Skip to main content
SLU publication database (SLUpub)

Conference paper2005

Detection of failed regeneration with multitemporal satellite remote sensing

Joyce, Stephen; Olsson, Håkan

Abstract

This study is an attempt to identify areas of failed forest regeneration from a time sequence of satellite imagery. All new clearfellings between 1990 and 1992 were detected and mapped from a pair of summer Landsat TM scenes within a 50 x 50km area in northern Sweden. A selection of these stands was visited in the field in 2004 to assess the current regeneration status. Spectral mean values in these stands were extracted from summer Landsat imagery from 2000 and 2004. The analysis showed that it was not possible to discriminate the failed regeneration areas in the image from 2000 (8 or 9 years after the previous stand was harvested), but there was significant separability in the image from 2004. The main factors that determined spectral separability were a combination of tree height and stocking density. It seems that during this early stage of regeneration, the re-colonization of the background vegetation and differences in site types dominate the satellite spectral signal before the trees make a significant contribution. Regeneration status was not apparent in the satellite imagery until the trees reached approximately 2.5m in height

Published in


Publisher: Skogsstyrelsen

Conference

ForestSat: Operational Tools in Forestry Using Remote Sensing Techniques