Analysis of echinoderm datasets for the Arabian Gulf and Red Sea region reveals lower species richness in the Gulf at 4 different spatial scales: Point, Sample, Large area and Biogeographical province. Species richness correlates positively with spatial scale in both regions, as does endemism, although levels of the latter are comparable for both regions.
Values of taxonomic distinctness (Δ*), a relatedness measure based on average properties, are very similar for the Gulf and the Red Sea over all spatial scales. This might seem surprising, given that (i) the Gulf's environment is highly stressful, and (ii) the Red Sea is one the world's top 10 coral reef high diversity areas, and the habitat associated with most echinoderm records for this region, but not for the Gulf. These findings may partly reflect the behaviour of Δ* to ‘saturate’ (reach maximal values as sampling/spatial scale increases) much sooner than species richness, endemism and other ‘species’ measures.
Our results suggest the Gulf may not be the lowspot of biodiversity generally acclaimed, especially if a broad set of measures is utilised. Biodiversity is an important constituent of ecosystem health and criterion for determining where conservation attention should be targeted. Until the functional/ecological significance of Δ* is better understood, management implications of our study's findings remain unclear. Further attention should also be given to β-diversity, the change in species composition along a gradient. We also stress that any conclusions drawn from the study are tentative, pending compilation and analysis of datasets for other groups.
The Arabian Gulf's biota is usually characterised as impoverished, with low species richness at both large area/biogeographical province and smaller spatial scales for many groups, including corals, algae and echinoderms (Sheppard et al., 1992). This is attributed to stressful environmental conditions, in particular marked fluctuations in sea temperatures (< 12– > 30°C), and high salinities (40 to > 60 ppt) (Sheppard et al., 1992), coupled with pronounced post-Pleistocene sea level fluctuations (Price, 1982a).
Recent studies have begun to question the contention that the Gulf is, unequivocally, a biodiversity lowspot. Price (2002) showed that while species richness of the benthos is low, its ß-diversity (change in species composition along a spatial gradient) is surprisingly high. The same pattern was evident for estuaries and hydrothermal vents, systems that are also environmentally stressful. In the case of hydrothermal vents, taxonomic distinctness (Δ*) was also marked. Unlike species richness and other ‘absolute’ biodiversity measures (e.g., Shannon Wiener), Δ* is an index based on ‘average’/assemblage properties. It is one of several measures of relatedness. Based on Δ*, a sample containing a sponge, a coral, an amphipod, a polychaete and an echinoid (5 species) would be much more diverse than a sample containing just polychaetes (5 species), particularly if all belonged to the same genus (see Price et al., 2002).
This paper examines the assertion that the Gulf is impoverished, by quantitative analysis of its echinoderm fauna, a group which has been well studied in the Gulf (Price, 1981, 1982a, b, 1983; Price and Rezai, 1996). The group is also a useful indicator of environmental conditions (Izsak et al., 2002). We report species richness, endemism and taxonomic distinctness as well as ß-diversity, at four different spatial scales. Comparison is made with the Red Sea, a very diverse region (Sheppard et al., 1992) and in fact one of the world's top 10 coral reef high diversity areas (Roberts et al., 2002). It is also a region whose echinoderm fauna has been studied comprehensively (Clark and Rowe, 1971; Price, 1982a). The paper concludes by considering the implication of the findings to conservation of the Gulf region.
Three biodiversity measures are used, all utilising presence/absence data: species richness, percent endemism and taxonomic distinctness (Δ*). The latter index was developed by Warwick and Clarke (1995) and used in subsequent studies (e.g., Price et al., 1999, 2002). Unlike earlier indices, such as Simpson's and Shannon Wiener, Δ* (and the similar, quantitative index Δ) takes into account the taxonomic level at which any two species are related. Hence, their identity matters. A simple linear weighting factor (wij) is applied, such that w = 1 (for species i and j within the same genus), w = 2 (for species within different genera but the same family), w = 3 (for species within different families but the same class)…etc.
We have used constant incremental path lengths (0,1,2, etc.) thoughout this study for taxonomic distinctness following the robustness of relative values of taxonomic distinctness to variation in the definition of path length (Clarke and Warwick, 1999). For echinoderms, 5 taxonomic levels were used (species, genus, family, order, class), hence the maximum path length is five.
