[Truncated abstract] Selection of priority areas for Marine Park conservation is often compromised by the lack of comprehensive biodiversity data and the resources and expertise necessary to gain this information directly by sampling. One cost effective alternative is the use of species groups or indicator species as surrogates for total biodiversity. However use of these surrogates requires an ecological understanding of how they reflect biodiversity gradients. A framework for unravelling these relationships has been suggested that involves relating species biodiversity to different and competing ecological models using appropriate statistical analysis. I use this framework to explore coral species biodiversity over a range of environmental gradients encompassing the North West Shelf of Australia and the Great Barrier Reef in North East Australia. ... I assessed physiological responses of corals to physical factors to corroborate crossshelf patterns in species biodiversity. Finally, I investigated to what extent coral cooccurrence based species groups (or guilds) can be used as surrogates for total coral biodiversity. The major findings of this thesis were: i) coral biodiversity along cross-shelf environments was highly correlated to specific gradients of abiotic reef conditions; ii) larval modelling indicates the potential for significant connectivity across continentalshelf environments such that differences in species distribution are not simply as a result of self seeding. iii) similar correlative patterns were demonstrated for coral species that occur along comparable abiotic gradients in reef areas of both Eastern and Western Australia, suggesting a causal relationship between the physical environment and coral biodiversity; iv) coral physiological parameters measured using lipid fractions independently corroborated the hypothesis that there is a biological basis for observed coral distributions; v) reef coral communities are not highly structured across abiotic physical gradients and biodiversity across the shelf increases as conditions become suitable for a wider range of species; vi) total coral biodiversity can be estimated very accurately (within r2 values ranging from 0.75 to 0.90) using a small number (15-30) of optimally chosen indictor species using the randomForest statistical method. These results suggest coral biodiversity over cross-shelf environments conforms most closely to the