and long term or one-off pulse events.
The fragmentation of seagrass habitat is expected to have consequences for estuarine fauna and so therefore numerous studies have investigated the effects of reduced patch size and increased edges on species abundance and interactions (Bell & Hicks 1991; McNeill & Fairweather 1993; Irlandi 1997; Eggleston et al. 1998; 1999; Frost et al. 1999; Irlandi et al. 1995, 1999; Bologna & Heck 2000; Bell et al. 2001; Bowden et al. 2001; Hovel & Lipcius 2001; 2002).
1.4 Importance of seagrass for estuarine fauna
Seagrasses alter benthic habitats into a unique ecosystem that produces organics matter and assimilates energy (Walker et al. 1999). The direct grazing of the seagrass has been thought to be unimportant for productivity in temperate Australia (Walker et al. 1999). Dietary studies have shown only a low proportion of the fish and invertebrates in seagrass actually consume large quantities of the grass (Klump et al. 1989) including the gar fish Hyporhamphus melanochir, the leatherjackets Meuschenia freycineti, Monacanthus chinensis, Meuschenia trachylepsis (Bell et al. 1978; Robertson and Klumpp 1983; Edgar & Shaw 1995a) and the crab Nectocarcinus integrifrons (Klumpp & Nichols 1983a; Edgar 1996). Instead it is thought that they provide shelter and habitat structure for the numerous fish and invertebrates, concurrent with food availability (Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984).
Seagrass beds are important habitats for the fauna in estuaries, especially for small inconspicuous fish, and the juveniles of larger fish and macroinvertebrate fauna (Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984; Jackson et al. 2001). The diversity of fish has been found to be higher in seagrass than in the surrounding unvegetated habitats (multiple seagrass species, Bell & Pollard 1989; Zostera, Ferrell & Bell 1991; Zostera, Connolly 1994; multiple species, Edgar & Shaw 1995a; Zostera, Gray et al. 1996; Posidonia, Jenkins et al. 1996; Heterozostera, Jenkins et al. 1997a; Zostera marina, Jackson et al. 2002) although this is
6 General Introduction not necessarily true for fish abundances or all sampling times (Zostera, Ferrell & Bell 1991; Posidonia, Jenkins et al. 1996; Zostera, Jenkins et al. 1997a; Zostera marina, Jackson et al. 2002) or spatial scales (multiple seagrass species, Edgar & Shaw 1995c). Furthermore the faunal assemblages between seagrass and unvegetated habitats also differ (Bell & Pollard 1989; Ferrell & Bell 1991; Jenkins & Wheatley 1998).
1.4.1 Importance of seagrass for fisheries
Many commercial species of fish and decapods are associated with seagrass during some stage of their life cycle (reviewed in Connolly et al. 1999). In the temperate east coast region of Australia, some of these species include (but are not limited to); juvenile tiger prawns Penaeus esculentus and P. semisulcatus, eastern king prawns P. plebejus, greasyback prawns Metapenaeus bennettae, school prawns M. macleayi, King George whiting Sillaginodes punctata, rock flathead Platycephalus laevigatus, bream Acanthopagrus australis, black bream A. butcheri, blue rock whiting Haletta semifasciata, garfish Hyporhamphus melanochir, luderick Girella tricuspidata, six-spine leatherjacket Meuschenia freychineti, sprat Hyperlophus translucidus and tarwhine Rhabdosargus sarba (Young & Carpenter 1977; Klumpp & Nichols 1983b; Robertson & Klumpp 1983; Ramm 1986; Ferrell & Bell 1991; Coles et al. 1993; Edgar & Shaw 1995a, 1995b; Jenkins et al. 1997a; MacArthur 1997; Jenkins & Wheatley 1998).
Small individuals of non-commercial species dominate the fish assemblages of seagrass, but these species can contribute to the diet of commercial species (Robertson 1982; Klumpp & Nichols 1983c; Edgar & Shaw 1995b). Also some commercial species that do not physically live in seagrass may benefit from seagrass production through the food chain (Edgar & Shaw 1995a, 1995b).
A decline in seagrass has been linked to a decline or change in fisheries production for that area. For example, in Western Port, Victoria, a 75% decline in seagrass over 15 years led to major changes in the local fisheries catches (MacDonald 1992). The species that declined were those that had strong associations with seagrass (habitat and dietary) while those that did
7 General Introduction not change or increased were not strongly associated (Jenkins et al. 1993; Edgar & Shaw 1995a, Jenkins et al. 1997a).
1.5 Structural features of seagrass habitat that promote the abundance and diversity of estuarine fauna
The structural features of seagrass habitat can refer to; 1) small scale features such as the shoot density and leaf length, 2) the larger-scale spatial structure (size and shape) of the seagrass beds, 3) the heterogeneity or patchiness of the seagrass beds, 4) the position of the seagrass beds within the estuary and 5) the proximity of seagrass beds to mangrove forests (interconnectedness of habitats).
1.5.1 Small scale structural features
It has been demonstrated that on a small scale the leaf density, leaf length and morphology can influence seagrass fish assemblages (Bell & Westoby 1986a; Bell et al. 1987; Worthington & Westoby 1991; Jenkins & Sutherland 1997). Bell & Westoby (1986a) found that when leaf height and density were manipulated at the scale of 25m2 plots abundances of decapods and small fishes changed significantly. This response, however, was not consistent when repeated at a larger scale of study, among seagrass beds within a bay (Bell & Westoby 1986b). A further experiment by these researchers, on the settlement of fish larvae in seagrass demonstrated that the eastern blue groper Achoerodus viridis did not discriminate between dense and sparse artificial seagrass units (Bell et al. 1987). Jenkins and Sutherland (1997) also compared artificial seagrass units with high with low complexity and found that some fish species were in greater abundances in the high complexity units. There were no differences, however, in the species richness between the two complexities.
Although the density and leaf height of a seagrass bed may be important on a small scale, on the larger scale of the whole estuary or bay, these patterns are masked by other features (See reviews by Connolly et al. 1999; Jelbart & Ross 2003).
8 General Introduction
1.5.2 Influence of size and shape of seagrass bed
The response of fauna to the size of a bed of seagrass has been investigated by a number of researchers (McNeill & Fairweather 1993; Irlandi et al. 1995, 1999; Irlandi 1996, 1997; Eggleston et al. 1998, 1999; Bologna & Heck 1999; Frost et al. 1999; Bell et al. 2001; Hovel & Lipcius. 2001, 2002). Most, but not all, of these studies investigated the response of invertebrates. Only three studies (McNeill & Fairweather 1993; Eggleston et al. 1999; Bell et al. 2001) considered the effects of bed size on fish.
McNeill and Fairweather (1993) found that a combination of two small seagrass beds consistently contained more fish species than one large bed of the same area. When using artificial seagrass beds, however, they found these results were not upheld. This could have been due to the small size of the artificial beds, which did not replicate the scale of natural beds. In contrast, Bell et al. (2001) found no consistent effects of bed size on resident fauna, including fish. Yet Bell et al. (2001) also found that the small beds of Halodule seagrass were not of poor quality in terms of fish densities; the densities of fish in the small and large beds were often similar and no species found in the large seagrass beds were missing from the small beds.
Eggleston et al. (1999) considered the response of fish and macro- invertebrates to different-sized plots of artificial seagrass. The only groups that responded to plot size were the grass shrimp (Palaemonidae) and other mobile crustacean (isopods and amphipods), which were in greater densities in the smaller plots. Other studies of invertebrates in seagrass habitat have had varied findings. Hovel and Lipcius (2001) found the survival of juvenile blue crab was higher in smaller beds of seagrass than in larger beds, whereas Eggleston et al. (1998) found the reverse. Hovel and Lipcius (2001) found that the density of seagrass shoots as well as bed size influenced predation on juvenile blue crabs. Further investigation found connectivity of beds to be more influential than bed size or structural complexity in determining rates of predation on the juvenile crabs (Hovel & Lipcius 2002), with crabs in isolated patches being more vulnerable. In contrast, in a mark–recapture experiment with juvenile hard clams (Irlandi 1997), larger beds (5–10 m across) of seagrass had higher survivorship than smaller beds
9 General Introduction
(~1 m across). With controls for below-ground biomass and shoot density (using artificial beds) there was no significant difference in the proportion of clams recovered live from large (4 x 4 m) and small (1 x 2 m) beds. In this experiment the scale of the artificial plots was similar to that of the natural beds surveyed. Similarly, Irlandi et al. (1999) found no long-term effects of patch size on the growth and survivorship of juvenile bay scallops. Generally, when considering the effect of seagrass bed size on invertebrates it seems other factors of seagrass beds may be more influential, such as shoot density, or connectivity of patches.
Larval processes are considered crucial for the abundance and diversity of juvenile and adult fish in most habitats. The advantage of many small seagrass beds of seagrass over just one bed is an increase in the probability of interception by larvae and recruits (Paine & Levin 1981; Sousa 1984; McNeill & Fairweather 1993). This increases the overall colonisation of the network of beds, as compared with a single larger seagrass bed (Bell et al. 1987; Sogard 1989; Worthington et al. 1992a; Eggleston et al. 1998). The recruitment process is variable in both time and space (Sogard 1989; McNeill et al. 1992; Worthington et al. 1992b) although the supply of larvae can be predictable spatially (McNeill et al. 1992; Jenkins et al. 1998). This suggests that many smaller beds will have the advantage over one large bed in terms of intercepting larvae and increasing overall recruitment. The size of a bed may indirectly affect recruitment through other mechanisms such as changing predator distribution, abundance and foraging behaviour (Irlandi 1997; Bologna & Heck 1999; Irlandi et al. 1999; Micheli & Peterson 1999; Hovel & Lipcius 2001, 2002) and modifying water flow (Eggleston et al. 1998).
The shape of a seagrass bed could also influence the diversity and abundance of fish by similar mechanisms. Some researchers propose that the high perimeter to area ratio of smaller habitats may offer more advantages than one large habitat with a low perimeter to area ratio (Paine & Levin 1981; Sousa 1984; McNeill & Fairweather 1993). A long, narrow bed (with a high perimeter to area ratio) may have an increased likelihood of intercepting more larvae than a rounder bed (with a low perimeter to area ratio).
10 General Introduction
Further investigation is required to determine if the smaller beds have greater numbers of prey organisms than larger beds. Seagrass beds are commonly used by macroinvertebrates. Dietary studies of fish assemblages in seagrass show that crustaceans are the major food item in fish diets (Burchmore et al. 1984; Pollard 1984; Robertson 1984; Edgar & Shaw 1995a, 1995b). Few fish species are capable of directly using plant material (Edgar & Shaw 1995a). Some amphipods, isopods and polychaetes will swim above seagrass at night in large numbers (Robertson & Howard 1978). They can also travel long distances from seagrass habitat (Virnstein & Curran 1986), so a small bed may attract macrofauna from a larger surrounding area. This macrofauna can then attract predator fish species to the bed or support larger numbers of fish than the size of the seagrass bed would suggest.
In summary, the gaps in research concern the effects of the size and shape of seagrass beds on the diversity and abundance of fish. This includes the recruitment processes of fish to seagrass beds and the interaction of invertebrate prey species with predator fish.
1.5.3 Are there edge effects in beds of seagrass for fish?
Ecotones, where two habitats meet such as sand and grass, are considered to be areas of high biodiversity. They provide two habitats for shelter, enhanced biotic interactions (such as predation or competition) and allow the mixing of two biotas from two separate habitats (Fox et al. 1997). An ecotonal effect occurs when the abundance of organisms changes about the edge of habitat (Lidicker 1999). The edges of seagrass beds may influence the abundance and diversity of fish. There are environmental changes in beds of seagrass as a function of the distance from the edge, such as a decrease in water flow from the edge to the centre (Fonseca et al. 1982; France & Holmquist 1997).
Numerous studies have considered the infaunal response to the edges of seagrass habitat (Summerson & Peterson 1984; Irlandi 1994; Bologna & Heck 1999; Bowden et al. 2001; Hovel & Lipcius 2002; Tanner 2004). Hovel and Lipcius (2002) found that the densities of juvenile blue crabs were
11 General Introduction greater in the interior of seagrass beds than at the edges (independent of shoot density). Tanner (in press) found that crustaceans tended to show a relatively strong edge effect (increased abundances) within 1 m of the edge; however, no distinctive edge-associated fauna was detected. Bowden et al. (2001) found some differences in the assemblage structure of small epifauna between the centre and edge of seagrass patches. In a review of the literature on faunal response to fragmentation in seagrass habitats, Bell et al. (2001) suggested a preferential use of the edge or interior by seagrass taxa although they could not detect this in their own research. Some studies found increased survival and growth for the taxa investigated on the edges of seagrass beds (Irlandi 1994; Bologna & Heck 1999) although the risk of predation was also greater (Bologna & Heck 1999; Hovel & Lipcius 2002). The edges of seagrass meadows were found by Sanchez-Jerez et al. (1999) to be relatively important for epifauna distribution, depending on the taxon and period of the year.
McNeill and Fairweather (1993) hypothesised that one of the reasons smaller beds have greater species richness than large beds is the increased likelihood of sampling an edge in a small bed. In fact, small seagrass beds could be considered to be all edge. An investigation of the existence of edge effects for fish in beds of seagrass is therefore required.
1.5.4 Habitat heterogeneity or spatial structure of a seagrass bed
Seagrass beds are considered to be simpler than some terrestrial ecosystems in terms of species diversity and structural complexity, and this simplicity may be useful for testing theories of habitat heterogeneity (Robbins & Bell 1994). Heterogeneous environments are considered to promote diversity (Heck & Orth 1980; Parrish 1989; Irlandi & Crawford 1997) and a positive correlation has been found between habitat heterogeneity and the number of fish species in shallow waters off the southern Bothnian sea in Sweden (Thorman 1986).
The structural complexity of seagrass can be a continuum from very sparse and patchy cover to dense prolific growth. At the end of the continuum some seagrass beds have an even grass cover and can be described as
12 General Introduction uniform. Other seagrass beds contain numerous patches of sand so that they appear to be broken up and can be described as patchy. The uniform seagrass beds could be considered to be a homogeneous environment (i.e. only one habitat), whereas the patchy seagrass beds could be considered to be a heterogeneous environment (i.e. containing two habitats: seagrass and sand). The spatial structure of a seagrass bed (such as patchy or uniform) may be important in determining fish abundance and diversity (Ferrell et al. 1992).
A mosaic of sediment and seagrass may directly or indirectly alter the assemblages of fish by numerous means (as outlined by Eggleston et al. 1999) including; 1. An alteration of predator distribution, abundance and foraging behaviour (e.g. Coen et al. 1981; Leber 1985; Main 1987; Bell & Hicks 1991; Danielson 1991; Edgar & Robertson 1992; Irlandi 1994; James & Heck 1994; Irlandi et al. 1995); 2. A modification of the hydrodynamics, which can influence settlement of larvae (Fonseca et al. 1982; Eckman 1983; Bell et al. 1995); 3. An influence on the accumulation of algae (Kulcycki et al. 1981; Holmquist 1994; Bell et al. 1995); and 4. The creation of changes in animal behaviour (reviewed by Heck and Crowder 1991). An organism’s response to habitat heterogeneity may depend on features specific to that organism such as body size and functional group (Eggleston et al. 1999).
Irlandi (1994) considered how the percentage cover and spatial arrangement or patchiness of a seagrass bed affected predation on Mercenaria mercenaria (hard clams). Predation was greater in the more patchy beds and in the beds with less percentage cover of seagrass. These findings were supported by further studies using another bivalve, the bay scallop Argopecten irradians (Irlandi et al. 1995). The pink shrimp Penaeus duorarum was found to be more abundant in low-energy, continuous seagrass beds than in high-energy, patchy seagrass beds (Murphey and Fonseca 1995). When investigating the differences in infaunal macroinvertebrates in patchy (fragmented) and unfragmented beds of Zostera marina, Frost et al. (1999) found there was no difference in the
13 General Introduction abundance of organisms or taxonomic groups, but there was a difference in the community composition. Similarly, the assemblage of mysids in shallow waters was strongly affected by the heterogeneity of the seagrass habitat (Barera-Cebrian et al. 2002), including two different species of seagrass and patches of sand.
Given the influences on the macroinvertebrate community from numerous studies, it is likely that the patchiness or heterogeneity of a seagrass bed could also influence the vertebrate community, i.e. fish. One study examining the fish found within patchy and continuous seagrass beds found that different fish species vary in their response to the patchiness of a seagrass bed (Crawford et al. 1995). More research is required to discern the influences of habitat heterogeneity on fish in seagrass beds.
1.5.5 Position of seagrass bed within the estuary
The position of a seagrass bed within an estuary can influence the abundance and diversity of fish within it (Bell et al. 1988; McNeill et al. 1992; Jenkins et al. 1996), although the measured effect of that influence can vary. For instance, between June and March of the years surveyed by McNeill et al. (1992), one seagrass bed was found to have up to 73 times the abundance of five species of fish than the other 15 beds surveyed. Yet, during the rest of the year, there was no significant difference between this seagrass bed and the others. The supply of larvae to this site was thought to have caused the recruitment of large numbers of individuals. In another study, there was a correlation between whiting (Sillaginodes punctata) abundance and distance from the bay entrance (Jenkins et al. 1996). Whiting spawn outside the bay, and hydrodynamic modelling demonstrated that a large amount of the variation in abundance at different sites could be attributed to two processes: the variation in currents delivering the larvae, and the exposure of the site to wave action that either kills or relocates the larvae (Jenkins et al. 1997b).
Bell et al. (1988) found that the location of the seagrass bed had a significant effect on the abundance of some fish species in the Pittwater estuary, Australia. They found that the fish were distributed in zones, with
14 General Introduction some species being more common close to the estuary mouth and others more common in the deeper reaches of the estuary. The estuary surveyed does not possess strong temperature or salinity gradients so they attributed this zoning to different patterns of spawning, larval dispersal and settling behaviour. Similarly, in another study (Hannan & Williams 1998), newly settled juveniles of ocean spawners were concentrated near the entrance of a marine lagoon. The distance of the seagrass beds from the ocean limited the larval distribution. In contrast, newly settled juveniles of lagoon spawners were widely distributed within the lagoon.
In southwestern, eastern and tropical Australia, marine species of fish dominate the fish assemblage in the lower parts of the estuary and with increasing distance from the estuary mouth; the fish assemblage is dominated by fish that can complete their life cycle within the estuary (Loneragan et al. 1986; Bell et al. 1988; Blaber et al. 1989; Loneragan & Potter 1990).
The ecological processes that ensure successful spawning, larval dispersal and recruitment need to be identified. Many estuarine fish spawn outside the estuary and use other habitats during stages of their life cycle (Boehlert & Mundy 1988; Hannan & Williams 1998). Hydrodynamic processes can influence seagrass landscape patterns (Fonseca & Bell 1998) and the transportation of larval to reach a bed (Boehlert & Mundy 1988; Jenkins et al. 1996; Hannan & Williams 1998; Etherington & Eggleston 2000; Smith & Suthers 2000). The hydrodynamic processes that influence a seagrass bed can be a function of its location within the estuary (Kjerfve et al. 1992; Kingsford & Suthers 1996; Smith & Suthers 2000). Further research is needed to determine how the hydrodynamic processes within an estuary influence fish abundance and diversity in seagrass beds.
1.5.6 Proximity to other habitats, e.g. mangroves
Numerous studies have found that the adjacent habitat can play an important role for species associated with seagrass (Heck 1979; Howard 1989; Sogard 1989: Ferrell & Bell 1991; Fortes 1991; Irlandi & Crawford 1997; Micheli & Peterson 1999; Nagelkerken et al. 2000; 2001; Cocheret de
15 General Introduction la Moriniere et al. 2002). The type of adjacent habitat and its distance from the seagrass bed can affect the diversity of seagrass fauna (Heck 1979; Sogard 1989).
Seagrass beds and mangrove forests in Australian temperate estuaries are often found within close proximity of one another and it may be predicted that interchange could occur in nutrients, sediments, larvae, post-larvae recruits or adult fish and invertebrates between the two habitats. Mangroves are considered to be important for many fish species as habitat for the post-larval and juvenile stages (Robertson & Duke 1987, 1990a, 1990b; Little et al. 1988; Laegdsgaard & Johnson 1995). Laegdsgaard and Johnson (1995) found that the majority of the juvenile fish in mangroves in summer were non-residents and therefore not confined to the mangrove habitat. One may predict that seagrass habitat close to mangrove could benefit in terms of species abundance and diversity from the proximity of the mangrove. Some researchers have found that the mangrove habitat had more fish and/or species of fish than the adjoining seagrass habitat (Robertson & Duke 1987; Thayer et al 1987; Laegdsgaard & Johnson 1995), but they compared the mangrove with seagrass only at high tide. It is conceivable that at low tide, when the mangrove forest floor is exposed, the fish might reside in an adjacent seagrass bed until the tide immerses the mangrove forest again. A study investigating the abundance of Caribbean reef fish found that seagrass beds close to mangrove forests had a greater species richness of nursery fish than beds with no adjacent mangrove forests (Nagelkerken et al. 2001). These authors suggested that the mangrove forests enhanced the species richness of the seagrass by an unknown interaction. It has been demonstrated that a few seagrass fish will forage in the adjacent mangroves as well as within the seagrass beds (Nagelkerken & van der Velde 2004a & 2004b).
The interconnectedness of habitats however, can have negative effects on some organisms. For instance, the species richness of macroinvertebrates on intertidal oyster reefs separated from seagrass and saltmarshes was higher than on reefs connected with either of these habitats. The seagrass and saltmarsh were shown to act as corridors for the movement of predatory blue crabs and hence facilitated greater rates of predation than in
16 General Introduction the reef habitat (Micheli & Peterson 1999). More studies are required which consider the interaction of seagrass beds with adjacent habitats, including the effect of varying distances between the two habitats.
1.6 Marine and Estuarine Protected Areas
The establishment of marine and estuarine protected areas (MEPAs) is generally regarded as a means to ameliorate the destruction of aquatic habitats (Roberts & Polunin 1991; Jones & Andrew 1992; Dugan & Davis 1993; Agardy 1994; Allison et al. 1998; García-Charton & Pérez-Ruzafa 1999). At present there is little consensus on criteria such as the size, shape and connectedness of protected areas (McNeill 1994). Nevertheless, there is evidence that some general features of a seagrass bed can favour the abundance and diversity of fauna. For example, within an estuary there can be a seagrass bed that contains greater diversity and abundance of fish and invertebrates than the rest of the beds in the estuary (McNeill et al. 1992). To provide adequate protection of estuarine fauna, information is required about the relationship between habitat characteristics and faunal assemblages. In the case of seagrass habitat, the influence of bed size, shape, spatial structure, location in the estuary and proximity of adjacent habitat (e.g. mangrove) on seagrass fauna is required. The effectiveness of a MEPA may also be influenced by the distance between habitats within the MEPA, their degree of interconnectedness, and the dispersal ability of individuals from other marine habitats (Goeden 1979; Sammarco & Andrews 1988).
1.7 This study
This study applies landscape ecology issues to the estuarine or seascape context. It investigates some of the large-scale spatial structural features of seagrass beds that can influence or be correlated with seagrass fauna. The results of this study will aid the implementation of marine-estuarine protected areas, by finding which spatial features of seagrass habitat provide the greatest contribution to seagrass faunal abundance and diversity.
17 General Introduction
In the first instance, the structural patterns of seagrass, such as the size and shape of the bed and the patchiness or heterogeneity of the seagrass cover were investigated for a correlation with the fish assemblages in seagrass beds. This included the large-scale spatial structural features, such as the distance of the seagrass beds from the estuary mouth and their proximity to mangrove forests.
The next level of investigation required a test of the hypothesis that edge effects results in greater fish species richness and densities (number per net) in smaller seagrass beds compared to larger seagrass beds. The fish and macroinvertebrates were sampled from the edges and interiors of small and large seagrass beds to test for differences in assemblage composition or abundances.
To further investigate whether the structural features of seagrass beds influence the fish and macroinvertebrate populations within them, artificial seagrass patch units were used to test the influences of patch size, perimeter length and perimeter to area ratio on seagrass fauna.
Finally, to test the theory that the connectivity of habitats can be advantageous for the abundance or diversity of fauna, seagrass beds located at varying distances from mangrove forests in two estuaries were sampled for fish and macroinvertebrates.
The results of this work can be applied to; general theories of landscape or seascape ecology, the design of marine estuarine protected area, and environmental management issues within estuaries. Most importantly, this study attempts to expand our ecological knowledge regarding the distribution and abundance of small fish and mobile macroinvertebrates in Zostera capricorni beds of temperate Australia.
18 General methods
2 General Methods
2.1 Study locations This study was done in the Pittwater and Brisbane Water estuaries, within
Broken Bay, just north of Sydney, NSW, Australia (Fig. 2.1; 33.6°S, 151.3°E). These estuaries are surrounded by considerable urban development (Geoscience Australia 2001) (Photo 2.1) although Kuring-gai Chase National Park is located on the western side of the Pittwater (Photo 2.2) and Brisbane Water National Park adjoins western sections of Brisbane Water. The Brisbane Water estuary also contains numerous small nature reserves and adjoins Bouddi National Park on the eastern side.
The Pittwater estuary is a tidal embayment with a maximum depth of 20 m although on average it is less than 5 m deep (Bell et al. 1988). It is an arm of the Hawkesbury River, but it receives little freshwater, so there is minimal variation in salinity and temperature throughout its length (Bell et al. 1988). Brisbane Water Estuary is a broad and shallow estuary with depths of generally 5 to 6m in the main water body (Geoscience Australia 2001). In each of these estuaries there are meadows or beds of Zostera capricorni (Photos 2.2 - 2.4) and Posidonia australis seagrass and mangrove forests of Avicennia marina (Photos 2.2 & 2.4).
2.2 Identification and mapping of the estuarine habitats Zostera capricorni seagrass was identified by visual examination and sampling of the vegetative structures. Seagrass beds that contained multiple species of plants or poorly delineated boundaries were not included in this study. Mangrove forests dominated by Avicennia marina (Photos 2.2 & 2.4) were often adjacent to the Z. capricorni seagrass beds. These habitats were mapped using a hand held GPS unit. The perimeter of each seagrass bed or mangrove forest was walked or boated around and every two metres the geographical position (longitude and latitude) was recorded. The perimeter,
19 General methods
Australia
Broken Bay, Sydney NSW
Brisbane Water
Broken Bay
Pittwater
Figure 2.1. A map of Broken Bay (-33.6°S, 151.3°E) showing the Pittwater estuary and the Brisbane Water estuary and its location within Australia.
20 General methods
Photo 2.1. The southern end of the Pittwater showing the extensive recreational use of the estuary (Geoscience Australia 2001).
Photo 2.2. The western side of the Pittwater estuary showing a Z. capricorni seagrass bed in the foreground and A. marina mangroves. In the background is the Kuring-gai Chase National Park.
21 General methods
Photo 2.3. A picture of a small and patchy seagrass bed in the Pitwater estuary.
Photo 2.4. A photo of a Z. capricorni bed and A. Marina mangroves in the Brisbane Waters estuary.
22 General methods
area and distance between the habitats were calculated using the GIS software ARC View® (ESRI 1996).
2.2.1 Description of Zostera capricorni Zostera capricorni Aschers is one of the most common seagrass species of the eastern coast of Australia (Conacher et al. 1994). It is mostly found in the sheltered bays, estuaries and lagoons of New South Wales, Queensland and also New Zealand (Kuo and McComb 1989). It is a monoecious perennial herb with linear leaf blades (usually 2-6) growing from shoots attached to creeping rhizomes that contain roots (Photo 2.4). This plant can be wholly submerged or intertidal and extend from brackish water to marine (Kuo and McComb 1989). It possesses flowers borne in spathes that enclose separate female and male flowers (den Hartog 1970).
2.2.2 The shoot density and blade length of Z. capricorni The density of the seagrass shoots in each seagrass bed was measured in situ in 25 x 25 cm quadrats using snorkel gear and visual examination. This quadrat size has been shown to be effective and efficient at estimating the density of seagrass shoots (Otway & Macbeth 1999). To determine how many quadrats were required to estimate shoot density, a cost benefit analysis was used. This involved calculating an accumulated mean and standard error with each subsequent quadrat. When the ratio of the mean to standard error is unchanged given subsequent sampling this number of quadrats is considered sufficient. It was found that for uniform or homogeneous seagrass beds six quadrats provided a small standard error to mean ratio, whilst minimising the sampling effort required (Fig. 2.2). In patchy seagrass beds, however, at least eight quadrats were required (Fig. 2.2). Thus, for future sampling, eight quadrats were used to estimate the shoot density of Z. capricorni in seagrass beds.
To determine the length of the blades of Z. capricorni, in each quadrat, ten shoots of Z. capricorni were cut off at the base and taken back to the
23 General methods
laboratory where the blades were measured to the nearest millimetre. This number of shoots per quadrat provided a small standard error to mean ratio by a cost benefit analysis (Fig. 2.3). The blades were thus sampled in each quadrat (n=8) after the shoot density was estimated.
60 Patchy seagrass bed
50
No. of 40 shoots
/quadrat30 number 20
10
0 123456789101112131415161718192021 Accumulated no. of quadrats
120 Uniform seagrass bed 100
No. of 80 shoots / quadrat 60 number
40
20
0 1234567891011121314 Accumulated no. of quadrats Figure 2.2. The number of seagrass shoots measured in accumulated quadrats in two types of seagrass beds (patchy or uniform) in the Pittwater estuary, January 2000.
24 General methods
Photo 2.5. The small seine net (8 x 2m, 1mm mesh) used for the surveys in Chapters 2,3,4 and 6.
20
18
16
14 Average 12 length of blades (cm) 10 8
6
4
2
0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Accumulated no. of shoots
Figure 2.3. The length (cm) of seagrass blades measured from sampling seagrass shoots in one quadrat in the Pittwater estuary, January 2000.
25 General methods
2.3 Fish survey techniques A trial of three different methods to sample the fish were undertaken in summer 1999-2000 using scuba underwater visual identification, a boom trawl net (1m in diameter, 2m drop 1mm mesh) and a small seine net (8 x 2m, 1mm mesh) (Photo 2.5). This size mesh was recommended (Rod Connolly pers. com.) because it effectively captures most of the small fish species found in seagrass beds (Guest et al. 2003). The underwater visual observation was too expensive and reliant on clear water conditions that were not frequent in some parts of the estuary. Furthermore visual observation would be ineffective for sampling fish during the night. The boom net proved to be unsuccessful at catching fish compared to the seine net, which consistently caught twice, if not three times the number of fish species per net haul. This was confirmed by a study conducted by Guest et al. (2003) that found seine netting to be more effective (collected more species and individuals) than beam trawls. For most of the subsequent sampling, a seine net was used to collect fish. At the start of every haul, the net was positioned in the seagrass bed fully open. The handlers moved into position twelve metres from the net walking wide of the sample area to avoid disturbing the fish. The sample area was then left undisturbed for a further five minutes. Finally, two people were used to haul the net, at a rate of 1m per second. No back net was used on the seine and few fish were seen to outrun the net as it was hauled. The bottom line of the net was weighed down with lead weights. Visual observation on numerous occasions confirmed that this kept the net down as it pulled through the seagrass. This bottom line momentarily flattened the seagrass, but it was not ripped or shredded by the net. Once the net was hauled it was placed into a large floating container that contained seawater to ensure the fish were submerged during the sorting process. Each fish and macroinvertebrate was then identified and measured (total length mm). Minimal handling of the fish and macroinvertebrates was encouraged and small hand nets were used to hold the fish. This was done so that fish could be released after identification back into the water, with minimal harm.
26 General methods
2.3.1 Identification of fish and macroinvertebrates Six months prior to starting the first pilot study, the technique of sampling with the seine net and identifying the fauna was practised. Over this sampling time numerous fish (over 500 individuals and 35 species) and macroinvertebrates were collected. The fauna that could not be identified in the field were taken back to the laboratory alive and placed into marine tanks where they could be observed until they grew into a stage where they could be identified (using Edgar 1997; Kuiter 2000). This technique ensured consistent and accurate identification of the fauna (especially those that were post-larval). After they were identified, the fauna were released into the estuary at the point of capture. Once the main part of the study commenced most fish and macroinvertebrates were identified in the field and released at another location within the estuary. There were still some species of fish, however, that could not be identified in the field (especially post-larval fish) so these individuals were taken back to the laboratory alive and placed in marine tanks until identified.
Initially, numerous specimens of macroinvertebrates were identified in the laboratory to the species level using a dissecting microscope. This identification was later confirmed by the Australian Museum. This method, however, required considerable amounts of time and included sacrificing the specimens. For these reasons it was decided that the macroinvertebrates would be identified to the order level only (eg. amphipoda, isopoda). The decapods were identified to the family level and the cephalopods were identified to the family or species level. This allowed for quick identification in the field.
2.4 Pilot study: Influence of tidal states Few studies have considered the influence of tidal regimes on fish assemblages in seagrass beds (but see Sogard et al. 1989). Instead most researchers limit their studies to a certain tidal state e.g. around the low tide or water depth. The aim of this pilot study was to determine if tidal levels affected the fish assemblages. If they were, this would indicate at which tidal state subsequent samples should be collected.
27 General methods
2.4.1 Method The sampling was done in winter (June to early September 2000) within an hour either side of the estimated low, mid or high tide time as published by Sydney Waterways (National Tidal Facility 2000). Two large seagrass beds (Careel Bay and Barrenjoey Beach) were used for this study to allow for independent sampling (each drag was no closer than ten metres to another). Sampling occurred over four days and three nights, but not within the same 24-hour period. A seine net (8 x 2m in size, 1mm mesh) was used to sample to the fish. This net was hand pulled through an area of seagrass that averaged 68 square metres. For every sampling period (n = 7) there were two hauls of the net for each of the three tidal states in each seagrass bed (twelve hauls in total). The water depth was recorded during each haul of the net.
A regression analysis was used to compare the number of fish species or fish individuals with the water depth during the time of sampling. To err on the side of caution the regression analyses were reported as correlations only and not a cause and effect direct relationship. Multivariate analysis was used to test for a difference in assemblage composition of fish among tidal samples. A Bray-Curtis similarity analysis among samples was performed using a square root transformation of the data to give a non-metric multi- dimensional scaling (MDS) plot (Clarke and Warwick 2001). A second similarity analysis was also performed using a presence-absence transformation. The first transformation gave an analysis of the composition and abundance of the fish assemblage whereas the second transformation gave an analysis of the composition alone. A two way crossed analysis of similarity (ANOSIM) was used to test for differences in species composition and abundance between day and night samples and tidal states (low, mid and high) or seagrass beds (Careel Bay or Barrenjoey Beach).
To determine how many hauls would ensure an adequate representation of the fish assemblage for each seagrass bed a species accumulation curve was constructed using the data collected.
