Remote sensing and geostatistics applied to post-stratification of eucalyptus stands

Abstract

Brazil has many rural properties with unmanaged eucalyptus stands. These plantations are heterogeneous, presenting different tree sizes, advanced ages, and large wood volumes that can be quantified using forest inventories. The prediction error of dendrometric variables, mainly in highly heterogeneous areas, can be associated with inadequate forest inventory procedures, i.e. low intensity of sampling plots. However, a larger number of plots increases the cost of inventorying. Therefore, a promising alternative is forest stratification into homogeneous sub areas. Accordingly, the aim of this study was to analyze the reduction of volume estimate errors by post-stratification procedures. We used the normalized difference vegetation index (NDVI) derived from Landsat 8 and Spot 6 images and geostatistical techniques, such as kriging the volume (V) and diameter at breast height (DBH). The most precise method to estimate the total volume was the stratified random sampling (STS), based on geostatistical interpolation, using the DBH (error lower than 10%).