Mapping fractional forest cover across the highlands of mainland Southeast Asia using MODIS data and regression tree modelling

• Published in: International journal of remote sensing 2007-01, Vol.28 (1), p.23-46
• Main Authors: Tottrup, C., Rasmussen, M. S., Eklundh, L., Jönsson, P.
• Format: Article
• Language: English
• Subjects: Animal, plant and microbial ecology; Applied geophysics; Biological and medical sciences; Earth and Related Environmental Sciences; Earth sciences; Earth, ocean, space; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Geovetenskap och miljövetenskap; Internal geophysics; Natural Sciences; Naturgeografi; Naturvetenskap; Physical Geography; Teledetection and vegetation maps
• ISSN: 0143-1161; 1366-5901; 1366-5901
• DOI: 10.1080/01431160600784218

Data from the moderate-resolution imaging spectroradiometer (MODIS) sensor, in combination with new mapping techniques, has the potential to improve regional research on tropical forest resources and land use dynamics. In this study, a supervised regression tree model was used to map fractions of (1) mature forest, (2) secondary forest, and (3) non-forest, using multi-temporal MODIS 250-m data as explanatory variables, and land cover information derived from high-spatial resolution image data as the response variables. From independent validation data, the overall mean absolute deviation of the resulting maps are estimated at 14.6% for mature forest, 21.6% for secondary forest, and 17.1% for non-forest cover. This study shows the increased potential of this new mapping technique to infer human imprints on forest cover across the highlands of mainland Southeast Asia, compared to other existing map sources.