ABSTRACT

Landscape ecologists have been successful in describing the structure and function of terrestrial environments using a variety of landscape level metrics and have used the information to facilitate their conservation efforts.  Habitat diversity as a measure of spatial heterogeneity within the landscape, for example, has proven useful as a predictor of high species abundance and biodiversity of plants, birds, butterflies, insects and even larger vertebrates. Landscape ecology principles, however, have not been widely applied to marine systems.  Here, we explore the use of landscape level metrics as they apply to coral reefs and their associated fishes. Landscape metrics (e.g., habitat diversity) were calculated for 10 coral reefs and their adjacent habitats around the island of St. John (USVI), to describe differences in the "reefscape".  Two metrics, habitat diversity and total area of seagrass around the reef, were then correlated with various measures of reef fish abundance and also species richness. Species richness was not strongly correlated with habitat diversity as calculated here.  Interestingly, the average fish abundance per census (without regard to species) was negatively correlated with habitat diversity and suggested that one habitat type might be largely responsible for the observed pattern. When individual groups of fishes were examined, we found similar results, i.e. the average numbers of serranids, lutjanids, haemulids, and scarids exhibited a negative relationship with habitat diversity. As a follow-up, we examined the relationship between total area of seagrass around a reef and average fish abundance and found that these two measures were often positively correlated. Although habitat diversity does not appear to be positively correlated with reef fish density, other landscape metrics (such as total area of critical habitats around reefs) may be useful in predicting some aspects of reef fish assemblage structure. The landscape approach is thus likely to aid coral reef resource managers in their efforts to better design and manage marine reserves.
 

INTRODUCTION

The focus of landscape ecologists is on spatial heterogeneity, pattern and process.  Their goals are to characterize patterns on a large-scale, i.e. a landscape level, determine the origin of the pattern and form an understanding of the processes involved in the maintenance of those patterns. In terrestrial systems, the spatial arrangement or configuration of distinct habitat types has been shown to influence the population and community structure of associated animals.  The insights gained from landscape level analysis have been extremely useful and widely applied to the conservation of terrestrial systems.  The landscape approach, however, has not been widely used in the study of marine systems.  Here, we use landscape level metrics as a way to describe coral reefs and their associated habitats and explore the potential use of these metrics as predictors of  reef fish assemblage structure.

A reefscape view of St. John, USVI

METHODS
 DESCRIBING THE REEFSCAPE WITH
LANDSCAPE ECOLOGY TOOLS

Reefscape metrics were calculated for 10 reefs around the island of St. John , US Virgin Islands using benthic habitat maps developed by NOAA's Biogeography Team (Table 1).  The reefscape was defined as a collection of reef-associated habitat patches that occur within a specified distance (100 m, 250 m or 500 m) from the edge of a focal reef. Each focal reef polygon was cut to a standardized size (25,000 m²) so calculations were not confounded by reef size. Additional reef area was simply considered as another habitat type. Calculations were made using a simplified classification scheme and spatial analysis tools in ArcView (ESRI 2000).  For each reef the following steps were employed:

1) Select the focal reef
2) Establish a distance from reef for calculations
3) Intersect buffer and benthic habitat theme
4) Calculate the area of each habitat type (e.g., m² of seagrass)
5) Calculate Habitat diversity index (H'), where
     H'= – p_i ln p_{i ,}where p= m² of each habitat type i.

DETERMINING REEF FISH ASSEMBLAGE STRUCTURE

Reef fish censuses were conducted at each of the 10 reefs using a point count method modified from Bohnsack and Bannerot (1986) during the summer of 1994.  At each reef, a total of 16 censuses per 2.5 hectares of reef, were completed by four observers. Total fish abundance and species richness and abundance of several families of fishes (e.g., serranids, lutjanids, haemulids and scarids) were calculated from the point count data (Table 2).
 

Table 1: REEFSCAPE METRICS

Landscape level metrics such as habitat diversity, richness, and dominance measures are various ways of determining the relative numbers of types, sizes, and arrangements of patches present in a mosaic (Frohn 1998).  These indices have been calculated for 10 reefs around the island of St. John. Results for several metrics (i.e. habitat diversity and total area of seagrass) have been calculated within 100m, 250m and 500m of the focal reef. Additional metrics which may prove to be of value in future research are also presented.

