Managed forests play crucial roles in ongoing climatic and environmental changes. Among other things, wood is capable of sinking and storing carbon in both standing timber and wood products. To promote these positive effects, more precise planning is required that will ensure sustainable forest management and maximal deposition of harvested wood for long-term applications. Information on wood properties plays a key role; i.e. the wood properties can impact the carbon stocks in forests and the suitability of wood for structural timber.
With respect to the theoretical background of wood formation, stem, crown, and branching constitute potential inputs (i.e. wood quality indicators) to allometric wood property, tree biomass, and wood quality models. Due to the complex nature of wood formation, measurements of wood quality indicators that could predict wood properties along the relevant directions of variation have previously been elusive in forest inventories. However, developments in laser scanning from aerial and terrestrial platforms support more complex mapping and modeling regimes based on dense three-dimensional point clouds.
The aim here was to determine how wood properties could be estimated in remote-sensing-aided forest inventories. For this purpose, methods for characterizing select wood quality indicators in standing timber, using airborne and terrestrial laser scanning (ALS and TLS, respectively) were developed and evaluated in managed boreal Scots pine (Pinus sylvestris L.) forests. Firstly, the accuracies of wood quality indicators resolved from TLS point clouds were assessed. Secondly, the results were compared with x-ray tomographic references from sawmills. Thirdly, the accuracies of tree-specific crown features delineated from the ALS data in predictive modeling of the wood quality indicators were evaluated.
The results showed that the quality and density of point clouds significantly impacted the accuracies of the extracted wood quality indicators. In the assessment of wood properties, TLS should be considered as a tool for retrieving as dense stem and branching data as possible from carefully selected sample trees. Accurately retrieved morphological data could be applied to allometric wood property models. The models should use tree traits predictable with aerial remote sensing (e.g. tree height, crown dimensions) to enable extrapolations.
As an outlook, terrestrial and aerial remote sensing can play an important role in filling in the knowledge gaps regarding the behavior of wood properties over different spatial and temporal extents. Further interdisciplinary cooperation will be needed to fully facilitate the use of remote sensing and spatially transferable wood property models that could become useful in tackling the challenges associated with changing climate, silviculture, and demand for wood.