Field methods for assessing coral reef benthic communities tend to focus on
biodiversity and/or reef benthic composition, rarely investigating the aspects of size- frequency dynamics or population structures within and between coral communities that add greatly to understanding.
Common methods for assessing reef benthic composition typically measure simple parameters such as relative coverage of major benthic groups and substrate types. These measures generally provide only a crude measure of ecological condition. For example, the widely-used intercept transect method assumes a reef to be a 2-
dimensional habitat, giving little indication of reef rugosity. Moreover, information on coral community composition or population structure is rarely captured using this approach.
Many studies use coral cover as a proxy for reef health, assuming that high coral cover generally relates to good reef health. However, such inferences of reef health may be extremely misleading. For example, a return of coral cover following a disturbance episode to pre-disturbance levels is usually considered to equate to recovery of the reef community to its previous condition. Yet this supposes that reefs with high coral cover are more ‘healthy’ than those with low coral cover, and fails to take into account both the natural spatial gradient of coral abundance that exists on reefs, and processes of succession. Large reef tracts typically show considerable variability in cover of stony and soft corals in response to site-specific characteristics (Devantier et al. 2000). Indeed, such natural variability is indicative of a healthy system, and should not be assumed to be indicative of differential levels of ecological health or disturbance. Moreover, studies have shown that apparent recovery of coral cover can occur without recovery of community structure (Berumen and Pratchett 2006). Ecologically-
important differences may exist between two coral communities with the same overall coral cover. For instance, the two communities might support different age structures of populations, with one community comprising a small number of large, structurally-
juvenile forms that have not yet made a significant contribution to reef architectural complexity. Benthic composition transects commonly overlook this ‘maturity discrepancy’; a potentially serious limitation given that recovery of absolute coral cover may precede recovery of rugosity and complexity by 10-20 years (Sheppard et al. 2008).
2.6.1 Challenges of surveying coral communities
Many problems commonly encountered when surveying coral communities are
associated with monitoring the smallest cohorts within coral assemblages, particularly the youngest, newly-metamorphosed coral recruits.
The challenge when incorporating very young colonies (which add almost nothing to total coral cover at this stage) is to obtain data that may be considered broadly representative of the overall juvenile coral community within a given reef area. Recruitment success of corals is non-random, and may be influenced by a range of environmental factors including depth and substrate type, orientation and morphology (Norstrom et al. 2007; Bak and Engel 1979; Mundy and Babcock 1998). Moreover, there can be great inter-annual variability in settlement rates (Wallace 1985). Young corals, typically 1 – 2 mm in diameter, normally remain cryptic, commonly settled in reef crevices or cracks, or hidden by larger epibenthic organisms, until they have attained a clearly visible size (Bak and Engel 1979). This results in under-
surveying of small juveniles in most visual surveys and a lag period of potentially up to 1-2 years occurring between larval settlement and the point at which corals can be accurately and reliably recorded by visual census. Photoquadrats, taken vertically, are unable to accurately record new corals < 1 – 2 cm in diameter, and likewise do not detect corals situated in a non-horizontal plane of orientation on the reef (Edmunds et al. 1998). Even larger colonies are often misidentified, or simply remain unseen, when surveyed retrospectively by photoquadrats, as a result of colony crypsis and difficulties of identification from 2-dimensional images.
These limitations to sampling often prohibit monitoring of newly-metamorphosed recruitsin situ, and mean that important growth and life history processes may be
omitted from field surveys. Thus, sampling strategies for studies of juvenile corals vary greatly between studies. Some studies sample only areas of substrate deemed suitable for the settlement of coral larvae (normally bare rock or carbonate surfaces free from macro-algae, live hard and soft coral, sessile invertebrates and excessive sediment). Other approaches sample juvenile corals randomly across all reef substrata.
In addition to the challenges associated with monitoring newly-recruited corals, sampling larger colonies also poses significant difficulties, in large part because of the considerable variation in colony sizes both within and between taxa. Differences in colony size, growth, longevity and partial mortality can all prevent useful comparison of size structures and distributions between populations of different species and/or genera.
