Skip to main content
. 2023 Oct 18;623(7986):340–346. doi: 10.1038/s41586-023-06642-z

Table 2.

Comparison of rubber-related deforestation estimates generated by this and other studies

Method Definition of ‘forest’ Time period Reference area Rubber-related deforestation in 1,000 ha yr−1
Total in reference area Indonesia Thailand Malaysia Cambodia
Ref. 4 Land balance model Tree cover greater than or equal to 25% (ref. 51) 2005–2017 135 tropical countries, including all chief rubber producers (except China and Laos) 53 22 9 5 0.1
Ref. 3 2005–2018 52 23 6 5 3
Ref. 1 Mix of spatially explicit data Tree cover greater than or equal to 30% (ref. 51) 2001–2015 Brazil, Cambodia, Cameroon, Democratic Republic of the Congo, India, Indonesia and Malaysia 140 64 NA 48 22
Ref. 2 Remote sensing Internal classifier 2003–2014 Mainland Southeast Asia 135 NA NA NA 69
437* 232*
Ref. 29 Remote sensing Tree cover greater than or equal to 10% (ref. 52) 2001–2015 Cambodia NA NA NA 34
This study Remote sensing ESA WorldCover 10 m 2020 v.100 (tree cover greater than or equal to 10%) 2001–2016 (baseline 1993) Southeast Asia 186 66 39 20 15
156* ± 22 NA NA NA NA

The dataset in bold (first row) has been used to guide deforestation policy7 and to calculate the imported deforestation of individual countries6,8. In this study, we use a conservative baseline of 1993. The earliest baseline in other studies is 2000 and hence other studies will include more plantation rotation. The different base lines also mean that our estimates cannot easily be set into the context of overall deforestation in Southeast Asia (estimated to be 3.22 million ha yr−1 between 2001 and 201918). At face value our rubber deforestation estimates account for 5–6% of that figure but this is very conservative as the overall figure is derived using a baseline of 2000 and hence includes more plantation rotation (of rubber and other types of tree cover). Sample-based area estimates for this study (following ref. 33) and for ref. 2 are indicated by an asterisk.