28 General methods
Table 2.1. The mean number of fish per net haul collected from the pilot study (Pittwater 2000) during the day and night sampling over three tidal states (low, mid and high). Family Species Day samples Night samples low mid high low mid Atherinidae Atherinomorus ogilbyi 2.13 3.81 0.11 10.17 0.58 Chandidae Ambassis jacksoniensis 0 0 0 0 0.25 Clinidae Cristiceps aurantiacus 0 0 0 0.08 0.25 Heteroclinus fasciatus 0 0 0 0.25 0.08 Heteroclinus perspicillatus 0.06 0 0 0.08 0 Heteroclinus whiteleggi 0 0 0 0 0.08 Diodontidae Dicotylichthys punctulatus 0 0 0.11 0 0 Girellidae Girella tricuspidata 0.63 0.38 2.33 0.25 0.33 Gobiidae Arenigobius frenatus 43.81 78.13 28.78 50.50 51.08 Bathygobius kreffti 1.00 2.25 0.33 5.92 5.92 Redigobius macroston 0.50 0 0 1.25 0.25 Hemiramphidae Hyporhamphus australis 2.06 0.81 0.33 0 0.08 Monacanthidae Acanthalutere spilomelanurus 0.19 0.38 0 6.33 8.67 Cantherhinus pardalis 0.06 0 0 0 0 Eubalichthys mosaicus 0.06 0 0 0.08 0 Meuschenia venusta 0.06 0.06 0 0.17 0.08 Scobinichthys granulatus 0 0 0 1.08 1.58 Mugilidae Myxus elongatus 0.06 0 0 0 0 Paralichthyidae Pseudorhombus jenynsii 0.06 0.13 0 0 0 Scorpaenidae Centropogon australis 0.50 1.19 0.56 2.42 2.83 Sillaginidae Sillaginodes maculata 0 0 0 0.33 0.17 Sillaginodes punctatus 0.06 0 0 0 0 Sparidae Rhabdosargus sarba 9.75 0.69 1.11 2.42 3.42 Sphraenidae Sphyraena obtusata 0.19 0.06 0.44 0 0 Syngnathidae Stigmatophora argus 0.19 0.06 0 0.33 1.25 Stigmatophora nigra 4.75 1.19 0 1.92 2.58 Urocampus carinirostris 4.56 7.19 2.56 3.92 3.42 Terapontidae Pelatus sexlineatus 9.69 5.50 2.56 2.58 1.92 Tetraodontidae Tetractenos hamiltoni 0.81 0.06 0 0.08 0.25
29 General methods
Table 2.2. The results of the analysis of similarity (two way crossed) comparing the fish assemblages; at day and night, during three tidal states and in two seagrass beds for the pilot study (Pittwater 2000). Square root Presence/absence Categories tested transformation transformation Global R P Global R P
Time: day and night 0.27 < 0.001 0.30 < 0.001 Tide: low, mid and high 0.12 < 0.01 0.11 < 0.01 Beds: Careel and Barrenjoey 0.09 < 0.01 0.11 < 0.01
Pairwise tests (tides) Low vs Mid 0.05 = 0.05 0.04 > 0.05 Low vs High 0.16 < 0.05 0.18 < 0.05 Mid vs High 0.31 < 0.01 0.24 < 0.01
2.4.2 Results and Discussion A total of 29 species and 5386 individuals of small fish were collected in the pilot study (Table 2.1). There was a greater number of fish species per net collected during the night sampling than the day (Fig. 2.4), although the number of fish individuals was similar. This suggested that all subsequent sampling of fish in seagrass beds must include sampling at night. There were less fish species and individuals collected during the high tide than the low or mid tide (Fig. 2.5). This was attributed to the sampling effectiveness of the seine net. In water depths over one metre it was difficult for the handlers to pull the net through the water effectively. As a result, with increasing water depth there was a decrease in the number of fish species and fish individuals collected. The number of fish species and the water depth were negatively correlated for the full depth range of 20 to 140 cm, low to high tide during samples collected during the day (R2 = 0.17, P < 0.01) (Fig. 2.6). Similarly, the abundance of individual fish and the water depth were negatively correlated at this depth range (R2 = 0.09, P = 0.05) (Fig. 2.6). However both of these regression values were low and should be considered with caution. When the data was limited to the depth range of 30 to 100 cm (low to early mid tide) there were no significant correlations
30 General methods between the number of fish species or individual fish per net and water depth (Figures 2.7 and 2.8). This indicated that sampling during the low to early mid tide (30-100cm) was required to ensure tidal variations would not influence the number of fish species and individuals collected.
There were different fish assemblages between the day and night sampling which was attributed to the composition of the fish assemblage and not the abundances (Fig. 2.9, Table 2.2). The fish assemblage also differed among tidal states. The fish assemblage collected during the high tide was significantly different from the fish assemblages collected during low and mid tides (Fig. 2.9, Table 2.2). The fish assemblages were not significantly different from one another during the low and mid tide (R = 0.049, P = 0.05), although this value was close to being significantly different. The presence-absence transformation, however, revealed no significant differences in composition of the fish assemblage (R = 0.041, P = 0.09). The global R-values were quite low for all factors investigated and the stress value (0.22) was high on the MDS plots. This suggests that one should be cautious to accept a significant difference in the fish assemblages for day and night sampling and the tidal states. However, for the purpose of this pilot study, the multivariate analysis supported the previous finding that sampling during the low to mid tide had no significant influence on fish assemblages in seagrass.
The species accumulation curve (Fig. 2.10) revealed a variation in the number of species accumulated per seagrass bed and between day and night sampling. The highest rate of accumulation of fish species was during the day at Careel Bay and at night at Barrenjoey Beach. During the day, six hauls of the net at Barrenjoey Beach and seven hauls at Careel Bay were required to represent over 80% of the maximum number of fish species collected from 16 and 18 hauls respectively. During the night, 10 hauls were required at Barrenjoey Beach and 9 at Careel Bay to represent over 80% of the total catch. For the first survey in this thesis (chapter 3) it was decided to use four hauls of a net in each seagrass bed for the day and night sampling (total eight) which would represent at least 55-70 % of the total catch. Although this seems a low estimate it was decided that the first part of this study would require more replication at the level of seagrass bed and
31 General methods less replication of net hauls. In the second survey (chapter four) the number of hauls in each seagrass bed was raised to nine for each season. In chapter six, the number of nets hauls was again dropped to four, because the high variability between seagrass beds required more replication at the level of seagrass bed.
10 Fish species 9 8 No. fish 7 6 species / 5 net 4 3 2 1 0 day night
110 Fish individuals 100 90 No. fish 80 70 individuals 60 / net 50 40 30 20 10 0 day night
Figure 2.4. The average number of fish species and fish individuals collected during the day and night sampling from each haul of the net in the pilot study (Pittwater 2000).
32 General methods
10 Fish species 9 8 No. fish 7 species /6 net 5 4 3 2 1 0 Low Mid High Low Mid
Day Night
140 Fish individuals
120
100 No. fish 80 individuals / net 60
40
20
0 Low Mid High Low Mid
Day Night
Figure 2.5. The average number of fish species and fish individuals collected during the three tidal states (low, mid & high) from day and night sampling in the pilot study (Pittwater 2000).
33 General methods
12 Day, Fish species 10 R2 = 0.17 No. fish 8 P < 0.01 species / 6 net 4 2 0 0 20 40 60 80 100 120 140 water depth (cm)
350 Day, Fish individuals 300 250 No. fish R2 = 0.09 200 individuals P = 0.05 150 / net 100 50 0 0 20 40 60 80 100 120 140 water depth (cm)
Figure 2.6. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the day. The results are shown for all water depths sampled (from 22 to 140 cm) in the pilot study (Pittwater 2000). .
34 General methods
12 Day, Fish species 10
No. fish 8 species /6 net 4 2 0 0 20406080100120 water depth (cm)
350 Day, Fish individuals 300 No. fish 250 individuals200 / net 150 100 50 0 0 20 40 60 80 100 120 water depth (cm)
Figure 2.7. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the day. The results are shown for water depths from 30 to 100 cm in the pilot study (Pittwater 2000). Note the absence of a correlation between the two factors.
35 General methods
16 Night, Fish species 14 No. fish 12 species /10 net 8 6 4 2 0 0 20406080100 water depth (cm)
200 Night, Fish individuals
No. fish 150 individuals 100 / net
50
0 0 20406080100 water depth (cm)
Figure 2.8. The plots of the average number of fish species or fish individuals per net and the water depth during sampling in the night. The results are shown for water depths from 30 to 100 cm in the pilot study (Pittwater 2000).
36 General methods
Day and night
Global R = 0.27 P < 0.001
Day
Night
Tidal states
Global R = 0.12 P < 0.01 Low tide
Mid tide
High tide
Seagrass beds
Global R = 0.09 P < 0.01
Careel Bay
Barrenjoey Beach
Stress = 0.22
Figure 2.9. The two-dimensional configurations for MDS ordinations of the fish assemblages collected in the pilot study (Pittwater 2000). The replicates are categorized into samples taken: during day and night sampling, during the three tidal states, low mid and high and from the two seagrass beds, Careel Bay and Barrenjoey Beach.
37 General methods
25 Careel Bay 20
Accumulated 15 no. of fish
species 10 day
5 night
0 0 5 10 15 20
Number of net hauls
25 Barrenjoey Beach 20 Accumulated 15 no. of fish species 10 day
5 night
0 0 5 10 15 20
Number of net hauls
Figure 2.10. The accumulation of fish species during the day and night sampling in the Careel bay seagrass bed and Barrenjoey beach seagrass bed in the pilot study (Pittwater 2000).
38 Testing the Species Area Relationship and Landscape Ecology Theories
3 A test of the species area relationship and landscape ecology theories 3.1 Introduction
Seagrass is considered an appropriate habitat to investigate the species-area relationship (McNeill & Fairweather 1993) and landscape ecology theories regarding spatial scale and patterns (Bell and Hicks 1991, Robbins and Bell 1994, Irlandi et al. 1995). Seagrass beds can be considered ‘islands’ of vegetation surrounded by a ‘sea’ of unvegetated sediment and are therefore appropriate for the application of the species-area relationship. They are also distributed within an estuary over different spatial scales while maintaining relative structural homogeneity (Robbins & Bell 1994; Turner et al. 1999).
3.1.1 The species-area relationship The species-area relationship proposes that larger islands or areas of habitat support more species diversity than small islands or areas (Arrhenius 1921; Gleason 1922; Williams 1964; Preston 1960; MacArthur & Wilson 1967). Three main models have been proposed as to why this pattern is expected. 1. The habitat diversity model. This proposes that larger islands have greater diversity of habitats and therefore greater species diversity (Gleason 1922; Williams 1964). 2. The area per se model (or equilibrium theory). This proposes that larger islands have more species because they have lower rates of local extinction and higher rates of successful immigration. Larger islands also offer greater targets for colonists than do smaller islands (Preston 1960; MacArthur & Wilson 1967). 3. The passive (random) sampling model. This suggests that there is a sampling artefact involved when sampling a large compared to a small islands. More area is sampled in a large island, and thus by chance alone, more species will be detected (Arrhenius 1921).
To test the first model the fish assemblages in seagrass beds containing uniform seagrass cover were compared to patchy seagrass beds that contained a mosaic of sand and seagrass. This provided a comparison
39 Testing the Species Area Relationship and Landscape Ecology Theories between structurally homogeneous and heterogeneous habitats. Heterogeneous environments are considered to promote diversity of species when compared to homogeneous environments (Heck & Orth 1980; Irlandi & Crawford 1997; Eggleston et al. 1999).
To test the area per se model or equilibrium theory the fish assemblages in seagrass beds are predicted to be dependent on the size of a seagrass bed i.e. larger seagrass beds are predicted to have greater species alpha diversity than smaller seagrass beds. Furthermore the perimeter to area ratios of the seagrass beds were also measured to test if the interception of colonists by the edge of the seagrass bed can contribute to the diversity of fish, as predicted by this model. Some researchers propose that the high perimeter to area ratio of smaller habitats may offer more advantages than one large habitat with a low perimeter to area ratio (Paine & Levin 1981; Sousa 1984; McNeill & Fairweather 1993). A long, narrow bed (with a high perimeter to area ratio) may have an increased likelihood of intercepting more larvae than a rounder bed (with a low perimeter to area ratio).
The area per se model has been supported in seagrass beds by Bowden et al. (2001) who found greater numbers of macroinvertebrate taxa in large patches of seagrass compared to small patches. In contrast, other researchers have found greater species richness or abundance of organisms in smaller than larger seagrass beds (McNeill & Fairweather 1993, Eggleston et al. 1999; Hovel & Lipcius 2001). In a review of the literature on the effect of seagrass patch size on seagrass fauna, Bell et al., (2001) suggested that the size of a seagrass bed does not have a consistent influence on the faunal assemblages. This conflicting information indicates that there are possibly other factors that need to be considered in addition to the size of habitat or that it may be reliant on the organism studied.
The passive (random) sampling model suggests that there is a sampling artefact involved when sampling large compared to a small islands. This can be avoided by sampling the same size area in large and small habitats, but this presents a second sampling artefact whereby a large proportion of a small seagrass bed will be sampled while only a small proportion of the large seagrass bed is sampled. Therefore, the number of species may
40 Testing the Species Area Relationship and Landscape Ecology Theories accumulate at a faster rate per sampling effort in smaller seagrass beds. Connor and McCoy (1979) suggested that the sampling artefact model must be proposed as the null hypothesis when testing the species-area relationship. In this study the presence of a sampling artefact was tested by determining if the rate of fish species captured in a seagrass bed was correlated with its area.
3.1.2 Location of the seagrass bed within the seascape Previous studies have revealed that the location of a seagrass bed within an estuary can influence the abundance and diversity of fish found in that bed (Bell et al. 1988, McNeill et al. 1992, Jenkins et al. 1996) although the measured effect of that influence could vary (Jenkins et al. 1996). Studies in Australia have found the fish assemblage in the lower parts of the estuary to be dominated by marine fish species and in the upper estuary to be dominated by fish that can complete their life cycle within the estuary (Loneragan et al. 1986; Bell et al. 1988; Blaber et al. 1989; Loneragan & Potter 1990). Therefore any survey of fish within an estuary must also account for the potential confounding influence of the location of a seagrass bed within the estuary i.e. the landscape or seascape context.
3.1.3 Study Aims The aim of this study was to test hypotheses arising from the alternative models explaining the species-area relationship. The habitat diversity model predicts that seagrass beds with more structural heterogeneity will have greater species richness (number of fish species collected per bed). The area per se model predicts the larger beds of seagrass will have greater species richness than the smaller beds and that species richness will be correlated with the perimeter-area ratio of the bed. The passive sampling hypothesis predicts that the species richness in each seagrass bed will be correlated with the sampling effort i.e. number of net hauls. An alternative seascape hypothesis predicts that the fish assemblage in beds closer to the estuary mouth will be different from the beds further from the estuary mouth.
The variables tested in this study were the number of fish species per net, the fish species richness, the abundance of individual fish per net, and the
41 Testing the Species Area Relationship and Landscape Ecology Theories fish assemblage of each seagrass bed. This study was concerned with alpha diversity i.e. how many fish species were collected per bed at a given time. The species-area relationship is traditionally concerned only with species richness but the other variables were measured to test the seascape ecology model.
3.2 Methods 3.2.1 Study area and description This study was done from September to early December 2000 (spring to early summer) in the Pittwater estuary north of Sydney, NSW, Australia (Fig. 3.1). Fifteen monospecific beds of Zostera capricorni were selected based on their distribution throughout the estuary (see Fig. 3.1) and similar water depths (30-80 cm at mean low tide (Chapter 2). Each bed was separated from one another by more than 100 m of bare sandy substratum and usually by much more. The seagrass varied in average shoot densities (range 514- 1166 shoots/m2, mean = 53.20 ± S.E. = 47.86) and average blade length (range 7.47-22.48 cm, mean = 13.19 ± S.E. = 1.11).
3.2.2 Seagrass bed categories A survey of the seagrass cover in each bed established that eight of the seagrass beds were patchy, while the other seven were uniform (Table 3.1). The seagrass beds were also categorised into three size groups: small (980 to 2300 m2), medium (3375 to 4090 m2) and large (5335 to 6630 m2). Both categories, seagrass cover (patchy and uniform) and sizes (small, medium and large) were spatially interspersed throughout the estuary (Fig. 3.1). The seagrass beds were further categorised by their distance from the mouth of the estuary i.e. either close or far. Although the distance of a seagrass bed from the mouth of the estuary could be considered a continuous variable, in this study there were few beds within the middle of the estuary (see Fig. 3.1). It was thus more appropriate to categorise the beds as those that were close to (between 1 and 4 km) or far from (between 8 to 10 km) the estuary mouth (Table 3.1).
42 Testing the Species Area Relationship and Landscape Ecology Theories
3.2.3 Estimation of habitat characters 3.2.3.1 Estimation of size and perimeter area ratio of beds The perimeter of each bed was walked or boated around. Every two metres the geographical position (longitude and latitude) was recorded using a hand held Garmin GPS unit. The perimeter and area of the beds were calculated using the GIS software ARC View. The perimeter to area ratio (PAR) of the seagrass beds ranged from 0.071 to 0.341 (Table 3.1).
3.2.3.2 Habitat heterogeneity or patchiness To estimate the patchiness or uniformity of the seagrass cover in a bed, 120 contiguous quadrats (25 x 25 cm) were arranged along a randomly placed 30 m transect. Three transects were used in each seagrass bed. The cover of the seagrass (measured as percentage cover / quadrat) was estimated in each quadrat by visual examination using snorkel gear.
It was predicted that the patchy seagrass beds would have significantly greater variability or dispersion of seagrass percentage cover / transect than the uniform seagrass beds. Variability in percentage cover of Z. capricorni in seagrass beds was detected by using Multivariate Dispersion Indices (Clarke & Warwick 2001). Although this is a multivariate analysis package, only a single variable was tested (i.e. seagrass cover). This method proved to be more robust than other univariate measures (eg. Paired quadrat method, Moran’s I coefficient). A measure of dispersion for transects from the patchy seagrass beds was 1.374 compared to 0.376 for transects from the uniform seagrass beds. This indicated that transects from the patchy beds contained significantly more variability in the measures of percentage seagrass cover than the uniform beds (Clarke & Warwick 2001).
43 Testing the Species Area Relationship and Landscape Ecology Theories
Pittwater
Sydney NSW
N Estuary P 15 mouth P 14
U P 13 11 U 12 U 10
9 U
Close to P 7 8 estuary mouth U Scotland Is.
U 6 4 U
3 Far from 5 P P estuary 2 P mouth 1km P 1
Figure 3.1. A map of Australia showing the location of the Pittwater estuary. The Z. capricorni beds are shown in their relative size and numbered according to the scheme in table 1. The demarcation between seagrass beds close and far from the estuary mouth is shown. Uniform beds = U, and patchy beds = P
44 Testing the Species Area Relationship and Landscape Ecology Theories
Table 3.1. A description of each Z.capricorni bed surveyed in Pittwater spring, 2000 and the number allocation for figure 3.1. Number Area Perimeter PA ratio Habitat Distance to on map (m2) (m) P/A structure estuary mouth 1 982 (small) 335 0.341 patchy far 2 5722 (large) 874 0.153 patchy far 3 3871 (med) 725 0.187 patchy far 4 5370 (large) 538 0.100 uniform far 5 4091 (med) 480 0.117 patchy far 6 3618 (med) 426 0.118 uniform far 7 2310 (small) 379 0.164 patchy far 8 5334 (large) 795 0.149 uniform not used 9 2168 (small) 357 0.165 uniform close 10 3971 (med) 544 0.137 uniform close 11 3373 (med) 637 0.189 patchy close 12 1258 (small) 237 0.188 uniform close 13 3812 (med) 683 0.179 uniform close 14 6627 (large) 473 0.071 patchy close 15 6050 (large) 391 0.065 patchy close
Table 3.2. The regression results comparing the total numbers of fish species or individuals with the measures of seagrass (shoot density, blade length percentage cover) and sand in each bed from day and night sampling in Pittwater 2000. The R2 values are shown and no significant relationships were found. Day Night Fish species Fish individuals Fish species Fish individuals Density of shoots 0.012 0.205 0.0001 0.185 P = 0.09 P = .0.11 Length of blades 0.058 0.072 0.013 0.001 Cover of seagrass 0.025 0.091 0.001 0.057 Amount of sand 0.006 0.013 0.001 0.019
45 Testing the Species Area Relationship and Landscape Ecology Theories
3.2.3.3 Shoot density and blade length of Z. capricorni The shoot density and leaf length of Z. capricorni was estimated in eight random quadrats in the uniform beds and in twelve random quadrats in the patchy beds (Chapter 2). The average shoot density, leaf length, percentage cover of seagrass and average sand cover per transect in each seagrass bed was tested for a relationship with the total number of fish species and fish individuals (from day and night sampling) using a regression analysis. There was no relationship found between these features (Table 3.2). There was, however a trend for the number of fish individuals to be positively correlated with the shoot density but this was not statistically significant. The tests indicated that the total number of fish species or fish individuals were not confounded by the amount of seagrass or sand present in each bed.
3.2.4 Fish Survey All seagrass beds were sampled for fish in the daytime and at night. To obtain an estimate of the assemblages of fish in seagrass beds, an 8x 2 m seine net (1 mm mesh diameter) was randomly placed and dragged through each bed, sampling an area approximating 68 m2 (mean = 68.17 ± 1.2, S.E.). In both the patchy and uniform beds the net dragged the same area, regardless of substratum i.e. seagrass or sand. A cost benefit analysis determined that four drags of the net gave an adequate estimation of the fish assemblage (Chapter 2). A small seine net captures small and juvenile fish more effectively than larger pelagic fish (Guest et al. 2003) and these smaller fish were the focus of the project. The sampling occurred during low to early mid tide (between 30 to 100 cm) to ensure that the tidal state or water depth did not confound the results (Chapter 2). After collection, the fish were identified, measured and released back in the estuary at a seagrass bed not sampled in this survey. Those fish that were difficult to identify, were taken back to the laboratory, placed in marine tanks until they could be identified. To separate juvenile from adult fish, individual fishes were measured to the nearest mm length to caudal fork (LCF) or total length (TL).
46 Testing the Species Area Relationship and Landscape Ecology Theories
3.2.5 Data analysis 3.2.5.1 Univariate analysis Analysis of variance (ANOVA) tests were performed on the number of fish species, fish individuals and the individuals of the most numerous fish species collected per net for each seagrass bed. Each analysis contained a combination of fixed (time of sampling, heterogeneity of the habitat, size of bed and distance from estuary mouth) and random (seagrass bed) factors. A Cochran’s test was used to detect heterogeneity of variances (Underwood
1997) and the data was transformed (Ln(x+1)) if the Cochran’s test proved to be significant. Student Newman Kuel's (SNK) tests were used to detect post- hoc differences among means.
The first ANOVA contained four levels: time of sampling (day or night), distance of the bed from estuary mouth (close or far); size of bed (small, medium or large) and seagrass beds (Fig. 3.2a). The second ANOVA was similar except that the distance of the bed from the estuary mouth was replaced by heterogeneity of habitat (patchy or uniform) (Fig. 3.2b). It was decided that these analyses required no adjustment of the significance level because the source of variation, heterogeneity of the seagrass beds was an unrelated hypothesis to the distance from the estuary mouth. Furthermore, it was considered to be more prudent to increase the probability of a type I error (erroneously rejecting the null hypothesis) compared to a type II error (erroneously accepting the null hypothesis) (Quinn 2002). The ANOVA results for just the first three levels (time, heterogeneity and size of bed) are shown for the individual species data for clarity.
The statistical program (GMAV5 for Windows) was used for the analysis of variance. This program only allows a balanced ANOVA so it was necessary to randomly select and drop some seagrass beds. This process of randomly dropping two medium sized beds and one large bed and analysing the data was done numerous times and the outcomes were the same for all combinations.
47 Testing the Species Area Relationship and Landscape Ecology Theories
Level 1 (n=2) Day Night
Level 2 (n=2) Close Far Close Far
Small beds Medium beds Large beds Small beds Medium beds Large beds Small beds Medium beds Large beds Small beds Medium beds Large beds Level 3 (n=3)
Bed 9 Bed 12 Bed 10 Bed 13 Bed 14 Bed 15 Bed 7 Bed 1 Bed 3 Bed 5 Bed 2 Bed 4 Level 4 (n=2) Replicates = four drags of the net Figure 3.2a. The design for the ANOVA comparing; day and night sampling (level 1, fixed orthogonal), seagrass beds close and far from the estuary mouth (level 2, fixed and orthogonal), and seagrass beds in three different size categories (level 3, fixed orthogonal) with seagrass beds a random factor nested in levels 2 and 3. See figure 1 and table 1 for a description of the beds.
Day Night Level 1 (n=2)
Small beds Medium beds Large beds Small beds Medium beds Large beds Level 2 (n=3)
Level 3 (n=2) Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy Uniform Patchy
Bed 9 Bed 12 Bed 1 Bed 7 Bed 6 Bed 14 Bed 5 Bed 11 Bed 4 Bed 8 Bed 2 Bed 15 Level 4 (n=2 ) Replicates =four drags of the net Figure 3.2b. The design for the ANOVA comparing; day and night sampling (level 1, fixed orthogonal), seagrass beds in three different size categories (level 2, fixed, orthogonal), heterogeneity of seagrass beds (level 3, fixed orthogonal) with seagrass beds a random factor nested in levels 2 and 3.
49
Testing the Species Area Relationship and Landscape Ecology Theories
Linear regressions were used to test for a relationship between the area of a bed and the total number of fish species (species richness) or the total number of fish individuals per bed. The former was performed to test for a species-area relationship. Linear regressions were also used to test for a relationship between the perimeter area ratio of the seagrass beds and the mean number of fish species and fish individuals per net.
To determine if there was a passive sampling artefact the accumulated number of fish species was plotted for each consecutive drag of the net for each seagrass bed. It was predicated that if there was a sampling artefact then the number of fish species should accumulate at a greater rate in small than large seagrass beds. The small seagrass beds would have a steeper slope on the regression line than the large seagrass beds. For each seagrass bed, the slope of the line was calculated (i.e. rate of species accumulation) and then plotted against the size of the bed. A regression analysis was then used to detect if the size of a seagrass bed was correlated with the rate at which fish species were accumulated during sampling. This was done using data collected in the day and night data separately.
3.2.5.1 Multivariate analysis A Bray-Curtis similarity analysis among samples was performed using a square root transformation of the data to give a non-metric multi- dimensional scaling (MDS) plot. A second similarity analysis was also performed using a presence absence transformation. The first transformation gave an analysis of the composition and abundance of the fish assemblage whereas the second transformation gave an analysis of the composition alone. A two way crossed analysis of similarity (ANOSIM) was used to test for differences in species composition and abundance between all patchy (n = 8) and uniform (n = 7) seagrass beds and all beds of different sizes (small = 4, medium = 6, large = 5). A second ANOSIM was also used to test for differences among seagrass beds of different sizes and distances from the mouth of the estuary (close = 7, far = 7). A similarity percentages (SIMPER) analysis was used to determine what species were contributing to the differences in assemblage composition detected by the ANOSIM (Clarke & Warwick 2001).
51 Testing the Species Area Relationship and Landscape Ecology Theories
3.3 Results There were 52 species of fish and 9350 individuals collected including six species of recreational and/or commercial importance (Table 3.3). The number of fish species per net were found to be significantly greater at night than during the day (Fig. 3.3; Table 3.4, SNK night > day) although there was no corresponding difference in the number of individual fish per net. Some species of fish that were more abundant during the night than the day were Arenigobius frenatus, Bathygobius kreffti, and Atherinomorus ogilbyi. Only one fish, Acanthalutere spilomelanurus, was more abundant during the day (Table 3.5).
3.3.1 Habitat heterogeneity The patchy and uniform seagrass beds had similar numbers of fish species and fish individuals per net (Table 3.4), but different fish assemblages (Fig. 3.4; Table 3.6). The ANOSIM was run twice, with two different transformations (square root and presence/absence). The low global R- values suggest a high degree of dispersion in the groupings, but the P value indicated a significant difference (Table 3.6). The low global R value for the square root transformation (R=0.060) compared to the higher R value (R=0.124) for the presence /absence transformation suggests the difference can be attributed to the composition of fish assemblages more than the abundances of fish. The fish species that provided the greatest discrimination between patchy and uniform beds were Arenigobius frenatus, A. spilomelanurus, B kreffti and Urocampus carinirostri. There was a trend for these species to be more abundant in the uniform beds than the patchy beds, although the univariate analysis was significant for only B. kreffti (Table 3.5). In contrast, the numbers of Girella tricuspidata and Pelates sexlineatus per net were significantly greater in the patchy beds, while Stigmatophora nigra was more numerous in the uniform beds (Table 3.5).
52 Testing the Species Area Relationship and Landscape Ecology Theories Fish species 14 12 10 8
6 4
No. fish species / net / species fish No. 2 0 Small Medium Large Small Medium Large Day Night
180 Fish individuals 160 140 120 100 80 60 40 20 No. fish individuals / net / individuals fish No. 0 Small Medium Large Small Medium Large Day Night
Figure 3.3. The number of fish species and fish individuals per net in the small, medium and large seagrass beds during the day and night sampling in Pittwater, spring 2000.
53 Testing the Species Area Relationship and Landscape Ecology Theories
Table 3.3. The fish collected from beds of Z. capricorni in Pittwater, spring 2000 using four drags of a seine net during the day and night. Species of recreational and/or commercial importance to fisheries are marked *.
FAMILY SPECIES DAY NIGHT Aplodactylidae Cheilodactylus vestitus 0 1 Atherinidae Atherinomorus ogilbyi 11 179 Batrachoididae Batrachomoeus dubius 0 3 Blenniidae Petroscirtes lupus 1 19 Callionymidae Repomucenus calcaratus 6 4 Chandidae Ambassis jacksoniensis 849 103 Clinidae Cristiceps argyropleura 0 1 Cristiceps aurantiacus 6 24 Heteroclinus fasciatus 6 21
Heteroclinus sp 4. 2 6 Clupeidae Hyperlophus translucidus 887 15 Clupeidae Spratelloides robustus 21 3 Diodontidae Dicotylichthys punctulatus 0 1 Diodon nichthemerus 0 2 Gerreidae Gerres subfasciatus 2 9 Girellidae Girella tricuspidata* 414 300 Gobiidae Arenigobius frenatus 842 1877 Bathygobius kreffti 79 471 Cristatogobius gobioides 0 1 Redigobius macroston 33 61 Hemiramphidae Hyporhamphus australis* 0 1 Labridae Achoerodus viridis 28 3 Halichoeres hortulanus 1 0 Stethojulis interrupta 0 1 Monacanthidae Acanthalutere spilomelanurus 439 267 Acanthaluteres vittiger 1 2 Brachaluteres jacksonianus 1 0 Cantherhinus pardalis 9 7 Eubalichthys mosaicus 10 7 Meuschenia trachylepis 6 4 Meuschenia venusta 51 72 Monacanthus chinensis 5 2 Scobinichthys granulatus 53 83 Mullidae Upeneichthys lineatus 2 19 Upeneus sp. 4 14 Upeneus tragula 0 6 Odacidae Neoodax balteatus 5 5 Paralichthyidae Pseudorhombus jenynsii 3 2 Scorpaenidae Centropogon australis 39 199 Sillaginidae Sillaginodes punctatus* 1 10 Sillago maculata* 0 10 Sillago flindersi* 1 1 Sparidae Rhabdosargus sarba* 276 447 Syngnathidae Festucalex cinctus 0 1 Filicampus tigris 7 10 Hippocampus whitei 4 4 Stigmatopora argus 4 3 Stigmatopora nigra 137 104 Urocampus carinirostris 176 178 Tetraodontidae Tetractenos hamiltoni 2 1 Terapontidae Pelates sexlineatus 132 230
54 Testing the Species Area Relationship and Landscape Ecology Theories
Table 3.4. The analysis of variance comparing the number of fish species and the number of fish individuals per net collected from seagrass beds during different times (day and night), from beds of different heterogeneity (patchy or uniform) and size (small, medium or large). The data for the fish individuals was transformed, Ln (x+1) *P< 0.05. Mean Source of Variation df F Mean squares F squares No. fish species / net No. fish individuals / net Time 1 90.1 8.05* 3.37 2.71 Heterogeneity (Het) 1 3.8 0.37 1.49 0.96 Size 2 95.7 9.46* 12.76 8.21* Beds (Het. x size) 6 10.1 1.42 1.55 2.84* Time x heterogeneity 1 0.5 0.05 0.07 0.06 Time x Size 2 1.9 0.17 1.23 0.99 Time x bed (Het. x Size) 6 11.2 1.57 1.24 2.27 Heterogeneity x Size 2 1.3 0.13 4.03 2.59 Time x Het. x Size 2 6.9 0.61 2.71 2.18 Residual 72 7.1 0.55
Table 3.5. The summary of analyses of variance results on the most abundant fish at different times (day and night), in seagrass beds of different heterogeneity (patchy and uniform) and size. * P<0.05, **P<0.01. Species Name Time Heterogeneity Size Acanthalutere spilomelanurus * ns ns Arenigobius frenatus * ns ns Atherinomorus ogilbyi * ns ns Bathygobius kreffti ** * ns Centropogon australis P = 0.056 ns ns Girella tricuspidata ns ** ** Pelates sexlineatus ns * ns Rhabdosargus sarba ns ns * Stigmatophora nigra ns * ns Urocampus carinirostris ns ns ns
55 Testing the Species Area Relationship and Landscape Ecology Theories
3.3.2 Size of the seagrass bed There were greater numbers of fish species and individual fish per net in the small seagrass beds compared to the medium and large beds (Fig. 3.3; Tables 3.4 & 3.7, SNK small > medium = large). For the number of fish individuals there was also a significant interaction at the level of seagrass beds. There were also differences in the composition of fish assemblages between small, medium and large beds although the R-values were quite low (Table 3.6). A pair-wise test found the greatest difference in assemblage composition was between the small and large seagrass beds, while the medium and large beds were statistically similar (Table 3.6). The small and large seagrass beds were also found to be different in species composition too, so their differences are not restricted to fish abundances (Table 3.6).
The species richness (number of species per bed) from the day sampling was not different in small, medium and large beds (Fig. 3.5). At night, however, an increase in bed size had a decrease in species richness (Fig. 3.5). A few species of fish were found to be more numerous in the smaller beds than the medium and large beds (SNK test P < 0.05) including Rhabdosargus sarba, Ambassis jacksoniensis, Girella. tricuspidata and Gerres subfasciata (Tables 3.5 & 3.8).
3.3.3 The perimeter to area ratio of a seagrass bed There was no relationship between the perimeter area ratio and the mean number of fish species in beds during the day or night (Fig. 3.6). There was, however, a significant relationship between the perimeter to area ratio and the mean number of individual fish in each bed (Fig. 3.6). An increase in the perimeter area ratio of the bed had a corresponding increase in the density of individual fish. This relationship was more pronounced during the night than day. This result needs, however, to be viewed with caution because the last point in the plot has considerable leverage on the regression line and when deleted the slope of the line was not significant (R2 = 0.10, P = 0.27).
3.3.4 Testing for the passive sampling hypothesis There was no relationship between the rate of fish species accumulated and the size of each seagrass bed during the day (R 2 = 0.16, P > 0.05) or during
56 Testing the Species Area Relationship and Landscape Ecology Theories the night (R 2 =0.01, P > 0.05). The species accumulation curves show that the small, medium and large seagrass beds accumulated fish species at a similar rate during day and night sampling (Fig. 3.7).
3.3.5 The distance of the seagrass bed from the estuary mouth There was no difference in the density of fish species or individuals between beds near to and far from the mouth of the estuary (Table 3.7). There was a significant interaction for the number of fish individuals at the time x estuary x size level. There were, however, differences detected in the species composition between beds at the two distances (Fig. 3.4; Table 3.6). The fish species contributing to the differences in assemblages were Bathygobius kreffti, Acanthalutere spilomelanurus, and Arenigobius frenatus. The mean abundances of these three fish species were greater in seagrass beds close to than far from the estuary mouth. Conversely, the mean abundances of next discriminating fish; Urocampus carinirostris, Pelates sexlineatus, and Rhabdosargus sarba were greater in seagrass beds far than close to the mouth of the estuary. The last discriminating species, Centropogon australis, was more abundant in beds close to the mouth of the estuary.
