Background: Forest understorey structure is an important component of forest ecosystems that affects forest-dwelling species, nutrient cycling, fire behaviour, biodiversity, and regeneration capacity. Mapping the structure of forest understorey vegetation with field surveys or high-resolution LiDAR data is costly. We tested whether landscape topography and underlying geology could predict the understorey structure of a 19 km2 area of wet eucalypt primary forest located at the Warra Long Term Ecological Research Supersite, Tasmania, Australia. In this study, we used random forest regressions based on twelve topographic attributes derived from digital terrain models (DTMs) at various resolutions and a geology variable to predict the densities of three understorey layers compared to density estimates from a high resolution (28.66 points/m2) LiDAR survey.
Results: We predicted the vegetation density of three canopy strata with a high degree of accuracy (validation root mean square error ranged from 8.97% to 13.69%). 30 m resolution DTMs provided greater predictive accuracy than DTMs with higher spatial resolution. Variable importance depended on spatial resolutions and canopy strata layers, but among the predictor variables, geology generally produced the highest predictive importance followed by solar radiation. Topographic position index, aspect, and SAGA wetness index had moderate importance.
Conclusions: This study demonstrates that geological and topographic attributes can provide useful predictions of understorey vegetation structure in a primary forest. Given the good performance of 30 m resolution, the predictive power of the models could be tested on a larger geographical area using lower density LiDAR point clouds. This study should help in assessing fuel loads, carbon stores, biomass, and biological diversity, and could be useful for foresters and ecologists contributing to the planning of sustainable forest management and biodiversity conservation.