“…To date, Landsat vegetation indices (VIs) and their combinations, such as the Normalized Difference Vegetation Index (NDVI) (Robinson et al, 2017;Tucker et al, 1986), Normalized Difference Vegetation Index (NDMI) (Grogan et al, 2015;Vogelmann and Rock, 1988), Land Surface Water Index (Wang et al, 2018;Xiao et al, 2004), Enhanced Vegetation Index (Dong et al, 2016;Huete, 1988), and Normalized Burn Ratio (NBR) (Frazier et al, 2018;García and Caselles, 1991;Li et al, 2015), have been used to detect and map rubber plantations. The relevant studies have focused on small-scale, regional mapping of rubber plantations in traditional planting regions like Indonesia and Malaysia (Razak and Shariff, 2018;Trisasongko, 2017) as well as emerging rubbergrowing areas such as Xishuangbanna (Fan et al, 2015;Kou et al, 2015;Sun et al, 2017;Xiao et al, 2019c), Hainan Island Dong et al, 2013) in southern China, northeast Thailand (Li and Fox, 2011), and the borders of China, Myanmar, and Laos (Liu et al, 2013(Liu et al, , 2016, the Cambodia-Vietnam border (Ye et al, 2018), and the China-Laos border Tang et al, 2019;Xiao et al, 2020a). However, few studies have been concerned with developing robust algorithms over larger areas, such as the entire MSEA (Chen et al, 2016).…”