Searching for rainforest understorey in wet Eucalyptus forest

Increasing pressure from anthropogenic and natural disturbances can lead to irreversible shifts in the composition and structure of vegetation. Fire-sensitive rainforest communities in the understorey of Tasmania’s wet Eucalyptus forests are particularly susceptible to these disturbances. Currently, forest managers have no means of comprehensively mapping these understorey rainforest communities; fieldwork techniques are costly and impractical at the landscape scale, and most remote sensing techniques are unable to effectively map sub-canopy habitats due to the blocking effect of the canopy. This presents a challenge for forest managers that must be addressed if these forests are to be managed sustainably. In this project, I examine two techniques that explore potential relationships between floristic and structural forest components as a means of locating rainforest understoreys. First, based on the premise that fire-driven succession of canopy and understorey strata follow parallel trajectories, I tested whether eucalypt canopy age-structure can be used to predict understorey floristics. I surveyed forty plots representative of the structural variation in the landscape, measuring the relative amounts of rainforest and old-growth eucalypts in each. From this, I generated a eucalypt old-growth and a rainforest variable, and compared these using Spearman’s rank correlation. While positive, the correlation between these variables was weak (ρ = 0.43). This result was unexpected, but may be explained by potentially independent effects of fire-disturbance in these two strata. Second, I used a LiDAR-based approach in an attempt to delineate rainforest understoreys based on structural characteristics. For this, I developed a suite of understorey LiDAR metrics that were tested against the amount of rainforest in the understorey using Spearman’s rank correlation and Random Forest analysis. Canopy density and height-based metrics showed significant positive correlations, but these were too weak for predictive purposes (ρ < 0.69). Similarly, the Random Forest analysis was unable to identify a predictive relationship (percent variance explained = 46 %). These analyses exposed notable structural overlap among the rainforest and wet sclerophyll understorey types. This project revealed a mismatch between canopy age-structure and understorey floristics, and between the structure and floristics within the understorey. I suggest that to better understand these relationships more research into fire-driven disturbances on understorey and canopy strata is required. This research, and the use of novel remote sensing approaches and technologies, will enable more effective management of understorey rainforest communities