Identifying regrowth forests with advanced mature forest values.

Highlights • Regrowth forest is important biodiversity habitat when little mature forest remains. • Identifying regrowth with mature forest elements can guide strategic reservation. • Stand maturity related weakly to remote-sensed and measured environmental variables. • Eucalypt crown senescence weakly indicates stands likely to contain habitat trees. • Timber production and conservation are not in conflict on more productive sites. Abstract In this study, we examined the associations between field-assessed floristic and structural habitat values for mature forest and GIS-derived variables to assess whether high conservation value forests could be predicted for strategic reservation at a landscape scale. We investigated the Eucalyptus regnans forests of the Victorian Central Highlands in south-eastern Australia, where several extensive wildfires in the last century have left little mature forest. We assessed the extent to which the floristic composition and a suite of habitat-related structural variables could be explained by two forest inventory GIS variables (percentage senescence and site productivity) and whether explanatory capacity improved from inclusion of additional environmental variables (climate, soils, topography, structure and spatial location). Results showed that the floristic composition was weakly related to productivity, but not to percentage senescence. Four habitat-related structural variables were positively related to percentage senescence (density of old-growth eucalypts, the quadratic mean diameter (QMD) of both live and dead eucalypts, and the maximum eucalypt form class (a proxy for tree hollows)) while the volume of CWD had a marginally significant positive relationship. Three structural variables were related to productivity (the maximum eucalypt form class, the QMD of dead understorey trees and of dead eucalypts). However, in all cases the explanatory power of percentage senescence and productivity was weak (proportion of deviance explained by the models <0.3). Inclusion of the other environmental variables did not substantially improve explanatory power in any case. Our results suggest that there is a high degree of stochasticity driving floristic and structural composition within these forests, making the detection of most mature forest values difficult without site visits. There may be some limited capacity for the presence of senescent trees visible in aerial photographs to highlight stands more likely to contain habitat trees, but there was no relationship with floristic maturity. Our results also suggest that more productive sites do not have substantially greater habitat values, indicating that the current timber harvesting approach of prioritising the harvest of stands on more productive sites is unlikely to have negative consequences for biodiversity. [ABSTRACT FROM AUTHOR]