A multi-scale approach for semi-automatic comparison assessment of large scale forest maps

Abstract

In terms of climate change analysis, information on large area forest cover distribution becomes increasingly important for studying terrestrial carbon cycle changes and its human impacts. Within the FOREST DRAGON 1 project large-area forest growing stock volume maps of Northeast and Southeast China based on ERS-1/2 tandem coherence data have been generated. For the validation of the large-area forest stem volume maps, a special cross-comparison design mainly based on freely available Earth Observation products had to be developed in consequence of lacking extensive in situ measurements. The sampling design, based on the FAO FRA2010 Sample Design and the Degree Confluence Project, uses a 1 degree sampling grid with 10 x 10 km sample plots. A reasonable agreement above 70 % between the forest growing stock volume maps and the land cover datasets in terms of forest/ non-forest could be achieved for a total area of about 4.5 million km².