8. Regionalisation Methodology

Traditional methods of determining biogeographic structure of faunas have been based on an investigation of the entire fauna of the region. Methods used herein aim to compare these traditional bottom up methods with a 'rapid assessment procedure' (RAP). In this procedure, a top down approach using a subset of highly informative genera, is used as a surrogate of the complete data set. The main method involved the determination of an index to identify this set of genera.

8.1 Biogeographic Information Index

The basis of the RAP approach was to objectively select a biogeographically informative subset of genera occurring in each biome (i.e. estuarine, coastal marine, shelf demersal and shelf pelagic) from the complete continental shelf data set. The species of these genera were defined as the key species. This data set was obtained using the information index (BI) where genera were ranked according to their potential information content. This index is defined by:

BI = C (2A + B)/log (n + 1)

where

BI = Biogeographic Information index

A = biogeographic potential

B = information content qualifier

C = value within each ecosystem

n = number of species in the genus

The biogeographic potential (A) incorporates both the species richness of the genus and the extent of the species' distribution (e.g. very restricted within Australia, subregional endemics (about half the Australian coastline), through to international species). Endemics are considered to be more informative than more widely distributed species in delimiting intraregional biogeographic patterns. Similarly, highly restricted endemics are likely to be more useful than those that occur more widely through the region so the range extents were considered when formulating the criteria (see below). The information qualifier (B) serves to include an index of the reliability of distributional information for species within the genus. Collective relevance of the species to each biome is modified according to variable 'C' - zero values of C result in zero BI values eliminating the genus from the candidate groups for that particular biome. The presence of the natural log function enabled size of the group concerned to be factored in. Consequently, genera with low BI scores are considered to be more informative than those with high values.

The biogeographic potential (A) of genera were assessed according to the following eight criteria:

1. High richness (10 or more species in the genus) and almost all members with very restricted distributions
2. Low/medium richness (9 or fewer species in the genus) and almost all members with very restricted distributions
3. High richness and mostly subregional species (diversity high to low) but not applying to 1 or 2
4. Low/medium richness and mostly subregional species (diversity high to low) but not applying to 1 or 2
5. Mixed genus, including endemics and international species, but a few species have very restricted Australian distributions
6. Mainly broadly distributed Australian endemics
7. Broad mix of international and Australian subregional endemics
8. Almost all species with international distributions

The information qualifier (B) takes into account the quality of distributional information available for each genus. Distributions were scored according to the following weightings:

1 Well defined for most species in the genus (i.e. easily determined or has been studied extensively)

2 Reasonably well known for most species in the genus (i.e. good baseline data exists but only general studies available)

5 An even mixture of both well defined and poorly known species within the genus

9 Poorly defined

0 Unknown

The relative value (C) of each genus to each of the four biomes was weighted according to the following criteria:

1 High potential value (i.e. most species occurring in the biome)

1.5 Medium value (i.e. half or so of species occurring in the biome)

3 Low potential value (i.e. few species occurring in the biome)

5 Genus unlikely to occur in the biome

Next Chapter: 9. Oceanographic Datasets and Attributes