The analysis of variance supported some of these findings (Table 3.8). Achoerodus viridus, Bathygobius kreffti, Centropogon australis and Stigmatophora nigra were also more abundant in beds close to the mouth of the estuary. Only two fish, Atherinomorus ogilbyi and Pelates sexlineatus were more abundant in seagrass beds far from the estuary mouth.
57 Testing the Species Area Relationship and Landscape Ecology Theories
Table 3.6. A two way crossed analysis of similarities comparing fish assemblages in seagrass beds with time of sampling (day or night) as the first factor, and the second factor being either; heterogeneity (grouped into patchy and uniform beds); size of the seagrass bed (grouped into large, medium and small beds); or distance from the estuary mouth (grouped into beds close and far groups) * P < 0.05, ** P < 0.01.
Square root Presence/absence Factor tested transformation transformation Global /paired R Global /paired R Time (averaged across spatial groups) 0.132 ** 0.096** Heterogeneity (averaged across time groups) 0.060 ** 0.124** Time (averaged across size groups) 0.152 ** 0.094 ** Size of beds (averaged across time groups) 0.119 ** 0.067** Pairwise tests: small vs large 0.128 ** 0.199** medium vs. large 0.042 ns 0.017ns small vs. medium 0.128 ** 0.029ns Time (averaged across estuary groups) 0.174 ** 0.133** Estuary distance (averaged across time groups) 0.317 ** 0.362**
58 Testing the Species Area Relationship and Landscape Ecology Theories
Table 3.7. The analysis of variance comparing the number of fish species and fish individuals per net from seagrass at different times (day and night), from beds of varying distance from the estuary mouth (close or far) and size (small, medium or large). The data for the fish individuals was transformed, Ln (x+1) *P< 0.05. Mean Mean Source of variation df squares F squares F No. fish species / net No. fish individuals / net Time 1 73.5 10.17* 1.9 9.64* Estuary distance 1 7.0 0.81 0.7 0.36 Size 2 51.2 5.86* 10.2 5.22* Beds (Est. x Size) 6 8.7 1.25 1.9 5.81** Time x Estuary 1 18.4 2.54 1.1 5.59 Time x Size 2 4.5 0.62 0.1 0.29 Time x Bed (Est. x Size) 6 7.2 1.04 0.2 0.58 Estuary x Size 2 2.2 0.25 3.5 1.81 Time x Estuary x Size 2 13.5 1.87 1.6 8.13* Residual 72 6.9 0.3
Table 3.8. The summary of the ANOVA comparing the most abundant fish in beds close to and far from the estuary mouth in Pittwater, spring 2000. *P > 0.05, ** P > 0.01, *** P > 0.001. Species Time Region in estuary Size of bed Achoerodus viridis ns * ns Ambassis jacksoniensis ns ns *** Arenigobius frenatus ns ns ns Atherinomorus ogilbyi ** * ns Bathygobius kreffti ns * ns Centropogon australis ns * ns Gerres subfasciata ns ns * Girella tricuspidata * ns * Pelates sexlineatus ns * ns Rhabdosargus sarba ns ns * Stigmatophora nigra * * ns Urocampus carinirostris ns ns ns
59 Testing the Species Area Relationship and Landscape Ecology Theories
Patchy seagrass beds
Uniform seagrass beds R = 0.060 P < 0.001
Seagrass beds far from estuary mouth
Seagrass beds close to estuary mouth
R = 0.317 P < 0.001
Figure 3.4. Two-dimensional configurations for MDS ordinations (square root transformation) of the composition and abundance of fish in seagrass beds in Pittwater, spring 2000. The samples in the plots represent patchy and uniform seagrass beds; or seagrass beds close and far from the estuary mouth.
60 Testing the Species Area Relationship and Landscape Ecology Theories
Day, species richness Night, species richness 30 30 R2 = 0.01 R2 = 0.30 25 25 P = 0.72 ns P < 0.05 20 20 15 15
10 10
5 5 No. fish species / bed / species fish No.
0 bed / species fish No. 0 0 2000 4000 6000 8000 0 2000 4000 6000 8000 2 2 Area of bed (m ) Area of bed (m )
Day, total number of fish Night, total number of fish
1400 700
1200 600 R2 = 0.38
1000 500 P < 0.025 R2 = 0.06 800 400 600 P = 0.39 ns 300 400 200 200 Total no. of fish / bed fish of / no. Total 100 Total no. of fish / bed / fish of no. Total 0 0
0 2000 4000 6000 8000 0 2000 4000 6000 8000 2 Area of bed (m ) Area of bed (m2)
Figure 3.5. Plot of the species richness and total number of fish individuals against the area of each seagrass bed from the day and night sampling in Pittwater, spring 2000.
61 Testing the Species Area Relationship and Landscape Ecology Theories
DAY, fish species NIGHT, fish species 16 16 14 14 12 12 10 10 8 8
R2 = 0.08 6 R2 = 0.02 6 4 P = 0.31 4 P = 0.65 No. fish species / net / species fish No. 2 net / species fish No. 2 0 0 0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4
perimeter area ratio perimeter area ratio
DAY, fish individuals NIGHT, fish individuals 350 180 300 160 250 140 120 200 100 R2 = 0.36 150 80 R2 = 0.26 P < 0.05 100 60 P < 0.05 50 40 fish individuals/No. net 20
0 fishnet individuals/ No. 0 0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 perimeter area ratio perimeter area ratio
Figure 3.6. Plots of the mean number of fish species and fish individuals per net against the perimeter area ratio of each seagrass bed during the day and night sampling, in Pittwater, spring 2000.
62 Testing the Species Area Relationship and Landscape Ecology Theories
Day sampling 18 16 14 12 small Accumulated 10 no. fish medium species 8 6 large 4 2
1 2 3 4
Number of net hauls
25 Night sampling
20
Accumulated small no. of fish 15 species medium 10 large
5
1 2 3 4
Number of net hauls
Figure 3.7 The species accumulation curves for day and night sampling in small, medium and large seagrass beds in Pittwater, spring 2000.
63 Testing the Species Area Relationship and Landscape Ecology Theories
3.4 Discussion
3.4.1 Comparison of day & night sampling The numbers of fish species per net were greater during the night sampling than the day. This finding supports the work or other researchers (Bell & Harmelin-Vivien 1982; Gray et al. 1998; Griffiths 2001) and it suggests that fish sampling programmes in estuaries need to incorporate night sampling to estimate fish assemblages in seagrass beds. One reason for this difference in sampling effectiveness may be that the fish can successfully detect and avoid the net during the day than the night. The fish may move to deeper water to avoid being exposed to avian predators during the day. This avoidance would not be necessary during the night low tide. Alternatively, more fish species may be nocturnal and therefore more likely to be caught at night.
3.4.2 Heterogeneity of seagrass and the assemblage of fish In this study there were different compositions of fish assemblages in the heterogeneous environment (patchy seagrass beds) than in the homogeneous environment (uniform seagrass beds). The heterogeneity of a seagrass bed however did not influence the abundance of fish species nor fish individuals. This is in contrast to the model that heterogeneous environments promote diversity of species (i.e. increase the number of species) when compared to homogeneous environments (Heck & Orth 1980; Irlandi & Crawford 1997; Eggleston et al. 1999). The results from this study do not suggest that increasing habitat diversity will increase the species diversity of an area for small fish in seagrass beds. It should be noted, however, that this survey compared only one habitat (seagrass) with two (seagrass and sand) and that even the uniform beds still contain some sand. A greater number of habitats such as other seagrass species or macro-algae species may have produced different results. For example, Knowles and Bell (1998) found a greater abundance of epifauna on drift algae in seagrass beds compared to seagrass habitat alone as have other researchers (Virnstein & Howard 1987).
The patchiness of seagrass has been found to influence biological interactions such as competition and predation (Coen et al. 1981; Irlandi
64 Testing the Species Area Relationship and Landscape Ecology Theories
1994; Hovel & Lipcius 2002) and community composition (Frost et al. 1999). In this study, the small-scale (1-4m) structure of the seagrass bed was correlated with the fish assemblages and it appeared that this favoured some fish species. In patchy seagrass beds, some mobile fish species such as G. tricuspidata and P. sexlineatus were more numerous while in uniform seagrass beds other more cryptic species such as B. kreffti and S. nigra were more numerous. The apparent lack of a relationship between the patchiness of a seagrass bed and number of fish species per net suggests that small fish are considerably mobile and more reliant on the macro- habitat (i.e. the whole bed) than the micro-habitat (i.e. cover of seagrass within the bed). This is supported by the work of Bell et al. (1987), who found that the settlement of juvenile fish was not based on physical complexity (i.e. leaf density) of seagrass, but had more to do with the number of larvae prepared to settle into any available shelter.
3.4.3 Small beds have greater fish densities and species richness The smaller beds of Z. capricorni (980 to 2300 m2) were found to have greater numbers of fish species and fish individuals per net than the medium (3375 to 4090 m2) and large beds (5335 to 6630 m2). The smaller seagrass beds also contained greater species richness (number of fish species per bed) than the medium and large beds, although this was statistically significant only during the night. These findings do not support the area -per se model, which proposed that larger seagrass beds have more species because they have lower rates of local extinction and higher rates of successful immigration. Other models may need to be considered for the assemblages of fish in seagrass beds.
McNeill and Fairweather (1993) hypothesised that one of the reasons for greater species richness of fish in smaller than large seagrass beds is the increased likelihood of sampling an edge in a small seagrass bed. Small beds could be considered all edge-habitat and perhaps the edges of seagrass beds contain greater abundances of fish or species richness than the interior regions. There are demonstrated changes in seagrass beds as a function of the distance from the edge such as a decrease in water flow from the edge to the centre (Fonseca et al. 1982). Evidence for the existence of edges having an affect on faunal assemblages is supported by
65 Testing the Species Area Relationship and Landscape Ecology Theories the positive relationship between the perimeter area ratio of a seagrass bed and the number of fish per net. The seagrass beds with high perimeter to area ratios (with a greater amount of edge habitat) had greater numbers of fish individuals per net than the beds with a low perimeter to area ratio. This relationship was, however, statistically weak and also difficult to support given that small seagrass beds will also have a high perimeter ratio so that the effect of habitat size and perimeter area cannot be separated. The detection of edge effects in small and large seagrass beds for fish species and their influence needs to be tested.
An alternative explanation for the greater numbers of fish species per net in a small seagrass bed is the limited area available for settlement. If a small and large seagrass beds receive similar amounts of fish larvae, then larvae may settle in the small bed in greater densities than in the large seagrass bed, particularly if the species have low dispersal rates. Fish larvae are known to migrate to seagrass beds within weeks or months after the initial recruitment into an estuary (Bell & Westoby 1986; Loneragan et al. 1986; Bell et al. 1987; Jenkins et al. 1996) and remain in the estuary for at least a year (Potter et al. 1983; Hannan & Williams 1998). The larvae of some commercial and recreational fish can also remain in seagrass beds for several months after settling from the plankton, before moving to other habitats (Middleton et al. 1984; Worthington et al. 1992). Thus during this period of low dispersal, fish larvae may concentrate in greater densities in smaller than in larger seagrass beds. This does not explain, however, why the smaller seagrass beds have greater species richness at night.
Some researchers have suggested that smaller patches may provide refuge for species that are poor competitors that would be out-competed in larger patches (Jackson 1977; Kay & Keough 1981; Schoener & Schoener 1981; Keough 1984). Shallow seagrass beds are known to have few large piscivorous fish, which increase their likely-hood of being sanctuary areas for juvenile fish (Whitfield & Blaber 1978; Blaber & Blaber 1980; Blaber 1980). Other studies have found the survival of macroinvertebrates to be greater as the size of a seagrass patch decreases (Hovel & Lipcius 2001). For example, Hovel & Lipcius (2001) found that the survival of juvenile blue crabs increased as the size of the seagrass patch decreased, because the
66 Testing the Species Area Relationship and Landscape Ecology Theories smaller beds of seagrass could not support its main predator the adult blue crab. The density of predators may thus be reduced in smaller than larger seagrass beds and therefore an abundance of small and juvenile fish can survive.
Alternatively, the density of prey organisms for fish may be greater in small than large seagrass beds. Crustaceans are the major food item in the diet of most seagrass fish (Burchmore et al. 1984; Robertson 1984; Edgar & Shaw 1995a, 1995b), because few fish are capable of using plant material (Edgar & Shaw 1995a, 1995b). If smaller seagrass beds have greater densities of crustaceans, then they may also support greater densities of fish.
The findings of this study are restricted to small fish because large fish were not captured nor were they the focus of this study. The smallest scale at which an organism responds to patch structure has been described as its ‘grain’ (Kotliar & Wiens 1990). It is thought that smaller organisms may have a smaller ‘grain’ than larger organisms. Larger organisms will functionally perceive a mosaic of patches as a single habitat whereas smaller organisms will generally perceive it as many habitats (Kotliar & Weins 1990). Perhaps only juvenile and small fish will respond to the size of a seagrass bed. An adult or larger fish would have the mobility to use numerous seagrass beds as habitat.
In fact, the different fish assemblages in the small beds compared to the larger beds indicates that there was no self-similarity across multiple scales (Robbins & Bell 1994). As the size of the seagrass bed increased, the assemblages of fish changed. This demonstrates that the landscape concept of similar patterns across multiple scales was not evident.
3.4.4 Passive sampling hypothesis The passive sampling hypothesis was tested to ensure that a sampling artefact was not involved when comparing the diversity of fish in large and small seagrass beds. In a study of the macroinvertebrates assemblages in Zostera marina seagrass beds of Devon UK, a relationship was found between the biomass of the seagrass and the number of macroinvertebrate species (Attrill et al. 2000). This indicated a species area relationship, but
67 Testing the Species Area Relationship and Landscape Ecology Theories one brought about by a random sampling artefact in which increasing the area of seagrass sampled had a concurrent increase in the proportion of the macroinvertebrate population randomly sampled.
In the current study, to ensure an equal sampling effort in all seagrass beds, the same number of samples was collected. This sampling regime, however, meant that the proportion of habitat sampled in a small seagrass beds was much higher than that sampled in a large seagrass bed, similar to the random sampling artefact encountered by Attrill et al. (2000). For this reason, species accumulation curves were performed for small, medium and large seagrass beds and no relationship was found between the size of a bed and the accumulation of fish species. Therefore, the passive or random sampling model can be refuted and other models can be examined.
3.4.5 Location of the bed in estuary influences the fish assemblage The assemblages of small fish in seagrass beds were significantly different in beds close than far from the estuary mouth, although there were no differences detected in the densities of fish individuals and fish species. Previous work in the same estuary (Bell et al. 1987) found that the abundance of juveniles of many species of fish were affected by the location of the bed within the estuary which they explained due to the combined effects of spawning location and the dispersal of eggs and larvae. Studies within other estuaries have also supported the zonation of certain seagrass fish species (McNeill et al. 1992; Jenkins et al. 1996; West & King 1996; Hannan & Williams 1998). In the present study, some species of fish were more numerous in seagrass beds near the mouth of the estuary such as Achoerodus viridis, Centropogon australis, Bathygobius kreffti, and Stigmatophora nigra. Other species were found to be more numerous in seagrass beds far from the estuary mouth such as Atherinomorus ogilbyi, and Pelates sexlineatus. These findings would suggest that the marine and freshwater processes affect the distribution and abundance of some species of fish. However, the Pittwater does not have strong salinity gradients. Thus the movement or dispersal of the eggs, larvae or juvenile fish within the estuary may be more important in determining the distribution of individual fish species (Bell & Westoby 1986b).
68 Testing the Species Area Relationship and Landscape Ecology Theories
3.4.6 Implications for the species-area relationship and marine conservation The three main models that attempt to explain the species-area relationship have all been refuted within this study. Habitat heterogeneity influenced the fish assemblages of a seagrass bed, but not the number of fish species. The relationship between size of habitat and species richness was the reverse of what was predicted by the species-area relationship. The passive sampling model was also not supported by this study and so alternative models need to be considered.
In conclusion, the size of a bed was the most influential feature on the number of small fish species. For this reason, small seagrass beds cannot be overlooked when designing protected areas in estuaries. This study does not recommend that small seagrass beds should be targeted more than large seagrass beds for protection, because it is not known if the larger beds are supplying recruits for the smaller beds. It does indicate, however, that small seagrass beds need to be included when designing protected areas in the estuarine environment. The fish assemblages were also related to the location of a seagrass bed within the estuary. This suggests that the seascape approach to estuarine conservation is more appropriate than attempting to examine the optimal size of a protected area. Seascape paradigms are required for the marine or estuarine environment based on water movement and the availability and survival of competent larvae as well as the provision of appropriate habitats. Most marine animals require numerous habitats for the completion of their life cycle and this also confounds the issue of optimal size of habitat for target species.
69 Do the edges of seagrass beds influence fauna?
4 Do the edges of seagrass beds influence small fish and macroinvertebrates? 4.1 Introduction Seagrass beds are an important habitat for fish and macroinvertebrates in estuaries (Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984), but they have become increasingly fragmented in urbanised areas of temperate Australia (Larkum & West 1990; Walker & McComb 1992). Human activities and other natural processes have reduced continuous seagrass beds into smaller patches or remnants isolated by sand (Thayer et al. 1975; Shepherd et al. 1989; Larkum & West 1990; Walker & McComb 1992), which is a structurally dissimilar matrix (Cox et al. 2003). This fragmentation is expected to have consequences for fish and macroinvertebrates (Irlandi et al. 1995; Irlandi 1997; Bell et al. 2001). In terrestrial habitats, smaller remnants can have lower species diversity and abundance (MacArthur & Wilson 1967; Diamond & May 1976; Simberloff & Abele 1976) however in the aquatic environment this may not follow.
In fact, small seagrass beds have been found to contain a greater abundance or density of organisms when compared to large beds and this has been attributed to the influence of edge effects (McNeill & Fairweather 1993; Eggleston et al. 1998, 1999; Irlandi et al. 1999; Chapter 3). The edges of seagrass patches have been found to contain greater abundances of some fauna than the interiors (Holt et al. 1983; Bologna & Heck 2000; Barbera-Cebrian et al. 2002; Hovel & Lipcius 2002, Tanner 2004) so researchers have considered the likelihood of sampling an edge to be the reason why small seagrass beds may contain more fauna (McNeill & Fairweather 1993).
McNeill and Fairweather (1993) consider there to be an increased likelihood of sampling an edge in a small seagrass bed. Indeed, small beds could be considered to be all edge because a decrease in the size of a seagrass bed increases its perimeter to area ratio and the proportion of edge. In the previous chapter smaller seagrass beds were found to contain greater numbers of small fish species per net than larger seagrass beds. Similarly, seagrass beds with a high perimeter to area ratio (and therefore a greater
70 Do the edges of seagrass beds influence fauna?
proportion of edge habitat) were found to have greater abundances of individual small fish than seagrass beds with a low perimeter to area ratio. Both results could be attributed to the existence of an edge-mediated effect.
Numerous studies have considered the influence of edge on the epifauna and infauna of seagrass beds (Holt et al. 1983; Summerson & Peterson 1984; Sanchez-Jerez et al. 1999; Bologna & Heck 1999; 2000; Bell et al. 2001; Bowden et al. 2001; Barbera’-Cebrian et al. 2002; Hovel & Lipcius 2002; Tanner 2004). Some studies have found fauna to be more abundant on the edges of seagrass beds than the interiors (the red drum fish, Holt et al. 1983; juvenile bivalves, Bologna & Heck 2000; mysids, Barbera-Cebrian et al. 2002; crabs, Hovel & Lipcius 2002; crustaceans, Tanner in press). Two studies also detected a unique epifaunal or infaunal assemblage associated with edge (Bowden et al. 2001; Bologna & Heck 2002) although another study on epifaunal assemblages did not support this finding (Sanchez-Jerez et al. 1999).
The edges of seagrass beds may impact seagrass fauna depending on the nature of the ecotone (a region where the two habitats meet) and the organisms that use it. An ecotone can be a region of enhanced biodiversity because it provides two habitats for shelter; it may enhance biotic interactions (such as predation or competition) or allow the mix of two different communities (Leopold 1933; Fox et al. 1997; Lidicker 1999). Conversely, ecotones have also been shown to decrease biodiversity because of the increased risk of predation and the loss of habitat from the invasion of exotic species (Andrèn et al. 1985; Yahner & Scott 1988; Fox et al. 1997). There are demonstrated changes in seagrass beds from the edge to the centre such as a decrease in water flow (Fonseca et al. 1982; France & Holmquist 1997) that may facilitate the influence of edge effects in seagrass beds.
The consequences of seagrass fragmentation on seagrass fauna were investigated by Bell et al. (2001) who reviewed the effects of seagrass patch size. In their own experimental work they could not identify a preferential use of the edge or interior by seagrass taxa. However, they made the assumption that if the fauna responded to the patch size of a seagrass bed
71 Do the edges of seagrass beds influence fauna? then this would infer a preferential use of the edge or interior. This assumption may be correct, but it needs to be tested by sampling the edges and interiors of different sized seagrass beds.
For this reason the aim of this study was to sample the edges and interiors of small and large seagrass beds for small fish and mobile macroinvertebrates. To test if the interactions of organisms are altered by edge effects a suite of seagrass fauna (fish and macroinvertebrates) was sampled so as to include predator and prey species.
The hypothesis tested is that there will be different numbers of fish species or macroinvertebrates in the edges compared to the inner regions of small and large seagrass beds. The edge and interior regions of the small seagrass beds are expected to be similar (i.e. all edge habitat) and different from the inner regions of the large seagrass beds. Furthermore, the relationship between the fish and macroinvertebrates is predicted to be different between the edge and inner regions of small and large seagrass beds.
4.2 Methods 4.2.1 Selection of the seagrass beds The Pittwater is classified as a modified estuary, with dense urban development on the eastern and southern shore and extensive boating and recreational use (Geoscience Australia 2001). This makes it a good choice for determining the effects of seagrass fragmentation on fauna.
Six mono-specific beds of Zostera capricorni were selected based on their area, location in the lower reaches of the estuary (Fig.4.1) and water depth (30-100 cm at mean low tide). In chapter 3 it was demonstrated that fish assemblages were correlated with the location of a seagrass bed within the estuary (ANOSIM, Global R = 0.316, P < 0.01). For this reason, the seagrass beds used in this study were from the same region, within 4 km of the estuary mouth. Three of the beds were grouped as small in area (0.22 – 0.40 ha) and the other three as large (0.66- 2.11 ha). All beds were uniform in seagrass cover as defined in Chapter 3. Each seagrass bed was separated from another bed by at least 200 m of bare sandy substratum and usually by
72 Do the edges of seagrass beds influence fauna?
much more. This maintained independence of sampling because it reduced the likelihood of faunal movement between seagrass beds.
N Estuary mouth The Pittwater
Scotland Is.
1km
Figure 4.1. A map of the Pittwater in Broken Bay NSW and the Zostera capricorni beds surveyed (blue = small, green = large) in autumn and spring 2001. The Z. capricorni beds not used for this survey are shown in yellow.
73 Do the edges of seagrass beds influence fauna?
4.2.2 Defining the edge, inner and central regions of the beds The edge of a seagrass bed was defined as the outer perimeter region, not including the sand. Peterson and Turner (1994) found the outer three metre region of saltmarsh to have the greatest densities of fish, while other studies (Bologna & Heck 2000; Tanner 2004) of infauna and epifauna defined the edge region (in seagrass beds) as the one metre outer perimeter. Given the high mobility of the fauna examined in this study, the edge was designated as four metres in width. Furthermore, the size of the small beds limited the size of the edge regions. The designation of a four metre edge was in proportion to the smallest bed (0.22 ha or 220 m2 ) and enabled an edge and inner region to be sampled within the bed. The size of the edge region was also defined in practice by the size of our seine net (8 m wide, but sampling 4 m when pulled). Seine nets are one of the more effective methods for capturing small fish in seagrass beds (Connolly 1994; Guest et al. 2004), and the net used here was ideal for comparisons of fish assemblages among beds. The deeper seaward edges of the seagrass beds could not be sampled because the water was over 100 cm deep, even on the low tide. All other edges, including those parallel and perpendicular to the shore, were sampled.
Within both small and large seagrass beds, two regions were designated edge and inner. In the large seagrass beds there was also a third (central) region because of the larger size of these beds (Fig. 4.2). A buffer zone of two metres separated all three regions (edge, inner and central). The edge region was 0 – 4 m from the outer perimeter of the seagrass bed; the inner region was 6 to 10 m from the perimeter and the central region of the large beds was 12 to 16 m in distance. Although called the central region, this region was not always consistently in the direct centre of any bed (Fig. 4.2). These regions were measured each sampling time from the perimeter of the seagrass bed.
4.2.3 Fish and macroinvertebrate survey All six beds were randomly sampled in time and space for small fish and macroinvertebrates during March to early May 2001 (autumn) and from September to November 2001 (spring). This was done during the evening
74 Do the edges of seagrass beds influence fauna?
(after dusk) because this was found to be the time when sampling for fish gave the greatest species richness (Chapter 3).
In chapter 2 it was found that sampling for fish on the low to mid tide (water depth 30-100cm) ensured that the water depth was not correlated with the number of fish species or individuals caught. Thus an 8 x 2 m seine net (1 mm mesh) was randomly placed and dragged through each seagrass bed on the low to mid tide, sampling an area approximating 35 m2. Each region of the seagrass bed was sampled three times per season. After each sample, a GPS location was recorded to ensure that the same area was not sampled again on a subsequent evening. This maintained independence of sampling and prevented violation of the assumptions of the factorial analysis of variance (WinGmav5) used in this study (Underwood 1997).
Edge region
0 to 4 m
Inner region 6 to 10 m
Large beds only Central region 12 to16 m
Figure 4.2. A diagram representing the regions (edge, inner and central) of the seagrass beds sampled in autumn and spring in the Pittwater, 2001. There was a two-metre buffer zone separating one region from another. There was no central region in the small seagrass beds. The numbers represent the distance of the region from the outer perimeter of the seagrass bed.
75 Do the edges of seagrass beds influence fauna?
Size Small beds Large beds
n = 2
Beds Bed 1 Bed 2 Bed 3 Bed 4 Bed 5 Bed 6 n = 3
Regions Edge Inner Edge Inner Centre n = 2, 3
Replicates = 3 hauls of net
Figure 4.3. A diagram representing the experimental design of the study. The first level of the design was time (autumn and spring) and is not shown. This was a fixed factor. The second level was the size of seagrass bed (fixed, orthogonal), the third level was seagrass bed (random, nested) and the fourth level was region of seagrass bed (fixed, orthogonal).
4.2.4 Data analysis Analysis of variance (ANOVA) tests were used to compare the number per net of fish species, fish individuals, macroinvertebrates and decapods individuals, and individuals of the most numerous fish species between the small and large seagrass beds and their regions. To ensure homogeneity of variances, all data was tested for heteroscedasticity and transformed Ln (x+1) if a Cochran’s C-test proved to be significant. Each analysis contained a combination of fixed (sampling time, size and region of bed) and random factors (beds nested in size). A Student Newman Kuels (SNK) test was used to detect post-hoc differences among the means (see Fig. 4.3 for the experimental design). Two separate ANOVAS were performed due to the unbalanced design of the survey. One ANOVA compared the edge and inner regions of the small and large seagrass beds over the two seasons. The second ANOVA compared the three regions (edge, inner and central) only in the large beds over both seasons.
A linear regression was used to test if there were a relationship between the number of invertebrate individuals and the number of fish species or fish
76 Do the edges of seagrass beds influence fauna?
individuals captured per net. The regression analyses examined the abundances during each season separately and together for the following; all the seagrass beds, for small and large beds separately and for the different regions of the seagrass beds (edge, inner, centre). A Bonferroni correction was used for these multiple tests. It was necessary to do these regression analyses separately because it was considered that a relationship may be limited to a region of a seagrass bed or a certain size bed or season. To exercise caution, in the results the regression analyses will calculate an R2 value but only a relationship or correlation will be concluded and not a direct cause and effect.
A non-metric multi-dimensional scaling (MDS) plot was obtained by performing a Bray-Curtis similarity measure between the samples after using a square root transformation of the data.
A multivariate analysis (Analysis of Similarity; ANOSIM) was used to determine any significant differences in the assemblage of fish and macroinvertebrates between sampling times, small and large beds and bed regions (Clarke and Warwick 2001). To detect which species of fish or macroinvertebrate groups were contributing to the differences in assemblage composition, a similarity percentages (SIMPER) analysis was used (Clarke & Warwick 2001).
4.3 Results
A total of 53 fish species and four orders of macroinvertebrates were collected in this study (Tables 4.1 and 4.2). The most numerous macroinvertebrate group was the decapods (Table 4.2). There were significantly greater numbers per net of fish species in the small compared to the large seagrass beds (Fig. 4.4; Table 4.3, SNK small > large) although this appears to be largely attributed to the density of fish species in spring. However the ANOVA did not detect differences between seasons across all beds, although one small seagrass bed had significantly more fish species per net in autumn than spring (Table 4.3, interaction time x bed(size)). No significant differences were detected in the number per net of fish species between; the edge and inner regions of the small and large beds; and the
77 Do the edges of seagrass beds influence fauna? edge, inner and central regions of the large seagrass beds (Fig. 4.5; Tables 4.3 & 4.4). The number per net of fish and macroinvertebrate individuals were also statistically similar between small and large seagrass beds and between regions of the beds (Fig. 4.5; Tables 4.3, 4.4, 4.5). However there was a trend for a greater number per net of fish individuals in autumn and a significant difference between beds.
One small seagrass bed had significantly less fish individuals than the two other small beds, and one large bed had significantly more fish individuals than another. There was also an interaction term (time x bed (size)) for the fish individuals whereby one large bed had greater numbers per net during spring than autumn and vice versa for another bed.
There was a significant difference between beds for the number per net of macroinvertebrate and decapod individuals whereby the smaller beds were significantly different from one another and one large bed had greater numbers per net than another. There was also an interaction term (time x size) that will be discussed further. Overall, there were no differences between regions (edge, inner and centre) detected for the number per net of fish species, fish individuals or macroinvertebrates.
The numbers per net of the macroinvertebrates and decapods in the small and large seagrass beds were variable (Fig. 4.4; Table 4.5 interaction time x size). The number per net of decapod individuals revealed identical patterns of distribution (Table 4.5) compared to the macroinvertebrates, suggesting this group was responsible for the overall detected patterns and for this reason only the decapod data was plotted (Figures 4.4, 4.5). During autumn the numbers of macroinvertebrates and decapods were greater in the small compared to the large seagrass beds. In contrast during spring, numbers were greater in the large compared to the small seagrass beds (Fig. 4.4). In the small beds there were greater numbers during autumn than spring, and in the large beds the numbers were greater in spring than autumn (Fig. 4.4).
78 Do the edges of seagrass beds influence fauna?
10 9 Fish species 8 No. fish 7 6 species 5 per net 4 a b a b 3 2 1 0 small beds large beds small beds large beds
autumn spring
140 Fish individuals 120 100 No. fish 80 individuals 60 per net 40
20
0 small beds large beds small beds large beds autumn spring
300 Decapod individuals 250
No. 200 decapods 150 per net a a 100
50 b b
0
small beds large beds small beds large beds
autumn spring
Figure 4.4. The mean abundance (±S.E.) of seagrass fauna in small and large seagrass beds during autumn and spring 2001. The letters (a & b) denote the groups that were statistically similar or different from one another.
79 Do the edges of seagrass beds influence fauna?
10 Fish species 9 No. fish 8 7 species 6 5 per net 4 3 2 1 0
edge inner edge edge inner edge inners inners centres centres small large small large
autumn spring 150 Fish individuals
100 No. fish
individuals 50 per net
0
edges inners edges inners edges inners edges inners centres centres
small large small large
autumn spring 400 Decapods
No. 300 decapods 200 per net 100
0
edges inners edges inners edges inners edges inners centres centres small large small large
autumn spring Figure 4.5. The mean abundance (±S.E.) of seagrass fauna in the different regions of small and large seagrass beds during autumn and spring in Pittwater 2001. No statistical differences were detected between the edge, inner and central regions for any faunal group.
80 Do the edges of seagrass beds influence fauna?
Table 4.1. The total abundance of fish species in the edges, inner and central regions of small (n=3) and large (n=3) seagrass beds in Pittwater, autumn and spring 2001.
Small Beds Large Beds Family Species Edge Inner Edge Inner Centre
Apogonidae Apogon cookii 44 0 0 0 Vincentia novaehollandiae 04 0 0 1 Atherinidae Atherinomorus ogilbyi 33 26 102 34 28 Blenniidae Petroscirtes lupus 10 0 0 0 Chandidae Ambassis jacksoniensis 0 0 33 115 451 Clinidae Cristiceps aurantiacus 00 3 2 0 Heteroclinus fasciatus 00 1 4 1 Heteroclinus whiteleggi 10 1 0 2 Clupeidae Hyperlophus translucidus 51 1 1 0 Diodontidae Dicotylichthys punctulatus 01 0 0 0 Gerreidae Gerres subfasciatus 0 0 34 87 43 Girellidae Girella tricuspidata 00 5 4 6 Gobiidae Arenigobius frenatus 292 273 742 522 562 Bathygobius kreffti 121 103 72 58 45 Cristatogobius gobioides 00 1 1 1 Redigobius macrostoma 10 8 10 2 Labridae Achoerodus viridis 00 0 1 0 Monacanthidae Acanthaluteres spilomelanurus 15 1 4 5 Cantherhinus pardalis 31 0 2 8 Eubalichthys mosaicus 36 0 0 0 Meuschenia trachylepis 10 0 0 0 Meuschenia venusta 20 0 0 0 Monacanthus chinensis 33 1 2 3 Scobinichthys granulatus 12 1 1 2 Mullidae Parupeneus signatus 11 1 0 0 Upeneichthys lineatus 30 0 0 0 Upeneus sp. 16 16 4 0 0 Upeneus tragula 22 28 3 5 3 Odacidae Neoodax balteatus 00 0 1 0 Ostraciidae Lactoria cornuta 00 0 0 1 Paralichthyidae Pseudorhombus jenynsii 01 0 0 0 Plotosidae Cnidoglanis macrocephala 00 1 1 0 Plotosus lineatus 00 0 1 0 Scorpaenidae Centropogon australis 159 120 136 118 94 Serranidae Epinephelus daemelii 00 1 0 0 Siganidae Siganus nebulosus 11 0 1 0 Sillaginidae Sillago ciliata 10 0 0 0 Sillago maculata 10 1 2 1 Sparidae Rhabdosargus sarba 12 8 36 40 32 Sphyraenidae Sphyraena obtusata 21 0 1 0 Syngnathidae Filicampus tigris 23 0 0 0 Hippocampus whitei 10 0 0 0 Stigmatopora argus 00 3 0 0 Stigmatopora nigra 24 17 7 6 10 Urocampus carinirostris 53 23 34 34 49 Vanacampus margaritifer 00 0 1 1 Tetraodontidae Tetractenos hamiltoni 00 1 1 0 Terapontidae Pelates sexlineatus 28 25 127 227 205
81 Do the edges of seagrass beds influence fauna?
Table 4.2. The total abundance of macroinvertebrates groups in the edges, inner and central regions of small (n=3) and large (n=3) seagrass beds in Pittwater, autumn and spring 2001.
