Design-based treatment of missing data in forest inventories using canopy heights from aerial laser scanning

Nonresponse is often a problem in forest inventories. It may arise when sample plots are inaccessible because of hazardous terrain. To account for this problem, the use of nonresponse calibration weighting is considered in a complete design-based framework, i.e., both nonresponse and survey variables are viewed as fixed characteristics of the plots. Information derived from remotely sensed data is exploited to compensate for the missing plots. Calibration is performed adopting canopy height from airborne laser scanning as an auxiliary variable. Conditions for approximate unbiasedness of the calibration estimator in two-phase inventories are derived, and some estimators of the sampling variance are proposed. Results from one-phase inventories are achieved as a particular case. Dummy variables are introduced in the presence of different forest types. Monte Carlo results support the validity of the procedure. An application to a forest survey carried out in Central Italy is performed.