DESIGNING A COST-EFFECTIVE INVERTEBRATE SURVEY - A TEST OF METHODS FOR RAPID ASSESSMENT OF BIODIVERSITY

We investigated three procedures that may lead to rapid and accurate a ssessment of epigaeic arthropod biodiversity. They are: (1) the identi fication of taxa whose diversity is correlated with that of others; (2 ) the identification of times and methods of sampling that produce est imates of diversity representative of more intensive sampling; and (3) the use of morphospecies inventories generated by non-specialists. An ts, beetles, and spiders were sampled from four forest types, in three seasons, using two collecting methods: pitfall trapping and extractio n from litter. Specimens were sorted by a non-specialist to morphospec ies and by specialist taxonomists to species. Richness (alpha-diversit y) and turnover (beta-diversity) were compared for different sampling regimes using morphospecies and species inventories. We found no signi ficant positive correlations between ant, beetle, and spider species r ichness but there was a strong negative correlation between ant and be etle richness. For beetles alone, richness within the families Carabid ae, Scarabaeidae, and Pselaphidae (i.e., avoiding taxonomically proble matic families) was significantly correlated with richness within all other families. Assessment of turnover revealed that: (1) the four for est types contained significantly different assemblages of ants and be etles but not spiders and (2) the four forests were less clearly discr iminated using species from the three beetle families Carabidae, Scara baeidae, and Pselaphidae when compared to species from all beetle fami lies pooled. Analyses of single sampling periods and methods revealed that summer and spring pitfall samples were most representative of mor e intensive sampling. That is: (1) the richness of ants and beetles in these samples was significantly positively correlated with the richne ss of all other samples and (2) turnover of beetles and ants among the four forests revealed by summer pitfall samples was similar to turnov er using all samples. The three beetle surrogate families recorded by pitfall samples in spring, and to a lesser extent summer, showed signi ficant correlations in richness with all other beetle species recorded in the same samples. However, the assessment of turnover was less acc urate when only surrogate families were used. The most accurate and co st-effective assessment of turnover was generated by a summer pitfall sample in which data for ants, carabid, and scarab beetles were combin ed and analyzed as a single data set. Results were largely consistent regardless of whether species or morphospecies were used, which sugges ts that monitoring and assessment of terrestrial invertebrate biodiver sity may be achieved by the careful use of morphospecies. Our results also suggest those invertebrate taxa, sampling methods, and sampling p eriods that yield the most consistent and reliable assessment of epiga eic invertebrate biodiversity in Australian temperate hardwood forests . However, empirical studies that follow the protocols discussed in th is paper are urgently required in different environments. These studie s may point the way to more representative monitoring and assessment o f terrestrial biodiversity.