Small Beds Large Beds Order Family / species name Edge Inner Edge Inner Centre
Amphipoda 32 34 157 67 102 Cephalopoda Idiosepiidae Idiosepius notoides 36 27 54 24 52 Loliginidae 0 1 3 2 3 Octopodidae 0 1 0 0 0 Sepiidae 0 5 0 0 0 Sepiolidae Sepioloidea lineolata 56 35 29 21 26 Decapoda Portunidae 11 8 31 34 22 Alpheidae, Penaeidae, 2649 3105 3071 2177 2327 Palaemonidae Isopoda 17 18 9 11 11
Table 4.3. An analysis of variance (ANOVA) comparing the number per net of fish species and fish individuals in seagrass beds during different sampling times (autumn and spring), from beds of different size (small or large) and from different regions within the beds (edge or inner) in Pittwater 2001. * P<0.05, **P<0.01, ***P<0.001. Source of Variation df Mean F Mean F Squares Squares No. of fish species No. of fish individuals per net per net Time 1 0.23 1.97 1.41 1.53
Size 1 0.14 10.73* 2.68 0.90 Beds(size) 4 0.01 0.32 2.99 17.11*** Region 1 0.01 0.10 0.14 0.33 Time x size 1 0.07 0.56 3.69 4.01 Time x beds(size) 4 0.12 2.78* 0.92 5.26**
Time x region 1 0.01 0.14 0.30 1.63 Size x region 1 0.19 3.95 0.24 0.57 Region x beds(size) 4 0.05 1.14 0.42 2.40 Time x size x region 1 0.03 0.64 0.14 0.77 Region x time x beds(size) 4 0.05 1.15 0.19 1.06 Residual 48 0.04 0.18 Total 71
82 Do the edges of seagrass beds influence fauna?
Table 4.4. An analysis of variance (ANOVA) comparing the number per net of fish species, fish individuals and invertebrate individuals collected from the large seagrass beds during different sampling times (autumn and spring), and the different regions (edge, inner and central) in Pittwater 2001. Source of Mean Mean Mean Variation df Squares F Squares F Squares F Number of fish Number of fish Number of macro- species / net individuals / net invertebrates / net Time 1 0.02 0.24 3.84 2.59 6.99 6.47 Beds 2 0.01 0.30 7.59 30.75*** 3.23 11.49*** Region 2 0.08 1.44 0.17 1.03 0.46 6.05 Time x beds 2 0.07 1.86 1.48 6.00** 1.08 3.84* Time x region 2 0.01 0.08 0.54 0.97 0.44 1.70 Beds x region 4 0.09 1.55 0.17 0.68 0.08 0.27 Time x beds x reg. 4 0.07 1.91 0.56 2.27 0.26 0.91 Residual 36 0.04 0.25 0.28 Total 53
Table 4.5. An ANOVA comparing the densities of macroinvertebrate individuals collected from seagrass beds during different times (autumn or spring), of different size (small or large) and from different regions within the beds (edge or inner) in Pittwater 2001. **P<0.01, ***P<0.001. Source of Variation df Mean F Mean F Squares Squares No. of macroinvertebrates / net No. of decapods / net Time 1 0.13 0.42 0.28 0.96 Size 1 0.86 0.32 1.04 0.37 Beds(size) 4 2.67 12.98*** 2.81 13.41*** Region 1 0.38 2.57 0.35 2.48 Time x size 1 10.60 33.93** 12.01 40.60** Time x beds(size) 4 0.31 1.52 0.30 1.41 Time x region 1 0.01 0.01 0.01 0.01 Size x region 1 0.53 3.59 0.55 3.89 Region x beds(size) 4 0.15 0.71 0.14 0.68 Time x size x region 1 0.68 0.68 0.74 1.74 Reg. x time x beds(size) 4 0.41 1.97 0.43 2.03 Residual 48 0.21 0.21 Total 71
83 Do the edges of seagrass beds influence fauna?
Analyses of the more numerous fish species found that the numbers per net of the eastern-striped trumpeter, Pelates sexlineatus, were greater in the inner than the edge regions of small and large seagrass beds, and greater in the central than edge regions of the large seagrass beds (Fig. 4.6; Table 4.6). This was the only fish to be found in different numbers per net
between regions. The cephalopod Idiosepius notoides (pygmy squid) was also found to be more numerous on the edges and central regions of the seagrass beds compared to the inner regions (Table 4.6).
16 Pelates sexlineatus 14 12 No. of 10 individuals 8 b per net 6 4 b 2 a 0 edge regions inner regions central regions
Figure 4.6. The number per net of the eastern-striped trumpeter Pelates sexlineatus in the edge, inner and central regions of all seagrass beds (large and small) in Pittwater, autumn and spring 2001. The letters (a & b) denote those groups that were statistically similar or different from one another.
84 Do the edges of seagrass beds influence fauna?
Table 4.6. Summary of the ANOVA comparing the abundances of single species of fish or macroinvertebrate groups from seagrass beds during different times (autumn and spring), from beds of different size (small or large) and from different regions within the beds (edge, inner or centre) in Pittwater. *P<0.05, **P<0.01, ***P<0.001. Species or Group Time Size Beds Regions Any further SNK results of (size) in bed significant bed interactions Arenigobius frenatus ns ns *** ns Timexbeds(size)*** Atherinomorus ogilbyi ns ns *** ns Timexbeds(size)*** Bathygobius kreffti ns ns *** ns Timexbeds(size)** Centropogon australis ns ns *** ns Timexbeds(size)** Pelates sexlineatus Timexbed(size)*** (Centre = inner) ns ns *** * Sizexregion** > edge Rhabdosargus sarba ns ns *** ns Timexbeds(size)*** Stigmatopora nigra Timexsizexregion* Spring>autumn, * * ns ns Small>large Upenius tragula & sp. Timexbeds(size)* Small>large ns * ns ns Regionxbeds(size)* Urocampus carinirostris ns ns *** ns Regionxbeds(size)* Amphipoda ns ns * ns Timexbeds(size)* Idiosepiidae (Edge = centre) ns ns ns * > inner Sepioloidea lineolata ns * ns ns Timexsize* Small > large
There was a significant negative correlation between the number of fish species and macroinvertebrate individuals in the large beds during autumn (Fig. 4.7; Table 4.7). In contrast, there was a positive correlation between the number of fish individuals and macroinvertebrates in the small beds during both sampling times (Fig. 4.7; Table 4.7). When this regression was tested for the inner or edge regions separately, this positive correlation was attributed to the inner regions of the small beds and not the edge regions (Fig. 4.7; Table 4.7). After a Bonferroni correction for multiple tests (Sokal and Rohlf 1995) the positive regression between fish and macroinvertebrates individuals in the inner regions of the small beds is still significant (adjusted P < 0.0063).
85 Do the edges of seagrass beds influence fauna?
Table 4.7. The regression values (R2) calculated from comparing the abundance of fish individuals or species with the abundance of macroinvertebrates in each haul of the net in Pittwater, 2001. See Fig. 4.7 for plots of the significant regressions. * P<0.05, **P<0.01. Time Seagrass beds or Fish species Fish individuals sampled regions sampled R2 value R2 value Autumn All seagrass beds 0.003 0.085 Autumn Small seagrass beds 0.023 0.301* Autumn Large seagrass beds 0.173* 0.041 Spring All seagrass beds 0.017 0.008 Spring Small seagrass beds 0.127 0.458** Spring Large seagrass beds 0.075 0.001 Both seasons Small beds, edge region 0.004 0.006 Both seasons Small beds, inner region 0.001 0.386**
There were seasonal differences in the fish and macroinvertebrate assemblages between autumn and spring (Fig. 4.8). Similarly, differences were detected when comparing the fish and macroinvertebrate assemblages in small and large seagrass beds (Fig. 4.8). However, the faunal assemblages in the different regions of the beds (edge, inner and central) were not significantly different (Fig. 4.8). A similarity percentages analysis revealed decapods to be the most discriminating group between the different assemblages in each season and size of seagrass bed. The average abundance of decapods per net haul was 131.33 individuals in autumn and 159.24 individuals in spring; 159.11 individuals in small seagrass beds compared 136.07 individuals in large beds.
86 Do the edges of seagrass beds influence fauna?
14 A 2 12 R = 0.30 10 P < 0.05 No. fish 8 6 species / 4 net 2
0 0A 50 100 150A 200 250 No. macroinvertebrates / net
100
90 B 80 70 60 No. fish R2 = 0.14 50 40 individuals P < 0.05 30 / net 20 10 0 0 100 200 300 400 500 600
No. macroinvertebrates / net
100 90 C 80
70 No. fish 60 R2 = 0.39 individuals 50 40 P < 0.01 / net 30 20 10 0 0 100 200 300 400 500 600 No. macroinvertebrates / net
Figure 4.7. Regression plots between A) the number of fish species and macroinvertebrate individuals in each net taken from large beds in autumn; B) the number of fish and macroinvertebrate individuals in each net taken from the small seagrass beds in both autumn and spring; C) the number of fish and macroinvertebrate individuals in each net taken from the inner regions of the small seagrass beds in both autumn and spring.
87 Do the edges of seagrass beds influence fauna?
A Global R = 0.32 P < 0.01 Autumn
Spring
B Global R = 0.27 P < 0.01
Small beds
Large beds
C Global R = -0.02 ns Edge regions
Inner regions
Central beds
Stress = 0.22
Figure 4.8. The two-dimensional configurations for MDS ordinations of the composition and abundance of fish and mobile invertebrates in seagrass beds. The replicates are categorized into samples taken; A) during autumn and spring, B) from small and large beds and C) from different regions (edge inner and central regions) of the seagrass beds in the Pittwater 2001.
88 Do the edges of seagrass beds influence fauna?
4.4 Discussion In both seasons, there was a greater number per net of fish species in the small compared to the large seagrass beds. Some researchers have considered the likelihood of sampling an edge to be the reason why small seagrass beds may contain more fauna (fish species richness; McNeill & Fairweather 1993). The edges of seagrass beds have also been found to contain greater abundances of some fauna than the interiors (Holt et al. 1983; Bologna & Heck 2000; Barbera-Cebrian et al. 2002; Hovel & Lipcius 2002; Tanner 2004). In this present study, however, there were no significant differences in the number per net of fish species (or individuals) in the different regions of the seagrass beds. Therefore the greater numbers per net of fish species in small seagrass beds cannot be attributed to the influence of edge effects. The only species of fish influenced by edge or regional effects was Pelates sexlineatus, which was found to be in greater numbers per net in the inner and central regions than the edges of the seagrass beds surveyed. Otherwise, there was a lack of detectable edge effects on fish and macroinvertebrates numbers in both large and small seagrass beds.
It has been proposed that smaller seagrass beds may be too small to support top predators (Hovel & Lipcius 2001). Shallow seagrass beds tend to have low numbers of large piscivorous fish, giving rise to the possibility that they act as sanctuary areas for juvenile fish (Whitfield & Blaber 1978; Blaber & Blaber 1980; Blaber 1980). Hovel and Lipcius (2001) found that small beds of seagrass contained more juvenile blue crabs (Callinectes sapidus) because the beds were too small to support its main predator, conspecific adult blue crabs. This may indicate the mechanism by which small seagrass beds can support greater densities of small fish species. Small fish in small seagrass beds may not only avoid larger predators but also have access to large numbers of potential macroinvertebrate prey.
Most of the macroinvertebrates collected were juvenile decapods, which are prey items for many of the fish species collected. Crustaceans have been found to be the most important food source for nearly all non-herbivorous fish (Burchmore et al. 1984; Hutomo & Peristiwady 1996). Seventy two
89 Do the edges of seagrass beds influence fauna?
percent of the fish species collected in this study are known to feed on crustaceans. The positive relationship between the fish (predators) and macroinvertebrates (prey) in the small beds found in this study could be the mechanism that explains the greater numbers per net of fish species in small seagrass beds. This must be viewed with caution because the smaller beds did not have a concurrent greater number per net of macroinvertebrates. However, macroinvertebrate abundances can be replenished by the incoming tide (Kneib 1994) and perhaps by the surrounding bare substratum so the densities of macroinvertebrates may not be limited by the size of the seagrass beds. Thus small seagrass beds may support greater numbers per net of fish species than would be expected from their area or density of prey items. Perhaps for small predators such as juvenile fish, the small seagrass beds facilitate rates of predation (Irlandi et al. 1999).
The lack of detectable changes in the abundance of fauna in the different regions of seagrass beds suggests that the dispersal of small fish and macroinvertebrates within a bed of seagrass is relatively homogeneous, or at least not biased towards the edge or interiors of the bed. The influence of regions within seagrass beds may be inconsequential compared to environmental variability, biotic interactions or stochastic events. Instead, the results suggest that the mobile fauna was evenly dispersed throughout the seagrass bed and large-scale processes influenced the assemblages. For instance, the larval supply is considered a major influence on the distribution of fish within an estuary and after the initial settlement period, post-larval fish may remain within a seagrass bed for several months before moving to other habitats (Middleton et al. 1984; Worthington et al. 1992b). This suggests that when fish larvae arrive at a small seagrass bed they would be required to settle in higher densities compared to lower densities if they arrived at a large bed. This is dependant on the delivery and survival of fish larvae but it may be one of the reasons as to why smaller seagrass beds have been found to have greater numbers per net of small fish than large seagrass beds.
Alternatively, the lack of edge-mediated changes in fauna abundances could suggest that the designation of a four-metre edge region was not an
90 Do the edges of seagrass beds influence fauna?
appropriate scale. Tanner (2004) found that the abundances of crustaceans in fragmented Zostera seagrass meadows changed (increased) within 0.25 – 1m from the edges. This result suggests that the designation of a four- metre edge region in the current study may not be small enough to detect a change in the abundances of macroinvertebrates.
The edges of the small beds in this study did not show a difference in the abundance and composition of the fauna assemblage, but they did have a different relationship between fish and macroinvertebrates. A positive correlation between the fish and macroinvertebrate individuals was detected in the inner regions of the small seagrass beds but not the edge regions. This could be considered an edge-mediated effect in the sense that some process operating in patch edges prevents the positive relationship between fish and macroinvertebrates. This could be attributed to the presence of pelagic predators that patrol the edges of the beds and a lack of shelter for both fish and macroinvertebrates from pelagic or avian predators.
Several researchers have concluded that the edges of seagrass facilitate predation. The patchiness of seagrass beds and its effects on bivalves has been thoroughly investigated in a series of experiments (Irlandi 1994, 1996 & 1997; Irlandi et al. 1995 & 1999), revealing that predation is higher in patchy seagrass compared to homogeneous seagrass. Bologna and Heck (1999) found increased survival and growth for macrofauna on the edges of seagrass beds but also an increased risk of predation. Further studies by Hovel and Lipcius (2002) found juvenile blue crabs to be more abundant in the interior of seagrass patches than at the edge, and they attributed this to predation being higher along the edges of seagrass beds. To test predation rates on fish in the edges of seagrass beds, manipulative experiments using cages would be effective, such as those designed by Bell (1986c) and Hindell et al. (2000).
This current study also detected an interaction between the season and size of bed when examining the numbers per net of macroinvertebrates in seagrass beds. In autumn the small beds had a greater numbers of macroinvertebrates than the larger beds but this pattern was reversed in spring. This could be attributed to the large recruitment of juvenile fish that
91 Do the edges of seagrass beds influence fauna? occurs in small seagrass beds during spring that may promote the movement of juvenile macroinvertebrates into the larger seagrass beds. The numbers of the macroinvertebrates in the small beds were greater during autumn compared to spring and the opposite was true for the large beds (greater during spring than autumn). In autumn there was also a weak negative correlation between the numbers per net of fish species and invertebrate individuals in large seagrass beds. This may suggest avoidance by the macroinvertebrates to the numbers of fish species or related factor in the large seagrass beds. Regardless of the complexity of macroinvertebrate abundances, it appears the size of a seagrass bed and the season can influence their distribution.
4.5 Conclusion The numbers per net of fish species in the small seagrass beds was consistently greater than that in larger seagrass beds, but this pattern of abundance could not be attributed to the greater amount of edge habitat in the smaller beds. In fact, there were no detectable regional (edge or interior) effects on the numbers of fish species, or fish and macroinvertebrate individuals in the seagrass beds. The only regional effect was a positive correlation between the abundance of fish and macroinvertebrate individuals in the inner regions of the small seagrass beds but not the edge regions. Edge effects in small beds of seagrass appear to influence the relationships between the fauna surveyed, but not their abundance or densities.
92 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
5 The influences of seagrass patch area, perimeter length and perimeter to area ratio on small fish and macroinvertebrates
5.1 Introduction
Seagrass beds of Zostera capricorni are a conspicuous component of estuarine landscapes along the temperate southeastern coast of Australia and are critical habitats for small inconspicuous fish, the juveniles of larger fish and macroinvertebrate fauna (Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984).
Fragmentation by natural and human processes has reduced large, continuous Z. capricorni seagrass beds into smaller patches that vary in size, shape and degree of isolation. Apart from decreasing the size or area of a seagrass bed, fragmentation also increases the amount of seagrass habitat that interfaces the surrounding sand, otherwise known as edge habitat. This is concurrent with an increase in the perimeter to area ratio (PAR) of the seagrass bed (i.e. more perimeter and less area).
A major concern with the fragmentation of seagrass beds is the consequences for seagrass fauna (Irlandi et al. 1995; Irlandi 1997; Bell et al. 2001), because generally more small fish occur in seagrass beds than in the surrounding sandy, unvegetated habitat (Orth et al. 1984; Bell & Pollard 1989; Edgar 1990; Ferrell & Bell 1991; Connolly 1994). Numerous researchers have considered the effects of seagrass fragmentation (with the concurrent decrease in patch size and increase in edge habitat) on the diversity and abundance of seagrass fauna (copepods, Bell & Hicks 1991; bivalves, Irlandi 1997; crustaceans, Eggleston et al. 1998, 1999; bivalves, Irlandi et al. 1999; bivalves, Bologna & Heck 2000; macroinvertebrate infauna, Bowden et al. 2001; blue crabs, Hovel & Lipcius 2001, 2002). Only a few studies, however, have considered the effects of patch size on abundances of fish and macroinvertebrates (McNeill & Fairweather 1993; Bell et al. 2002; Uhrin & Holmquist 2003; Chapters 3 & 4).
93 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
McNeill and Fairweather (1993) found greater fish species richness (number of fish species per bed) in small compared to large seagrass beds. Similarly, in the current study (Chapters 3 and 4) greater densities (numbers per net) of fish species were found in small compared to large seagrass beds. In both studies, the influences of the structural features of seagrass beds such as the area, perimeter length and the perimeter to area ratio (PAR) could not be tested independently of one another. For example, a small compared to a large seagrass bed will usually have a greater perimeter to area ratio. Large seagrass beds usually have a low perimeter to area ratio, but the amount of perimeter can vary. Small habitats with a high perimeter to area ratio are proposed by some researchers to collect more fauna than larger habitats because of the increased interception of recruits and migrating species (Paine & Levin 1981; Sousa 1984; Eggleston et al. 1999). This idea, however, confounds perimeter to area ratio; a measure of habitat shape, with perimeter length; a measure of the habitat boundary with the outside matrix. In Chapter 3, a positive relationship was found between the densities of fish individuals and the perimeter to area ratio of natural seagrass beds. This supports the model that the shape of a seagrass bed can influence the number of fish per bed. The smaller seagrass beds (with a high perimeter to area ratio), however, also had greater numbers of fish species per net. These results could be attributed to either the perimeter to area ratio or the area of a seagrass bed. For this reason a test was required whereby the perimeter to area ratio, the perimeter and the area of a seagrass bed could be manipulated separately from one another.
To test the influence of seagrass bed perimeter length it is necessary to hold constant the area of a seagrass bed while changing the perimeter length. Conversely, to test the alternative hypothesis that there will be no influence of the area of a seagrass bed, the perimeter length of the seagrass bed must be unaltered while the area of the seagrass bed is varied.
The physical manipulation of naturally occurring seagrass beds into smaller patches suitable for this experiment would have been logistically difficult and unwise for conservation. Therefore artificial seagrass patch units were used to manipulate the area and the perimeter length of a seagrass patch or
94 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna bed. Artificial seagrass attracts fauna similar to that found in natural seagrass (fish and macroinvertebrates: Bell et al. 1985; fishes and decapods: Sogard 1989; macrofauna: Edgar 1990), although sometimes with fewer individuals than natural seagrass beds (Bell et al. 1985). The artificial seagrass patch units were designed to either be large in area and short in perimeter, large in area and long in perimeter, or small in area and short in perimeter (Fig. 5.1). Manipulating area independently of perimeter was used to determine which feature was important in determining the density and abundance of small fish and mobile macroinvertebrates in seagrass beds.
Shape Area Perimeter P:A
2 3.6 m (m ) (m) ratio
a) Large a 3.6 m square 12.96 14.40 1.11
b) Large 10.8 m rectangle 12.96 24.00 1.85 b 1.2m
c) Small 6 m rectangle 7.20 14.40 2.00 c 1.2m
Figure 5.1. The area, perimeter length and shape of the three artificial seagrass patch units. There were two replicates of each patch design at the two locations (a total of four replicates for each patch and 12 patches in all). To test which feature influenced the abundance of fauna the predictions were as follows: if area then a = b ≠ c; if perimeter then a = c ≠ b; if perimeter area ratio then a ≠ b = c.
95 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
5.2 Methods 5.2.1 Construction of the artificial seagrass units. The artificial seagrass patch units were designed to mimic the dominant seagrass (Zostera capricorni) in the Pittwater estuary. The bases of the artificial seagrass units were constructed out of plastic garden mesh (5x5 cm mesh size) which was 1.2 metres wide and as long as required for each unit. To produce a fringe of seagrass blades, 30 cm by 25 cm squares of black commercial plastic was cut into a seagrass comb or fringe that mimicked the natural size and shape of Z.capricorni blades (5 mm wide and 20 cm long).
Fifteen of these squares were attached with cable ties for every metre of the plastic mesh base to create artificial seagrass with density of 525 blades.m2, which was within the density range of natural seagrass (514 to 1166 blades.m2) in the Pittwater estuary (Chapter 2). The arrangement of the artificial seagrass blades was haphazard to ensure the artificial units were not too uniform in cover compared to the natural seagrass beds and gaps in the canopy were no larger than 20 x 20 cm. The artificial seagrass was transported as a plastic mesh roll (in separate segments totalling 110.4 metres in length) to the study location in the Pittwater estuary.
5.2.2 Size, shape and arrangement of the artificial seagrass units. Three patch designs of artificial seagrass units were used; two were large in area (13 m2) and one was small (7.2 m2); with four replicates of each. The large units were either square or rectangular in shape with a relatively low (14.4 m) or high (24 m) perimeter length respectively (Fig. 5.1). The smaller patches were rectangular in shape with the same perimeter length as the large square patches (i.e. 14.4 m). To test the influence of seagrass patch perimeter length on fauna, the assemblages of fish and macroinvertebrates in the large square and small rectangular units (perimeter length = 14.4 m) were compared to the large rectangular units (perimeter length = 24.0 m). To test the influence of seagrass patch area on fauna the assemblages of fish and macroinvertebrates in the large rectangular and square units (area = 13 m2) were compared with the small rectangular units (area = 7.2 m2). To test the influence of perimeter to area ratio, the assemblages of fish and
96 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna macroinvertebrates in the large and small rectangular units with a similar PAR (2.0 and 1.85 respectively) were compared to the large square units (PAR = 1.1). This design un-confounds the influences of seagrass patch area, perimeter length and the perimeter to area ratio on the abundance and diversity of fish and mobile macroinvertebrates (see Fig. 5.1).
The artificial seagrass was rolled out at two locations in the Pittwater estuary (Fig. 5.2) at similar water depths (30-80 cm at mean low tide) on sand bars that were adjacent to shore and near Z. capricorni in the deeper water regions. They were fixed in place with tent pegs and sand. The order and orientation of the patch units in space was determined randomly (Fig. 5.3). Each artificial seagrass patch unit was more than 15 metres away from the nearest unit and all were placed 20 metres from a natural seagrass bed. This distance was chosen because epifauna associated with seagrass beds were found to extensively colonise clumps of artificial seagrass positioned 0 to 15 m from a seagrass meadow (Virstein & Curran 1986). This design ensured that the artificial seagrass patch units would be colonised by fish and macroinvertebrates independently of the features of the surrounding seagrass and maintained independence in the data (Underwood 1997). The artificial seagrass patch units were assembled at location one on 14 December and at location two on 17 December 2001. Each seagrass unit was checked weekly to ensure that they were free from sand and had not been vandalised. In the third and fourth week of deployment sand was cleared from all units, however all units received the same amount of clearing time and effort.
5.2.3 Sampling of fauna The seagrass fauna in the artificial seagrass patch units were sampled in location one on the evening of 22 –23 January 2002 and in location two on the evening of the 24 –25 January 2002, 39-40 days and 38-39 days after deployment, respectively. In previous studies, the time of sampling of fish in artificial seagrass units varied from: after a few days (Virnstein & Curran 1986), after 12 days, (Levin et al. 1997), after 6 weeks (Bell et al. 1985; McNeill & Fairweather 1993) and every week for 5 months (Sogard 1989). A period of six weeks was considered long enough to represent several settlement and post-settlement events (Sogard 1989; Kenyon et al. 1999).
97 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
This period of time was also chosen to avoid the deposition of excessive amounts of sand on the artificial seagrass and to limit the prolific growth of algae on the artificial seagrass, each of which could have confounded the experiment.
The sampling of seagrass fauna occurred in the evening within two hours of the low tide (water depth varied from 40 - 85cm, mean = 62cm). Previous work had demonstrated this water depth and tidal state to have no detectable influences on abundances of fauna (Chapter 2). In the afternoon preceding a sampling night, plastic posts were deployed around the edge of the units and a drop net constructed out of 1mm mesh (1.2 metres in height and weighed down with lead on the lower edge) was suspended above the water line. The drop net was left for up to three hours until after dark whereupon the net was dropped rapidly and fastened into the sand with pegs. A scoop net (3.6 metres wide for the square patch units and 1.2 metres wide for the rectangular patch units) was then dragged through the entire unit four times to collect all the fauna, which were then pooled. In the pilot stage of this project four drags were found to collect all the fauna. The fauna were identified and enumerated on site and released.
To confirm the effectiveness of the artificial seagrass patch units, bare sand was sampled twice using the same drop net method and collected only two fish and three mobile macroinvertebrates the first time and one fish and six mobile macroinvertebrates the second time. This indicated that seagrass fauna were utilising or attracted to the artificial seagrass. Over the period of the experiment, a variable amount of algae grew on the surface of all artificial seagrass units. Knowles and Bell (1998) found a high abundance of epifauna on drift algae compared to seagrass and Bologna and Heck (2000) found that the density of bivalves was significantly greater in artificial seagrass units that were fouled by a community of epiphytes than non- fouled units. To ensure that the amount of algae on the artificial seagrass units did not confound the experiment, the alga was collected from the units and weighed. A regression analysis was used to test for a relationship between the abundance of fauna and the dry weight of algae harvested from each unit.
98 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
5.2.4 Data analysis To test for differences in the abundance of fauna among artificial seagrass units a two-way orthogonal analysis of variance was used. The first source of variation was sampling location (one or two), which was fixed and orthogonal. The second source of variation was unit design (n=3), which was fixed and orthogonal. There were two replicates in each group. This ANOVA was used to compare the number and density (number per m2) of fish species, fish individuals, and macroinvertebrate individuals among artificial seagrass units. All the data were tested using Cochran’s C-test and transformed where necessary to ensure homogeneity of variances (Underwood 1997). The fish abundances (both fish species and fish individuals) required a transformation (Ln (x+1)), while the macroinvertebrate abundances did not.
In chapter 4, there was a relationship between fish and macroinvertebrates abundances in the inner regions of the small seagrass beds, but not in the large seagrass beds. In this chapter, linear regressions were used to test for a relationship between the abundances and densities (number per m2) of fish and mobile macroinvertebrates. The tests were performed for 1) all patch units, 2) the large patch units, 3) the patch units with a short perimeter and; 4) for the patch units with a high PAR. These multiple regressions were performed under the assumption that size, perimeter or PAR may influence fauna relationships.
A linear regression was also used to test for a relationship between the total number and density of; fish species, fish individuals, macroinvertebrate individuals and dominant single species and the amount of algae collected from each artificial seagrass patch unit.
99 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
N
2 1 1km
Figure 5.2. Locations of the study sites 1 & 2 in the Pittwater estuary, 2002.
100 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
1 Location one
seagrass
sand
shore 10m
Location two
seagrass
sand
shore
10m
Figure 5.3. The arrangement of the artificial seagrass units at location one and location two in the Pittwater estuary, 2002.
101 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
5.3 Results There were a total of 19 species of juvenile fish (404 individuals) and 6 groups of mobile macroinvertebrates (755 individuals) captured in the artificial seagrass patch units over the four sample nights (Table 5.1). The most abundant species of fish were the estuary catfish, Cnidoglanis macrocephala (115 individuals), the eastern-striped trumpeter Pelates sexlineatus (70 individuals), the half-bridled goby Arenigobius frenatus (66 individuals), and the bar-tailed goatfish Upeneus tragula (61 individuals). Five species of leatherjackets (Monocanthidae) were also collected (39 individuals). The most abundant group of mobile macroinvertebrates was the decapods with 624 individuals of which Penaeidae was the most numerous (252 individuals).
There was no difference in the mean total number of fish species, fish individuals and macroinvertebrate individuals collected from the three different patch designs; large rectangle, large square and small rectangle (Figures 5.4 and 5.5; Tables 5.2 and 5.3). There were, however, differences in the densities (number per m2) of fish species among artificial seagrass units (Fig. 5.4, Table 5.2). There were significantly greater densities of fish species in the small rectangular patches than in the large square patches with the same perimeter (SNK; small rectangular > large square). This difference was a magnitude of almost a two-fold increase in density (Fig. 5.4). There was no difference in the density of fish species between the small and large rectangle patches, although there was a trend for greater densities in the small seagrass patches (Fig. 5.4). There were no differences in the densities of the fish individuals and macroinvertebrate individuals between the three patch designs (Figures 5.4 and 5.5; Tables 5.2 and 5.3).
There was a positive correlation between the abundance and density of fish species and mobile macroinvertebrates in the artificial seagrass patch units that had a shorter perimeter length i.e. the large squares and small rectangles (Fig. 5.6; Table 5.4). This relationship was not evident in the larger rectangular patches.
102 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Some species of fish were correlated with the amount of epiphytic algae collected from each artificial seagrass patch unit. These included Pelates sexlineatus (positive correlation), Cnidoglanis macrocephala (positive correlation) and Upenius sp. (negative correlation) (Fig. 5.7; Table 5.5). No other components of the fauna including the number of fish or mobile macroinvertebrates were correlated with the weight of epiphytic algae collected from the artificial seagrass.
5.4 Discussion The area of the artificial seagrass patch was the main feature that influenced the fauna in this study. There were greater densities of fish species in the small compared to the large artificial seagrass patches. The total abundances of fish and mobile macroinvertebrates, however, were not influenced by the area, perimeter or PAR of the patches. In fact, the total abundance of juvenile fish in each artificial seagrass patch unit was similar. Only when the densities of the fish species were considered were there significant differences detected.
Other research (McNeill & Fairweather 1993; Chapters 3 & 4) has found greater densities of fish species or species richness in small than large natural seagrass beds. McNeill and Fairweather (1993) speculated that the mechanism producing this effect could be the increased likelihood of sampling an edge in small seagrass beds. Our study found that the greater densities of fish species in small seagrass beds was not explained by the greater proportion of edge habitat (high PAR) in small artificial seagrass units. The densities of fish species were similar in artificial seagrass patch units with a low and high PAR. This corroborates the findings in Chapter 4 where the abundances of small fish and mobile macroinvertebrates on the edges and inner regions of natural seagrass beds were similar.
103 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Table 5.1. The fauna collected from the artificial seagrass units from four replicates of each design (large rectangle, large square and small rectangle) in the Pittwater, January 2002. The bare sediment fauna was collected from two replicate samples.
Large Large Small Bare Fish family name Species name rectangle square rectangle sediment Apogonidae Apogon molucensis 1 0 0 0 Atherinidae Atherinomorus ogilbyi 4 1 0 0 Batrachoididae Batrachomoeus dubius 6 2 0 0 Girellidae Girella tricuspidata 0 0 1 0 Gobiidae Arenigobius frenatus 25 26 12 3 Bathygobius kreffti 2 0 6 0 Monocanthidae Acanthalutere spilomelanurus 1 2 0 0 Cantherhinus pardalis 2 3 2 0 Eubalichthys mosaicus 5 4 11 0 Monacanthus chinensis 1 1 0 0 Scobinichthys granulatus 4 3 0 0 Mullidae Upeneus sp. 10 4 1 0 Upeneus tragula 20 29 12 0 Scorpaenidae Centropogon australis 2 0 1 0 Serranidae Epinephelus daemelii 1 1 0 0 Siluriformes Cnidoglanis macrocephala 22 41 52 0 Syngnathidae Hippocampus whitei 1 0 1 0 Urocampus carinirostris 6 0 2 0 Tetrapontidae Pelates sexlineatus 28 18 24 0
Macroinvertebrate Class or order Family or species name Amphipoda 18 23 22 0 Isopoda 0 1 1 0 Decapoda Penaeidae 80 80 92 3 Palaemonidae 59 14 32 0 Alphidae 9 24 16 1 Hippolytidae 15 21 4 0 Portunidae 70 50 58 3 Asteroidea 3 0 0 1 Cephalopoda Sepioloidea lineolata 3 1 6 0 Idiosepius notoides 22 5 15 0 Polychaeta 6 1 4 1
104 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Table 5.2. An analysis of variance comparing the abundance and density of fish species and fish individuals from two locations in the Pittwater estuary, Jan. 2002 and from three different artificial patch unit designs. *P < 0.05.
Source of variation df Mean Square F
Total number of fish individuals per unit
Location 1 0.001 0.00
Unit design 2 0.046 0.37
Location x design 2 0.128 1.01
Residual 6 0.126 Total 11 Density (number per m2) of fish individuals per unit
Location 1 0.000 0.00
Unit design 2 0.131 1.54
Location x design 2 0.085 1.00
Residual 6 0.085 Total 11 Total number of fish species per unit
Location 1 0.014 0.24
Unit design 2 0.101 1.76
Location x design 2 0.068 1.17
Residual 6 0.058 Total 11 Density (number per m2) of fish species per unit
Location 1 0.010 1.00
Unit design 2 0.051 5.15*
Location x design 2 0.013 1.29
Residual 6 0.010 Total 11
105 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Table 5.3. An analysis of variance comparing the abundance and density of macroinvertebrates and decapods from two locations in the Pittwater estuary, Jan. 2002 and from three different artificial patch unit designs.
