Document Type : Research Paper
Authors
1 Department of Animal and Poultry Sciences, Aburaihan College of Agricultural Technology, University of Tehran, Tehran, Iran
2 Department of Animal and Poultry Sciences, Aburihan Faculty of Agricultural Technology, University of Tehran, Tehran, Iran
3 Department of Animal Science, Faculty of Agriculture, Malayer University, Malayer, Iran
Abstract
Objective: Accurate evaluation of the chemical composition and nutritional quality of feedstuffs, particularly forage crops, is critical for formulating balanced rations, enhancing livestock performance, and minimizing production costs. Among the available analytical techniques, near-infrared reflectance (NIR) spectroscopy has gained prominence as a rapid, non-destructive, and cost-effective alternative to conventional wet chemistry methods. The NIR ability to analyze samples with no chemical reagents and minimal sample preparation makes it particularly attractive for routine applications. This study aimed to compare the accuracy of NIR with standard laboratory procedures in estimating chemical constituents, fractions of protein and carbohydrate based on the Cornell Net Carbohydrate and Protein System (CNCPS), and nutritional attributes of four legume forages.
Method: Forage samples from four species including two cultivars of common vetch (Vicia sativa) and hairy vetch (Vicia villosa), one cultivar of forage pea (Pisum arvense), and second-year alfalfa (Medicago sativa, used as the control crop) were analyzed for organic matter (OM), ash, acid detergent lignin (ADL), crude protein (CP), neutral detergent fiber (NDF), acid detergent fiber (ADF), starch, and CNCPS-based fractionation of protein and carbohydrates,. Nutritional indicators such as potential dry matter intake (DMI), total digestible nutrients (TDN), digestible energy (DE), metabolizable energy (ME), and quality index (QI) were also estimated. All analyses were conducted in parallel using NIR and the reference wet chemistry methods. Statistical agreement and precision between the two methods were assessed based on mean bias, root mean square error (RMSE), concordance correlation coefficient (CCC), and Bland–Altman limits of agreement (LOA), providing a comprehensive evaluation of methodological consistency.
Results: The NIR results showed high accuracy and strong correlation with wet chemistry methods for key components such as CP, OM, starch, total carbohydrates, and fraction B1 (B1), with CCC values exceeding 0.85 and no statistical differences (P > 0.05). The method also demonstrated acceptable precision in predicting energy-related parameters including TDN, DE, and ME, which are critical for ration formulation. However, for structural constituents such as ADL, NDF, protein fractions (ADIP, NADIP) and carbohydrates (B2, B3, and C), the accuracy and concordance declined, and statistically significant differences were observed. These findings suggest that the spectral sensitivity of NIR is limited when evaluating slowly degradable or indigestible fractions of carbohydrate and protein, making it less reliable for the parameters of dynamic nutritional models such as CNCPS.
Conclusions: Owing to its unique advantages, particularly speed, ease of operation, and compatibility with field analyses, NIR can serve asa useful tool for rapid screening, feed quality monitoring, and routine proximate analysis in feed laboratories. However, for the most accurate evaluation of CNCPS model components, particularly those resisting digestion, the use of concentional chemical methods offers greater advantages. Integrating NIR as a complementary tool for classical approaches may offer a logical cost-effective strategy for extensive feed analyses.
Keywords
)