Quantifying oil palm expansion in Southeast Asia from 2000 to 2015: A data fusion approach

• Published in: Journal of land use science 2022-01, Vol.17 (1), p.26-46
• Main Authors: Wagner, Melissa, Wentz, Elizabeth A., Stuhlmacher, Michelle
• Format: Article
• Language: English
• Subjects: accuracy assessment; Agricultural land; Classification; Data fusion; Data integration; Deforestation; Information processing; Land cover; land cover change; Land management; Land use; land use change; oil palm; Radar; Radar data; Radar imaging; remote sensing; Sustainability management; Vegetable oils
• ISSN: 1747-423X; 1747-4248
• DOI: 10.1080/1747423X.2021.2020918

The fusion of optical imagery with radar data can provide more accurate land cover change analysis of deforestation and tree-based agriculture. Radar data is limited temporally with most geographic areas not covered prior to 2007. This paper presents a new methodology to classify land cover change related to oil palm expansion that takes historic data limitations into account. Our approach utilizes Hansen's Global Forest Cover data, optical imagery, and texture information, to extract land cover information in Sumatra and Western Malaysia, where historical data is absent. Our method demonstrates how to accurately classify oil palm without radar data with overall accuracies for optical only experiments within 4.4% of optical plus radar classifications. Our results show agricultural land use was the primary driver of land cover change with the largest increase due to oil palm expansion (6.1%). Better estimations of oil palm expansion could be used in sustainable land management policies.