A meta-review of remote sensing for rubber plantations

Abstract

The rapid expansion of rubber plantations has led to significant ecological impacts, including deforestation and reduced carbon storage capacity, such as rubber-induced 4.1 million ha of forest loss in Southeast Asia (SEA). Remote sensing, essential for monitoring rubber expansion and supporting initiatives like REDD+ and EU deforestation regulations. However, due to the heterogeneity, dynamism, and complexity of rubber plantations, developing universal mapping algorithms remains challenging, and there is a lack of all-around reviews on remote sensing research specific to rubber plantations. Our meta-review systematically analyzes 305 peer-reviewed papers (2000–2024), synthesizing advancements in remote sensing techniques. The findings show that research is primarily focused on SEA (82 %), while regions like Africa and South America are underexplored. Optical data remains dominant (68 %), but the use of SAR has tripled, achieving up to 89 % accuracy when combined with phenological features. Additionally, deep learning improved classification accuracy by 15–20 %, especially in detecting young plantations under six years old. However, discrepancies and gaps in global plantation maps persist due to inconsistent validation methods and resolution limitations (>30 m). Our review highlights the need for standardized global datasets and provides insights into future research directions, including improved feature selection, algorithm transferability, and better integration of multi-source data to support sustainable plantation management and accurate carbon accounting.