Source of variation df Mean Square F
Total number of macroinvertebrate individuals / unit
Location 1 374.08 0.20
Unit design 2 264.58 0.14
Location x design 2 2256.08 1.19
Residual 6 1890.25 Total 11 Density (number per m2) of macroinvertebrate individuals / unit
Location 1 0.24 0.02
Unit design 2 20.93 1.66
Location x design 2 19.65 1.56
Residual 6 12.63 Total 11 Total number of decapod individuals / unit
Location 1 96.33 0.08
Unit design 2 127.75 0.10
Location x design 2 1749.08 1.38
Residual 6 1270.33 Total 11 Density (number per m2) of decapod individuals / unit
Location 1 0.010 0.00
Unit design 2 12.27 1.48
Location x design 2 14.54 1.75
Residual 6 8.31 Total 11
106 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Table 5.4. The regression values for comparing the abundances or densities of the fish and mobile macroinvertebrates collected from the artificial seagrass patch units in Pittwater, Jan. 2002. See figure 5.6 for plots of the significant regressions. *P<0.01, **P<0.025, ***P<0.001. Fauna compared Patches R2
Number of Number of All units 0.02 fish invertebrate Large area (large rectangles & individuals individuals squares) 0.03 Short perimeter (large squares & small rectangles) 0.001 High PAR (large and small rectangles) 0.09
Number of Number of All units 0.09 fish species invertebrate Large area (large rectangles & individuals squares) 0.06 Short perimeter (large squares & small rectangles) 0.60** High PAR (large and small rectangles) 0.001
Density of Density of All units 0.02 fish invertebrate Large area (large rectangles & individuals individuals squares) 0.03 Short perimeter (large squares & small rectangles) 0.06 High PAR (large and small rectangles) 0.001
Density of Density of All units 0.35* fish species invertebrate Large area (large rectangles & individuals squares) 0.06 Short perimeter (large squares & small rectangles) 0.80*** High PAR (large and small rectangles) 0.17
107 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
Table 5.5. Regression values for comparing the abundance of fauna with the amount of epiphytic algae that was removed from each artificial seagrass patch unit in Pittwater, Jan. 2002. See figure 5.7 for plots of the significant regressions. *P< 005, **P<0.025. Fauna tested R2 Slope of regression line Fish species 0.23 Fish individuals 0.07 Arenigobius frenatus 0.22 Cnidoglanis macrocephala 0.36* positive Monocanthidae 0.22 Pelates sexlineatus 0.35* positive Syngnathidae 0.19 Upeneus sp. (inc. tragula) 0.42** negative
Macroinvertebrate individuals 0.10 Decapod individuals 0.05
108 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna 45 Number of fish individuals 40 35 No. of fish 30
25 individuals 20 / unit 15 10
5 0 Large rectangle Large square Small rectangle 6 Density of fish individuals 5 No. of fish 4 individuals
2 3 / m / unit 2 1
0 Large rectangle Large square Small rectangle 12 Number of fish species 10 No. of fish 8 species / unit 6
4
2
0 Large rectangle Large square Small rectangle
1.2 Density of fish species
1 No. of fish 0.8 species / 0.6 m2/unit 0.4
0.2
0 Large rectangle Large square Small rectangle Figure 5.4. The mean number or densities of fish (± S.E.) collected from the three artificial patch designs (n = 4) in the Pittwater, 2002.
109 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
120 Number of macroinvertebrates
100
No. of 80 invertebrate 60 individuals / unit 40
20
0 Large rectangle Large square Small rectangle 12 Densities of macroinvertebrates 10 Number of 8 invertebrate individuals 6 /m2/unit 4
2
0 Large rectangle Large square Small rectangle 100 Number of decapods 80 No. of decapod 60 individuals / unit 40
20
0 Large rectangle Large square Small rectangle
10 Densities of decapods
8 Number of 6 decapod individuals 4 /m2/ unit 2
0 Large rectangle Large square Small rectangle Figure 5.5. The mean number or densities of macroinvertebrates (± S.E.) collected from the three patch unit designs (n = 4) in the Pittwater, 2002.
110 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
12 a 10 8 6 R2 = 0.60 4 P < 0.025 2
Number of fish species fish of Number 0
0 50 100 150 Number of macroinvertebrates 1.2 b 1.0 0.8
0.6 R2 = 0.35 0.4 P < 0.05 0.2 Density of fishof species Density 0.0
051015 Density of macroinvertebrates
1.2 c
1.0
0.8
0.6 R2 = 0.80 0.4 P < 0.01 0.2
fishof species Density 0.0 0 5 10 15 Density of macroinvertebrates
Figure 5.6. The plots of the significant regressions comparing: a) The abundance of fish species and macroinvertebrate individuals in the units with a shorter perimeter length (small rectangles and large squares). b) The density of fish species and macroinvertebrate individuals in all units. c) The density of fish species and macroinvertebrate individuals in the units with a shorter perimeter length.
111 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
16 Pelates 14 sexlineatus 12
10 R2 = 0.35 8 P < 0.05 6 individuals No. 4 2
0 0 20406080100120 algae (grams)
45 Cnidoglanis 40 macrocephala 35 R2 = 0.36
30 P < 0.05 25 20 15 10 individuals No. 5
0 0 50 100 150 algae (grams)
15 Upenius sp.
2 10 R = 042 P < 0.025
5 No. individuals No.
0 0 20 40 60 80 100 120 algae (grams)
Figure 5.7. The plots of the significant regressions comparing the abundance of some individual fish species with the amount of epiphytic algae that was removed from each artificial seagrass unit in Pittwater 2002
112 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
The assumption that the density of organisms within a habitat is proportional to the area or amount of space available (Arhenius 1921; Gleason 1925; McArthur & Wilson 1967) was also not supported by the findings in this study. Perhaps this is because this model is more applicable to terrestrial habitats where there are limited dispersal mechanisms when compared to the aquatic environment. It has been suggested (Burgess 1988) that the predictions made about animal movements in systems with low permeable boundaries, such as islands (Macarthur and Wilson 1967), may not apply to those systems with highly permeable boundaries, such as aquatic habitats.
The perimeter length of the artificial seagrass patch units may be related to a positive relationship between fish and mobile macroinvertebrates. In units with a shorter perimeter length (the small rectangular and large square patches) there was a positive correlation between the density of fish species and macroinvertebrate individuals. In the units with a longer perimeter length (the large rectangular patches) this correlation was not detected. This suggests that faunal interactions or relationships may be influenced by the perimeter of a seagrass patch. This corroborates previous findings (Chapter 4) where the abundance of small fish and macroinvertebrates was correlated in the inner regions of small seagrass beds, but not in the edge regions.
A mechanism explaining the relationship between fish and macroinvertebrates in the inner region may be the presence of pelagic predators that patrol the edges of the seagrass beds and inhibit the interaction or movement of juvenile fish and mobile macroinvertebrates. Numerous researchers have found the edges of seagrass facilitate predation (Irlandi 1994, 1996 & 1997; Irlandi et al. 1995 & 1999; Bologna & Heck 1999; Hovel & Lipcius 2002). The patchiness of seagrass beds and its effects on bivalves has been thoroughly investigated by Irlandi and her coworkers (Irlandi 1994, 1996 & 1997; Irlandi et al. 1995 & 1999) and they reveal that predation is greater in patchy compared to uniform seagrass beds. Thus on the edges of seagrass beds or in seagrass patches with long perimeters, the
113 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna relationship between fish and macroinvertebrates could be inhibited by predation.
Bologna & Heck (1999) also found predation rates on scallops living on the edges of seagrass beds to be greater (>20%) than those living in the bed interiors. Similarly, Hovel and Lipcius (2002) found juvenile blue crabs to be more abundant in the interior of seagrass patches than at the edge and they attributed this to predation being greater along the edges of seagrass beds.
In this study, the abundances of fish and macroinvertebrates were not influenced by the amount of edge or perimeter but their relationship was. In patches with a lower perimeter the fish and mobile macroinvertebrates relate to one other or a common factor (such as abundance of plankton). Although there was no reduction in the abundance of small fish and mobile macroinvertebrates in the artificial seagrass patches with a longer perimeter length, this does not necessarily indicate an absence of predation. It has been observed that significant predation rates can occur on a small scale that cannot be detected (i.e. no decrease in prey abundances) on the larger scale of whole habitat (Kneib 1994). Research on the fish Fundulus heteroclitus and its harpacticoid copepod prey in intertidal marshes found the predator to have a significant local effect on prey abundance during the incoming tide, but not on the outgoing tide (Kneib 1994). The copepods were replenished by tidal redistribution so a reduction in the population of the copepod prey was not found regardless of the predator prey interactions observed.
In this study, the positive correlation found between the abundance of fish species and mobile macroinvertebrates could be attributed to predation by juvenile fish on macroinvertebrate prey. An analysis of the gut contents of seagrass fish by Hutomo and Peristiwady in Lombok, Indonesia (1996) showed crustaceans to be the most important food source (86.69% diet composition) for all non-herbivorous fish. In another study from subtropical Queensland, juvenile prawns were an important food source for 11 species of juvenile fish (Blaber & Blaber 1980). Decapods were the most numerous group of macroinvertebrates collected from the artificial seagrass patches and 58% of the fish species and 86.5% of the fish individuals collected prey
114 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna on this group. The positive correlation between the fish and mobile macroinvertebrates in the artificial seagrass units with a shorter perimeter length (large square and small rectangular patches) suggests that perimeter length of a seagrass patch may influence predator prey relationships. Other studies have considered the influence of patch size on predator prey interactions, but not the perimeter length of the habitat. Wellenreuther and Connell (2002) investigated reef fish predators (Cheilodactylus nigripes) and their invertebrate prey in boulder reefs of South Australia and found the density of prey within a patch rather than the size of the patch itself, strongly influenced the functional response of the predator fish. This occurred even when they manipulated the density of the prey in small and large patches. Research on small teleosts (toadfish) preying on oyster beds also found predation to be independent of patch size (Connell and Anderson 1999). None of these studies, however, considered the shape and the size of the patch on predator prey relationships.
The amount of epiphytic algae collected from the artificial seagrass units was correlated with the abundance of three species of fish and may reflect their habitat or prey selection. Juvenile P. sexlineatus individuals were positively correlated with the amount of epiphytic algae. Juveniles of this species are known to congregate around floating algae (Kuiter 2000). The juvenile catfish, C. macrocephala, was also positively correlated with the amount of epiphytic algae. This fish eats macroinvertebrates and other fish, but may be attracted to the shelter provided by algae as a juvenile. In contrast, the fishes Upenius sp. had a negative correlation with the amount of epiphytic algae. These fish hunt for invertebrate prey in the sand and therefore may avoid epiphytic algae.
The results of this study contribute to the debate regarding why small seagrass beds contain greater densities of fish species. The ‘edge effects’ model was not supported by this study. Other mechanisms thus need to be considered to explain the greater densities of juvenile and small fish in small compared to large seagrass beds. Using artificial seagrass units, features such as blade density, time of deployment and location within the estuary were equal. Therefore the only features that varied among the artificial seagrass patch units were the area and perimeter.
115 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna
One likely explanation for this may be the “settle and stay” model (Bell and Westoby 1986b), which proposed that pelagic larvae of seagrass fauna will settle into the first seagrass bed encountered, regardless of its size and shape. Therefore small seagrass beds will contain greater densities (numbers per m2) of fish species because of the limited space or area available compared to larger seagrass beds. Larvae are known to migrate to seagrass beds within weeks or months after the initial recruitment into an estuary (Bell & Westoby 1986b; Loneragan et al. 1986; Bell et al. 1987) and remain in the estuary for at least a year (Potter et al. 1983; Hannan & Williams 1998). Post larval fish are thought to remain within the bed until reaching a competent level of development. The larvae of some commercial and recreational fish can remain in seagrass beds for several months after settling from the plankton, before moving to other habitats (Middleton et al. 1984; Worthington et al. 1992b). Bell et al. (2001) showed experimentally that patches of artificial seagrass that were similar in size, shape and time of deployment captured different amounts of macro-algae when placed at different sites. This suggests that the passive delivery and/or interception of organisms are a major means of movement in marine systems.
In an experiment using artificial seagrass units, Sogard (1989) found a concentration effect where increased numbers of individual fish accumulated onto units that were far from seagrass compared to units adjacent to seagrass. For this reason, the current experiment was designed with all the units placed equidistant (20m) from natural seagrass beds to control for the influences of nearby habitats. Sogard (1989) found larger juveniles and even adult fish colonised the seagrass units, so the results of her study contradicted the “settle and stay” model, but supported the model of lateral migration of already settled juveniles and adults from surrounding habitats. Similar to Sogard (1989), the findings of the current study suggest a concentration effect where small fish will accumulate onto artificial seagrass patches, regardless of patch size. In contrast to Sogard (1989), but similarly to Bell et al. (1986) who also worked around NSW, the artificial patches in this present study collected post-larval fish and not many large juveniles and adult fish. Sogard (1989) attributed the differences between her experiment and Bell et al. (1986) to the larger pool of potential
116 Influences of Seagrass Patch Area, Perimeter Length and PAR on Fauna planktonic recruits available in NSW seagrass beds. The “settle and stay model” should therefore only be applied to habitats that are supplied by recruits from the plankton. Furthermore, the combination of all three experiments (Bell et al. 1986, Sogard 1989, and this current study) suggests that the model should be changed to “settle and stay for a while”. Perhaps post-larval and juvenile fish will settle into patches of seagrass regardless of its quality (i.e. size and shape) but can move to other habitats when reaching a certain life stage as demonstrated by Sogard (1989).
5.5 Conclusion The results of this study indicate the importance of small seagrass beds for juvenile fish and mobile macroinvertebrates. In particular, the “settle and stay” model proposed by Bell and Westoby (1986b) may be one mechanism explaining the greater densities of fish species in small compared to large artificial seagrass units. To test this model more thoroughly, would require artificial seagrass patch units to be orientated to intercept the oceanic currents in the estuary. Further studies investigating the effects of habitat perimeter length on faunal interactions, such as predation, could also reveal the mechanism behind the positive correlation between the fish species and mobile macroinvertebrates.
117 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6 Fish assemblages in seagrass beds can be correlated with proximity of mangrove
6.1 Introduction Mangrove forests are an important habitat for post larval and juvenile fish (Odum & Heald 1972; Bell et al. 1984; Little et al. 1988; Robertson & Duke 1987, 1990a, 1990b; Rooker & Dennis 1991; Laegdsgaard & Johnson 1995, 2001). Similarly, seagrass beds are important habitats for juvenile fish and small inconspicuous fish (Middleton et al. 1984; Orth et al. 1984; Pollard 1984; Bell and Pollard 1989). The fish assemblages of both habitats have been found to share some similarities in species composition, although the mangroves have been demonstrated to contain greater species richness than the adjacent seagrass beds (Robertson & Duke 1987; Thayer et al. 1987; Laegdsgaard & Johnson 1995). Therefore one might assume that seagrass beds close to mangrove forests may benefit in greater fish abundance and diversity.
Numerous studies have concluded that fish (especially during ontogenetic changes) can move between different marine habitats that are located close to one another (Austin 1971; Odum & Heald 1972; Jones & Chase 1975; Weinstein & Heck 1979; Martin & Cooper 1981; Ogden & Gladfelter 1983; Parrish 1989; Rooker & Dennis 1991; Robertson & Blaber 1992; Nagelkerken et al. 2000; Cocheret de la Moriniere et al. 2002). Most of these studies considered the role that estuarine habitats provide for coral reef fishes. However, research that demonstrates direct connectivity among estuarine habitats from the movement of fauna is scarce (but see Sheridan 1992, Irlandi & Crawford 1997, Michelli & Peterson 1998, Nagelkerken et al. 2001). Sheridan (1992) found that of the five species of fish captured in the flooded red mangrove habitat of Florida, three were also captured in adjacent seagrass habitat. The seagrass, however, contained a further 11 species of fish that were not captured in the mangrove forest. Irlandi and Crawford (1997) found that pinfish were twice as abundant in saltmarsh adjacent to seagrass beds than bare unvegetated sediments. Micheli & Peterson (1998) found benthic macroinvertebrates to be more abundant on oyster reefs that were separated from seagrass beds and saltmarshes. This was due to the
118 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests blue crab using the vegetated habitats as a corridor to the reefs where upon they would prey on the macroinvertebrates.
Adjacent habitats can play an important role in maintaining the diversity of fish species associated with seagrass (Ferrell & Bell 1991; Howard, 1989; Irlandi & Crawford 1997). However, the few studies examining the connectivity of seagrass beds and mangroves are based in the tropical regions of the world (e.g. Robertson & Duke 1987; Sheridan 1992; Sedberry & Carter 1993; Nagelkerken et al. 2001).
A study in the Caribbean found the species richness of juvenile coral fish was greater in seagrass beds adjacent to mangrove than seagrass beds in bays that did not contain mangroves (Nagelkerken et al. 2001). The authors attributed this to the coral fish utilizing the seagrass beds and mangrove as nursery habitats (Nagelkerken et al. 2001). Other studies on juvenile coral reef fish (Nagelkerken et al. 2000; Cocheret de la Moriniere et al. 2001) have indicated that individuals of some species shelter in the mangrove forests during the day and forage in the seagrass beds at night.
At present, only a few studies have examined the fish assemblages in mangroves within the temperate regions of southeastern Australia (see Bell et al. 1984; Clynick and Chapman 2002; Hindell & Jenkins 2004) but none have addressed the influence of mangrove forests on the fish associated with the adjacent seagrass habitat. In southeastern Australia, seagrass beds are often found near mangroves forests.
This study aims to investigate the relationship between the fish assemblages in seagrass beds and the proximity of mangroves. To do this, three surveys were conducted over different periods in two estuaries of temperate southeastern Australia. The estuaries contain Zostera capricorni seagrass beds that are of varying distances from Avicennia marina mangrove forests. It was predicted that seagrass beds close to mangroves would have greater species abundance and diversity than seagrass beds further from mangroves. Secondly, the fish assemblages in seagrass beds adjacent to mangroves were predicted to be different from seagrass beds located far from mangroves. Thirdly, the fish species in the mangroves were sampled
119 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests and it was predicted that the densities of these fish species in the seagrass beds would be in proportion to the bed’s proximity to mangroves i.e. greater densities of these fish species in seagrass beds closer to mangrove forests.
6.2 Methods 6.2.1 Study areas and descriptions The surveys were done in the Pittwater and Brisbane Water estuaries, within the Hawkesbury River system just north of Sydney, NSW, Australia (Fig. 6.1;
33.6°S, 151.3°E). The preliminary study was done in the Pittwater estuary during spring (Sep-Nov 2000). The main studies were done in Pittwater estuary during autumn (Mar-Apr 2002) and in Brisbane Water during autumn (Mar-Apr 2003).
6.2.2 Experimental design 6.2.2.1 Preliminary study To determine the influence of proximity to mangroves on the fish assemblages, six beds of Zostera capricorni were chosen from the lower reaches of the Pittwater estuary, where most of the mangrove forests are located (Fig. 6.1). Three seagrass beds were close to mangroves (< 200 m from major mangrove forest stands) and three seagrass beds were far from mangrove forests (> 500 m from all mangrove forests except isolated trees). Each seagrass bed was sampled four times during the day and night (eight times total) to discern the appropriate diurnal sampling regime for the main studies.
6.2.2.2 Main studies one and two In the main studies, the distance of seagrass beds from mangrove forests was treated as a continuous variable. This allowed more seagrass beds to be sampled (and so increased the bed replication) and did not impose a categorisation on the seagrass beds. To determine the effect of distance from mangrove forests on fish density and diversity in seagrass, eleven beds were selected in the Pittwater estuary that ranged from 1 to 1695 m in distance from mangrove forests (Fig. 6.1). Each seagrass bed was sampled for fish four times during the day because in the preliminary study the day sampling was when the differences in densities of fish species were more
120 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests pronounced (a reduced P value) between seagrass beds close and far from mangroves. The beds were also grouped into two size classes, large and small, because in previous research (Chapters 3 & 4) the size of a seagrass bed was found to influence the density of fish species. The fish assemblages of three mangrove forests of lower Pittwater were also sampled during the high tide.
In study two, thirteen beds of varying distance (3 - 1332 m) from mangrove forests in Brisbane Water were sampled. The mangrove forests in this estuary occur near the mouth as well as in the upper estuary and are thus more evenly distributed than the mangrove forests in the Pittwater estuary (Fig. 6.1). This estuary therefore provides a more rigorous test of the importance of proximity to mangrove forests for seagrass fish. However this also introduced another variable, the distance of a seagrass bed from the estuary mouth. This variable can also influence the fish assemblage in a seagrass bed (Bell et al. 1988; West & King 1996; Chapter 3) and had to be included in the data analysis. Sampling was done after dusk to avoid interference from the general public (and on request from the licensing body NSW Fisheries) with three samples taken in each seagrass bed. The difference assemblages in seagrass beds close and far from mangroves were detected during both day and night sampling in the preliminary study.
For all studies, sampling of the seagrass fauna happened on the low to mid tide at similar water depths (30-100 cm at mean low tide). This ensured that the water depth or tidal state did not influence the number of fauna collected (Chapter 2). To maintain independence of samples, all seagrass beds were separated by over 500 m of bare sandy substratum. To quantify the area of each seagrass bed and mangrove forest, the perimeter of each habitat was mapped by taking GPS positions every 2 m. The area of the seagrass beds and mangrove forests and their distance from mangrove forests were calculated using the GIS software ARC View (ESRI 1996).
121 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Brisbane Water
Brisbane Water
Broken
Bay
Avicennia marina forests
Zostera capricorni beds
The Pittwater
Pittwater
1 km
Figure 6.1. A map of Broken Bay with the estuaries Brisbane Water and the Pittwater. The mangrove forests and seagrass beds sampled are labeled as shown in the legend.
122 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.2.3 Fish Survey 6.2.3.1 The preliminary study and study 1 (Pittwater 2000 & 2002) To quantify the abundance and diversity of the fish assemblage, an eight by two metre seine net (1 mm mesh diameter) was randomly placed and dragged through each seagrass bed. This sampled approximately 68 m2 of the seagrass bed (mean = 68.17, S.E. = ± 1.2). A cost benefit analysis determined that four drags of the net gave an adequate estimation (see Chapter 2). This method was designed to capture small and juvenile fish more effectively than large pelagic fish, because seagrass beds are considered to be areas of importance for juvenile fish (Heck & Thoman 1984; Middleton et al. 1984). During sampling, the fish were identified, measured and released back into a seagrass bed within the estuary that was not included in this survey. To separate juvenile from adult fish, the total length of individual fishes were measured.
The fish sampled in the mangrove forest were collected using the seine net mentioned above and fish traps (deployed for three hours). The seine net was dragged through the mangrove forest, in the spaces between the Avicennia marina trees and always under the forest canopy. It was observed repeatedly that the pneumatophores were bent down by the bottom edge of the net but not broken. The fish traps were 60 x 35 x 40cm in dimensions (5mm mesh size) and were designed to catch large juveniles and sub-adult fish. They were deployed in areas within the forests where the seine net could not access (under trees, roots etc). For both methods, the sampling was conducted around the high tide to ascertain which fish species utilised the mangrove forests.
6.2.3.2 Study two (Brisbane Water 2003) The fish assemblage was sampled using a seine net that was 1.5 metres wide (1mm mesh) and dragged 10 metres through the seagrass bed. This sampled 10 m2 of seagrass. This net was found to collect a similar fish assemblage as the larger seine net but in reduced abundance corresponding with the smaller area sampled. Three drags of the net were used to estimate the density of the small fish captured. As in the previous study, individual fish were categorised as adult or juvenile based on their size.
123 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.2.4 Univariate data analysis 6.2.4.1 Preliminary study (Pittwater 2000) An analysis of variance (ANOVA) test compared the number of fish species per net and the number of fish individuals per net in beds close to mangrove forests with beds far from mangrove forests. All the data were tested using Cochran’s C-test and transformed where necessary to ensure homogeneity of variances (Underwood 1997). The ANOVA was a three-way, mixed model design. The first source of variation was the time (day or night) of sampling (fixed, orthogonal); the second was proximity (close or far) to mangrove forests (fixed orthogonal) and the third was seagrass beds (random, nested). Single species analysis on the juvenile stage of development for the dominant fish species was also done using an analysis of variance.
6.2.4.2 Study one (Pittwater 2002) A linear regression was used to test for a relationship between the distance of the seagrass bed to mangrove forests and the mean number of; all fish species, mangrove fish species (those that were caught in the mangrove forests), non-mangrove fish species (those that were not caught in the mangrove forests), juvenile fish species, all fish individuals, mangrove fish individuals, non-mangrove fish individuals and juvenile fish individuals per net. The species richness of all fish species, mangrove fish species and non- mangrove fish species per seagrass bed was also tested for a correlation with the seagrass bed distance to mangrove forests. The abundance of some dominant individual fish species was also used as a variable.
6.2.4.3 Study three (Brisbane Water 2003) A linear regression was used to test for a relationship between the distance of the bed to mangrove forests and the mean number of fish species, fish individuals, juvenile fish species, and juvenile fish individuals per net. The abundance of some dominant individual fish species was also used as a variable. The fish assemblages in the mangrove were not sampled in this study so the categorisation of fish into mangrove forests and non-mangrove utilising fish was not possible.
124 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.2.5 Multivariate data analysis 6.2.5.1 All studies A Bray-Curtis similarity analysis between samples was performed using a square root transformation of the data to give a non-metric multi- dimensional scaling (MDS) plot (Clark and Warwick 2001). A second similarity analysis was performed using a presence absence transformation. The first transformation gave an analysis of the composition and abundance of the fish assemblage whereas the second gave an analysis of the composition alone.
A two way crossed analysis of similarity (ANOSIM) was used to test for differences in species composition and abundance between beds close and far from mangrove forests. The continuous variable (distance from mangrove) was categorised into beds far and close from mangrove forests for studies one and two because of the limitations of the multivariate analysis that would not allow for the analysis of a discrete variable nested within a continuous variable. In the preliminary study and study one, the close beds were 1-80 metres away from mangrove forests and the far beds were over 220 metres away (up to 900 m metres in study one and 1695 metres in study two). In study two, the close beds were 3 to14m from mangrove forests and the far beds were 100 to 1332m away from mangrove forests. This difference in distance from mangrove forests in the two estuaries was unavoidable because of the difference between the two estuaries. In Brisbane Water there is an abundance of mangrove forests making it difficult to find sites without nearby mangroves.
For the preliminary study, the second factor in the two-way crossed analysis was the two sampling times (day and night). In study one, the second factor was size of the bed (small or large) and in study two the second factor was the location of the bed within the estuary (top or bottom). A similarity percentages (SIMPER) analysis was used to determine what species were contributing to the differences in assemblage composition detected by the ANOSIM (Clarke & Warwick 2001).
125 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.3 Results There was a total of 50 fish species and 9532 fish individuals collected from the seagrass beds in the three studies; 36 fish species and 2507 fish individuals from the preliminary study (Pittwater 2000); 27 fish species and 1242 fish individuals from study one (Pittwater 2002) and 29 fish species and 5783 fish individuals from study two (Brisbane water 2003) (Table 6.1). There was a total of 11 fish species and 792 fish individuals collected from the mangrove forests in Pittwater 2002 (Table 6.2). This difference in fish abundances for seagrass and mangrove habitat can be attributed to the different sampling efforts; in the seagrass beds the sampling was quantitative and extensive compared to the sampling effort in the mangroves which was qualitative and was only required to indicate which fish species were utilising the habitat.
6.3.1 Univariate analysis: abundance and density of fishes 6.3.1.1 Preliminary study (Pittwater 2000) There were significantly greater densities (numbers per net) of fish species in the beds closer to mangrove than far and during the night than the day (P < 0.05, Fig. 6.2, Table 6.3). Concurrently, the total species richness (number of species per bed) was greater in the beds close to the mangrove (33 fish species in total) than beds far from mangrove (24 species). There was no difference in the density of the fish individuals between beds close and far from mangrove, although there was a trend for greater densities of fish individuals in beds close to than far from the mangrove forests.
There were significantly greater densities of juvenile fish species and juvenile fish individuals in beds closer to mangrove forests than further away (Table 6.4). The density of juvenile fish individuals was also significantly greater in one of the seagrass beds close to mangrove than the other two beds close to mangrove (Table 6.4). There were also significant differences at the level of single species. The density of the goby Bathgobius kreffti was significantly greater in seagrass beds that were far compared to close to mangrove forests. In contrast the tarwhine Rhabdosargus sarba was in significantly greater densities in seagrass beds close compared to far from mangrove forests (Table 6.5).
126 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
20 Fish species
15 No. of fish species / 10 net 5
0
close far close far
day night
500 Fish individuals
400
No. of fish 300 individuals 200 / net 100
0
close far close far
day night
7 Juvenile fish species 6 No. of juvenile 5 fish species / 4
net 3 2 1 0 close far close far day night
60 Juvenile fish individuals 50 40 No. of juvenile fish individuals 30
/ net 20 10 0 close far close far
day night
Figure 6.2. The mean number per net of fish species, fish individuals, juvenile fish species and juvenile fish individuals in seagrass beds close and far from mangroves during the day and night in the preliminary study (Pittwater 2000).
127 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.1. The densities of fish per 10m2 in seagrass beds close and far from mangrove forests in the preliminary study (Pittwater 2000), study one (Pittwater 2002) and two (Brisbane Water 2003).
Preliminary Study one Study two Family Species Close Far Close Far Close Far Apogonidae Apogon limenus 0 00000.01 Atherinidae Atherinomorus ogilbyi 3.64 1.40 0.06 0 34.14 15.61 Batrachoididae Batrachomoeus dubius 0.29 00000 Blennidae Petroscirtes lupus 0 0.15 0 0.02 0 0.06 Chandidae Ambassis jacksoniensis 8.75 0.66 0.06 0.01 35.48 62.56 Clinidae Cristiceps aurantiacus 0.20 0.29 0 0.01 0 0.06 Heteroclinus sp 4 0 0.15 0 0 0 0 Clupeidae Hyperlophus translucidus 0.96 00000 Spratelloides robusta 0 0 0.02 0.01 0 0 Diodontidae Diodon nichthemerus 0.15 00000 Dicotylichthys punctulatus 0 0000.050 Eleotrididae Philypnodon grandiceps 0 0 0 0 0.29 0.61 Gerreidae Gerres subfasciatus 0.39 0 0 0 13.33 18.72 Girellidae Girella tricuspidata 8.01 0.51 0.01 0 0.52 0.17 Gobiidae Arenigobius frenatus 7.30 4.39 0.39 0.30 11.10 7.17 Bathygobius kreffti 0.38 1.37 0.02 0.06 0.19 0.28 Cristatogobius gobioides 0 0000.330 Redigobius macroston 1.18 0.15 2.13 0.03 21.95 45.22 Hemiramphidae Hyporhamphus australis 0 0.15 0 0 0.14 0.28 Lethrinidae Lethrinus laticaudis 0 00.01 0 00 Monocanthidae Acanthalutere spilomelanurus 4.04 2.89 0.06 0.06 0 0.22 Acanthaluteres vittiger 0 0 0.05 0.19 0 0 Cantherhinus pardalis 0.29 0 0.10 0.05 0.62 1.22 Eubalichthys mosaicus 0 0 0 0.01 0 0 Meushenia flavolineata 0 00000.11 Meuschenia trachylepis 0.29 00000 Meuschenia venusta 0.22 0.15 0 0 0 0 Monacanthus chinensis 0.20 0.15 0.20 0.54 0.52 0.44 Scobinichthys granulatus 0.81 0.44 0 0 0 0 Monodactylidae Monodactylus argenteus 0 00.04 0 00 Mullidae Upeneus sp. 0.29 0.20 0.01 0.07 0 0 Upeneus tragula 0 0.15 0.02 0.02 0 0 Odacidae Neoodax balteatus 0.15 0.59 0 0 0 0 Paralichthyidae Pseudorhombus jenynsii 0 0.15 0 0 0 0 Poeciliidae Gambusia holbrooki 0 0000.190 Scorpaenidae Centropogon australis 0.34 0.63 0 0.05 0.81 1.56 Sillaginidae Sillaginodes maculata 0.37 00000 Sillaginodes punctatus 0.15 0.34 0 0 0 0 Sillago flindersi 0.15 00000 Sparidae Acanthopagrus australis 0 0 0 0 0.14 0.11 Rhabdosargus sarba 6.54 0.81 1.00 1.05 0.05 1.72 Sphyraenidae Sphyraena obtusata 0 00.01 0 00 Syngnathidae Filicampus tigris 0.15 0.15 0 0 0 0 Hippocampus whitei 0.29 0 0 0.02 0.05 0.06 Stigmatophora argus 0 0.15 0 0 0 0.17 Stigmatophora nigra 0.44 0.71 0 0.11 0 0.11 Urocampus carinirostris 2.60 2.25 0.17 0.37 3.57 4.67 Vanacampus margaritifer 0 0 0 0 0.10 0.17 Tetraodontidae Tetractenos hamiltoni 0.15 0.15 0 0.01 0.19 0.67 Tetrapontidae Pelatus sexlineatus 4.58 1.65 1.17 2.05 1.48 2.00
128 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.2. The total abundance of fish collected from four mangrove forests in study one (Pittwater 2002) using a seine net and fish traps. Family Species Abundance Atherinidae Atherinomorus ogilbyi 100 Blennidae Omobranchus anolius 1 Chandidae Ambassis jacksoniensis 422 Gerreidae Gerres subfasciatus 55 Gobiidae Arenigobius frenatus 6 Gobiidae Redigobius macroston 6 Monodactylidae Monodactylus argenteus 12 Mugilidae Myxus elongatus 14 Sillaginidae Sillago ciliata 34 Sparidae Rhabdosargus sarba 130 Tetraodontidae Tetractenos hamiltoni 12
Table 6.3. The analysis of variance comparing the densities (numbers per net) of fish species and individuals during different sampling times (day and night); and in seagrass beds adjacent to mangrove with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). No transformation of data was required. *P < 0.05, **P < 0.01, ***P < 0.001. Source of variation Mean squares df F Number of fish species / net Time 108.00 1 75.13** Proximity 85.33 1 7.57* Beds (proximity) 11.27 4 2.13 Time x Proximity 3.00 1 2.09 Time x Bed (proximity) 1.44 4 0.27 Residual 5.30 36
Number of fish individuals / net Time 336.02 1 0.21 Proximity 28178.52 1 2.74 Beds (proximity) 10299.39 4 13.65*** Time x Proximity 1530.02 1 0.96 Time x Bed (proximity) 1591.64 4 2.11 Residual 754.56 36
129 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.3.1.2 Study one (Pittwater 2002) There was no significant relationship between the density of all fish species, individual fish, juvenile fish species, juvenile individual fish and the dominant single species and the distance of the seagrass bed from mangrove forests (Table 6.6). There was, however, a significant relationship for the density of fish species caught in the mangroves and the density of fish species not caught in the mangroves with the distance of the seagrass bed from mangrove forests. This relationship revealed that the density of fish species caught in the mangrove forests and their species richness increased as the distance of the seagrass bed to mangrove forests decreased. In contrast, the density of fish species of fish not caught in the mangrove forest and their species richness decreased as the distance of the seagrass bed from the to mangrove decreased (Fig. 6.3, Table 6.6).
6.3.1.3 Study two (Brisbane Water 2003) There was no relationship detected between the density (number per net) of all fish species, individual fish, juvenile fish species, juvenile fish individuals and the dominant single species and the distance of the seagrass bed from the mangrove forest (Table 6.7).
6.3.2 Multivariate analysis: composition and abundance of fish assemblages There were significant differences in the fish assemblages between seagrass beds close and far from mangrove forests, in all studies in each estuary (Fig. 6.4, Table 6.8). This difference was detected even when the data was adjusted by a presence absence transformation. This indicated that the composition of the assemblage (and not the abundances of fish species) was the contributing to this difference.
For all three studies Arenigobius frenatus was consistently found in greater abundance in seagrass beds close to mangrove forests than far (Table 6.9). In the Pittwater studies, Rhabdosargus sarba and Pelates sexlineatus were found in greater abundance in seagrass beds close to mangrove forests, while the pattern was different from one study to the next for Urocampus carinirostris. Similarly, Redigobius macroston was found to be more abundant in seagrass beds close than far from mangrove forests in the
130 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Pittwater estuary (2002) and more abundant in seagrass beds far from mangrove forests in the Brisbane Water estuary (2003). Arenigobius frenatus, Rhabdosargus sarba, and Redigobius macroston were caught in the mangrove forests in study two and the simper results confirm their status as species whose densities were correlated with distance from mangrove forests (Table 6.9).
