Biomass Prediction Using Sentinel-2 Imagery and an Artificial Neural Network in the Amazon/Cerrado Transition Region

• Published in: Forests 2024-09, Vol.15 (9), p.1599
• Main Authors: Faria, Luana Duarte de, Matricardi, Eraldo Aparecido Trondoli, Marimon, Beatriz Schwantes, Miguel, Eder Pereira, Junior, Ben Hur Marimon, Oliveira, Edmar Almeida de, Prestes, Nayane Cristina Candido dos Santos, Carvalho, Osmar Luiz Ferreira de
• Format: Article
• Language: English
• Subjects: aboveground biomass; aerosols; Algorithms; Aluminum; Anthropogenic factors; Artificial intelligence; Artificial neural networks; Biodiversity; Biomass; Biomes; Brazil; Carbon; Carbon dioxide; Carbon dioxide emissions; cattle; cerrado; Correlation coefficient; Correlation coefficients; Deep learning; Deforestation; Dry season; Ecosystem services; Ecosystems; ecotones; Error analysis; Estimation; Forest biomass; forests; habitat destruction; Habitat loss; indigenous species; Machine learning; natural resources conservation; Neural networks; normalized difference vegetation index; Normalized difference vegetative index; Parameter sensitivity; Plant ecology; prediction; Remote sensing; Root-mean-square errors; Satellite imagery; Soils; t-test; Tropical environment; Tropical environments; Vegetation
• ISSN: 1999-4907; 1999-4907
• DOI: 10.3390/f15091599

The ecotone zone, located between the Cerrado and Amazon biomes, has been under intensive anthropogenic pressures due to the expansion of commodity agriculture and extensive cattle ranching. This has led to habitat loss, reducing biodiversity, depleting biomass, and increasing CO2 emissions. In this study, we employed an artificial neural network, field data, and remote sensing techniques to develop a model for estimating biomass in the remaining native vegetation within an 18,864 km2 ecotone region between the Amazon and Cerrado biomes in the state of Mato Grosso, Brazil. We utilized field data from a plant ecology laboratory and vegetation indices from Sentinel-2 satellite imagery and trained artificial neural networks to estimate aboveground biomass (AGB) in the study area. The optimal network was chosen based on graphical analysis, mean estimation errors, and correlation coefficients. We validated our chosen network using both a Student’s t-test and the aggregated difference. Our results using an artificial neural network, in combination with vegetation indices such as AFRI (Aerosol Free Vegetation Index), EVI (Enhanced Vegetation Index), and GNDVI (Green Normalized Difference Vegetation Index), which show an accurate estimation of aboveground forest biomass (Root Mean Square Error (RMSE) of 15.92%), can bolster efforts to assess biomass and carbon stocks. Our study results can support the definition of environmental conservation priorities and help set parameters for payment for ecosystem services in environmentally sensitive tropical regions.