Forestcheck: the response of macro-invertebrates to silviculture in jarrah (Eucalyptus marginata) forest

Abstract

Summary The response of invertebrates to silvicultural treatments (including timber harvesting and post-harvest burning) was examined as part of the FORESTCHECK monitoring project in the jarrah (Eucalyptus marginata) forest of south-west Western Australia. Invertebrates were collected twice yearly (spring and autumn) at 48 sample grids in four jarrah forest ecosystems, which included two different vegetation complexes within the same ecosystem. Grids represented examples of unharvested forest (reference) and forest subject to gap release and shelterwood/selective cut treatments. Collection methods included light trapping, pitfall trapping and hand sampling. A total of 56 705 invertebrate specimens comprising 1497 morphospecies were collected. Within each forest ecosystem, species richness did not differ significantly between gap release, shelterwood/selective cut and reference forest. Estimated total richness of macro-invertebrates varied with silvicultural treatment and collection method: 630 species for pitfall trapping and 1095 for light trapping. Estimated richness from light trapping of grids in unharvested forest was 892 species, compared with 780 species in shelterwood/selective cut and 665 species in gap release treatments. Collection method influenced the direction and magnitude of changes in estimated richness following timber harvesting and post-harvest burning, indicating that multiple collection methods are needed to assess invertebrate response to disturbance. Based on combined capture methods, estimated whole bioregional richness of macro-invertebrate morphospecies was highest in reference forest (1783), intermediate in forest subject to shelterwood/selective cut treatment (1652) and lowest in forest subject to gap release (1527 morphospecies). Greater richness estimates in reference forest are attributable to a greater proportion of macro-invertebrate species that occurred in only one grid. However, the robustness of comparisons between treatments of these bioregional estimates of richness is compromised by inadequate sampling. In contrast, for equivalent sampling effort, species accumulation was slightly greater in silviculturally treated forest than in reference forest, implying that silviculturally treated grids had a more homogenous species composition. Data were analysed using non-metric multidimensional scaling ordination and canonical analysis of principal co-ordinates (CAP) analysis. Invertebrate assemblage composition and abundance were strongly influenced by biogeographic pattern, as represented by forest ecosystem/year of collection (these factors were confounded in the design of the monitoring project). No significant effects of silvicultural treatments on morphospecies composition were detected using CAP analysis of morphospecies abundance. Retention of patches of unharvested forest adjacent to harvested forest will assist in maintaining invertebrate biodiversity at the landscape scale. The data suggest that indicator species may be of limited value for monitoring recovery after disturbance in the jarrah forest bioregion because of substantial species turnover between different forest ecosystems and years of sampling. Monitoring of invertebrates should be continued to demonstrate whether harvested coupes are sufficiently heterogeneous, so that their contribution to bioregion species diversity is equivalent to that of unharvested areas.