Table 6.4. The analysis of variance comparing the densities of juvenile fish individuals and juvenile fish species during different sampling times (day and night); and in seagrass beds adjacent to mangrove with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). The data was transformed (Sqrt (x+1)) to give a non-significant (P > 0.05) Cochran’s value. *P < 0.05, **P < 0.01 Source of variation Mean squares df F Number of juvenile fish species / net Time 0.9564 1 19.20** Proximity 3.1015 1 9.68* Beds (proximity) 0.3204 4 2.21 Time x Proximity 0.1602 1 3.22 Time x Bed (proximity) 0.0498 4 0.34 Residual 0.1450 36
Number of juvenile fish individuals / net Time 0.7424 1 1.38 Proximity 101.4968 1 8.11* Beds (proximity) 12.5114 4 3.87** Time x Proximity 0.6827 1 1.27 Time x Bed (proximity) 0.5361 4 0.17 Residual 3.2297 36
131 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.5. The summary of an analysis of variance comparing individual fish species collected during different times (day and night) and from seagrass beds adjacent to mangrove forest with seagrass beds far from mangrove in the preliminary study (Pittwater 2000). The P values are shown for the first three levels.
Proximity to SNK test Species Time mangrove P < 0.05 Acanthalutere spilomelanurus < 0.05 ns Day > night Arenigobius frenatus ns ns Atherinomorus ogilbyi < 0.01 ns Night > day Bathygobius kreffti <0.01 < 0.05 Night > day, far > close Centropogon australis ns ns Pelates sexlineatus ns ns Rhabdosargus sarba ns < 0.05 Close > far Urocampus carinirostris ns ns
132 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.6. Regressions of the abundance of different groups of fish and the dominant single species of fish with the distance of the seagrass bed from the mangrove forests for study one (Pittwater 2002). The densities are calculated by averaging the number caught per net per bed. The species richness is calculated from the total number of species caught per bed. *P<0.05 Group tested R2 value Density of all fish species 0.09 Density of mangrove fish species 0.37* Density of non-mangrove fish species 0.17 Density of juvenile fish species 0.01 Species richness (all fish species) 0.02 Species richness (mangrove species) 0.41* Species richness (non-mangrove species) 0.38* Density of all fish individuals 0.29 Density of mangrove fish individuals 0.21 Density of non-mangrove fish individuals 0.01 Density of juvenile fish individuals 0.09 Arenigobius frenatus 0.08 Monacanthus chinensis 0.01 Pelates sexlineatus 0.02 Redigobius macroston 0.24 Rhabdosargus sarba 0.06 Urocampus carinirostris 0.01
133 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.7. Regressions of the abundance of fish species, individual fish and the dominant single species with the distance of the seagrass bed from the mangrove forests for study two (Brisbane Waters 2003). All were non- significant values. Group or single species tested R2 Density of fish species 0.10 Density of juvenile fish species 0.11 Density of fish individuals 0.19 Density of juvenile fish individuals 0.02 Arenigobius frenatus 0.17 Atherinomorus ogilbyi 0.04 Centropogon australis 0.13 Gerres subfasciatus 0.01 Monacanthus chinensis 0.03 Pelates sexlineatus 0.03 Redigobius macroston 0.05 Rhabdosargus sarba 0.27 Urocampus carinirostris 0.05
Table 6.8. The results of the analysis of similarity (two way crossed) for all studies. Square root Presence / absence Study and categories tested transformation transformation Global R Global R Preliminary study (Pittwater 2000) 0.19 ** 0.21 ** Time: day and night 0.34 ** 0.27 ** Distance: close and far Study one (Pittwater 2002) Size: large and small 0.09 0.19 ** Distance: close and far 0.46 ** 0.23 ** Study two (Brisbane Water 2003) Estuary location: close and far 0.26 ** 0.40*** Distance: close and far 0.24 ** 0.35***
134 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Table 6.9. Simper (similarity percentages) results for the three studies selecting the most discriminating species. Study and species name Average Average Ratio Contrib abund. in abund. in Dissim / -ution beds close beds far SD % Prelim. study (Pittwater 2000) Arenigobius frenatus 12.42 7.46 1.12 15.40 Urocampus carinirostris 4.42 3.83 1.07 6.64 Acanthalutere spilomelanurus 6.46 4.92 1.01 9.76 Girella tricuspidata 13.63 0.29 0.99 13.86 Rhabdosargus sarba 11.13 0.92 0.90 12.30 Pelates sexlineatus 7.79 2.33 0.90 11.27 Study 1 (Pittwater 2002) Rhabdosargus sarba 6.53 1.10 1.12 14.29 Redigobius macroston 14.41 0.07 1.11 31.48 Pelates sexlineatus 8.35 7.48 1.09 21.56 Arenigobius frenatus 2.47 1.03 1.04 6.26 Urocampus carinirostris 1.18 1.55 1.00 4.88 Study 2 (Brisbane Water 2003) Ambassis jacksoniensis 38.92 62.47 1.44 30.85 Redigobius macroston 20.50 52.20 1.12 23.27 Gerres subfasciatus 18.13 12.13 1.01 11.17 Arenigobius frenatus 12.21 4.60 0.95 6.02
135 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
4 Density of mangrove fish species 3 R2 = 0.37 Mean number of P < 0.05 2 fish species
/ net 1
0 0 500 1000 1500 2000
Distance to mangrove forests (m)
6 Species richness of
5 mangrove fish
No. species / 4 R2 = 0.41
seagrass bed 3 P < 0.05
2
1 0 0 500 1000 1500 2000 Distance to mangrove forests (m)
Species richness for 12 non-mangroves fish 10 8 No. species / 6 R2 = 0.38 seagrass bed 4 P < 0.05
2
0 0 500 1000 1500 2000 Distance to mangrove forests (m) Figure 6.3. The regressions of; the densities of mangrove fish; the species richness of mangrove fish and; the species richness of non-mangrove fish, with the distance of each seagrass bed in study 1 (Pittwater 2002) from the mangrove forests.
136 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
Study 1 Pittwater 2000
R = 0.34 P < 0.01
Study 2 Pittwater 2002 Samples from seagrass beds close to mangrove
Samples from R = 0.46 seagrass beds P < 0.01 far from
mangrove
Study 3 Brisbane Waters 2003
R = 0.24 P < 0.01
Figure 6.4. Two-dimensional configurations for MDS ordinations of the composition and abundance of fish in seagrass beds in Pittwater 2000, Pittwater 2002, and Brisbane Waters 2003.
137 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
6.4 Discussion For all three studies there was a relationship between the fish assemblages in seagrass beds and the proximity of mangrove forests. Seagrass beds that were near the mangrove forests had different fish assemblages than seagrass beds far from mangrove forests. These differences were attributed to the composition of the fish species, and not their abundances. This correlation was detected regardless of other differences such as the time of sampling (preliminary study), the size of the bed (study one) and the location of the bed (study two) within the estuary. This finding supports the work of other researchers who have suggested that mangrove and seagrass habitats are utilised by juvenile fish and the combination of habitats influences fish assemblages (Nagelkerken et al. 2000, 2001; Cocheret de la Moriniere et al. 2001).
The fish assemblages collected from seagrass beds between the night and day sampling were significantly different. Different fish assemblages in the night compared to day has been supported by other studies (Bell & Harmelin-Vivien 1982; Gray et al. 1998; Griffiths 2001). There were also different fish assemblages found between seagrass beds located closer to the estuary mouth than those located further down the estuary and this has also been found in other estuaries (Bell et al. 1988; McNeill et al. 1992; Jenkins et al. 1996; Chapter 3).
In the preliminary study, the density of all fish species was significantly greater in seagrass beds close than far from mangrove forests, for both sampling times day and night. Therefore it was predicted in the later studies that the density of fish species would increase as the distance between the seagrass beds and mangrove forest decreased. This did not occur in either study. However, in Pittwater estuary (2002), the density and species richness of those fish species found in the mangroves decreased in the seagrass beds as the distance of that bed increased from the mangrove forests. In contrast, the species richness of non-mangrove fish species increased as the distance between the seagrass bed and mangrove forest increased. Therefore the patterns of fish distribution seen in the first study (Pittwater 2000) could have been attributed to those fish species that use
138 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests the mangrove forest. In study one, the seagrass beds closer to the mangrove forests in Pittwater had a greater abundance of Rhabdosargus sarba, Girella tricuspidata and Atherinomorus ogilbyi, although not all of these results were statistically significant. Morton (1990) and Laegdsgaard & Johnson (1995) revealed that these fish are commonly found in mangrove forests. In the present studies, R. sarba and A. ogilbyi were also found in the mangrove forests in great abundance. Arenigobius frenatus, R. sarba and Pelates sexlineatus were also found to be in greater abundances in beds closer to mangroves than far. However, in studies one (Pittwater 2002) and two (Brisbane Water 2003), the single species regression did not detect significant changes in abundances of these three species as the mangrove forests to seagrass bed distance increased.
One possible mechanism for the greater density of fish species in seagrass beds close to mangrove forests in study one was thought to be that the mangrove forests were elevating the carrying capacity for fish species in the proximal seagrass beds. The results from the later studies in Pittwater and Brisbane waters estuaries do not support this hypothesis. In fact, in an inter-tidal bay of Mozambique, the productivity of seagrass beds in terms of nutrient outputs (carbon, nitrogen and phosphorus) was greater than that of mangrove forests (de Boer 2000). It may be that seagrass beds are enhancing the biota of the adjacent mangrove forests. Nagelkerken et al. (2001) suggested that the enhancement of the seagrass fish assemblage by the presence of mangrove could be mutual (i.e. seagrass enhancing the fish assemblage of mangrove forests) although at this stage this hypothesis is untested.
Mangroves are considered to be important habitat for post larval & juvenile fish (Bell et al. 1984; Little et al. 1988; Robertson & Duke 1987, 1990a, 1990b; Laegdsgaard & Johnson 1995, 2001). Juvenile fish are attracted to the decomposition of mangrove leaves in tropical mangroves (Rajendran and Kathiresan 1999). Similarly, seagrass is important for these fish life history stages. It would seem intuitive to assume that the combination of these two habitats would be advantageous for the biodiversity of fish in estuaries. The analyses, however, of the juvenile fish assemblages produced different results among the three studies. In the preliminary study (Pittwater 2000)
139 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests the densities of juvenile fish species and fish individuals were greater in seagrass beds closer to mangrove forests. In the first (Pittwater 2002) and second studies (Brisbane Water 2003), however, the proximity of mangrove forests did not contribute to the abundance of juvenile fish in the adjacent seagrass beds. In contrast, other researchers in tropical regions of the world (Nagelkerken et al. 2001) have found that the seagrass beds near mangrove forests had greater species richness of juvenile coral reef fish than seagrass beds without mangrove forests. These results may not apply to the temperate fish species in this study.
Other researchers have found mangrove habitats to contain more fish and/or fish species than the adjoining seagrass habitat (Robertson & Duke 1987; Thayer et al. 1987; Laegdsgaard & Johnson 1995). Furthermore, Laegdsgaard & Johnson (1995) found in summer that the majority of the juvenile fish in mangroves were non-residents and therefore not confined to the mangrove habitat. They compared the biodiversity of fish in the mangroves with seagrass beds during the high tide. One could hypothesise that during the low tide when the mangrove forests were exposed, fish would reside in the adjacent seagrass beds until the mangrove forests were immersed again on the incoming tide. The results of our studies support this model because the sampling of fish in the seagrass beds was done during low tide and the different assemblages between seagrass beds close and far from mangrove were in part attributed to some of the fish species that were found in the mangrove forests. Furthermore, in seagrass beds closer to mangrove forests, there were greater densities of fish species and species richness of those fish found in the mangrove forests. Nagelkerken et al. (2001) also found that the abundances of fish and species richness of nursery species in the Caribbean were greater on mudflats that were located near seagrass and mangrove forests compared to mudflats with neither habitat nearby. The movement of fish between and among seagrass, mangrove and mudflats was considered to be an explanation of this pattern of distribution. Cocheret de la Moriniere et al (2002) suggested that some species of coral reef fish prefer the mangrove forests during the day and forage in the seagrass beds at night. Some fish are thought to utilise both mangrove forests and seagrass but at different life stages (Nagelkerken et al. 2000; Cocheret de la Moriniere et al. 2002). Similarly, Thayer et al.
140 Fish Assemblages in Seagrass Beds can be Correlated with Mangrove Forests
(1987) found that the gray snapper Lutjanus griseus was more likely to forage in the prop root fringe of the mangrove forests as a juvenile and forage in adjacent habitats including seagrass as an adult. Sheridan (1992) in a comparison of adjacent intertidal seagrass, mangroves and open water habitats found that the mangrove forests contained the lowest abundances and biomass of decapods and fishes. The flooded mangroves, however, were at times used by both resident fish and crabs in similar densities as that found in the seagrass and the open water habitats.
In contrast, the results of others working in the temperate regions (Clynick & Chapman 2002) found that mudflats adjacent to mangrove forests (within 50m) did not contain different assemblages of fish from mudflats some distance (approximately 200m) away from mangrove forests. This survey was conducted in the winter when there are low numbers of juvenile fish. This may have contributed to a non-significant result. Furthermore, the distance categories (50 to 200m) may not have been significantly great enough to detect differences in mobile fauna such as fish. In the present studies the distance categories were greater and varied from 0.3m to 1695m in Pittwater and 3m to 1332m in Brisbane Water. The results of our studies corroborate the work of researchers in the tropical regions (Nagelkerken et al. 2000, 2001; Cocheret de la Moriniere et al. 2002). The dearth of studies conducted within the temperate regions of the world suggests that more temperate studies are required before general conclusions about tropical and temperate mangrove forests and the fish assemblages they potentially support can be made.
6.5 Conclusion The presence of mangrove forests was correlated with the fish assemblages of Zostera capricorni beds in the two temperate estuaries. In particular, those fish that were found to use mangrove forests during the high tide were found in greater species densities and species richness in the seagrass beds close to the mangrove forests during the low tide than seagrass beds further away. The results suggest that habitat connectivity (i.e. mangroves and seagrass) can be correlated with the fish assemblages in seagrass beds.
141 General Discussion
6 General Discussion
6.1 Seagrass beds contain small fish and macroinvertebrates This study confirms that Z. capricorni seagrass beds contain abundant small fish and macroinvertebrates. During the seven seasons of sampling in the Pittwater and Brisbane Water estuaries, forty species of juvenile fish and twenty-four species of small adult fish (total of 64 fish species) were collected. Four orders of macroinvertebrates were also collected; mainly cephalopods and decapods, and 99% of the penaeids were juveniles. Numerous other studies have found seagrass beds to be important habitats for fish and macroinvertebrates in estuaries (Heck & Thoman 1984; Middleton et al. 1984; Orth et al. 1984; Pollard 1984). In particular, these studies have demonstrated that the fish assemblages in seagrass beds consist mainly of small inconspicuous adult fish and the juveniles of larger fish (Middleton et al. 1984; Pollard 1984; Bell & Pollard 1989; Beck et al. 2001). This present study also confirms that seagrass contains these groups of fish although it does not exclude the importance of other habitats. For example, mangroves have also been found to be important habitat for post larval & juvenile fish (Odum & Heald 1972; Bell et al. 1984; Little et al. 1988; Robertson & Duke 1987; 1990a; 1990b; Laegdsgaard & Johnson 1995; 2001).
6.2 Fish assemblages in seagrass beds support landscape ecology theories Landscape ecology investigates the relationship between spatial / temporal patterns and ecological processes (Mazerolle & Villard 1999). It attempts to relate the dynamics and development of spatial heterogeneity in a landscape (Risser et al. 1984; Turner 1989; Robbins & Bell 1994) with its biotic and non-biotic features.
Seagrass habitats are ideal for the application of research considering landscape ecology theories. Seagrass beds occur in meadows that can extend over kilometre-wide areas (i.e. historically defined landscape) and it
142 General Discussion is at this scale of patchiness that marine studies are relatively scarce (Robbins and Bell 1994). In particular some species of seagrass grow as monospecific stands in beds or meadows with clearly defined boundaries, often surrounded by bare substrata (Bowden et al. 2001). Seagrass beds are also relatively simple in structure and floristic compositions compared to other marine or terrestrial habitats and are distributed within an estuary over different spatial scales, while maintaining relative structural homogeneity (Robbins & Bell 1994; Turner et al. 1999).
Research in landscape ecology measures landscape structure, function and change (Robbins & Bell 1994). The principal aim of this current study was to find a correlation between seagrass landscape structural patterns and seagrass faunal assemblages. Landscape structural patterns within an estuary can refer to; 1) the spatial structure (size and shape) of the seagrass beds, 2) position of the seagrass beds within the estuary, 3) the proximity of seagrass beds to mangrove forests (interconnectedness of habitats) or 4) the heterogeneity or patchiness of the seagrass beds. This thesis has supported all of the above landscape structural features as correlatives with the fish assemblages in seagrass beds. Other landscape ecology structural features, however, were not correlated with fish assemblages in seagrass beds. These include the predictions that there will be 5) edge effects in seagrass beds and 6) evidence of self-similarity across multiple scales.
6.2.1 The spatial structure (i.e. size and shape) of seagrass beds This research (in three separate studies) has confirmed that small Z.capricorni beds contain greater densities of fish species than large beds in the Pittwater estuary. This occurred regardless of the position of the seagrass bed within the estuary and the season or the year sampled. Furthermore, the results were independent of the perimeter length or the perimeter to area ratio of the seagrass bed. The spatial patterns of seagrass beds, including their size, shape and perimeter and the influence of these features on seagrass fauna has been investigated by other researchers (see Bell & Hicks 1991; McNeill & Fairweather 1993; Irlandi 1997; Eggleston et al. 1998; 1999; Frost et al. 1999; Irlandi et al. 1999; Bologna & Heck 2000; Bowden et al. 2001; Hovel & Lipcius 2001; 2002) although only a few of these studies considered fish and macroinvertebrates. Similar to the
143 General Discussion findings of the present study, McNeill and Fairweather (1993) also found greater fish species richness in small compared to larger seagrass beds.
In the current study, the total abundances of fish species were similar in small and large seagrass beds, but because of the limited space available in the smaller seagrass beds they contained greater densities of fish species. This could be considered a concentrating or limiting effect of habitat size and corroborates the findings of Sogard (1989). Furthermore, the limitations on the available space in smaller beds also changed the composition of the fish assemblage in the seagrass bed. The possible mechanisms for this phenomenon will be considered in detail later.
6.2.2 The position of a seagrass bed within the estuary The assemblages of small fish were significantly different in seagrass beds closer to the estuary mouth than beds located far from the mouth in the Pittwater and Brisbane Water estuaries. There were, however, no significant differences in the abundances of fish. Another study conducted in Pittwater estuary also found that the juveniles of some fish species were distributed in zones, with some species being more common closer to the estuary mouth, while others were more abundant further away from the estuary mouth (Bell et al. 1988). They proposed that the mechanism behind this pattern was a combination of spawning location and the dispersal of eggs and larvae of the fish species. Studies within other estuaries have also supported that certain species of fish in seagrass beds are found more abundantly in some locations within the estuary than others (McNeill et al. 1992; Jenkins et al. 1996; Hannan & Williams 1998) although the measured affect of this pattern could vary in time (Jenkins et al. 1996). The results indicate that the location of a seagrass bed within the estuary (i.e. seascape spatial patterns) can be correlated with the fish assemblages.
To test if the delivery of competent larvae and eggs to a seagrass bed determines the density of fish species, artificial seagrass units could be placed within the path of the oceanic currents flowing into an estuary that presumably carry larvae. The artificial seagrass patch units would need to be placed in regions of different degrees of oceanic currents i.e. exposed directly to the currents and varying degrees of a lack of exposure. These
144 General Discussion artificial seagrass units would need to be sampled for post-larvae fish over numerous periods to coincide with delivery larvae.
6.2.3 The proximity of seagrass beds to mangrove forests For all three studies, the fish assemblages in the seagrass beds were correlated with proximity to mangrove forests. The fish assemblages were different in seagrass beds that were close to mangroves compared to seagrass beds that were far from mangroves. This difference was attributed to the composition of the fish assemblage and not the abundances of fish individuals. This result is supported by the work of other researchers who have suggested that both mangrove forests and seagrass beds are used by juvenile fish (Nagelkerken et al. 2000; 2001; Cocheret de la Moriniere et al. 2002). Furthermore, the current study found there was an increase in the density and species richness of fish that use mangroves in the seagrass beds that were close to mangrove forests. Some fish are thought to utilise both mangrove forests and seagrass beds but at different stages in their life history (Thayer et al.1987; Nagelkerken et al. 2000; Cocheret de la Moriniere et al. 2002). Some species of coral reef fish also utilise mangrove forests during the day, but forage in the seagrass beds at night (Cocheret de la Moriniere et al. 2002).
The biological correlations between estuarine habitats that are proximal to one another, or ‘linked’ has been under investigation by other researchers and for other estuarine ecosystems such as; nekton in marshes (Rozas & Odum 1987), nekton in saltmarsh (Hettler 1989), fish in saltmarsh, (Irlandi & Crawford 1997), mobile shrimp and fishes in macroalgae (Holmquist 1994), blue crabs in saltmarshes and reefs (Micheli & Peterson 1999), amphipods on drift algae (Brooks & Bell 2001) and gastropods in mangrove forests (Ross & Yerman 2003). The outcomes of these studies support the theory that faunal assemblages are correlated with habitat connectivity.
This current study demonstrated connectivity not just by proximity (structural connectivity) but also functional connectivity, whereby the dispersal of organisms potentially connects the habitats and populations’ therein (With et al. 1999) i.e. similar fish species were found in both seagrass and mangroves.
145 General Discussion
To further test if mangrove forests are contributing to the diversity of fish in seagrass beds (or vice versa) the movement of fauna between mangrove forests and seagrass beds needs to be measured. This could be achieved by using artificial seagrass units positioned at varying distances from mangrove forests and sampled over time. It would also be interesting to position the artificial seagrass patches between real seagrass beds and mangrove forests as ‘stepping-stones’ or connecting patches between the two habitats.
6.2.4 The patchiness or heterogeneity of Z. capricorni beds The spatial scale of landscape studies in the marine environment ranges from metres to thousands of metres. In this study, the heterogeneity of seagrass beds occurred on a scale of metres (1- 4 m sand patches) compared to the scale of whole beds (20-100’s of metres). The composition of the fish assemblages was different in heterogeneous or patchy seagrass beds compared to the homogeneous or uniform seagrass beds. Other studies have found that heterogeneous environments promote diversity by a number of mechanisms (see introduction) compared to homogeneous environments (Heck & Orth 1980; Irlandi & Crawford 1997; Eggleston et al. 1999). The patchiness of seagrass has also been found to influence biological interactions such as competition and predation (Coen et al. 1981; Irlandi 1994; Hovel & Lipcius 2002). It is therefore not surprising to find different fish assemblages between patchy and uniform seagrass beds.
There was, however, no difference in fish abundances between patchy and uniform seagrass beds. This suggests that fish are considerably mobile and more reliant on the macro-habitat (i.e. the whole bed) than the micro-habitat (i.e. cover of seagrass within the bed). This could indicate to the “grain” or scale of response of small fish, whereby the fish utilise a patchy seagrass bed as an entire habitat (Kotliar & Wiens 1990). Small-scale influences, however, were strong enough to select or favour certain species over others. For example, more individuals of cryptic fish species were found in the uniform seagrass beds compared to the patchy seagrass beds.
146 General Discussion
6.2.5 Edge effects in seagrass beds Other researchers have found a greater abundance or density of organisms in small compared to large seagrass beds. This has been attributed to the influence of edge effects (McNeill & Fairweather 1993; Eggleston et al. 1998; 1999). McNeill and Fairweather (1993) considered there to be an increased likelihood of sampling an edge in a small bed. In contrast, during two seasons of sampling, there were no differences detected in the abundance of fish, fish species or macroinvertebrates between the edge and inner regions of seagrass beds. Other studies have found the fauna of seagrass patches to be more abundant on the edges than the interiors (the red drum fish, Holt et al. 1983; juvenile bivalves, Bologna & Heck 2000; mysids, Barbera-Cebrian et al. 2002; crabs, Hovel & Lipcius 2002; crustaceans, Tanner in press). Similar to the current study, some researchers have not detected faunal assemblages to be specifically associated with the edge of seagrass beds (epifauna, Sanchez-Jerez et al. 1999; Bell et al. 2001; bivalves, Tanner 2004) while another has (Bowden et al. 2001).
Tanner (in press) found that crustaceans were more numerous within 1m of the edge of a seagrass patch. This result suggests that the current study did not detect changes in macroinvertebrate abundances between edges and inners because the scale of the survey (4m edge region) was not small enough. Tanner (2004) found that other groups of macroinvertebrates such as bivalves generally did not show a response and only a few polychaete taxa responded to seagrass patch edge. It appears that changes in faunal response to edge regions of seagrass beds are taxa specific or dependent on the scale of the study.
The lack of detectable changes in the abundance of fauna in the current study for the different regions of seagrass beds suggests that the dispersal of small fish and some mobile macroinvertebrates within a bed of seagrass is relatively homogeneous, or at least not biased towards the edge or interiors of the seagrass bed. The regions of a seagrass bed thus may not be the major influence of fish assemblages in seagrass beds compared to environmental variability, biotic interactions or stochastic events. Instead, the results suggest that the fish were evenly dispersed throughout the bed
147 General Discussion and large-scale processes influenced the assemblages such as larvae delivery and survival.
The absence of edge-mediated influences in seagrass beds could suggest that it may be incorrect to assume a model developed in a terrestrial habitat can be applied with similar consequences in marine habitats with aquatic fauna. The increased negative impacts of physical fluxes (eg. radiation, wind, water) in edges of small terrestrial habitats (Saunders et al. 1991) may not be relevant for aquatic habitats. There are greater limitations on dispersal of organisms in terrestrial compared to aquatic habitats (Eggleston et al. 1998). Furthermore as demonstrated by numerous researchers (Holt et al. 1983; Bologna & Heck 2000; Barbera-Cebrian et al. 2002; Hovel & Lipcius 2002; Tanner in press) the edges may in fact enhance abundances of some aquatic organisms, however, the present study suggests this is not the case for fish assemblages. The detection of edge-mediated effects in seagrass beds will be further expanded later in this chapter.
6.2.6 Self-similarity across multiple scales Another landscape ecology model that was not supported by this study was the presence of self-similarity across multiple scales (Robbins & Bell 1994). This was indicated by the different assemblages of fish in the small beds compared to the larger beds. As the size of a seagrass bed increased the fish assemblages changed and so the landscape concept of similar patterns across multiple scales (i.e. small and large seagrass beds) was not evident. Similarly, the assemblages of fish in seagrass beds were different between patchy and uniform beds. This indicated that at a scale of a few metres compared to tens of metres, there was little similarity.
6.2.7 Seascape approach required This study suggests that the landscape / seascape model should be considered when attempting to conserve the fauna of an estuary (Robbins & Bell 1994; Micheli & Peterson 1999; Bell et al. 2001). The landscape model must be modified to incorporate ‘seascape ecology’ (a term devised by Bartlett & Carter 1991) which incorporates features such as water movement, oceanic currents, tidal fluctuations, up-welling processes and the
148 General Discussion marine dispersal processes (Robbins & Bell 1994), not to mention the ontogenetic changes of marine organisms.
6.3 Patch size and the species-area relationship One of the oldest theories in ecology still being tested today is the relationship between the available (or sampled) area or habitat and the number of species (Arhenius 1921; Gleason 1925; reviewed by Connor & McCoy 1979; McGuinness 1984a, 1984b). Numerous hypotheses have been put forward to explain why in general, larger habitats (or islands) have greater species diversity than small habitats. These include: the random placement hypothesis (Arhenius 1921); the habitat diversity hypothesis (Williams 1943, Connor & McCoy 1979); and the equilibrium theory of island biogeography (MacArthur & Wilson 1967). This present study did not support the species area relationship and found the opposite to be true, whereby there was greater species richness in smaller than larger habitats (seagrass beds).
Numerous researchers have investigated these theories and their application for the marine environment (Paine & Levin 1981; McGuinness 1984a, 1984b; Keough 1984; Sousa 1984; Raimondi 1990; Navarrete & Castilla 1990; Kim & DeWreede 1996; Underwood & Skilleter 1996; Minchinton 1997; Anderson 1998; Bowden et al. 2001). McGuinness (1984a; 1984b) used intertidal boulders to investigate if the island biogeography theory (especially the species area relationship) applied to the intertidal environment. He proposed that an interaction of disturbance, habitat area and heterogeneity, and other biotic and non-biotic factors were responsible for the assemblages found on intertidal boulders. Other researchers have investigated the influence of habitat size on the recruitment of barnacles on rocky shores and conflicting results have been observed. Some have found that there was increased recruitment of barnacles in smaller than large patches of habitat (Paine & Levin 1981; Keough 1984; Sousa 1984; Jeffrey 2000). Others found the reverse (Kim & DeWreede 1996) or that the size of the habitat i.e. diameter of the pool, had little influence on the colonisation by organisms (Underwood & Skilleter 1996). Paine and Levin (1981) found the recolonisation of mussels into patches cleared within existing mussel beds was proportional to the perimeter and shape of the clearing, because it
149 General Discussion relied on the encroachment of adjacent conspecifics. Similarly, Sousa (1984) found that the density of limpets in a cleared patch was related to the perimeter of the patch and not the area. Raimondi (1990) found the recruitment of the gregarious barnacle, Chaemiosphio ansipoma, was related to the perimeter of a cleared patch, as did Navarrete and Castilla (1990) with recruiting mussels. Minchinton (1997) also found that the recruitment of the tubeworm Galeolaria caespitosa to cleared patches on rocky intertidal shores was related to the proximity of conspecific adults and therefore the perimeter of the patch and not a function of its area. All of these studies, however, considered sessile and gregarious intertidal organisms.
The colonisation of patches of habitat by mobile fauna such as fish and macroinvertebrates may not follow the same pattern whereby perimeter dictates the colonisation of the area. In fact, the results from this current study suggest that the perimeter of a seagrass patch does not determine the assemblage or abundance of seagrass fauna.
Anderson (1998) suggested that the timing of the experiments in regard to recruitment and succession of the organisms studied influenced the species area relationship. The spatial, temporal and successional stochasticity of biological and physical processes has been attributed (by Bowden et al. 2001) for the inconclusive or contradictory evidence for the existence of species area relationship and edge effects in marine assemblages (e.g. Schoener & Schoener 1981; Keough 1984; Tsuchiya & Nishihira 1985; Svane & Ompe 1993; Svane & Setyobudiandi 1996; Anderson 1998). Bowden et al. (2001) also suggested that the environmental noise at progressively smaller scales may mask any species-area effects and only within the larger landscape, regional or biogeographically scales do these patterns become discernable or measurable. Yet perhaps the problem is not the scale of measure but the actual model, which has been designed from a terrestrial understanding of environmental processes. This study did not support the species-area model and so alternative aquatic-based models were considered to explain the greater densities of fish in the small compared to the large seagrass beds.
150 General Discussion
6.4 Models to explain the greater densities of small fish in small compared to larger seagrass beds 6.4.1 The “settle and stay” model The “settle and stay” model (Bell & Westoby 1986b) proposes that pelagic larvae of seagrass fauna will settle into the first seagrass bed they encounter, choose a suitable micro-site within that bed and remain within the bed until reaching a competent level of development. In this model the delivery of larvae to a seagrass bed determines the density of fish and studies of the distribution of post larval fish support this model (Potter et al. 1983; Middleton et al. 1984; Bell & Westoby 1986a; Loneragan et al. 1986; Bell et al. 1987; Worthington et al. 1992b; Hannan & Williams 1998). Most of the fish species collected in this present study were planktonic post- juvenile settlers. This “settle and stay” model provides a possible mechanism for why the smaller seagrass beds had greater densities of small fish. The post-larval fish arrive within the estuary in ichthy-planktonic episodes and settle into the nearest seagrass bed independent of its area. A small seagrass bed would then accumulate greater densities of fish species than a larger seagrass bed where the larval can disperse into a larger area of seagrass. The oceanic currents that flow into the Pittwater estuary circulate counter-clockwise. It cannot be assumed that the seagrass beds closest to the estuary mouth are the first to receive the oceanic waters (with the ichthy- plankton). So unfortunately comparing the density of fish in seagrass beds close to the estuary mouth with beds further away would not be a satisfactory test of this model. In the Pittwater estuary the seagrass bed that consistently had the greatest species diversity was, however, one of the first seagrass beds to receive the oceanic currents in the counter-clockwise current circulation.
Planktonic post-larval fish dominated the fish species collected in this current study, similar to Bell et al. (1986b) who also worked around New South Wales, Australia. In contrast, Sogard (1989), collected larger juveniles and even adult fish in an experiment using artificial seagrass units in New Jersey, U.S.A. Her results contradict the ‘settle and stay’ model but support the model of lateral migration of already settled juveniles and adults from surrounding habitats. Sogard (1989) attributed the differences between her
151 General Discussion experiment and Bell et al. (1986b) to the larger pool of potential planktonic recruits available in N.S.W. seagrass beds. In New Jersey, the fish species were found to leave the natural seagrass beds and cross expanses of predation risky sand to reach the artificial seagrass (Sogard 1989). Perhaps the model should be changed from “settle and stay” to “settle and stay for a while”.
6.4.2 The ‘grain’ or scale of response of small fish The smallest scale at which an organism responds to patch structure is defined as its ‘grain’ (Kotliar & Wiens 1990). It is thought that smaller organisms will have a smaller ‘grain’ than larger organisms. Larger organisms will functionally perceive a mosaic of patches as a single habitat whereas smaller organisms will generally perceive it as many habitats (Kotliar & Weins 1990). Therefore, the abundance and diversity of small organisms may be more influenced by the size of a habitat than larger organisms (Eggleston et al. 1999). This model suggests that the results of the current study are only applicable to juvenile and small fish. An adult or larger fish would have the mobility to use numerous seagrass beds as habitat. The results of this study are therefore limited to the small fish assemblages of seagrass habitat and should not be applied to the large fish assemblages in estuaries.
6.4.3 The absence of large predators in small seagrass beds It has been proposed that smaller habitats may be too small to support top predators (Jackson 1977; Kay & Keough 1981; Schoener & Schoener 1981; Keough 1984; Anderson 1998). Shallow seagrass beds have been found to have an absence of large piscivorous fish, which increases their likelihood of being areas of low predation and sanctuaries for juvenile fish (Whitfield & Blaber 1978; Blaber & Blaber 1980; Blaber 1980). Anderson (1998) proposed that smaller patches of habitat could act as refuges for species that are poor competitors (such as small fish) who would be out-competed in larger patches (Jackson 1977; Kay & Keough 1981; Schoener & Schoener 1981; Keough 1984).
Perhaps for small predators such as juvenile fish, the small seagrass beds facilitate rates of predation (Irlandi et al. 1999) concurrent with providing
152 General Discussion protection from larger predators (Orth et al. 1984; Orth 1992). Hindell et al. (2000) suggested that the habitat complexity provided by seagrass could mediate predation on small fish, although their results varied with the time of sampling.
Other studies have shown that the abundances of small fish were negatively correlated with the abundance of piscivorous predators in coral reef systems (Hixon 1991; Connell & Kingsford 1998; Connell 1998) and this was corroborated by dietary analysis (Connell & Kingsford 1998; Connell 1998). This implies that predation by fish or other piscivores has the potential to influence the structure of small fish assemblages.