Δ* is the average (weighted) path length, that is, between any two species in a sample/assemblage and, when used with presence-absence data, may be defined as follows:
We have selected 4 different and progressively larger spatial scales, adopting as closely as feasible the categories and notation of Gray (2000), for measurement of species richness and other biodiversity measures, as follows: Point (P): a single sampling unit; Sample (S): a number of sampling units from a site of defined area; Large area (L): a large area which includes a variety of habitats and assemblages; Biogeographical province (B): a biogeographical province, following the designation of the Gulf and Red Sea, including the Gulfs of Suez and Aqaba, as 2 major subdivisions of the Indo-West Pacific (Clark and Rowe, 1971; Figure 1). Inevitably, there is some arbitrariness, even confusion, over the boundaries of different spatial scales (Gray, 2000); for example, the distinction between Point (P) and Sample (S) scales. As applied to this study, both show some affinities with α diversity. Similarly, Large area (L) shows some equivalence, but not perfect concordance, with γ -diversity (Ellingsen and Gray, 2002).
The results of biodiversity analysis of echinoderm datasets for the Gulf and Red Sea region are summarised in Table 1. Echinoderm species richness in the Gulf is consistently lower than in the Red Sea at small, intermediate and large spatial scales. Levels of endemism are comparable at small and intermediate spatial scales, and at the largest (biogeographic) scale levels are also similar in the Gulf to the Red Sea region. Taxonomic distinctness (Δ*) values for the Gulf and the Red Sea region are very similar over all spatial scales. Our findings suggest the Gulf may not be the lowspot of biodiversity generally acclaimed, especially if a broad suite of measures is utilised.
Our analyses of echinoderm data demonstrate the influence of both biodiversity measure and spatial scale on observed patterns of biodiversity. This reinforces similar recent findings in the marine literature (e.g., Izsak and Price, 2001; Ellingsen and Gray, 2002; Price, 2002). The study exposes an apparent paradox of the Gulf being both a biodiversity highspot and lowspot, at least for one group, depending on the measure used.
Especially striking is that Δ* values for both the Gulf and the Red Sea region are very similar over all spatial scales. This may in part reflect the behaviour of Δ*, and other relatedness measures based on average properties, to ‘saturate’ (reach maximal values as sampling/spatial scale increases) much sooner than species richness and most other ‘species’ measures (Warwick and Clark, 1995; Izsak and Price, 2001). On the other hand, Red Sea echinoderm data relate mainly to coral reefs, a prevalent and highly ‘diverse’ ecosystem (Sheppard et al., 1992; Roberts et al., 2002). Yet Δ* for the Red Sea is not particularly marked. Biodiversity data for the Gulf, in contrast, relate to mixed habitats, coral reefs in this region being much less extensive than in the Red Sea. Perhaps surprising, therefore, is that Δ* is as high in the Gulf as the Red Sea.
Echinoderm species richness shows positive association with spatial scale for both regions. Ellingsen and Gray (2002) cite several studies showing local richness to be directly proportional to regional richness, although they found no clear association between mean alpha and gamma diversity in the Norwegian continental shelf. In our study, echinoderm species richness for both regions, is around 10% (Point), 20 to 25% (Sample), 50 to 75% (Large area) of the Biogeographic scale species richness. Also relevant is that γ -diversity is a consequence of two factors: the richness of each of the smaller areas that compose it (α diversity), and the turnover in species composition among them (β -diversity; Willig et al., 2003).
The echinoderm fauna of the Red Sea is more speciose than that of the Gulf at all spatial scales. Greater sampling focus on coral reefs in one region (Red Sea) than the other (Gulf), as noted above, is a partial explanation. But the more stressful environmental conditions in the Gulf, which is known to depress species richness (Price, 1982a, b; Sheppard et al., 1992), may be more critical. On the other hand, the Gulf has 542 known fish species (Krupp et al., 2000), the second highest number for any enclosed or semi-enclosed sea in the world (Price et al., 2002). Ocypodid crabs are another, recently demonstrated example of high diversity in the Gulf, with 14 taxa recorded in Kuwait alone (Al-Khayat and Jones, 1996). Hence, even in terms of species richness, the Gulf is not necessarily impoverished.
Like species richness, endemism is correlated (positively) to spatial scale. Unlike species richness, but like Δ*, levels of endemism in the Gulf and Red Sea are comparable for echinoderms.
Biodiversity is an important constituent of ecosystem health, and one criterion for determining where conservation attention should be targeted (Price, 2002). Until the functional/ecological significance of taxonomic distinctness is better understood, management implications our study's findings remain unclear. We also stress that any conclusions about biodiversity drawn from one biotic group must be tentative, pending compilation and analysis of comprehensive datasets for other fauna and flora.