Predicting bulk density in Brazilian soils for carbon stocks calculation: a comparative study of multiple linear regression and Random Forest models using continuous and categorical variables.

WHARLEY PEREIRA DOS SANTOS; CARLOS MANOEL PEDRO VAZ, CNPDIA; LADISLAU MARTIN NETO, CNPDIA; ADRIANO ANSELMI, BAYER; JAVIER TOMASELLA, INSTITUTO NACIONAL DE PESQUISAS ESPACIAIS; FALBERNI DE SOUZA COSTA, CPAF-AC; JACKSON ADRIANO ALBUQUERQUE; QUIRIJN DE JONG VAN LIER, UNIVERSIDADE DE SÃO PAULO; RAFAEL GALBIERI, INSTITUTO MATO-GROSSENSE DO ALGODÃO; FABIANO JOSE PERINA, CNPA.2026-07-07http://hdl.handle.net/123456789/463121The present study aims to develop pedotransfer functions for estimating ρb using stepwise multiple linear regression applied to soil properties from available databases, which are considered suitable predictors (PTF-1).openAccessQuímica del sueloDensidad aparenteCarbono orgánico del sueloDisponibilidad de nutrientesFunciones de edafotransferanciaPredicciónAnálisis estadísticoAnálisis de regresiónQuímica do SoloCarbonoEstoqueDisponibilidade de NutrienteEstimativaAnálise EstatísticaRegressão LinearSoil chemistryBulk densitySoil organic carbonNutrient availabilityPedotransfer functionsPredictionStatistical analysisRegression analysisPredicting bulk density in Brazilian soils for carbon stocks calculation: a comparative study of multiple linear regression and Random Forest models using continuous and categorical variables.Artigo de periódico