Evaluation of criteria for describing crop residue degradability in a model of carbon and nitrogen turnover in soil

Good predictions of plant litter carbon and nitrogen turnover in soil depends heavily on a reliable quantification of litter quality. In this work, eight common agricultural crop residues were described by their intrinsic chemical properties and incubated in a sandy soil (15 degrees C and -10 kPa water potential). Measured C mineralization varied greatly between the plant materials and was well predicted by a simulation model in which litter C is subdivided into three pools according to the results of stepwise chemical digestion (Van Soest analysis). Nevertheless, there was a significant lack of fit for some of the materials. This was caused by differences in the specific decay rates of holocellulose-like substances (k(SPM)) as subsequently estimated for each plant material by fitting the model to data on remaining holocellulose. Even though the model takes account of microbial N deficiency, the optimized k(SPM) values were significantly correlated with initial N (r(2) = 0.93) but not with lignin concentration. To test the predictive value in our model of indices and quantities thought to be related to litter degradability, we investigated whether they were correlated with the kinetically defined pool of readily decomposable plant constituents as estimated by fitting the model to measured C mineralization rates. Neutral detergent-soluble C (Van Soest analysis) was best correlated with the estimated pool (r(2)=0.78) followed by water-soluble C (r(2)=0.69) and C digestible in vitro in rumen fluid (r(2) =0.66). Measured C to N ratio in holocellulose-like substances was highly correlated with the overall C to N ratio of the plant materials (r(2) = 0.96). In the model, we describe the degradability of litter N on the basis of the measured C to N ratios of the litter pools. The simulated microbial N requirement is governed by a successive replacement of rapidly growing organisms with a low C to N ratio by more slowly growing organisms with a slightly higher C to N ratio, reflecting the commonly observed increase along the decay continuum in the fungal contribution to microbial activity. This model feature was supported by a measured tendency to an increasing biomass C to N ratio. The model gave unbiased simulations of N mineralization and microbial biomass N. This indicates that the descriptions of litter N availability and microbial N requirement in the model were reasonable. However, significant discrepancies between simulated and measured values occurred for some of the plant materials during the first few days of decomposition, emphasizing the need for more accurate knowledge for this very dynamic phase. Our results suggest that an a priori characterization of litter degradability is possible in our model by the use of stepwise chemical digestion for subdivision of litter C and N combined with measurements of initial N to set the decay rate constant of holocellulose-like materials. (C) 1999 Elsevier Science Ltd. All rights reserved.