Machine learning and satellites reveal carbon-storing power of mango orchards
image: Machine learning technique for carbon sequestration estimation of mango orchards area using Sentinel-2 Data
Credit: Gitsada Panumonwatee, Sittichai Choosumrong, Savent Pampasit, Rudklow Premprasit, Tatsuya Nemoto & Venkatesh Raghavan
A team of researchers has developed a powerful new method to measure how much carbon mango orchards can store, using artificial intelligence and satellite imagery. Their approach could help farmers and policymakers better understand the role of fruit trees in tackling climate change.
The study, published in Carbon Research, combines data from the European Space Agency’s Sentinel-2 satellites with machine learning models to estimate how much carbon is locked away in mango trees across Phitsanulok Province, Thailand. The researchers found that the Random Forest model they developed could predict carbon storage with remarkable precision, achieving an accuracy of 97 percent.
“Mango orchards are an important part of Thailand’s agricultural economy, but their potential as a natural carbon sink has not been well quantified,” said lead author Gitsada Panumonwatee of Naresuan University. “By integrating satellite data with artificial intelligence, we can now estimate carbon sequestration across large areas quickly and cost-effectively.”
The team collected data from 49 field plots across seven districts and analyzed 12 vegetation indices that describe plant health and biomass, such as the Normalized Difference Vegetation Index (NDVI) and the Modified Triangular Vegetation Index (TVI-2). These indices were used to train a Random Forest model that learns the relationships between satellite signals and the actual carbon measured on the ground.
The results showed that mango orchards can store an average of 40.6 tons of carbon per hectare, with individual sites ranging from 4.1 to 218.6 tons depending on tree age, density, and management practices. Among the tested indicators, TVI-2 was the most influential, followed by NDVI and GNDVI, highlighting the importance of canopy structure and leaf area for carbon storage.
“This approach provides a robust tool for monitoring carbon sequestration in fruit tree plantations,” said co-author Sittichai Choosumrong. “It also supports carbon credit programs and sustainable farming strategies that contribute to climate change mitigation.”
Because the method relies on freely available satellite imagery and open-source machine learning tools, it offers a scalable and low-cost solution for estimating carbon stocks in agricultural systems worldwide. The researchers hope their framework can be extended to other fruit crops and integrated into climate-smart land management programs.
“Our model can help link farmers’ practices to measurable climate benefits,” Panumonwatee added. “It’s a step forward in using digital technology to make agriculture part of the climate solution.”
===
Journal Reference: Panumonwatee, G., Choosumrong, S., Pampasit, S. et al. Machine learning technique for carbon sequestration estimation of mango orchards area using Sentinel-2 Data. Carbon Res. 4, 33 (2025). https://doi.org/10.1007/s44246-025-00201-z
===
About Carbon Research
The journal Carbon Research is an international multidisciplinary platform for communicating advances in fundamental and applied research on natural and engineered carbonaceous materials that are associated with ecological and environmental functions, energy generation, and global change. It is a fully Open Access (OA) journal and the Article Publishing Charges (APC) are waived until Dec 31, 2025. It is dedicated to serving as an innovative, efficient and professional platform for researchers in the field of carbon functions around the world to deliver findings from this rapidly expanding field of science. The journal is currently indexed by Scopus and Ei Compendex, and as of June 2025, the dynamic CiteScore value is 15.4.