Building a case for lidar-derived structure stratification for Australian softwood plantations.

Conventional resource inventories in Australian softwood plantations usually utilise Geographic Information System (GIS) thematic layers to stratify the resource prior to field sampling. Airborne lidar can offer a viable alternative for stratification but there are no standardised methods for plantation managers. This paper explores issues with current thematic stratification approaches and argues that relatively basic metrics extracted from a lidar-derived canopy height model (CHM) are suitable for constructing better stratification options. The case is supported with findings from an airborne lidar inventory undertaken in a pine plantation in New South Wales (NSW), Australia. Lidar stand level metrics including mean height, mean above mean height, mean dominant height, predicted stocking, canopy cover percentage, occupied volume and height variance were tested as surrogates for plantation structure. The study demonstrated that lidar metrics can predict stand attributes such as age class (R 2 = 0.91, RMSE 1.9), thinning treatment (89% accuracy), mean height, (R 2 = 0.95, RMSE 4%), stocking (R 2 = 0.82, RMSE 26%), basal area (R 2 = 0.67, RMSE 19%) and total stand volume (R 2 = 0.8, RMSE 19%) across a range of stand structures. Since the metrics tested were highly correlated with survey data it is argued that they could provide a valid basis for a developing a new structure stratification approach to improve sampling design in future plantation resource inventories.