Description of the carbon mineralization of swine manure and oat straw in the soil through nonlinear models

Abstract

Agricultural management is a viable way for recycling animal residues in feedlots. The substances that make up organic residues change the dynamics of the organic matter decomposition in the soil. Information on carbon mineralization curves allows seeking improvements in soil quality and, consequently, in crop productivity. The Stanford & Smith Nonlinear Model is the most used to describe C mineralization of organic residues in the soil. This model considers organic residues are composed of substances that are mineralized exponentially. The Cabrera Model considers two fractions, one composed of substances that are mineralized exponentially and other composed of more resistant substances with constant mineralization. The objective of this work was to compare nonlinear models that describe carbon mineralization, considering residues on surface or incorporated into the soil. The data evaluated were from an experiment with oat straw, liquid swine manure, and swine litter bedding. The Stanford & Smith and Cabrera Models were used considering structure of first order autoregressive errors - AR(1), when necessary. The fittings were compared using the Akaike Information Criterion (AIC). The Cabrera Model was more adequate to describe C mineralization in four treatments (soil + incorporated liquid swine manure; soil + oat straw on surface + liquid swine manure on surface; soil + incorporated straw; and soil + straw on surface). The Stanford & Smith Model was better in three treatments (soil + incorporated straw + incorporated liquid swine manure; swine litter bedding on surface; and incorporated swine litter bedding). None of the models described the treatment soil + liquid swine manure on surface.