ARS | Publication request: Determinants and prediction of C:N ratio in five diverse crop plants

Submitted to: Communications in Soil Science and Plant Analysis
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 28, 2008
Publication Date: N/A

Interpretive Summary: The C:N ratio is used frequently as a quality index of crop residues and as an important factor in determining residue decomposition. Residue decomposition is an important function of returning nutrients that the crop has used back to the soil for the use of subsequently grown crops. In order to develop a better understanding of how the C:N ratio and ultimately decomposition processes can be predicted for various crops we developed and validated statistical models for the C:N ratios in stems, leaves and roots of two traditional (corn and soybean) and three alternative (alfalfa, cuphea and switchgrass) crops. One of the difficulties encountered in evaluating crop residue in the field is the contribution and processing of below-ground biomass. The major outcome of this examination is that we were able to establish that stem C:N could be used to reliably predict the C:N ratio of roots. Results of this study provide insights into the interrelationships among biochemical composition and C:N ratios necessary for agronomists and farmers to predict the crop residue decomposition and to design more diverse and viable crop rotations that can be important in reducing the need for additional inputs, such as chemical fertilizers.

Technical Abstract: In order to fully predict the fertilizing potential of crop residues, insight into the interrelationships among their biochemical composition and C:N ratio is required. The C:N ratio is frequently used as a quality index of crop residues without taking into consideration the large within-crop variation and co-variation of biochemical constituents. We examined multivariate relationships in and statistical moments of eight biochemical constituents and their impact on estimating C:N ratio in organs (stems, leaves and roots) of alfalfa, corn, soybean, cuphea and switchgrass residues as candidates in diverse crop rotations. The portion of explained variance (R2), the Root Mean Square Error (RMSE) in the Partial Least Squares (PLS) prediction and validation models, and results of sensitivity analyses using Artificial Neural Networks (ANN) indicated that (1) equal portions of variation (R2=35.0) in C:N were explained by differences among crops and by differences among organs; however, the largest variation in N (R2 =53.3%) and C (R2 =70.9%) were explained by differences among crops and among organs within crops, respectively. (2) Variation in N, but not in C or N+C, content explained the greatest variance (R2>73.0 for crops, and >62.0 for organs) in C:N ratios. (3) On a multivariate scale, stems were closer to roots than to leaves; hence the large portion of variation in C:N ratio in roots explained by variation in biochemical constituents in stems and leaves (R2 =61.0%), and in stems only (R2 =58.0%). (4) Statistical moments, other than mean values of biochemical constituents, significantly impacted C:N ratio estimates and the reliability of these estimates (i.e., RMSE), both of which were positively correlated (r=0.64, p<0.001).