model

Maintaining animal health and performance relies on the availability of an appropriate diet. For herbivores, accurate assessment of forage nutrient quality is critical for appropriate diet formulation and rationing, including potential supplementation. Near-Infrared Reflectance Spectroscopy (NIRS) is a rapid method that is used in place of traditional chemical methodologies (wet chemistry) to predict analyte contents in forage samples. The method relies on scanning a sample with near-infrared light and predicting the analyte content by comparing the reflected spectra to a model which has been developed with samples of known analyte content measured by wet chemistry. The purpose of this study was to examine the accuracy of four NIRS-based methods on haylage from seven farm holdings compared with wet chemistry (the control). We analysed 64 samples for a range of analytes (dry matter (DM), pH, ash, acid detergent fibre expressed inclusive of residual ash (ADF), neutral detergent fibre assayed with a heat stable amylase and expressed inclusive of residual ash (aNDF), crude protein and water-soluble carbohydrate (WSC)) commonly assessed for haylage quality in equid nutrition. We compared results obtained by wet chemistry to corresponding NIRS-based predictions from four commercially available NIRS services. The results revealed large discrepancies amongst all five methods. For DM, average bias (mean±SD) for three reported methods was -15.5±188.4, -10.1±50.4, 12.9±33.8 g/kg respectively and for WSC reporting positive bias from four methods of 26.9±51.3, 24.8±38.2, 26.2±50.1 and 14.5±45.2, g/Kg respectively. The extent of these discrepancies from the wet chemistry also varied by analyte where for example, predictions for DM were more reliable than those for WSC and results demonstrated that predictions obtained by NIRS could result in feeding forage outside of target nutritional values.