The feasibility of using near infrared spectroscopy (NIRS) for rapid prediction of the composition of cattle and pig slurries was investigated. Samples with a total solids content from <1 to 15% were collected during a 3-yr period and used for calibration and validation. Test samples were sealed in plastic bags and measured in a sample cell moved vertically during measurements. Reflectance data in the range 1200 to 2400 nm were used for calibration based on partial least square regression. Dry matter (DM), N, NH4–N, and P could be determined with r2 values of 0.97, 0.94, 0.92, and 0.87, respectively. The ratios between analyte variation range standard deviation and the root mean square error of prediction (RMSEP) obtained on a calibration independent test set were 6.2 (DM), 4.3 (N), 3.8 (NH4–N), and 3.6 (P). Total C and plant-available N (PAN) could be determined by near infrared spectroscopy with r2 values of 0.94 and 0.89, but the same correlation was obtained by calculation from DM and NH4–N results, respectively. The applicability of NIRS for K, Mg, Ca, Na, S, Cu, and Zn analysis was also investigated. The r2 values were in the range 0.41 to 0.82, with the poorest results for Na, Zn, and K. The corresponding SD/RMSEP ratios were in the range 1.2 to 3.5. We concluded that the applied NIRS methodology is suitable for rapid routine analysis of DM, C, N, NH4–N, P, and PAN in both cattle and pig slurries
source; http://soil.scijournals.org/cgi/content/abstract/71/4/1398
