The survey of small mammal communities in forest environments of Australia

Abstract

In an analytical review of a sample of past Australian faunal surveys, the survey of small mammal communities was identified as requiring most immediate attention. Variables in survey design were found to be inconsistent and most of the results were not comparable. The most important variables that were used inconsistently were sampling configuration, intensity and size. Almost i all surveys were conducted without reference to the relationships between methods and small scale environmental heterogeniety. Field trials were established in the Brindabella Range (ACT) and Bodalla State Forest (N.S.W.) to evaluate the above key survey variables. The first site (Brindabella Range) consisted of a discrete area of small mammal habitat containing a well studied small mammal community which provided a stable, known situation with which the results of all methods could be compared. Three other sites (Bodalla S.F./ represented a successional gradient following major forestry disturbance, which allowed the methods to be evaluated under a range of environmental conditions. Capture data were used to compute indices of diversity to provide a basis for evalcating methods in terms of their capability of sampling community structure. Characteristic responses of diversity indices to small number data sets were studied·by modelling hypothetical data sets. Analysis of the indices showed that methods used in a grid configuration were sensitive to changes in sampling intensity and frequently misrepresented community structure. In contrast, trapline (index-line) methods were found to be more consistent and accurate in measuring community structure. When consistency under different envirorunencal conditions, field effort and species representation are taken into consideration, trap spacings of 1Om are preferred. Sample sizes for index-line methods were analysed by progressively accumulating capture data. Most sites were adequately sampled by 120m of index-line. Values of indices for the richest site (5 species) were still increasing up to a length of l70m, the full extent of the line at this site. Further extensions to the line may have returned additional information. A supplementary length of 50m is suggested for any index-line to account for variations caused by the end traps. Similar analyses for grid methods revealed that only the 5m spacing method was adequately sampling the community. Other spacings may have returned additional information with increased frequency of sampling. Five environmental attributes were found to be significantly correlated with ; indices of community diversity at the first site. These were used to evaluate the ways in which the different methods responded to environmental patchiness. ' Neither grid nor index-line methods were found to differ amongst themselves, however there was a significant difference between the two configurations in ii their ability to sample fern cover, one of the key attributes found to be correlated with species in the community. Index-line methods sampled the dense patches more efficiently than did the grid methods. The potential is recognised for developing a model of environmental complexity and small mammal community diversity against which survey methods may be calibrated to sample a range of environments in a comparable way. Show more