634
Susan S. Bell, Mark S. Fonseca, and Nathaniel B. Stafford
landscape features were altered (Frost et al., 1999).
On the other hand, Bowden et al. (2001) found that
infauna communities in Z. marina in England varied with patch size and within-patch location, but
concluded that the most likely factor causing the observed differences was small differences in physical parameters. Clearly, then, effects of landscape
structure on infauna and epifauna tend to be variable
and species-specific, emphasizing the importance of
considering landscape parameters in the design of
seagrass and subtidal experiments.
4. Mobile Fauna and Patchiness
Large mobile organisms are also common in seagrass ecosystems and recently studies have begun
to examine their abundance and movement with reference to seagrass patch structure. MacArthur and
Hyndes (2001) reported that fish (Family: Odacidae)
community composition varied with degree of seagrass patchiness, although the seagrass species composition also varied with patch density. Salita (2000)
examined fish abundances in the Philippines and
determined that the first two components of a Principal Components Analysis [‘continuity of vegetation’ (including patch size and number) and fractal
dimension] explained 77% of the variation in fish
abundance measured [see also Salita et al. (2003)].
In an examination of intertidal and subtidal landscape mosaics composed of seagrass, mangroves,
and unvegetated sediments, Pittman (2002) found
evidence that species of fish and prawns had variable
responses to the reduction of seagrass within these
mosaics. However, he did note that below a cover of
15–20%, most seagrass patches in fragmented areas
ceased to support seagrass-associated fish. Thus, information is starting to emerge to suggest that large
nekton can also respond to landscape-scale features
of the seagrass habitat mosaic.
Seagrass beds have most commonly been studied
at the single patch scale or in landscapes made up of
a single patch type, but some studies have adopted
a broader view by investigating landscape mosaics
composed of multiple patch types (see Fig. 2). By
illustrating that nekton displayed a strong preference
for benthic mosaics that contained an abundance of
seagrass patches, Pittman (2002) reinforced and extended the importance of seagrasses in benthic landscapes. Moreover, Irlandi and Crawford (1997) reported that adjacency of seagrass habitat influenced
faunal abundance in salt marshes. Specifically, pinfish (Lagodon rhomboides) were substantially more
abundant in salt marshes located adjacent to seagrass
beds than in marshes spatially distant from seagrass
beds. The authors suggested that pinfish movement
could be a mechanism for transfer of secondary production between marsh and seagrass, and argued that
consideration be given to preserving and restoring
such closely linked systems as a composite rather
than either habitat alone (Irlandi and Crawford, 1997;
see also Bell et al., 1997). Finally, Micheli and
Peterson (1999) suggested that seagrass beds may
act as corridors for the movement of predatory blue
crabs between large oyster reefs and smaller, separate reefs, postulating that fragmentation of the intervening seagrass beds could alter the community
composition of benthic macroinvertebrates on oyster
reefs.
5. Multiple Spatial Scales and
the Landscape Approach
A uniquely ‘landscape ecology’ methodology is to
conduct investigations over multiple spatial scales.
Turner et al. (1999) examined infauna in seagrass
sites dominated by Z. novazelandica in New Zealand
at both the patch and the landscape scale, finding that
the landscape level analysis explained significantly
more of the variation in faunal abundance than the
patch scale analysis. Notably, the authors suggested
that a landscape metric, fractal dimension (D) of the
habitat, was a primary factor underlying the variation in infaunal abundance. Further, landscape level
effects on abundance were not necessarily related
to effects observed at a smaller scale (Turner et al.,
1999). In another multi-scale study Bell et al. (1995)
illustrated that, for some ecological processes, seagrass beds cannot be treated as larger versions of
seagrass patches. In an experiment with ASUs of
different sizes, large (4 m
2 ) patches collected disproportionately greater amounts of drift algae than
small (1 m
2 ) patches. These studies suggest that it
may be inappropriate to infer large scale features
of seagrass landscapes from measurements made at
small, local scales (sensu Fonseca et al., 2000).
Finally, the landscape scale may not be the largest
scale of interest in all cases. A recent study by Hovel
et al. (2002) highlights the complex nature of the relationship between seagrass landscape structure and
fauna in a North Carolina, USA system. They found
that landscape features (% cover and total linear edge
of beds) could explain little of the local patterns of
Susan S. Bell, Mark S. Fonseca, and Nathaniel B. Stafford
landscape features were altered (Frost et al., 1999).
