چکیده (انگلیسی):

Biological assessments of forest systems often involve a single ground-invertebrate sampling method that may ignore the biological component of the non-sampled canopy. Pitfall trapping for groundactive arthropods is a widely implemented technique for biological assessment in forested and open habitats. Although much evidence highlights the biases of pitfall trapping, this evidence typically comes from open-habitat crop and grassland systems. In forest systems where much of the biodiversity is found within the above-ground structure, management recommendations based solely on ground sampling may not represent the diversity within the three dimensional forest habitat. We provide evidence from combined ground and canopy sampling of three major forest types within the study region. We use canopy insecticide fogging to compare with more traditional ground-based pitfall trapping, and use spiders as a comparative species-rich biota that is able to colonise most terrestrial habitats and is strongly affected by changes in environmental condition.
We identified 3933 spiders from 109 species from the 18 forest patches sampled. Both types of sampling defined differences in community composition between forest types in a similar manner; hence, either method could be used to evaluate differences or test management regimes in well-replicated experiments of forest type. However, the association in community composition between ground and canopy assemblages at the individual site-based level was weak; we found low correlation between the two data sets indicating that surrogacy between methods was not supported at this level. Furthermore, disparities in spider habitat association, body size, hunting guild and vertical stratification of spider families indicates that where detailed species and family-based information is required, or if inventorying is necessary, then multiple targeted surveys are essential.