528
Bronwyn M. Gillanders
have been conducted at the landscape scale, even
though seagrasses are ideal systems to use to address landscape scale questions (Robbins and Bell,
1994). Further studies are clearly needed to investigate how the configuration of the habitat (e.g. patch
size, patch isolation, proximity to edge), amount
of habitat (e.g. habitat cover), and adjacent habitats (e.g. non-vegetated areas, mangroves) may influence not only abundance of organisms within seagrass but also their growth, survival, and movement
patterns. Structural characteristics of the seagrass
habitat (e.g. shoot density, shoot biomass) often vary
at the among patch and landscape scale; for example,
structural complexity is often positively correlated
with seagrass patch size (Hovel, 2003; Bell et al.,
Chapter 26). In addition, seagrass cover and configuration of habitat also may covary (Hovel, 2003).
Covariation of variables at a range of spatial scales
creates problems for comparisons of the effects of
different variables. Such problems can be overcome
either by using artificial habitats (e.g. McNeill and
Fairweather, 1993) or by statistically removing the
covariation (e.g. Hovel, 2003). Use of artificial habitats may, however, not be possible at the landscape
scale.
Elements of seagrass habitat structure at more
than one scale may influence abundance, growth,
and survival of organisms, thus multi-scale approaches are required (see Bell et al., Chapter 26).
It will also be necessary to determine both the direction and magnitude of influence of different variables and how these vary at regional and biogeographic scales. Simultaneous sampling and experiments over broad spatial and temporal scales are
required. In addition, the scale at which studies have
been conducted has varied (Villard et al., 1999;
Hovel, 2003). Future studies should consider biological information on the species being investigated
(e.g. dispersal capabilities, foraging behavior, home
range) and factor this information into the choice
of scales at which to measure responses to habitat
structure. For example, studies on wide-ranging
species should be conducted at a larger scale than
those focusing on infauna of seagrass beds.
C. Regional and Biogeographic Scales
At regional and biogeographic scales, a variety
of sampling designs and methods have been utilized making comparisons among studies difficult.
A global approach, where researchers around the
world conduct experiments across similar temporal
and spatial scales using the same design and methodology, although ambitious, could be extremely useful for determining generality of patterns. Such an
approach has recently been used to determine the
impact of natural UV-radiation on shallow marine
hard-bottom assemblages (Wahl, Kiel University,
Germany; personal communication).
D. Seagrass Loss and Fisheries
Few clear examples exist of the demise in fisheries
catches when seagrass has been lost from an area.
The great wasting disease of eelgrass in the North Atlantic in the 1930s, and its effects on fisheries are still
the subject of much debate (see Duarte et al., Chapter 11, Moore and Short, Chapter 16 and Kenworthy
et al., Chapter 25). As seagrass declines within an
area, fish may use alternative habitats. Examining the
link between habitat and fisheries requires monitoring of both catch and seagrass cover prior to episodes
of seagrass loss and following recovery. Data on both
catch and seagrass cover are currently unavailable
for many areas. In addition, changes in organisms
inhabiting seagrass before and after seagrass loss
are poorly known (Connolly et al., 1999). Areas that
have lost seagrass may also go through a range of
habitat successions, but further information is required on what habitats occur in areas where seagrass
is lost and how the assemblage structure may change
as a consequence. Baseline studies within seagrass
beds should encompass not only sufficient spatial
scales, but also sufficient temporal scales such that
natural site-to-site and annual variability can be considered enabling useful comparisons after seagrass
loss (Hughes et al., 2002). Events that result in loss of
seagrass should be seen as opportunities for natural
experiments. Few datasets of a sufficient time frame
exist to enable adequate modeling of the relationship between fisheries catches and seagrass cover (or
some variable encompassing the amount of seagrass
in an area). Further work in this area is required.
Few studies have attempted to model the effects
of seagrass removal on fishery species. Despite a
lack of studies, it should be possible to use past data
and potentially describe future states. Initial model
fitting should adequately describe the data, but will
also require independent tests of the assumptions
and the importance of particular variables (Haddon,
2001).