The presence or absence of larger predators in small seagrass beds could be measured using video surveillance techniques. A stratified sampling regime would be required to test if there was any particular region (such as the edge region) that the predators were more likely to occupy or hunt in. This technique may also reveal why the abundances of macroinvertebrates and fish species were correlated within the inner regions of small seagrass beds.
6.4.4 Life-history strategies (r and K selection) Smaller patches of seagrass have been demonstrated to experience greater environmental disturbance from water movement and wave action than larger patches of seagrass (Fonseca et al. 1982; Irlandi et al. 1995). This has led some researchers to hypothesise that the fauna found in small and large seagrass beds may differ in terms of life history attributes such as their colonising and competitive abilities (Keough 1984; Bowden et al. 2001). The smaller (seagrass) patches with greater disturbance regime will contain species that are poor competitors with wide dispersal and rapid colonising life history (r strategists, Pianka 1974). In contrast, the large (seagrass) patches will contain species (K strategists) with limited dispersal and good competitive abilities (Keough 1984; Bowden et al. 2001). It is difficult to support or refute this model when considering the assemblages of fish collected from the small and large seagrass beds in this study, especially given that numerous fish species contain attributes of both r and K strategists. Furthermore the smaller seagrass beds within this study did not experience a greater disturbance regime than the larger beds in the time
153 General Discussion period of this study. It may be that r-K life history concepts are inappropriate for assemblages of fish. For example, Steele (1985) considered that the different feeding and mortality patterns between the larvae and adult fish phase makes the r-K life history concepts inapplicable for fish populations.
6.4.5 The source or sink model Habitats may be classified as source or sink, depending on their resource abundance for single species and assemblages. A source habitat provides a high quality habitat for the assemblage, which produces a surplus of offspring. A sink habitat is low in resource quality and does not allow the assemblage to produce enough offspring to make up for mortality (Pulliam 1988; Danielson 1991; Pulliam & Danielson 1991). Given that different densities of fish species were found in small and large seagrass beds it could be suggested that the small seagrass beds act as source habitats for the large beds or vice versa. Therefore, one type of seagrass bed (small or large) could be providing the recruits and overall productivity for the other. Furthermore, seagrass beds may act as source or sink habitats for other habitats such as mangrove forests or the surrounding unvegetated sediments. Small seagrass beds may be invaluable habitats for smaller fish, but larger seagrass beds are required for larger organisms (Hovel and Lipcius 2001). For this reason the source sink model may rely on the size of the organisms in question. Small seagrass beds may only be suitable source habitats for small seagrass fauna.
6.5 Reasons for correlations of fish and macroinvertebrate in the inners of the small beds In the inner regions of small seagrass beds there was a positive correlation between the fish and macroinvertebrates. This was not detected in the outer regions of the small seagrass beds, nor in any regions of the large beds. Furthermore, this correlation was also detected in the artificial seagrass units with a short perimeter length but not in those units with a longer perimeter length. It thus appears that the perimeter or edge of a seagrass bed and artificial seagrass unit may be deterring correlations between fish and macroinvertebrates. This could indicate a possible edge-
154 General Discussion mediated effect that is related to restricting the interactions of seagrass fauna.
There are four classes of edge-mediated effects on species interactions as outlined by Fagan et al. (1999). 1. Edges of habitats may restrict or facilitate the movement of organisms or propagules within the landscape; 2. Edges of habitats may differentially influence the mortality of species; 3. “Cross boundary subsidies” (Janzen 1986) may occur whereby the population of some species are maintained at a high rate in one patch, which then leads to their dispersal into another patch, where the invaded residents are greatly impacted, especially those nearer the edge; 4. The edges of habitats can be unique within their own right with interactions between species that would not otherwise occur.
These classes of edge-mediated effects are not mutually exclusive and all four mechanisms may occur at the edges of seagrass beds.
The findings of this current study suggest that the movement or interaction of the fish and macroinvertebrates could be restricted at the edges of seagrass beds (edge effect number one). The correlation between the fish and macroinvertebrates in the inner regions was absent on the edges of the small seagrass beds. This suggests that the fish and macroinvertebrates may be attracted to similar features such as an influx of plankton or attracted to each other in the inner regions of small seagrass beds. It is well known that crustaceans are the major food item in the diet of most fish in seagrass beds (Burchmore et al. 1984; Robertson 1984; Edgar & Shaw 1995a; 1995b). Few fish are capable of directly using plant material (Edgar & Shaw 1995a; 1995b). Most of the macroinvertebrates collected in this study were juvenile decapods, which are also the prey items for sixty seven percent of the fish species collected.
Furthermore the significant correlation between the fish (predators) and macroinvertebrates (prey) in the inner regions of small seagrass beds could indicate the mechanism which is responsible for the greater densities of fish species in small compared to larger seagrass beds. If smaller seagrass beds
155 General Discussion have greater prey items than larger seagrass beds, then they can potentially support greater densities of fish species. The smaller seagrass beds, however, did not have concurrently a greater density of macroinvertebrates. Yet a study of predation by fish on copepods in saltmarsh found significant predation rates at the small scale (10 cm) that were not detected at the large scale of whole habitat (i.e. saltmarsh). Kneib (1994) suggested that the replenishment rate of copepod individuals by the incoming tide was high and potentially masked any effects of predation on the size of the population. If this is true for seagrass beds, then predation on macroinvertebrates by fish or other fauna will not be related to the size of the macroinvertebrate population because of replenishment from other sources. In conclusion, any relationship between fish and macroinvertebrates within seagrass beds needs to be investigated at small spatial scale using manipulative experiments such as those conducted by Kneib (1994), Irlandi (1994) with bivalves, Hindell et al. (2000) with fish and Hovel & Lipcius (2001) with crabs.
Numerous researchers have concluded that there is an increased risk of predation on the edges of seagrass beds for bivalves (Irlandi 1994, 1996 & 1997; Irlandi et al. 1995 & 1999), macrofauna (Bologna & Heck 1999) and juvenile blue crabs (Hovel & Lipcius 2002). The commercial fish, red drum, Sciaenops ocellatus, was found in significantly greater abundances on the edges of seagrass beds (Holt et al. 1983). This was thought to reflect the greater success of catching prey at the edges of seagrass beds, concurrent with the need for protection from predators. If larger predators are patrolling the edges of seagrass beds then the smaller fish may be deterred from preying on the macroinvertebrates in this region. This does not, however, entail a concurrent decline in fish abundances at edge regions of the seagrass beds, just a restriction on their movement (edge effect number one).
The designation of the edge region within seagrass beds could be confirmed by a survey whereby each successive region of numerous seagrass beds (measured in 1m increments from the outer perimeter to the centre) is sampled for fish and macroinvertebrates. This would indicate the exact region, if any, where the seagrass fauna change in either species
156 General Discussion composition or abundance. Similarly, hydrodynamic features such as water movement would be quantified in these different regions to test for a correlation between hydrodynamic factors and the seagrass fauna.
6.6 No change in the abundances of macroinvertebrates between small and large seagrass beds The density of macroinvertebrate species was similar in small and large seagrass beds. This was in contrast to the greater density of fish species in smaller seagrass bed. Other researchers (Eggleston et al. 1999) have found greater densities of grass shrimps Palaemondidae, isopods and amphipods in smaller than larger patches of seagrass beds. Bowden et al. (2001) found a greater total number of taxa of infaunal macroinvertebrates in large (over 30m diameter) compared to small (less than 15 m diameter) seagrass beds, although the total number did not differ significantly between patch sizes. In contrast, numerous other studies on macroinvertebrates (clams, Irlandi 1997; scallops, Irlandi et al. 1999; macrofauna, Frost et al. 1999; amphipods and polychaetes Bell et al. 2001) have not detected differences in the assemblages of macroinvertebrates between beds or seagrass patches that vary in size. It could be that the abundances of macroinvertebrates in seagrass beds are replenished by the incoming tide (Kneib 1994) with individuals from the surrounding bare sediments.
6.7 Implications for seagrass regeneration projects Coastal regions are among the most rapidly urbanising places on earth (Crooks and Turner 1999; Ehrenfeld 2000) and this has resulted in unprecedented habitat loss. The destruction of seagrass habitat is a worldwide phenomenon (Walker & McComb 1992) and in Australia alone there is an estimated decline of over 45,000 ha (Walker & McComb 1992). Much of this loss has occurred relatively recently (Larkum et al. 1989) and can be attributed to numerous anthropogenic factors such as propeller scaring and vessel groundings (Sargent et al. 1995), sediment and nutrient loading (Kirkman 1978; Orth & Moore 1983; West 1983; King & Hodgson 1986; Dennison et al. 1993) and land reclamation and changes in land use (Kemp et al. 1983; Larkum & West 1990). Efforts have been made to restore seagrass beds by seeding or transplanting (Fonseca et al. 1997;
157 General Discussion
West et al. 1990), but this is an expensive and time consuming operation with highly variable success rates (unsuccessful; Larkum & West 1983; West et al. 1990; but see Thorhaug 1983; McLaughlin et al. 1983; Thorhaug 1985 for successful results and West et al. 1990 which was successful until destroyed by storms).
The restoration or revegetation of a habitat attempts to restore ecosystem services including the biological interactions of inhabitants and the resilience of the previous habitat (Moberg & Rönnbäck 2003) and not simply replace the lost habitat. To ensure the biodiversity of fauna within seagrass beds, the results in this current study and those of other researchers (Turner et al. 1999; Bell et al. 2001, McNeill & Fairweather 1993) suggest large patches of revegetated seagrass in restoration projects may not be the most important criteria for fish species. Instead, small seagrass patches of equivalent area re-established within areas identified as important for links within the seascape context (Bell et al. 2001) may be more effective in terms of fish species and a better use of restoration dollars. The selection of critical sites for restoration of seagrass beds has demonstrated successful establishment of target marine fauna in mixed habitats in two studies so far (Irlandi & Crawford 1997; Micheli & Peterson 1999).
6.8 Implications for fragmentation of seagrass beds The fragmentation of habitat reduces a continuous habitat into smaller remnants that are isolated usually by a structurally dissimilar matrix (Cox et al. 2003). Not only is the size of the habitat reduced by fragmentation, but also the isolation of that fragment from other habitats increases. Numerous studies have considered the implications of declining habitats for marine organisms by investigating the effects of habitat size on species abundance, recruitment and interactions (Bell & Hicks 1991; McNeill & Fairweather 1993; Irlandi 1997; Eggleston et al. 1998; 1999; Irlandi et al. 1999; Bologna & Heck Jr. 2000; Bowden et al. 2001; Hovel & Lipcius 2001; 2002). Fewer aquatic studies have considered the degree of isolation between habitats and the nature of the surrounding matrix (but see Sogard 1989; Irlandi & Crawford 1997) in aquatic habitats.
158 General Discussion
The effects of seagrass fragmentation on the seagrass fauna may depend more on the landscape context of the habitat, rather than an actual reduction of the available habitat (Bell et al. 2001; Hovel 2003). For example, in a terrestrial study, Collinge (1998) found that large sized species of grassland insects were not reduced in fragmented habitats because the interspace distance between the fragments was small and the insects were using the fragmented patches as one habitat.
Seagrass can exist naturally as a mosaic of patches within a sand matrix because of the hydrodynamic setting (Fonseca and Bell 1998). In this case even anthropogenic-induced fragmentation may not produce spatial patterns that are unnatural (Bell et al. 2001). This current study has found, however, different assemblages of fish in patchy beds compared to uniform beds. This indicates that fragmented seagrass beds contain different fauna to non-fragmented seagrass beds.
The greater density of fish species in the small compared to the large seagrass beds is not evidence that the fragmentation of large into small seagrass beds will produce greater fish biodiversity. This study considered the assemblage of small or juvenile fish in seagrass beds and therefore cannot make any conclusions about the recruitment of juvenile fish into the adult population. Furthermore, the fragmentation of seagrass involves a disturbance regime at the time of impact that can be deleterious for the seagrass fauna. In this current study the seagrass beds did not experience a disturbance regime at the time of sampling.
6.9 Conservation and protection of seagrass beds for estuarine fauna The implementation of marine reserves is still a relatively new conservation strategy based on the idea that protected areas will provide refuge from harvesting or deleterious impacts (Roberts & Polunin 1991; Jones & Andrew 1992; Dugan & Davis 1993; Agardy 1994; García-Charton & Pérez-Ruzafa 1999). Although this body of research is growing it is still tempting to draw experiences from terrestrial conservation but erroneous (Allison et al. 1998). The seascape approach has been considered the most appropriate model
159 General Discussion when designing marine reserves and the literature investigating the functional links between habitats is expanding (Steele 1989; Sogard 1989; Fairweather & Quinn 1992; Robbins & Bell 1994; Irlandi et al. 1995; Bell et al. 1995; Irlandi & Crawford 1997; Hovel & Lipcius 2001; Hovel 2003; Moberg & Rönnbäck 2003) often with fish or other fauna described as vectors or the mobile agents by which habitats are connected. The incorporation of features such as the size, number and distribution of habitats is still unknown for any given target species or ecosystems in the aquatic environment (McNeill 1994; Allison et al. 1998). Furthermore these features may be difficult to define because of the natural fluctuations and variability of marine assemblages. There is no doubt that marine reserves are essential to marine conservation, although their potential effectiveness is limited by large-scale processes that inevitably originate from outside the reserve area (Agardy 1994; Allison et al. 1998). Research that can demonstrate conservation outcomes, such as the type of habitat required by the target organism is essential. Targeting areas such as nursery grounds, spawning grounds and regions of greater species diversity (Allison et al. 1998), of which seagrass beds can be considered all three, are important criteria for the planning of marine reserves. Seagrass beds receive a variable but considerable influx of propagules such as post-larval and juvenile fauna and therefore should be most targeted.
Furthermore in this study a total of ten commercial and recreational fish species were collected from the seagrass beds. Small individuals of non- commercial species dominated the fish assemblages of seagrass, but these species can contribute to the diet of some commercial species (Robertson 1982; Klumpp & Nichols 1983c; Edgar & Shaw 1995b). Also some commercial species that do not physically live in seagrass may benefit from seagrass production through the food chain (Edgar & Shaw 1995a, 1995b).
Ferrell, Worthington, McNeill and Bell (1992) considered that small patchy seagrass beds could provide a variety of habitats for fish that is out of proportion to their size when compared with large uniform seagrass beds. The results of the current study certainly support this. It has been suggested that many small beds of seagrass can increase the probability of interception by larvae or recruits. This increases the overall colonisation of
160 General Discussion the network of beds, as compared to a single larger bed or patch of seagrass (Bell et al. 1987; Sogard 1989; Worthington et al. 1992a; Mc Neil & Fairweather 1993; Eggleston et al. 1998). The recruitment process has been shown to be very variable in time and space (Sogard 1989; Worthington et al. 1992b; McNeill et al. 1992), although the supply of larvae has been shown to be quite predictable spatially (Jenkins et al. 1998; McNeill et al. 1992). This suggests that many smaller beds will have the advantage over one large bed in terms of intercepting larvae and increasing overall recruitment. This many also effect the survival of larvae and recruitment processes. In conservation terms it would be erroneous to consider smaller seagrass beds more conservation worthy than larger seagrass beds. The results in this present study do, however, suggest that small beds should be protected from human impacts and included in marine reserves.
6.10 Final Comment Landscape ecology theories can be applied to the seascape context, and this thesis has supported numerous landscape ecology models in the estuarine seascape. They include the correlation between fauna and; the spatial structure of habitat, the location of a habitat within the greater landscape context and the connectivity of habitats. However, a greater understanding of the influences of marine processes on estuarine faunal assemblages is required, including the delivery of larvae and recruits by tides, currents and up-welling events. These processes can then be incorporated into a seascape model that is unique from the terrestrial understanding of landscape ecology. Furthermore, to manage our coastal ecosystems effectively we require an understanding of the seascape context.
161 References
6 References
Agardy, M.T., (1994). Advances in marine conservation: the role of marine protected areas. Trends in Ecology and Evolution 9: 267-270. Allison, G.W., Lubchenco, J., Carr, M.H., (1998). Marine reserves are necessary but not sufficient for marine conservation. Ecological Applications 8: 579-592. Anderson, M.J., (1998). Effects of patch size on colonisation in estuaries: revisiting the species-area relationship. Oecologia 118: 87-98. Andrèn, H., and Angelstam, P., (1985). Elevated predation rates as an edge effect in habitat islands: Experimental evidence. Ecology 69: 544- 547. Arrhenius, O., (1921). Species and area. Journal of Ecology. 9: 95-99. Attrill, M.J., Strong, J.A., and Rowden, A.A., (2000). Are macroinvertebrate communities influenced by seagrass structural complexity? Ecography 23: 114-121. Austin, H.M., (1971). A survey of the ichthyofauna of the mangroves of western Puerto Rico during December, 1967-August, 1968. Caribbean Journal of Science 11: 27-39. Barbera-Cebrian, C., Sanchez-Jerez, P., and Ramos-Espla, A.A., (2002). Fragmented seagrass habitats on the Mediterranean coast, and distribution and abundance of mysid assemblages. Marine Biology 141: 405-413. Bartlett, D., and Carter, R.W.D., (1991). Seascape ecology: the landscape ecology of the coastal zone. Ecology 10(1): 43-53. Beck, M.W., Heck, K.L., Childers, D.L., Eggleston, D.B., Gillanders, B.M., Halpern, B., Hays, C.G., Hoshino, K., Minello, T.J. Orth, R.J., Sheridan, P.F., Weinstein, M.R., (2001). The identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. Bioscience 51: 633-641. Bell, J.D., Burchmore, J.J., and Pollard, D.A. (1978). Feeding ecology of three sympatric species of leatherjackets (Pisces: Monacanthidae) from a Posidonia seagrass habitat in New South Wales. Australian Journal of Marine Freshwater Research 29: 631-643.
162 References
Bell, J.D., and Harmelin-Vivien, M.L., (1982). Fish fauna of French Mediterranean Posidonia oceanica seagrass meadows. 1. Community Structure. Tethys 10:337-347. Bell, J.D., and Pollard, D.A., (1989). Ecology of fish assemblages and fisheries associated with seagrasses. In Biology of the seagrasses: A treatise on the biology of seagrasses with special reference to the Australian region. (Eds A.W.D. Larkum, A.J. McComb, and S.A. Shepherd) Elsevier, Amsterdam, pp. 565-609. Bell, J.D., Pollard, D.A., Burchmore, J.J., Pease, B.C., and Middleton, M.J. (1984). Structure of a fish community in a temperate tidal mangrove creek in Botany Bay, New South Wales. Australian Journal of Marine and Freshwater Research 35: 33-46. Bell, J.D., Steffe, A.S., and Westoby, M., (1985). Artificial seagrass: how useful is it for field experiments on fish and macroinvertebrates. Journal of Experimental Marine Biology and Ecology 90: 171-177. Bell, J.D., Steffe, A.S., and Westoby, M., (1988). Location of seagrass beds in estuaries: affects on associated fish and decapods. Journal of Experimental Marine Biology and Ecology 122:127-146. Bell, J.D., and Westoby, M., (1986a). Importance of local changes in leaf height and density to fish and decapods associated with seagrasses. Journal of Experimental Marine Biology and Ecology 104: 249-274. Bell, J.D., and Westoby, M., (1986b). Variation in seagrass height and density over a wide spatial scale: effects on common fish and decapods. Journal of Experimental Marine Biology and Ecology 104: 275-295. Bell, J.D., and Westoby, M., (1986c). Abundance of macrofauna in dense seagrass is due to habitat preference, not predation. Oecologia, 68: 205-209. Bell, J.D., Westoby, M., and Steffe, A.S., (1987). Fish larvae settling in seagrass beds of different leaf density? Journal of Experimental Marine Biology and Ecology 111: 134-144. Bell, S.S., Brooks, R.A., Robbins, B.D., Fonseca M.S., Hall. M.O., (2001). Faunal response to fragmentation in seagrass habitats: implications for seagrass conservation. Biological Conservation 100: 115-123. Bell, S.S., Hall, M.O., Robbins, B.D., (1995). Toward a landscape approach in seagrass beds: using microalgal accumulation to address questions of scale. Oecologia 104: 163-168.
163 References
Bell, S.S., Hall, M.O., Soffian, S., Madley K., (2002). Assessing the impact of boat propeller scars on fish and shrimp utilizing seagrass beds. Ecological Applications 12(1): 206-217. Bell, S.S., and Hicks, G.R.F., (1991). Marine landscapes and faunal recruitment: a field test with seagrasses and copepods. Marine Ecology Progress Series 73: 61-68. Bell, S.S., Robbins, B.D., and Jensen, S.L., (1999). Gap dynamics in a seagrass landscape. Ecosystems 2: 493-504. Blaber, S.J.M., (1980). Fish of the Trinity Inlet system of north Queensland with notes on the ecology of fish faunas of tropical Indo-Pacific estuaries. Australian Journal of Marine and Freshwater Research 31: 137-146. Blaber, S.J.M., and Blaber, T.G., (1980). Factors affecting the distribution of juvenile estuarine and inshore fish. Journal of Fish Biology 17: 143- 162. Boehlert, G.W., and Mundy, B.C., (1988). Roles of behavioural and physical factors in larval and juvenile fish recruitment to estuary nursery areas. Amer.Fish.Soc.Symp. 3: 51-67. Bologna, P.A.X., and Heck Jr., K.L., (1999). Differential predation and growth rates of bay scallops within a seagrass habitat. Journal of Experimental Marine Biology and Ecology 239: 299-314. Bologna, P.A.X., and Heck Jr., K.L., (2000). Impacts of seagrass habitat architecture on bivalve settlement. Estuaries 23(4): 449-457. Bologna, P.A.X., and Heck Jr., K.L., (2002). Impact of habitat edges on density and secondary production of seagrass-associated fauna. Estuaries. 25: 1033-44. Bowden, D.A., Rowden, A.A., and Attrill, M.J., (2001). Effect of patch size and in-patch location on the infaunal macroinvertebrate assemblages of Zostera marina seagrass beds. Journal of Experimental Marine Biology & Ecology 259: 133-154. Brooks, R.A., and Bell, S.S., (2001). Mobile corridors in marine landscapes: enhancement of faunal exchange at seagrass / sand ecotones Journal of Experimental Marine Biology and Ecology 264: 67-84. Burchmore, J.J., Pollard, D.A., and Bell, J.D., (1984). Community structure and trophic relationships of the fish fauna of an estuarine Posidonia
164 References
australis seagrass habitat in Port Hacking, NSW. Aquatic Botany 18: 71-87. Burgess, R.L., (1988). Community organization: effects of landscape fragmentation. Canadian Journal of Botany 66: 2687-2690. Cambridge, M.L., and McComb, A.J., (1984). The loss of seagrasses in Cockburn Sound, Western Australia. 1. The time course and magnitude of seagrass decline in relation to industrial development. Aquatic Botany 20: 229-243. Clarke, S.M., and Kirkman, H., (1989). Seagrass dynamics, In: Larkum, A.W.D., McComb, A.J., and Shepherd, S.A. (eds) Biology of Seagrasses. Elsevier Science Publishers, Amsterdam pp. 304-345. Clarke, K.R., and Warwick, R.M., (2001). Change in Marine Communities: an approach to statistical analysis and interpretation. 2nd ed, Plymouth Marine Laboratory, UK. Clynick, B., and Chapman, M.G., (2002). Assemblages of small fish in patchy mangrove forests in Sydney Harbour. Marine Freshwater Research 53: 669-677. Cocheret de la Moriniere, E., Pollux, B.J.A., Nagelkerken, I., and van der Velde G. (2002). Post-settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuarine Coastal Shelf Science 55: 309-321. Coen, L., Heck Jr, K.L., Abel, L.G., (1981). Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62: 1484- 1493. Coles, R.G., Lee Long, W.J., Watson, R.A., and Derbyshire, K.J., (1993). Distribution of seagrasses, and their fish and penaeids prawn communities in Cairns Harbour, a tropical estuary, northern Queensland, Australia. Australian Journal of Marine and Freshwater Research 44: 193-210. Collinge, S.K., (1998). Spatial arrangement of habitat patches and corridors: clues from an ecological field experiment. Landscape Urban Planning 42: 157-162. Conacher, C.A., Poiner, I.R., and Donohue, M.O., (1994). Morphology, flowering and seed production of Zostera capricorni Aschers in subtropical Australia. Aquatic Botany 49: 33-46.
165 References
Connell, S.D., (1998). Effects of predators on growth, mortality and abundance of a juvenile reef-fish: evidence from manipulations of predator and prey abundance. Marine Ecology Progress Series 169: 251-261. Connell, S.D., and Anderson, M.J., (1999). Predation by fish on assemblages of intertidal epibiota: effects of predator size and patch size. Journal of Experimental Marine Biology and Ecology 241: 15-29. Connell, S.D., and Kingsford, M.J., (1998). Spatial, temporal and habitat related variation in the abundance of large predatory fish at One Tree Reef, Australia. Coral Reefs 17: 49-57. Connolly, R.M., (1994). A comparison of fish assemblages from seagrass and unvegetated areas of a southern Australian estuary. Australian Journal of Marine and Freshwater Research. 45: 1033-1044. Connolly, R.M., Jenkins, G., and Loneragan, N., (1999). Seagrass dynamics and fisheries sustainability. In Butler, A., and Jernakoff, P. (eds) Seagrass in Australia CSIRO Pub., Collingwood Vic. Aust. pp. 25-59. Connor, E.F., and McCoy, E.D., (1979). The statistics and biology of the species-area relationship. The American Naturalist 113, 791-833. Cox, M.P., Dickman, C.R., and Hunter, J., (2003). Effects of rainforest fragmentation on non-flying mammals of the eastern Dorrigo Plateau, Australia. Biological Conservation 115: 175-189. Crawford, M.K., Crowder, L.B., and Fonseca, M.S., (1995). Spatial sampling of fish in meadows. Twenty-third Benthic Ecology Meeting 1995. Crooks, S., and Turner, R.K., (1999). Integrated coastal management: sustaining estuarine natural resources. Advan.Ecol.Res. 29: 241-289. Danielson, B.J., (1991). Communities in a landscape: the influence of habitat heterogeneity on the interactions between species. The American Naturalist 138(5):1105-1120. de Boer, W.F., (2000). Biomass dynamics of seagrasses and the role of mangrove and seagrass vegetation as different nutrient sources for an intertidal ecosystem. Aquatic Botany 66: 225-239. den Hartog, C., (1970). The Seagrasses of the World. North Holland Publishing Co. Amsterdam. Dennison, W.C., Orth, R.J., Moore, K.A., Stevenson, J.C., Cater, V., Dollar, S., Bergstrom, P.W., and Batiuk, R.A., (1993). Assessing water quality with submerged aquatic vegetation. Bioscience 43: 86-94.
166 References
Diamond, J.M., & May, R.M., (1976). Island biogeography and the design of natural reserves. In, Theoretical Ecology (ed) May, R.M., Philadelphia, Saunders. pp.163-86. Dugan, J.E., and Davis, G.E., (1993). Applications of marine refugia to coastal fisheries management. Canadian Journal of Fisheries and Aquatic Sciences 50: 2029-2042 Eckman, J.E., (1983). Hydrodynamic processes affecting benthic recruitment. Limnology and Oceanography 28: 241-257. Edgar, G.J., (1997). Australian Marine Life Reed Pub., Kew, Victoria. Edgar, G.J., and Robertson, A.I., (1992). The influence of seagrass structure on the distribution and abundance of mobile epifauna: pattern and process in a Western Australian Amphibolis bed. Journal of Experimental Marine Biology and Ecology 160: 13-31. Edgar, G.J., and Shaw, C., (1995a). The production and trophic ecology of shallow-water fish assemblages in southern Australia. I. Species richness, size-structure and production of fishes in Western Port, Victoria. Journal of Experimental Marine Biology and Ecology 194: 53- 81. Edgar, G.J., and Shaw, C., (1995b). The production and trophic ecology of shallow-water fish assemblages in southern Australia. II. Diets of fish and trophic relationships between fishes and benthos at Western Port, Victoria. Journal of Experimental Marine Biology and Ecology 194: 83- 106. Edgar, G.J., and Shaw, C., (1995c). The production and trophic ecology of shallow-water fish assemblages in southern Australia. III. General relationships between sediments, seagrasses, invertebrates and fishes. Journal of Experimental Marine Biology and Ecology 194, 107- 131. Edgar, G.J., and Robertson, A.I., (1992). The influence of seagrass structure on the distribution and abundance of mobile epifauna: pattern & process in a Western Australian Amphibolis bed. Journal of Experimental Marine Biology and Ecology 160:13-31. Edgar, G.J., (1990). The influence of plant structure on the species richness, biomass and secondary production of macrofaunal assemblages associated with Western Australian seagrass beds. Journal of Experimental Marine Biology and Ecology 137: 215-240.
167 References
Edgar, G.J., (1996). The distribution and diets of crabs associated with seagrass and unvegetated habitats in Western Port, southeastern Australia. In Kuo, J., Phillips, R.C., Walker, D.I., and Kirkman, H., (Eds) Seagrass Biology: Proceedings of an International Workshop Faculty of Sciences, The University of Western Australia, Perth pp. 225-232. Eggleston, D.B., Elis, W.E., Etherington, L.L., Dahlgren, C.P., and Posey, M.H., (1999). Organism responses to habitat fragmentation and diversity: Habitat colonization by estuarine macrofauna. Journal of Experimental Marine Biology and Ecology 236: 107-132. Eggleston, D.B., Etherington, L.L., and Elis, W.E., (1998). Organism response to habitat patchiness: Species and habitat-dependent recruitment of decapod crustaceans. Journal of Experimental Marine Biology and Ecology 223: 111-132. Ehrenfeld, J.G., (2000). Evaluating wetlands within an urban context. Ecological Engineering 15: 253-265. Environmental Systems Research Institute, Inc., (1996). Arc View GIS. Environmental Systems Research Institute, Inc, CA. USA. Etherington, L.L., and Eggleston, D.B., (2000). Large scale blue crab recruitment: linking post-settlement planktonic dispersal, and multiple nursery habitats. Marine Ecology Progress Series 204: 178- 198. Fagan, W.F., Cantrell, R.S., Cosner, C., (1999). How habitat edges change species interactions. American Naturalist 153(2): 165-182. Fairweather, P.G., and Quinn, G.P., (1992). Seascape ecology, the importance of linkages. In Battershill, C.N., Schiel, D.R., Jones, G.P., Creese, R.G., MacDiarmid, A.B., (eds.). Proceedings of the Second International Temperate Reef Symposium, 7-10 January 1992, Auckland, New Zealand, NIWA Marine, Wellington NZ, pp. 77-83. Ferrell, D.J., and Bell, J.D., (1991). Differences among assemblages of fish associated with Zostera capricorni and bare sand over a large spatial scale. Marine Ecology Progress Series 72: 15-24. Ferrell, D.J., Worthington, D.G., McNeill, S.E., and Bell, J.D., (1992). Assemblages of fish and macro-invertebrates associated with seagrass In Jervis Bay Marine Ecology Study (final report) Fisheries Research Institute, NSW Fisheries Cronulla NSW.
168 References
Fonseca, M.S., and Bell, S.S., (1998). Influence of physical setting on seagrass landscapes near Beaufort, North Carolina, USA. Marine Ecology Progress Series 171: 109-121. Fonseca, M.S., Fisher, J.S., Zieman, J.C. and Thayer, G.W., (1982). Influence of the seagrass, Zostera marina, on current flow. Estuarine Coastal and Shelf Science 15: 351-364. Fonseca, M.S., Kenworthy, W.J., and Thayer, G.W., (1997). Transplanting of the seagrass Halodule wrightii, Syringodium filiforme and Thalassia testudinum for sediment stabilisation and habitat development in the south-east region of the United States. Technical Report EL-97-8 for the US Army Corps of Engineers, South-eastern Fisheries Center, National Marine Fisheries Service, Beaufort, NC. Fortes, M.D., (1991). Seagrass-mangrove ecosystems management: a key to marine coastal conservation in the ASEAN region. Marine Pollution Bulletin EMECS`90. 23:113-116. Fox, B.J., Taylor, J.E., Fox, M.D., and Williams, C., (1997). Vegetation changes across edges of rainforest remnants. Biological Conservation 82: 1- 13. France, R.L., and Holmquist, J.G., (1997). C13 variability of macroalgae: effects of water motion via baffling by seagrass and mangroves. Marine Ecology Progress Series 149: 305-308. Frost, M T, Rowden, A A, and Attrill, M J (1999). Effect of habitat fragmentation on the macroinvertebrate infaunal communities associate with the seagrass Zostera marina L. Aquatic Conservation: Marine Freshwater Ecosystem 9: 255-263. García-Charton, J.A., and Pérez-Ruzafa, Á., (1999). Ecological heterogeneity and the evaluation of the effects of marine reserves. A mini-review. Fisheries Research 42: 1-20. Geoscience Australia, (2001). Oz Estuaries Database at http://www.ozestuaries.org/ accessed 23 August 2004. Gleason, H.A., (1922). On the relationship between species and area. Ecology 3: 158-62. Gleason, H.A., (1925). Species and area. Ecology 6, 66-74. Goeden, G.B., (1979). Biogeographic theory as a management tool. Environmental Conservation 6: 27-32.
169 References
Gray, C.A., Chick, R.C., and McElligott, D.J., (1998). Diel changes in assemblages of fishes associated with shallow seagrass and bare sand. Estuarine Coastal and Shelf Science 46: 849-859. Gray, C.A., McElligott, D.J., and Chick, R.C., (1996). Intra- and inter-estuary differences in assemblages of fishes associated with shallow seagrass and bare sand. Marine and Freshwater Research 47: 723-735. Griffiths, S.P., (2001). Diel variation in the seagrass ichthyofaunas of three intermittently open estuaries in south-eastern Australia: implications for improving fish diversity assessments. Fish Management & Ecology 8: 123-140. Guest, M., Connolly, R.M., and Loneragan, N.R., (2003). Seine nets and beam trawls compared in the day and night for sampling fish and crustaceans in shallow seagrass habitat. Fisheries Research 64: 185- 196 Hannan, J.C., and Williams, R.J., (1998). Recruitment of juvenile marine fishes to seagrass habitat in a temperate Australian estuary. Estuaries 21: 29-51. Harlin, M.M., (1975). Epiphyte-host relations in seagrass communities. Aquatic Botany 1: 125-131. Harrison, P.G., and Mann, K.H., (1975). Detritus formation from eelgrass (Zostera marina L.): the relative effects of fragmentation, leaching and decay. Limnology and Oceanography 20: 924-934. Heck Jr., K.L., (1979). Some determinants of the composition and abundance of motile macroinvertebrate species in tropical and temperate turtlegrass (Thalassia testudinum) meadows. Journal of Biogeography 6, 183-200. Heck Jr., K.L., and Crowder, L.B., (1991). Habitat structure and predator prey interactions in vegetated aquatic systems. In: Bell, S.S., McCoy, E.D., and Mushinsky, H.R., (Eds.), Habitat Structure: The Physical Arrangement of Objects in Space, Chapman & Hall NY, pp.47-68. Heck Jr., K.L., and Orth, R.J., (1980). Seagrass habitats: the role of habitat complexity, competition and predation in structuring associated fish & motile macroinvertebrate assemblages. In: Kennedy, V.S., (ed) Estuarine Perspectives. Academic Press, NY, pp.449-461.
