An Application of Endogenous Post-Stratification Using Forest Inventory and Lidar Data

• Activity Number: 561
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Ronald E McRoberts*+ and Erik Næsset and Terje Gobakken
• Companies: U.S. Forest Service and Norwegian University of Life Sciences and Norwegian University of Life Sciences

Abstract

Endogenous post-stratification has been shown to be effective for increasing the precision of estimates of forest attributes such as forest area when the strata are constructed using satellite imagery such as Landsat. However, optical sensors such as Landsat respond primarily to passively reflected sunlight which, for forestry applications, originates in the canopy. Thus, Landsat-based stratifications are less effective for increasing the precision of estimates of sub-canopy attributes such as tree volume. In recent years, a strong relationship has been demonstrated between lidar metrics and sub-canopy forest attributes. Lidar metrics are derived from responses to reflected light that has been emitted light from an airborne platform and can penetrate the forest canopy. Stratifications constructed by aggregating continuous lidar-based model predictions of tree volume into 2-8 strata are shown to reduce variances of estimates of tree volume by factors as great as 6.