A recent study published in the Journal of Remote Sensing introduces a groundbreaking method for estimating above-ground carbon (AGC) at the individual tree level, particularly in semi-arid regions. This approach, developed by Lobelia Earth S.L., leverages very high-resolution (VHR) satellite imagery and machine learning algorithms to provide a more precise tool for measuring carbon sequestration. The study, titled 'Unlocking precision in carbon stock mapping: a new AI-driven approach,' details an Artificial Neural Network (ANN) model trained on over 400 individual tree crowns, achieving AGC estimates with an R² of 0.66 and a relative RMSE of 78.6%.
The research team constructed a comprehensive AGC reference database from on-the-ground tree measurements, using species-specific allometric equations to convert these measurements into biomass. Deep learning models were then employed to segment individual tree crowns and extract spectral information from VHR imagery, which was used to train and validate the ANN model. The result was a highly accurate model, with a tree-level RMSE of just 373.85 kg, demonstrating its robustness in predicting AGC from remote sensing data.
Martí Perpinyana-Vallès, the study's lead author, highlighted the importance of this development, noting its potential to significantly improve our understanding of carbon sequestration dynamics and enhance global land management practices. The study utilized Pléiades Neo satellite imagery, known for its exceptional 30cm native resolution, to achieve unprecedented precision in Earth observation. This precision, combined with deep learning algorithms for crown extraction and ANN models for AGC prediction, allowed for the accurate geolocation of individual trees, addressing longstanding limitations in carbon stock estimation.
The implications of this technology are vast, promising to improve global carbon cycle assessments, optimize land use, and enhance reforestation initiatives. It could also provide essential data for climate change mitigation strategies, aiding policymakers in addressing environmental challenges. As the method gains wider adoption, it has the potential to harmonize carbon estimation discrepancies, offering invaluable support for international climate agreements and global sustainability efforts.
Published under the DOI10.34133/remotesensing.0359, the study was conducted as part of the Jeunesse Sahelienne pour l'Action Climatique (JESAC) project, with funding from Intermon Oxfam Spain, and under Industrial PhD grants AGAUR and DIN2020-010982 financed by MCIN AEI and the European Union 'NextGenerationEU/ PRTR'. Additional support came from the ESA Network of Resources Initiative. This research marks a significant advancement in the fight against climate change, providing a more accurate means of quantifying carbon within trees and enabling better tracking of mitigation efforts' effectiveness.
