C and N Mineralization Dynamics in Composts: Prediction of Soluble Organic Carbon by Multiple Nonlinear Regression

Abstract

Urban biosolids present a considerable concentration of nutrients, which are currently wasted and deposited in landfills causing environmental contamination. In the present study, a dimensionality reduction technique is used to select indicators with a higher relationship in their variability. Subsequently, a multivariate nonlinear regression process is used to establish an equation that allows predicting the behavior of the soluble organic carbon indicator. The indicators with the greatest relationship with the variability of the data analyzed were N-NO3-, N-NH4+/N-NO3- and IES. The resulting model presented a correlation of 30% with the soluble organic carbon indicator in the composting systems.