Bronwyn M. Gillanders
have been conducted at the landscape scale, even
though seagrasses are ideal systems to use to address landscape scale questions (Robbins and Bell,
1994). Further studies are clearly needed to investigate how the configuration of the habitat (e.g. patch
size, patch isolation, proximity to edge), amount
of habitat (e.g. habitat cover), and adjacent habitats (e.g. non-vegetated areas, mangroves) may influence not only abundance of organisms within seagrass but also their growth, survival, and movement
patterns. Structural characteristics of the seagrass
habitat (e.g. shoot density, shoot biomass) often vary
at the among patch and landscape scale; for example,
structural complexity is often positively correlated
with seagrass patch size (Hovel, 2003; Bell et al.,
Chapter 26). In addition, seagrass cover and configuration of habitat also may covary (Hovel, 2003).
Covariation of variables at a range of spatial scales
creates problems for comparisons of the effects of
different variables. Such problems can be overcome
either by using artificial habitats (e.g. McNeill and
Fairweather, 1993) or by statistically removing the
covariation (e.g. Hovel, 2003). Use of artificial habitats may, however, not be possible at the landscape
scale.
Elements of seagrass habitat structure at more
than one scale may influence abundance, growth,
and survival of organisms, thus multi-scale approaches are required (see Bell et al., Chapter 26).
It will also be necessary to determine both the direction and magnitude of influence of different variables and how these vary at regional and biogeographic scales. Simultaneous sampling and experiments over broad spatial and temporal scales are
required. In addition, the scale at which studies have
been conducted has varied (Villard et al., 1999;
Hovel, 2003). Future studies should consider biological information on the species being investigated
(e.g. dispersal capabilities, foraging behavior, home
range) and factor this information into the choice
of scales at which to measure responses to habitat
structure. For example, studies on wide-ranging
species should be conducted at a larger scale than
those focusing on infauna of seagrass beds.
C. Regional and Biogeographic Scales
At regional and biogeographic scales, a variety
of sampling designs and methods have been utilized making comparisons among studies difficult.
A global approach, where researchers around the
world conduct experiments across similar temporal
and spatial scales using the same design and methodology, although ambitious, could be extremely useful for determining generality of patterns. Such an
approach has recently been used to determine the
impact of natural UV-radiation on shallow marine
hard-bottom assemblages (Wahl, Kiel University,
Germany; personal communication).
D. Seagrass Loss and Fisheries
Few clear examples exist of the demise in fisheries
catches when seagrass has been lost from an area.
The great wasting disease of eelgrass in the North Atlantic in the 1930s, and its effects on fisheries are still
the subject of much debate (see Duarte et al., Chapter 11, Moore and Short, Chapter 16 and Kenworthy
et al., Chapter 25). As seagrass declines within an
area, fish may use alternative habitats. Examining the
link between habitat and fisheries requires monitoring of both catch and seagrass cover prior to episodes
of seagrass loss and following recovery. Data on both
catch and seagrass cover are currently unavailable
for many areas. In addition, changes in organisms
inhabiting seagrass before and after seagrass loss
are poorly known (Connolly et al., 1999). Areas that
have lost seagrass may also go through a range of
habitat successions, but further information is required on what habitats occur in areas where seagrass
is lost and how the assemblage structure may change
as a consequence. Baseline studies within seagrass
beds should encompass not only sufficient spatial
scales, but also sufficient temporal scales such that
natural site-to-site and annual variability can be considered enabling useful comparisons after seagrass
loss (Hughes et al., 2002). Events that result in loss of
seagrass should be seen as opportunities for natural
experiments. Few datasets of a sufficient time frame
exist to enable adequate modeling of the relationship between fisheries catches and seagrass cover (or
some variable encompassing the amount of seagrass
in an area). Further work in this area is required.
Few studies have attempted to model the effects
of seagrass removal on fishery species. Despite a
lack of studies, it should be possible to use past data
and potentially describe future states. Initial model
fitting should adequately describe the data, but will
also require independent tests of the assumptions
and the importance of particular variables (Haddon,
2001).