170 References
Heck Jr., KL., and Thoman T.A., (1984). The nursery role of seagrass meadows in the upper and lower reaches of the Chesapeake Bay Estuaries 7 (1), 70-92. Heck Jr., K.L., and Wetstone, G.S., (1977). Habitat complexity and invertebrate species richness ad abundance in tropical seagrass meadows. Journal of Biogeography 4: 135-142. Hemminga, M.A., Harrison, P.G., and Vanlent, F., (1991). The balance of nutrient losses and gains in seagrass meadows. Marine Ecology Progress Series 71: 85-96. Hettler Jr., W.F., (1989). Nekton use of regularly flooded saltmarsh cordgrass habitat in North Carolina, USA. Marine Ecology Progress Series 56: 111-118. Hindell, J.S., and Jenkins, G.P., (2004). Spatial and temporal variability in the assemblage structure of fishes associated with mangroves (Avicennia marina) and intertidal mudflats in temperate Australian embayments. Marine Biology 144: 385-395. Hindell, J.S., Jenkins, G.P., and Keough, M.J., (2000). Evaluating the impact of predation by fish on the assemblage structure of fishes associated with seagrass (Heterozostera tasmanica) (Martens ex Ascherson) den Hartog, and unvegetated sand habitats. Journal of Experimental Marine Biology and Ecology 255: 153-174. Hixon, M.A., (1991). Predation as a process structuring coral reef fish communities. In Sale, P.F., (ed) The ecology of fishes on coral reefs. Academic Press, San Diego, p. 475-508. Holmquist, J.G., (1994). Benthic macroalgae as a dispersal mechanism for fauna: Influences of a marine tumbleweed. Journal of Experimental Marine Biology and Ecology 180: 235-251. Holt, S.A., Kitting C.L., & Arnold, C.R., (1983). Distribution of young red drums among different seagrass meadows. Transactions of the American Fisheries Society 112: 267-271. Hovel, K.A., and Lipcius, R.N., (2001). Habitat fragmentation in a seagrass landscape: patch size and complexity control blue crab survival. Ecology 82: 1814-1829. Hovel, K.A., and Lipcius, R.N., (2002). Effects of seagrass habitat fragmentation on juvenile blue crab survival and abundance. Journal of Experimental Marine Biology and Ecology 271: 75-98.
171 References
Hovel, K.A., (2003). Habitat fragmentation in marine landscapes: relative effects of habitat cover and configuration on juvenile crab survival in California and North Carolina seagrass beds. Biological Conservation 110: 401-412. Howard, R.K., (1989). The structure of a nearshore fish community of Western Australia: diel patterns and the habitat role of limestone reefs. Environmental Biology of Fishes 24: 93-104. Hutomo, M., and Peristiwady,T., (1996). Diversity, abundance and diet of fish in the seagrss beds of Lombok Island, Indonesia. In Seagrass Biology: Proceedings of an International Workshop, Rottnest Island, Western Australia. (Eds) J. Kuo, R.C. Phillips, D.I. Walker & H. Kirkman, pp. 205-212. Irlandi, E.A., (1994). Large- and small-scale effects of habitat structure on rates of predation: How percent coverage of seagrass affects rates of predation and siphon nipping on an infaunal bivalve. Oecologia 98: 176-183. Irlandi, E.A., (1996). The effect of patch size and energy regime on growth of an infaunal bivalve. Journal of Marine Research 54: 1-26. Irlandi, E.A., (1997). Seagrass patch size and survivorshop of an infaunal bivalve. Oikos 78: 511-518. Irlandi, E.A., Ambrose, W.G., and Orlando, B.A., (1995). Landscape ecology and the marine environment: how spatial configuration of seagrass habitat influences growth and survival of the bay scallop Oikos 72: 307-313. Irlandi, E.A., and Crawford, M.K., (1997). Habitat Linkages: the effect of intertidal saltmarshes and adjacent subtidal habitats on abundance, movement, and growth of an estuarine fish Oecologia 110: 222-230. Irlandi, E.A., Orlando, B.A., and Ambrose, Jr., W.G., (1999). Influence of seagrass habitat patch size on growth and survival of juvenile bay scallops Argopecten irradians concentricus (Say). Journal of Experimental Marine Biology and Ecology 235: 21-43. Jackson, E.L., Rowden, A.A., Attrill, M.J., Bossy, S.F., Jones M.B., (2001). The importance of seagrass beds as a habitat for fishery species. Oceanography and Marine Biology 39: 269-303. Jackson, E.L., Rowden, A.A., Attrill, M.J., Bossy, S.F., Jones M.B., (2002). Comparison of fish and mobile macroinvertebrates associated with
172 References
seagrass and adjacent sand at St. Catherine Bay, Jersey (English Channel): Emphasis on commercial species. Bulletin of Marine Science 71(3): 1333-1341. Jackson, J.B.C., (1977). Habitat area, colonisation, and development of epibenthic community structure. In: Keegan, B.F., Ceidigh, P.O., Boaden, P.J.S., (eds) Biology of Benthic Organisms. Pergamon, London, pp. 349-358. James, P.L., and Heck Jr., K.L., (1994). The effects of habitat complexity and light intensity on ambush predation within a simulated seagrass habitat. Journal of Experimental Marine Biology and Ecology 176: 187-200. Janzen, D.H., (1986). The external threat. In; Soule, M.E., (ed.) Conservation Biology; the science of scarcity and diversity. Sinauer, Mass. pp. 286- 303. Jeffery, C.J., (2000). Settlement in different-sized patches by the gregarious intertidal barnacle Chamaesipho tasmanica Foster and Anderson in New South Wales. Journal of Experimental Marine Biology and Ecology 252: 15-26. Jelbart, J.E., and Ross, P.M., (2003). What features matter when designing protected areas for fish in beds of seagrass: A review. In Beumer, J.P., Grant A., and D.C. Smith (eds). World Congress on Aquatic Protected Areas proceedings, Cairns, Australia August 2002. University Queensland Press, St Lucia Qld. Australia. Jenkins, G.P., Black, K.P., Wheatley, M.J., and Hatton, D.N., (1997b). Temporal and spatial variability in recruitment of a temperate, seagrass associated fish is largely determined by physical processes in the pre- and post-settlement phases. Marine Ecology Progress Series 148: 23-35. Jenkins, G.P., Edgar, G.J., May, H.M.A., and Shaw, C., (1993). Ecological basis for parallel declines in seagrass habitat and catches of commercial fish in Western Port Bay, Victoria. In Sustainable Fisheries through Sustaining Fish Habitat. Australian Society for Fish Biology Workshop, Victor Harbour, SA, 12-13 August 1992, (Ed. D.A. Hancock). Bureau of Resource Sciences Proceedings (AGPS, Canberra). Jenkins, G.P., Keough, M.J., and Hamer, P.A., (1998). The contributions of habitat structure and larvae supply to broad-scale recruitment
173 References
variability in a temperate zone, seagrass associated fish. Journal of Experimental Marine Biology and Ecology 226: 259-278. Jenkins, G.P., May, H.M., Wheatley, A.M.J., and Holloway, M.G., (1997a). Comparison of fish assemblages associated with seagrass and adjacent unvegetated habitats of Port Phillip Bay and Corner Inlet, Victoria, Australia with emphasis on commercial species. Estuarine Coastal and Shelf Science 44: 569-588 Jenkins, G.P., and Sutherland, C.R., (1997). The influence of habitat structure on nearshore fish assemblages in a southern Australian embayment: Colonisation and turn-over rate of fishes associated with artificial macrophyte beds of varying physical structure. Journal of Experimental Marine Biology and Ecology 218: 103-125. Jenkins, G.P., and Wheatley, M.J., (1998). The influence of habitat structure on nearshore fish assemblages in a southern Australian embayment: Comparison of shallow seagrass, reef algal, and unvegetated habitats, with emphasis on their importance to recruitment. Journal of Experimental Marine Biology and Ecology 221: 147-172. Jenkins, G.P., Wheatley, M.J., and Poore, A.G.B., (1996). Spatial variation in recruitment, growth and feeding of post settlement King George whiting, Sillaginodes punctata, associated with seagrass beds of Port Phillip Bay, Australia. Canadian Journal of Fisheries & Aquatic Sciences 53: 96-105. Jones, G.P., and Andrew, N.L., (1992). Temperate reefs and the scope of seascape ecology. In Battershill, C.N., Schiel, D.R., Jones, G.P., Creese, R.G., MacDiarmid, A.B., (eds.). Proceedings of the Second International Temperate Reef Symposium, 7-10 January 1992, Auckland, New Zealand, NIWA Marine, Wellington NZ, pp. 63-76. Jones, R.S., and Chase, J.A., (1975). Community structure and distribution of fishes in an enclosed high island lagoon in Guam, Micronesica 11: 127-148. Kay, A.M., and Keough, M.J., (1981). Occupation of patches in the epibenthic communities on pier pilings and the bivalve Pinna bicolour at Edithburg, South Australia. Oecologia 48: 123-130. Kemp, W.M., Boynton, W.R., Twilley, R.R., Stevenson, J.C., and Means, J.C., (1983). The decline of submerged vascular plants in Upper
174 References
Chesapeake Bay: summary of results concerning possible causes. Mar. Tech. Soc. J. 17: 78-89. Kenyon, R.A., Haywood, M.D.E., Heales, D.S., Loneragan, N.R., Pendrey, R.C., and Vance, D.J., (1999). Abundance of fish and crustacean postlarvae on portable artificial seagrass units: daily sampling provides quantitative estimates of the settlement of new recruits. Journal of Experimental Marine Biology and Ecology 232: 197-216. Keough, M.J., (1984). Effects of patch size on the abundance of sessile marine invertebrates. Ecology 65: 423-437. Kim, J.H., and De Wreede, R.E., (1996). Effects of size and season of disturbance on algal patch recovery in a rocky intertidal community. Marine Ecology Progress Series 133: 217-228. King, R.J., and Hodgson, B.R., (1986). Aquatic angiosperms in coastal saline lagoons of New South Wales. IV. Long-term changes. Proceedings Linn. Society N.S.W. 109: 51-60. Kingsford, M.J., and Suthers, I.M., (1996). The influence of tidal phase on patterns of ichthyoplankton abundance in the vicinity of an estuarine front, Botany Bay, Australia. Estuarine Coastal and Shelf Science 43: 33-54. Kirkman, H., (1978). Decline of seagrass in northern areas of Morton Bay, Queensland. Aquatic Botany 5: 63-76. Kjerfve, B., Seim, H.E., Blumberg, A.G., and Wright, L.D., (1992). Modelling of the residual circulation in Broken Bay and the Lower Hawkesbury River, NSW. Australian Journal of Marine and Freshwater Research 43: 1339-57. Klumpp, D.W., and Nichols, P.D., (1983a). Utilisation of the seagrass Posidonia australis as food by the rock crab Nectocarcinus integrifons (Labreille) (Crustacea: Decapoda: Portunidae). Marine Biology Letters 4: 331-339. Klumpp, D.W., and Nichols, P.D., (1983b). Nutrition of the southern sea garfish Hyporhamphus melanochir: gut passage rate and daily consumption of two food types and assimilation of seagrass components. Marine Ecology Progress Series 12: 207-216. Klumpp, D.W., and Nichols, P.D., (1983c). A study of food chains in seagrass communities II. Food of the rock flathead, Platycephalus laevigatus
175 References
Cuvier, a major predator in a Posidonia australis seagrass bed Australian Journal of Marine and Freshwater Research 34: 745-754. Klumpp, D.W., Howard, R.K., and Pollard, D.A., (1989). Trophodynamics and nutritional ecology of seagrass communities. In: Larkum, A.W.D., McComb, A.J., and Shepherd, S.A. (eds) Biology of Seagrasses. Elsevier Science Publishers, Amsterdam pp. 394-457. Kneib, R.T., (1994). Spatial pattern, spatial scale, and feeding in fishes. In Theory and Application in Fish Feeding Ecology. (Eds. D.J. Stouder, K.L. Fresh, R.J. Feller). Uni. Of South Carolina Press, South Carolina USA, pp. 171-185. Knowles, L.L., and Bell, S.S., (1998). The influence of habitat structure in faunal habitat associations in a Tampa Bay seagrass system, Florida. Bulletin Marine Science 62(3): 781-794. Kotliar, N.B., and Wiens, J.A., (1990). Multiple scales of patchiness and patch structure: A hierarchical framework for the study of heterogeneity. Oikos 59: 253-260. Kuiter, R.H., (2000). Coastal Fishes of South Eastern Australia, Gary Allen, Sydney Australia. Kulcycki, G.R., Virnstein, R.W., and Nelson, W.G., (1981). The relationship between fish abundance and algal biomass in a seagrass-drift community. Estuarine Coastal and Shelf Science 12: 341-347. Kuo, J., and McComb, A.J., (1989). Seagrass taxonomy, structure and development. In Larkum, A.W.D., McComb, A.J., and Shepherd, S.A. (eds) Biology of Seagrasses. Elsevier Science Publishers, Amsterdam pp. 6-73. Laegdsgaard, P., and Johnson, C.R., (1995). Mangrove habitats as nurseries: unique assemblages of juvenile fish in subtropical mangroves in eastern Australia. Marine Ecology Progress Series 126: 67-81. Laegdsgaard, P., and Johnson, C.R., (2001). Why do juvenile fish utilise mangrove habitats? Journal of Experimental Marine Biology and Ecology 257: 229-253. Larkum, A.W.D., McComb, A.J., and Shepherd, S.A., (1989). Seagrass Ecosystems – an Australian perspective, Elsevier Scientific Pub. Co., Amsterdam. Larkum, A.W.D., and West, R.J., (1983). Stability, depletion and restoration of seagrass beds. Proceedings Linn. Soc. N.S.W. 106: 201-212.
176 References
Larkum, A.W.D., and West, R.J., (1990). Long-term changes of seagrass meadows in Botany Bay, Australia. Aquatic Botany 37: 55-70. Leber, K.M., (1985). The influence of predatory decapods, refuge, and microhabitat selection on seagrass communities. Ecology 66: 1951- 1964. Leopold, A., (1933). Game Management. Charles Scribner Sons, New York. Levin, P., Petrik, R., and Malone, J., (1997). Interactive effects of habitat selection, food supply and predation on recruitment of an estuarine fish. Oecologia 112: 55-63. Lidicker, W.Z., (1999). Responses of mammals to habitat edges: an overview. Landscape Ecology 14: 333-343. Lidicker, W.Z., and Peterson, J.A., (1999). Responses of small mammals to habitat edges. In Landscape ecology of small mammals. (Eds. G.W. Barrett and J.D. Peles). Springer-Verlag, New York. pp. 211-227. Little, M.C., Reay, P.J., and Grove, S.J., (1988). The fish community of an eastern African mangrove creek. Journal of Fish Biology 32: 729-747. Loneragan, N.R., Potter, I.C., Lenanton, R.C.J., and Caputi, N., (1986). Spatial and seasonal differences in the fish fauna in the shallow of a large Australian estuary. Marine Biology 92: 575-86. Loneragan, N.R., and Potter, I.C., (1990). Factors influencing the community structure and distribution of different life-cycle categories of fishes within the shallow waters of a large Australian estuary. Marine Biology 106: 25-37. MacArthur, L., (1997). Distributions, size compositions and diets of different species of the Odacidae in south-western Australia: evidence for resource partitioning among species. Honours Thesis, Murdoch University, pp. 77. MacArthur, R., and Wilson, E.O., (1967). The Theory of Island Biogeography. Princeton Uni Press Princeton NJ. MacDonald, C.M., (1992). Fluctuations in seagrass habitats and commercial fish catches in Western Bay and the Gippsland Lakes, Victoria. In Recruitment Processes. Australian Society for Fish Biology Workshop, Hobart, 21 August 1991, (Ed. D.A. Hancock). Bureau of Resource Sciences Proceedings. No. 16. (AGPS, Canberra). Main, K.L., (1987). Predator avoidance in seagrass meadows: Prey behavior, microhabitat selection, and cryptic coloration. Ecology 68: 170-180.
177 References
Martin, F.D., and Cooper, M., (1981). A comparison of fish faunas found in pure stands of two tropical Atlantic seagrasses, Thalassia testudinum and Syringodium filiforme. North-east Gulf Science 5: 31-37. Mazerolle, M.J., and Villard M.A., (1999). Patch characteristics and landscape context as predictors of species presence and abundance: A Review. Ecoscience 6(1): 117-124. McGuinness, K.A., (1984a). The species-area relations of the communities on intertidal boulders: testing the null hypothesis. Journal of Biogeography 11: 439-456. McGuinness, K.A., (1984b). Equations and explanations in the study of species-area curves. Biological Review 59: 423-440. McLaughlin, P.A., Treat, S.F., Thorhaug, A., and Lemaitre, R., (1983). A restored seagrass (Thalassia bed) and its animal community. Environmental Conservation 10: 247-254. McNeill, S.E., (1994). The selection and design of marine protected areas: Australia as a case study. Biodiversity and Conservation 3: 586-605. McNeill, S.E., and Fairweather, P.G., (1993). Single large or several small marine reserves? An experimental approach with seagrass fauna. Journal of Biogeography 20: 429-440. McNeill, S.E., Worthington, D.G., Ferrell, D.J., and Bell, J.D., (1992). Consistently outstanding recruitment of five species of fish to a seagrass bed in Botany Bay, NSW. Australian Journal of Ecology 17: 359-365. Micheli, F., and Peterson, C.H., (1999). Estuarine vegetated habitats as corridors for predator movements. Conservation Biology 13(4): 869- 881. Middleton, M.J., Bell, J.D., Burchmore, J.J., Pollard, D.A., and Pease, B.C., (1984). Structural differences in the fish communities of Zostera capricorni and Posidonia australis seagrass meadows in Botany Bay, NSW. Aquatic Botany 18: 89-109. Minchington, T.E., (1997). Life on the edge: conspecific attraction and recruitment of populations to disturbed habitats. Oecologia 111: 45- 52. Moberg, F., and Rönnbäck, P., (2003). Ecosystem services of the tropical seascape: interactions, substitutions and restoration. Ocean and Coastal Management 46: 27-46.
178 References
Morton, RM., (1990). Community structure, density and standing crop of fishes in a subtropical Australian mangrove area. Marine Biology 105: 385-394. Murphey, P.E., and Fonseca, M.S., (1995). Role of high and low energy seagrass beds as nursery areas for Penaeus duorarum in North Carolina. Marine Ecology Progress Series 121: 91-98. Nagelkerken, I., Kleijnen, S., Klop, T., van den Brand, R.A.C.J., de la Moriniere, E.C., and van der Velde, G., (2001). Dependence of Caribbean reef fishes on mangroves and seagrass beds as nursery habitats: a comparison of fish faunas between bays with or without mangroves/seagrass beds. Marine Ecology Progress Series 214: 225- 235. Nagelkerken, I., van der Velde, G., Gorissen, M.W., Meijer, G.J., van’t Hof, T., den Hartog, C., (2000). Importance of mangrove forests, seagrass beds and the shallow coral reef as a nursery for important coral reef fishes using a visual census technique. Estuarine Coastal Shelf Science 51: 31-44. Nagelkerken, I. and van der Velde, G., (2004a). Are Caribbean mangroves important feeding grounds for juvenile reef fish from adjacent seagrass beds? Marine Ecology Progress Series 274: 143-151. Nagelkerken, I. and van der Velde, G., (2004b). Relative importance of interlinked mangroves and seagrass beds as feeding habitats for juvenile reef fish on a Caribbean Island. Marine Ecology Progress Series 274: 153-159. National Tide Facility, (2000). NSW Tides 2000-2001. National Tide Facility Flinders University, South Australia. Navarette, S.A., and Castilla, J.C., (1990). Barnacle walls as mediators of intertidal mussel recruitment: effects of patch size on the utilisation of space. Marine Ecology Progress Series 68: 113-119. Odum, W.E., and Heald, E.J., (1972). Trophic analysis of an estuarine mangrove community. Bulletin of Marine Science 22: 671-738. Ogden, J.C., and Gladfelter, E.H., (eds.) (1983). Coral reefs, seagrass beds and mangroves: their interaction in the coastal zones of the Caribbean. UNESCO Reports in Marine Science 23: 1-333. Olesen, B., and Sand-Jensen, K., (1994a). Patch dynamics of eelgrass Zostera marina. Marine Ecology Progress Series 106: 147-156.
179 References
Olesen, B., and Sand-Jensen, K., (1994b). Biomass-density patterns in the temperate seagrass Zostera marina. Marine Ecology Progress Series 109: 283-291. Orth, R.J., Heck Jr., K.L., and van Montfrans, J., (1984). Faunal communities in seagrass beds: a review of the influence of plant structure and prey characteristics in predator-prey relationships. Estuaries 7: 339-350. Orth, R.J., (1992). A perspective on plant-animal interactions in seagrasses: physical and biological determinants influencing plant and animal abundance. In: John, D.M., Hawkins, S.J., Price, J.H., (eds.), Plant- Animal Interactions in the Marine Benthos. Association Special Volumes, Vol. 46. Clarendon Press, Oxford, pp. 147-164. Orth, R.J., and Moore, K.A., (1983). Chesapeake Bay: An unprecedented decline in submerged aquatic vegetation. Science 22: 51-52. Orth, R.J., and Moore, K.A., (1986). Seasonal and year to year variations in the growth of Zostera marina L. (eelgrass) in the lower Chesapeake Bay. Aquatic Botany 24: 335-341. Otway, N.M., and Macbeth, W.G., (1999). Physical affects of hauling on seagrass NSW Fisheries Final Report Series No.15, NSW Fisheries Research Institute Cronulla, Sydney. Paine, R.T., and Levin, S.A., (1981). Intertidal landscapes: Disturbance and the dynamics of pattern. Ecological Monographs 51:145-178. Parrish, J.D., (1989). Fish communities of interacting shallow water habitats in tropical oceanic regions. Marine Ecology Progress Series 58: 151- 160. Peterson, C.G.L., (1918). The sea bottom and its production of fish foods; a survey of the work done in connection with valuation of Danish waters from 1883-1917. Report of the Danish Biological Station 21: 1- 62. Peterson, G.W., and Turner R.E., (1994). The value of salt marsh edge vs interior as a habitat for fish and decapod crustaceans in a Louisiana tidal marsh. Estuaries 17(1B): 235-262. Pianka, E.R., (1974). Evolutionary Ecology. Harper and Row, London Pollard, D.A., (1984). A review of ecological studies on seagrass-fish communities, with particular reference to recent studies in Australia. Aquatic Botany 18: 3-42.
180 References
Potter, I.C., Loneragan, N.R., Lenanton R.C.J., Chrystal, P.J., Grant, C.J., (1983). Abundance, distribution and age structure of fish populations in a Western Australian estuary. Journal of Zoology 200: 21-50. Preston, F.W., (1960). Time and space and the variation of species. Ecology 41: 611-627. Pulliam, H.R., (1988). Sources, sinks and population regulation. The American Naturalist 132: 652-661. Pulliam, H.R., and Danielson, B.J., (1991). Sources, sinks and habitat selection: a landscape perspective on population dynamics. The American Naturalist 137: S50-S66. Quinn, G.P., (2002). Experimental design and data analysis for biologists. Cambridge University Press, U.K. Raimondi, P.T., (1990). Patterns, mechanisms, consequences of variability in settlement and recruitment of an intertidal barnacle. Ecological Monographs 60: 283-309. Rajendran, N., and Kathiresan, K., (1999). Do decomposing leaves of mangrove forests attract fishes? Current Science 77(7), 972-976. Ramage, D.L., and Schiel, D.R., (1999). Patch dynamics and response to disturbance of the seagrass Zostera novazelandica on intertidal platforms in southern New Zealand. Marine Ecology Progress Series 189: 275-288. Ramm, D.C., (1986). An ecological study of the icthyoplankton and juvenile fish in the Gippsland Lakes, Victoria. PhD Thesis, University of Melbourne, Victoria, Australia. Risser, P.G., Karr, J.R., Forman, R.T.T. (1984). Landscape Ecology: Directions and Approaches. Special Publ. No. 2. Natural History Survey, Champaign Illinios. Robbins, B.D., and Bell, S.S., (1994). Seagrass landscapes: a terrestrial approach to the marine subtidal environment. Trends in ecology and evolution 9(8), 301-304. Roberts, C.M., and Polunin, N.V.C., (1991). Are marine reserves effective in management of reef fisheries? Rev. Fish. Biol. Fish. 1: 65-91. Robertson, A.I., (1982). Population dynamics and feeding ecology of juvenile Australian Salmon (Arripis trutta) in Western Port, Victoria. Australian Journal of Marine and Freshwater Research 33: 369-375.
181 References
Robertson, A.I., (1984). Trophic interactions between the fish fauna and macrobenthos of an eelgrass community in Western Port, Victoria. Aquatic Botany 18:135-153. Robertson, A.I., and Blaber, S.J.M., (1992). Plankton, epibenthos and fish communities, Coast Estuar Stud. 41: 173-224. Robertson, A.I., and Duke, N.C., (1987). Mangrove forests as nursery sites: comparisons of the abundance and species composition of fish and crustaceans in mangrove forests and other nearshore habitats in tropical Australia. Marine Biology 96:193-205. Robertson, A.I., and Duke, N.C., (1990a). Mangrove fish-communities in tropical Queensland, Australia: spatial & temporal patterns in densities, biomass & community structure. Marine Biology 104: 369- 379. Robertson, A.I., and Duke, N.C., (1990b). Recruitment, growth & residence time of fishes in a tropical Australian mangrove system. Estuarine Coastal and Shelf Science 31: 723-743. Robertson, A.I., and Howard, R.K., (1978). Diel trophic interactions between vertically-migrating zooplankton and their fish predators in an eelgrass community. Marine Biology 48: 207-213. Robertson, A.I., and Klumpp, D.W., (1983). Feeding habits of the southern Australian garfish Hyporhamphus melanochir; a diurnal herbivore and nocturnal carnivore. Marine Biology 10: 197-201. Rooker, J.R., and Dennis, G.D., (1991). Diel, lunar and seasonal changes in a mangrove fish assemblage off southwestern Puerto Rico. Bulletin Marine Science 49: 684-698. Ross, P.M., and Yerman, M., (2003). Landscape issues for the macrofauna in temperate urban mangrove forests. In Urban Wildlife, more than meets the eye. (eds) S. Burgin and D. Lunney, Royal Zoological Society of NSW, Sydney.pp. 205-210. Rozas, L.P., Odum, W.E., (1987). The role of submerged aquatic vegetation in influencing the abundance of nekton on contiguous tidal fresh- water marshes. Journal of Experimental Marine Biology and Ecology 114: 289-304. Sammarco, P.W., and Andrews, J.C., (1988). Localised dispersal and recruitment in Great Barrier Reef Corals: the helix experiment. Science 239: 1422-1424.
182 References
Sanchez-Jerez, P., Cebrian, C.B., Espla, A.A.R., (1999). Comparison of the epifauna spatial distribution in Posidonia oceanica, Cymodocea nodosa and unvegetated bottoms: Importance of meadow edges. Acta Oecologica-International Journal of Ecology. 20(4): 391-405. Sargent, F.J., Leary, T.J., Crewz, D.W., and Kruer, C.R., (1995). Scarring of Florida’s seagrasses: Assessment and Management Options. Florida Department of Environmental protection, FMRI Technical Report TR1, Florida Marine Resource Institute, St Petersburg Florida, 33701, USA: 37 pp. Saunders, D.A., Hobbs, R., Margules, C.R. (1991). Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5: 18-32. Schoener, A., and Schoener, T.W., (1981). The dynamics of the species-area relation in marine fouling systems. 1. Biological correlates of changes in the species-area slope. American Naturalist 118: 339-360. Scoffin, T.P., (1970). The trapping and binding of subtidal carbonate sediments by marine vegetation in Bimini Lagoon, Bahamas. J. Sedim. Petrol. 40: 249-273. Sedberry, G.R., and Carter, J., (1993). The fish community of a shallow tropical lagoon in Belize, Central America. Estuaries 16: 198-215. Shepherd, S.A., McComb, A., Bulthuis, D., Neverauskus, V., and West, R.J., (1989). Decline in some Australian seagrass beds. In (Eds. A.W.D. Larkum, A.J. McComb and S.A. Shepherd). Seagrass Ecosystems – An Australian Perspective. Elsevier: Amsterdam, pp. 346-393. Sheridan, F.F., (1992). Comparative habitat utilization by estuarine macrofauna within the mangrove ecosystem of Rookery Bay, Florida. Bulletin Marine Science 50: 21-39. Short, F.T., and Wyllie-Echeverria, S. (1996). Natural and human induced disturbances of seagrasses. Environmental Conservation 23: 17-27. Simberloff, D.S., and Abele, L.G., (1976). Island biogeography theory and conservation practice. Science 191: 285-286. Smith, K A, and Suthers, I.M., (2000). Consistent timing of juvenile fish recruitment to seagrass beds within two Sydney estuaries. Marine and Freshwater Research 51: 765-776. Sogard, S.M., (1989). Colonisation of artificial seagrass by fishes and decapod crustaceans – Importance of proximity to natural eelgrass. Journal of Experimental Marine Biology and Ecology 133: 15-37.
183 References
Sogard, S.M., Powell, G.V.N., and Holmquist, J.G., (1989). Utilisation of fishes of shallow seagrass covered mudbanks in Florida Bay. 2. Diel and tidal patterns. Environmental Biology of Fishes 24: 81-92. Sokal, R.R., and Rohlf, F.J., (1995). Biometry: The principles and practice of statistics in biological research, Freeman & Co. New York. Sousa, W.P., (1984). Intertidal mosaics: Patch size, propagule availability, and spatially variable patterns of succession. Ecology 65: 1918-1935. Steele, J.H., (1985). A Comparison of terrestrial and marine ecological systems. Nature 313: 355-358. Steele, J.H., (1989). The ocean ‘landscape’. Landscape Ecology 3: 185-192. Summerson, H.C., and Peterson, C.H., (1984). Role of predation in organising benthic communities of a temperate-zone seagrass bed. Marine Ecology Progress Series 15: 63-78. Svane, I., and Ompe, M., (1993). Patch dynamics in beds of the blue mussel Mytilis edulis L: effects of site, patch size and position within a patch. Ophelia 37: 187-202. Svane, I., and Setyobudiandi, I., (1996). Diversity of associated fauna in beds of the blue mussel Mytilis edulis L: effects of location, patch size and position within a patch. Ophelia 45: 139-153. Tanner, J.E., (2004). Edge effects on fauna in fragmented seagrass meadows. Austral. Ecology.30(2): 210-218. Thayer, G.W., Colby, D.R., and Hettler, W.F., (1987). Utilisation of the red mangrove prop root habitat by fishes in south Florida. Marine Ecology Progress Series 35: 25-38. Thayer, G.W., Wolfe, D.A., and Williams, R.B., (1975). The impact of man on seagrass systems. American Science 63: 288-296. Thorhaug, A., (1983). Habitat restoration after pipeline construction in a tropical estuary: seagrasses. Marine Pollution Bulletin 14: 422-425. Thorhaug, A., (1985). Large-scale seagrass restoration in a damaged estuary. Marine Pollution Bulletin 16, 55-62. Thorman, S., (1986). Physical factors affecting the abundance and species richness of fishes in the shallow waters of the Southern Bothnian sea Sweden. Estuarine Coastal and Shelf Science 22(3): 357-370. Townsend, E.C., and Fonseca, M.S., (1998). Bioturbation as a potential mechanism influencing spatial heterogeneity of North Carolina seagrass beds. Marine Ecology Progress Series 169: 123-132.
184 References
Tsuchiya, S.J., and Nishihira, M., (1985). Islands of Mytilus as a habitat for small inter-tidal animals: effect of island size on community structure. Marine Ecology Progress Series 25: 71-81. Turner, M.G., (1989). Landscape Ecology: the effect of pattern on process. Annu. Rev. Ecol. Syst. 20:171-197. Uhrin, A.V., and Holmquist, J.G., (2003). Effects of propeller scarring on macrofaunal use of the seagrass Thalassia testudinum. Marine Ecology Progress Series 250: 61-70. Underwood, A.J., (1997). Experiments in Ecology: Their logical design and interpretation using analysis of variance. Cambridge University Press, Cambridge. Underwood, A.J., and Skilleter, G.A., (1996). Effects of patch size on the structure of assemblages in rock pools. Journal of Experimental Marine Biology and Ecology 197: 63-90. Virnstein, R.W., and Howard, R.K., (1987). The motile epifauna of marine macrophytes in the Indian River Lagoon, Florida: II. Comparisons between drift algae and three species of seagrasses. Bulletin Marine Science 41: 13-26. Colonisation of artificial seagrass versus time and distance from source. Marine Ecology Progress Series 29: 279-288. Virnstein, R.W., and Curran M.C., (1986). Colonisation of artificial seagrass versus time and distance from source. Marine Ecology Progress Series 29: 279-288. Walker, D., Dennison, W., Edgar, G. (1999). Status of Australian seagrass research and knowledge. In Butler, A., and Jernakoff, P. (eds). Seagrass in Australia CSIRO Publishing, Collingwood Vic. Aust. pp. 1- 24. Walker, D.I., and McComb, A.J., (1992). Seagrass degradation in Australian coastal waters. Marine Pollution Bulletin 25: 191-195. Weinstein, M.P., and Heck, K.L., Jr. (1979). Ichthyofauna of seagrass meadows along the Caribbean coast of Panama and in the Gulf of Mexico: composition, structure and community ecology. Marine Biology 50: 97-107. Wellenreuther, M., and Connell, S.D., (2002). Response of predators to prey abundance: separating the effects of prey density and patch size. Journal of Experimental Marine Biology and Ecology 273: 61-71.
185 References
West, R.J., (1983). The seagrasses of New South Wale’s estuaries and embayments. Wetlands (Australia) 3, 34-44. West, R.J., Jacobs, N.E., and Roberts, D.E., (1990). Experimental transplanting of seagrasses in Botany Bay, Australia. Marine Pollution Bulletin 21 (4), 197-203. West, R.J., and King, R.J., (1996). Marine, brackish, and freshwater fish communities in the vegetated and bare shallows of an Australian coastal river. Estuaries 19:31-41 Whitfield, A.K., and Blaber, S.J.M., (1978). Food and feeding ecology of piscivorous fishes at Lake St Lucia, Zululand. Journal of Fish Biology 13: 675-691. Williams, C.B., (1943). Area and the number of species. Nature 152: 264- 267. Williams, C.B., (1964). Patterns in the balance of nature. Academic Press, London. With, K.A., Cadaret, S.J., and Davis C., (1999). Movement responses to patch structure in experimental fractal landscapes. Ecology 80(4): 1340- 1353. Worthington, D.G., Ferrell, D.J., McNeill, S.E., and Bell, J.D., (1992a). Effects of the shoot density of seagrass on fish and decapods: Are they evident over larger spatial scales? Marine Biology 112: 139-146. Worthington, D.G., Ferrell, D.J., Mc Neill, S.E., and Bell, J.D., (1992b). Growth of four species of juvenile fish associated with seagrass Zostera capricorni in Botany Bay, NSW. Australian Journal of Marine and Freshwater Research 43: 1189-98. Worthington, D.G., and Westoby, M., (1991). Fish larvae settling in seagrass: Effects of leaf density and epiphytic alga. Australian Journal of Ecology 16: 289-293. Yahner, R.H., and Scott, D.P., (1988). Effects of forest fragmentation on depreciation of artificial arboreal nests. Journal of Wildlife Management 52: 158-161. Young, P.C., and Carpenter, S.M. (1977). Recruitment of postlarval Penaeid prawns to nursery areas in Moreton Bay, Queensland. Australian Journal of Marine and Freshwater Research 28: 745-773.
186