On the other hand, Bowden et al. (2001) found that
infauna communities in Z. marina in England varied with patch size and within-patch location, but
concluded that the most likely factor causing the observed differences was small differences in physical parameters. Clearly, then, effects of landscape
structure on infauna and epifauna tend to be variable
and species-specific, emphasizing the importance of
considering landscape parameters in the design of
seagrass and subtidal experiments.
4. Mobile Fauna and Patchiness
Large mobile organisms are also common in seagrass ecosystems and recently studies have begun
to examine their abundance and movement with reference to seagrass patch structure. MacArthur and
Hyndes (2001) reported that fish (Family: Odacidae)
community composition varied with degree of seagrass patchiness, although the seagrass species composition also varied with patch density. Salita (2000)
examined fish abundances in the Philippines and
determined that the first two components of a Principal Components Analysis [‘continuity of vegetation’ (including patch size and number) and fractal
dimension] explained 77% of the variation in fish
abundance measured [see also Salita et al. (2003)].
In an examination of intertidal and subtidal landscape mosaics composed of seagrass, mangroves,
and unvegetated sediments, Pittman (2002) found
evidence that species of fish and prawns had variable
responses to the reduction of seagrass within these
mosaics. However, he did note that below a cover of
15–20%, most seagrass patches in fragmented areas
ceased to support seagrass-associated fish. Thus, information is starting to emerge to suggest that large
nekton can also respond to landscape-scale features
of the seagrass habitat mosaic.
Seagrass beds have most commonly been studied
at the single patch scale or in landscapes made up of
a single patch type, but some studies have adopted
a broader view by investigating landscape mosaics
composed of multiple patch types (see Fig. 2). By
illustrating that nekton displayed a strong preference
for benthic mosaics that contained an abundance of
seagrass patches, Pittman (2002) reinforced and extended the importance of seagrasses in benthic landscapes. Moreover, Irlandi and Crawford (1997) reported that adjacency of seagrass habitat influenced
faunal abundance in salt marshes. Specifically, pinfish (Lagodon rhomboides) were substantially more
abundant in salt marshes located adjacent to seagrass
beds than in marshes spatially distant from seagrass
beds. The authors suggested that pinfish movement
could be a mechanism for transfer of secondary production between marsh and seagrass, and argued that
consideration be given to preserving and restoring
such closely linked systems as a composite rather
than either habitat alone (Irlandi and Crawford, 1997;
see also Bell et al., 1997). Finally, Micheli and
Peterson (1999) suggested that seagrass beds may
act as corridors for the movement of predatory blue
crabs between large oyster reefs and smaller, separate reefs, postulating that fragmentation of the intervening seagrass beds could alter the community
composition of benthic macroinvertebrates on oyster
reefs.
5. Multiple Spatial Scales and
the Landscape Approach
A uniquely ‘landscape ecology’ methodology is to
conduct investigations over multiple spatial scales.
Turner et al. (1999) examined infauna in seagrass
sites dominated by Z. novazelandica in New Zealand
at both the patch and the landscape scale, finding that
the landscape level analysis explained significantly
more of the variation in faunal abundance than the
patch scale analysis. Notably, the authors suggested
that a landscape metric, fractal dimension (D) of the
habitat, was a primary factor underlying the variation in infaunal abundance. Further, landscape level
effects on abundance were not necessarily related
to effects observed at a smaller scale (Turner et al.,
1999). In another multi-scale study Bell et al. (1995)
illustrated that, for some ecological processes, seagrass beds cannot be treated as larger versions of
seagrass patches. In an experiment with ASUs of
different sizes, large (4 m
2 ) patches collected disproportionately greater amounts of drift algae than
small (1 m
2 ) patches. These studies suggest that it
may be inappropriate to infer large scale features
of seagrass landscapes from measurements made at
small, local scales (sensu Fonseca et al., 2000).
Finally, the landscape scale may not be the largest
scale of interest in all cases. A recent study by Hovel
et al. (2002) highlights the complex nature of the relationship between seagrass landscape structure and
fauna in a North Carolina, USA system. They found
that landscape features (% cover and total linear edge
of beds) could explain little of the local patterns of