Application of Near Infrared Reflectance Spectroscopy to Forage Analysis and Prediction of TDN Contents

JARQ : Japan Agricultural Research Quarterly
ISSN 00213551
書誌レコードID(総合目録DB) AA0068709X

Near infrared reflectance spectroscopy (NIRS) method was used to predict the chemical composition and total digestible nutrients (TDN) content of forage. A total of 388 out of 503 samples were used to develop the calibration equation. The samples consisted of temperate grass in the form of hay (n = 126), silage (n = 120), and its hay-silage mixture (n = 60), and corn silage (n = 142). The remaining 115 forages samples including temperate grass hay (n = 58), silage (n = 39), and corn silage (n = 18) were used to validate the developed calibration equations. These samples were analyzed in duplicate for moisture, crude protein (CP) and crude fiber (CF) contents by the proximate method. Organic cellular contents (OCC), organic cell wall (OCW), low digestible fiber fraction (organic b fraction: Ob), and high digestible fiber fraction (organic a fraction: Ob) were determined by an enzymatic method, while acid detergent fiber (ADF) was determined by a detergent method. Accuracy of the predictions was assessed by the correlation coefficients (r) and standard error of prediction (SEP). Two prediction methods of the TDN by NIRS were compared. In the first method, NIRS values were extrapolated to the TDN value calculated by the following equations (indirect method): (1) equation in Adams method, (2) equation in Martin method, (3) Donker equations, (4) equations based on enzymatic analysis, and (5) equation for detergent method. In the other method, the in vivo TDN was directly calibrated (direct method). In vivo value was obtained from digestion trials carried out in sheep. In the prediction by NIRS method, the r and SEP values of hay were 0.98 and 0.96 for CP, 0.79 and 2.95 for CF, 0.92 and 2.44 for OCC, 0.91 and 3.01 for OCW, 0.81 and 4.80 for Ob, 0.74 and 2.44 for ADF, respectively. Similar values were obtained for grass silage. However, lower r and higher SEP values were observed in corn silage compared with the hay samples used in this study. These results indicated that the moisture, CP, CF, OCC, OCW, Ob and ADF contents in forage samples can be predicted by using NIRS techniques with an acceptable accuracy. For the TDN, in the case of the indirect method, an r value of over 0.7 and a standard error (Se) of 1.5-3.8 were obtained from calibration based on the equations of Adams, Martin, and enzymatic analysis in hay. However, for corn silage a very low r value, below 0.5, and Se values of 0.8-3.8 were obtained. In the direct method of TDN calibration, correlation coefficients over 0.8 and standard error of 2.3-3.4 for all the feed groups were recorded.

作成者 Masahiro AMARI Akira ABE
国立情報学研究所メタデータ主題語彙集(資源タイプ) Journal Article
開始ページ 55
終了ページ 63
言語 eng