Because colony growth for perhaps most species therefore rarely follows a linear pattern with coral age, in some cases, size increases more rapidly the larger the colony, while some, such as tabular and ramoseAcroporaspp. are encrusting for a few years before commencing vertical growth. Similarly, a loss of colony diameter through partial coral mortality of, for example, 3 cm, would have a profound impact on a small juvenile colony but a far less significant biological impact on a large adult colony. Many species vary enormously in growth form. Morphological plasticity is such that the same species may exhibit different growth forms in different biotopes, or display different genotypic morphotypes. Both these factors can show substantial geographic variation (Devantier et al. 2000). Corals also vary considerably in size both within and between species, some attaining several metres in diameter, having grown from a newly settled polyp of only 1-2 mm in size. While some species grow to over 100,000 cm2, coral populations are invariably numerically dominated by small area classes (Soong 1993). These characteristics of coral populations present analytical problems for studies comparing area frequency distributions of species on a linear scale, since any particular increase in colony area will have a vastly different relative effect on overall colony area for a newly recruited juvenile than for a large coral (Vermeij and Bak 2003).
Of the few studies that have investigated population structures of Indopacific hard corals in recent years, sampling approaches have tended to group corals into size-class categories (Devantier et al. 2000; Obura 2009). Although this provides a means of obtaining a broad understanding of approximate population structures, size categories are generally classified widely, at a resolution that prevents discrimination of changes and dynamics between the smaller juvenile, vulnerable size stages.
These factors are important to understanding reef community dynamics, especially in the context of recovery from the massive 1998 mortality in the Indian Ocean. The methods and measurements used here attempt to capture this information which, in most studies, has been omitted.
METHODS AND MATERIALS
3 Site selection and sampling
Field research was carried out between 2006 and 2009 in the Chagos archipelago, the Seychelles, Madagascar and Saudia Arabia. These locations reflect different levels of anthropogenic disturbance, latitude and biogeography as well as different temperature environments, particularly in terms of variation of mean temperatures, exposure to past bleaching-related mortality and possible acclimation to temperature.
These differences are expected to have affected the abilities of corals to adapt to thermal stress events (Ateweberhan and McClanahan 2010, in press). Differing responses of coral communities between reefs are, in turn, likely to affect reef resilience to anthropogenic as well as ongoing climatic disturbance.
It is hypothesised that over recent years physical and temperature differences have resulted in differing reef composition between regions. Such differences are
anticipated both at the macro level of the structure and composition of the entire reef community, as well as more specifically within the communities and populations of scleractinia.
Within each region, surveys were carried out at a number of sites considered to be representative of a range of reef conditions. Surveying was not focused on sites considered to be either particularly ‘healthy’ or degraded, but aimed to cover an objective, representative range of sites in each region. Replicate sample numbers differed between regions as a consequence of survey programmes. A total of 108 reef sites was examined in the Indian Ocean. Sites were located in the following areas within the 5 survey regions (Figure 3):
Chagos
Surveys were carried out between February and March 2006 at 19 sites in 5 atolls of the archipelago. From north to south the atolls visited were (with numbers of survey sites in brackets): Peros Banos (4), Salomon (5), Great Chagos Bank (3), Egmont (2) and Diego Garcia (5) (figure 1).
Granitic Seychelles
Surveys were carried out at 20 sites in 7 survey areas around the islands of Mahe and Praslin in the Granitic Seychelles in April 2008. With the exception of one survey area, three sites were surveyed within each area. Following the sites and categorisation of Jennings (1996), reefs were classified as either ‘patch’, ‘carbonate’, and ‘granite’, with one of each reef type represented at each area. 2 survey areas (comprising a total of 6 sites) are managed as no-take marine reserves: St Anne Island at Mahe; and Cousin island at Praslin.
Madagascar - Andavadoaka
Surveys were carried out at 14 sites in Andavadoaka, southern Madagascar, between June and August 2008. Sites were distributed across three geomorphological reef types within the region: near shore fringing reefs, offshore fringing reefs, and lagoonal patch reefs following the geomorphological categorisation of reef types by Nadonet al.
(2007) and Gillibrand & Harris (2007). At the time of surveying, none of these sites were protected from any form of fisheries gear restrictions or management, although 3 lagoonal patch sites experiencedde factoprotection on account of their remote
Madagascar – Toliara
Surveys were carried out at 4 sites on the outer reef slope of theGrand Récifbarrier reef adjacent to the city of Toliara in southwestern Madagascar in July 2008. Sites were selected to be as close as possible to those examined and described in detail by Pichon (1978)(Mara, 2008, pers. comm.).
Southern Red Sea - Farasan Banks, Saudi Arabia
Surveys were carried out at 52 reef sites across the Farasan Banks, southern Saudi Arabia, from a total of 58 sites visited by the Living Oceans Foundation expedition in April 2009.