Model-assisted estimation of forest resources with generalized additive models

bration. Multi-phase surveys are often conducted in forestry, with the goal of estimating tree characteristics and volume over large regions. Design-based estimation of such quantities, based on information gathered during ground visits of sampled plots, can be made more precise by incorporating auxiliary information available from remote sensing. The exact relation-ship between the ground visit measurements and the remote sensing variables is not known, hence it is modelled using generalized additive models. Non-parametric estimators for these models are discussed and applied to forest data collected in northeastern Utah in the United States. By using these model predictions in a model-assisted survey estimation procedure for tree volume and related variables, we improve the accuracy of the survey estimates com-pared to currently available estimation procedures. The procedures described in this article are applica-ble to many other survey contexts.