Evaluating sample plot imputation techniques as input in forest management planning

Abstract

Recent advances in the use of data from airborne laser scanners have produced results that are potentially useful for forest-management planning. In this study, the results from recently developed imputation techniques using laser scanner and satellite data were evaluated as input in a timber-oriented forestry planning context. Evaluation comprised a cost plus loss analysis in which the data cost for a specific method is added to the expected loss arising from nonoptimal forestry activities caused by erroneous forest descriptions. Forest data from sample plot imputations based on laser scanner data, satellite data, or a combination of both were available for 64 stands in southern Sweden. For comparison, sample plot field inventories of 5 and 10 plots were simulated for each stand. Different stand areas and real interest rates were tested. The best performing imputation method, using both laser scanner and satellite data, produced the lowest total cost plus loss in the smallest stands when using the highest interest rate. In all other cases, the sample plot methods performed better. Operative phase considerations, altering the original plan, would likely mitigate the effect of nonoptimal forest-management decisions, improving the competitiveness of the imputation methods. Further analysis should include such owner-specific considerations.