Aplicação de Técnicas de Aprendizado de Máquina na Determinação de Estoque de Carbono no Solo
Resumo
Os solos representam o mais significativo estoque de carbono orgânico (SOC) nos ecossistemas terrestres, sublinhando a importância crítica de estimar com precisão o carbono orgânico do solo para garantir a preservação das funções do solo e a mitigação das alterações climáticas globais. Este estudo emprega uma metodologia baseada em dados para estimar os estoques de carbono em solos brasileiros, comparando técnicas de aprendizado de máquina com diversas estratégias de otimização de hiperparâmetros. Os resultados demonstram o papel fundamental da seleção e processamento de dados, juntamente com a otimização de hiperparâmetros, na resolução deste problema, resultando em melhorias notáveis nas análises do erro médio absoluto (MAE) e da raiz do erro quadrático médio (RMSE).
Palavras-chave:
Solos, Estoque de Carbono, Aprendizado de Máquina, Hiperparâmetros
Referências
Al-Qinna, M. and Jaber, S. (2013). Predicting soil bulk density using advanced pedotransfer functions in an arid environment. Transactions of the ASABE, 56(3):963–976.
Ceddia, M. B. et al. (2015). Spatial variability of soil carbon stock in the urucu river basin, central amazon-brazil. Science of the Total Environment.
Ceddia, M. B. et al. (2016). The use of pedotransfer functions and the estimation of carbon stock in the central amazon region. Scientia Agricola.
Ferreira, A. C. S. et al. (2023). Predicting soil carbon stock in remote areas of the central amazon region using machine learning techniques. Geoderma Regional - Elsevier.
Gomes, L. C. et al. (2019). Modelling and mapping soil organic carbon stocks in brazil. Geoderma, 340:337–350.
Haddad, D. B. et al. (2017). A first approach using neural network to estimating soil bulk density of urucu basin in central amazon-brazil. IEEE - Institute of Electrical and Electronic Engineers.
Haddad, D. B. et al. (2018). Brazilian soil bulk density prediction based on a committee of neural regressors. IEEE - Institute of Electrical and Electronic Engineers.
Japa, L. et al. (2023). A population-based hybrid approach for hyperparameter optimization of neural networks. IEEE Access, 11:50752–50768.
Kumar, S. et al. (2023). Potential impact of data-centric ai on society. IEEE Technology and Society Magazine, 42(3):98–107.
Mahmoudzadeh, H. et al. (2020). Spatial prediction of soil organic carbon using machine learning techniques in western iran. Geoderma Regional - Elsevier.
Mousavi, S. R. et al. (2022). Three-dimensional mapping of soil organic carbon using soil and environmental covariates in an arid and semi-arid region of iran. Journal of the International Measurement Confederation - Elsevier.
Silatsa, F. B. et al. (2020). Assessing countrywide soil organic carbon stock using hybrid machine learning modelling and legacy soil data in cameroon. Geoderma - Elsevier.
Song, J. et al. (2022). Estimation of soil organic carbon content in coastal wetlands with measured vis-nir spectroscopy using optimized support vector machines and random forests. Remote Sensing - MDPI.
Szatmari, G. et al. (2023). Countrywide mapping and assessment of organic carbon saturation in the topsoil using machine learning-based pedotransfer function with uncertainty propagation. Catena - Elsevier.
Tranter, G. et al. (2007). Building and testing conceptual and empirical models for predicting soil bulk density. Soil Use and Management, 23(4):437–443.
Wadoux, A. M.-C. et al. (2020). Machine learning for digital soil mapping: applications, challenges and suggested solutions. Earth-Science Reviews - Elsevier.
Wadoux, A. M. J.-C. et al. (2023). Shapley values reveal the drivers of soil organic carbon stock prediction. Soil.
Ye, Z. et al. (2021). Using machine learning algorithms based on gf-6 and google earth engine to predict and map the spatial distribution of soil organic matter content. Sustainability - MDPI.
Ceddia, M. B. et al. (2015). Spatial variability of soil carbon stock in the urucu river basin, central amazon-brazil. Science of the Total Environment.
Ceddia, M. B. et al. (2016). The use of pedotransfer functions and the estimation of carbon stock in the central amazon region. Scientia Agricola.
Ferreira, A. C. S. et al. (2023). Predicting soil carbon stock in remote areas of the central amazon region using machine learning techniques. Geoderma Regional - Elsevier.
Gomes, L. C. et al. (2019). Modelling and mapping soil organic carbon stocks in brazil. Geoderma, 340:337–350.
Haddad, D. B. et al. (2017). A first approach using neural network to estimating soil bulk density of urucu basin in central amazon-brazil. IEEE - Institute of Electrical and Electronic Engineers.
Haddad, D. B. et al. (2018). Brazilian soil bulk density prediction based on a committee of neural regressors. IEEE - Institute of Electrical and Electronic Engineers.
Japa, L. et al. (2023). A population-based hybrid approach for hyperparameter optimization of neural networks. IEEE Access, 11:50752–50768.
Kumar, S. et al. (2023). Potential impact of data-centric ai on society. IEEE Technology and Society Magazine, 42(3):98–107.
Mahmoudzadeh, H. et al. (2020). Spatial prediction of soil organic carbon using machine learning techniques in western iran. Geoderma Regional - Elsevier.
Mousavi, S. R. et al. (2022). Three-dimensional mapping of soil organic carbon using soil and environmental covariates in an arid and semi-arid region of iran. Journal of the International Measurement Confederation - Elsevier.
Silatsa, F. B. et al. (2020). Assessing countrywide soil organic carbon stock using hybrid machine learning modelling and legacy soil data in cameroon. Geoderma - Elsevier.
Song, J. et al. (2022). Estimation of soil organic carbon content in coastal wetlands with measured vis-nir spectroscopy using optimized support vector machines and random forests. Remote Sensing - MDPI.
Szatmari, G. et al. (2023). Countrywide mapping and assessment of organic carbon saturation in the topsoil using machine learning-based pedotransfer function with uncertainty propagation. Catena - Elsevier.
Tranter, G. et al. (2007). Building and testing conceptual and empirical models for predicting soil bulk density. Soil Use and Management, 23(4):437–443.
Wadoux, A. M.-C. et al. (2020). Machine learning for digital soil mapping: applications, challenges and suggested solutions. Earth-Science Reviews - Elsevier.
Wadoux, A. M. J.-C. et al. (2023). Shapley values reveal the drivers of soil organic carbon stock prediction. Soil.
Ye, Z. et al. (2021). Using machine learning algorithms based on gf-6 and google earth engine to predict and map the spatial distribution of soil organic matter content. Sustainability - MDPI.
Publicado
14/10/2024
Como Citar
PARDELINHA, Alexandre; CEDDIA, Marcos Bacis; GERVASIO, Roberto; BELLOZE, Kele; AGUILAR, Carolina de L.; ASSIS, Laura; BRANDÃO, Diego.
Aplicação de Técnicas de Aprendizado de Máquina na Determinação de Estoque de Carbono no Solo. In: SIMPÓSIO BRASILEIRO DE BANCO DE DADOS (SBBD), 39. , 2024, Florianópolis/SC.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2024
.
p. 673-679.
ISSN 2763-8979.
DOI: https://doi.org/10.5753/sbbd.2024.243218.
