The Influence of Field Sample Data Location on Growing Stock Volume Estimation in Landsat TM-Based Forest Inventory in Eastern Finland

Abstract

Forest inventory methods that utilize Landsat TM data mostly assume that all sample plots within the image can be used as ground reference data for any subarea of the image. The assumption about the suitability of the field sample data within a satellite image area was evaluated in this study. The best overall results were obtained when plots within a 20-km range were used in point estimation of plot growing stock volume within a 7×8-km field data network. The estimation procedure was also tested in the calculation of area-level statistics. When the restriction of a 20-km distance was used in volume estimation, the error was always smaller than the sample plot variation and was always smaller than results received from estimates with no distance limit. The best benefits from incorporating satellite data were obtained for the smallest areas using a “20-km distance” restriction.