ESTIMATING CO2 EMISSIONS FROM TILLED SOILS THROUGH ARTIFICIAL NEURAL NETWORKS AND MULTIPLE LINEAR REGRESSION1

ABSTRACT Quantifying soil gas emissions is costly, since it requires specific methodologies and equipment. The objective of this study was to evaluate modeling by nonlinear regression artificial neural networks (ANN) estimate CO2 caused managements. were evaluated in two different management systems: no-tillage minimum tillage. Readings flow carried out an automated closed system chamber; temperature, water content, density, total organic carbon also determined. model the ANN models adjusted based on correlation variables measured areas where managed with tillage data emission. Artificial are more accurate determine correlations between content than linear regression.