RESULTS  I: HABITAT DIVERSITY AS A PREDICTOR
OF REEF FISH ASSEMBLAGE STRUCTURE

Habitat diversity (H') as defined by a simplified classification scheme using the following habitat types: reef, pavement, bedrock, seagrass, sand, mangrove, mud, macroalgae and deep unknown, was not strongly correlated with species richness. Similarly, when H' was calculated using only the area within 100m of the focal reef there was no strong relationship between H' and mean number of fish per census. Interestingly, the strength of the relationship increased with a concomitant increase in area around the focal reef, but the relationship was strongly negative suggesting that a single critical habitat type might be driving the pattern (see panel in upper right).  At the family level, the average  number of serranids (groupers), lutjanids (snappers), and haemulids (grunts) per census decreased with increasing habitat diversity. Acanthurids (surgeonfishes), on the other hand, increased with increasing habitat diversity. These patterns were similar whether H' was calculated using distances away from the focal reef of either 100 m, 250 m or 500m.  The relationship was strongest for all families at a distance of 250m, and although the strength of the relationship changes the direction does not. Habitat diversity, as defined here, does not attribute any value to a particular habitat type.  A reef surrounded by sand, mud and bedrock may have the same H' value as a reef surrounded by seagrass, additional reef and colonized pavement.  The importance of critical habitats such as seagrass is explored in Results II.

RESULTS II: SEAGRASS HABITAT AS A PREDICTOR
OF REEF FISH ASSEMBLAGE STRUCTURE

The total area of seagrass habitat near a coral reef was positively correlated with average abundance of lutjanids (snappers), serranids (groupers) and haemulids (grunts).  The average abundance of scarids (parrotfishes) was not strongly correlated with the total area of seagrass near the focal reef.

DISCUSSION AND CONCLUSIONS

Landscape ecology can be used to describe patterns in coral reef ecosystems at large spatial scales (100's of meters).  With the increasing availability of benthic habitat maps, landscape ecology methods offer promise for resource managers charged with the design and management of marine protected areas.  Large-scale metrics, which do not rely on costly in-water measurements, may provide new approaches for scientists who seek to understand the functional linkages between tropical marine habitats.

Habitat diversity has been shown to be a positive predictor of assemblage structure in terrestrial systems and small-scale studies of reef fishes often indicate that high spatial heterogeneity affects patterns of reef fish abundance and influences assemblage structure.  However, results from this study suggest that this relationship in reef systems may not scale-up. Based on these preliminary findings on fringing reefs in the USVI, resource managers are advised to be cautious in using habitat diversity as a metric to design networks of marine reserves as H' was not strongly positively correlated with either fish density or species richness.  Specific habitats are important in controlling coral reef fish distribution and in controlling the timing, rate, pathway, and distance of fish migration (Appeldoorn et al 1997). Critical habitat area, e.g. seagrass, therefore, may be an important predictor of reef fish assemblage structure (especially for those targeted by exploitation).


REFERENCES

Appeldoorn RS, Reckseick CW, Hill RL, Pagan FE, Dennis GD (1997) Marine protected areas and reef fish movements: the role of habitat in controlling ontogenetic migration, 8th Intl Coral Reef Symposium Proc., Vol. 2, 1917-1922.

Bohnsack J, Bannerot SP (1986) A stationary visual census technique for quantitatively assessing community structure of coral reef fishes.  NOAA Tech Report, 18 pp.

ESRI (1996) ARCView Spatial Analyst. Redlands CA.

Frohn RC (1998) Remote Sensing for landscape ecology.  New metric indicators for monitoring, modeling, and assessment of ecosystems.  Lewis Publishers 99pp.

ACKNOWLEDGEMENTS

Funding for the research was provided by the Student Cooperative Education Program of the Biological Resources Division of the USGS, Florida Integrated Science Center.  Field work and all logistical support was conducted thanks to the assistance of the Virgin Islands National Park and Biosphere Reserve. The support and friendship of Dr. Caroline Rogers are greatly appreciated. NOAA's Biogeography Team generated the benthic habitat map and discussions in the field have proven valuable in developing research ideas. Dr. Alan Friedlander and Jim Beets conducted most of the reef fish censuses.