Low-Cost NIR Spectroscopy Versus NMR Spectroscopy for Liquid Manure Characterization

Area of Science:

• Agricultural Science
• Analytical Chemistry
• Spectroscopy

Background:

• Accurate characterization of liquid manure properties (DM, TN, NH4-N, TP) is crucial for effective agricultural nutrient management.
• Traditional laboratory methods can be time-consuming and costly for on-site analysis.

Purpose of the Study:

• To investigate the potential of near-infrared spectroscopy (NIRS) for predicting liquid manure properties.
• To compare the predictive accuracy of NIRS with nuclear magnetic resonance (NMR) spectroscopy.
• To evaluate the impact of advanced spectral pre-processing and machine learning techniques on NIRS performance.

Main Methods:

• 51 liquid manure samples were analyzed for reference properties using laboratory methods.
• Samples were scanned using NIRS (941-1671 nm) and NMR spectroscopy.
• NIRS data underwent pre-processing including two- and three-band index transformations and feature selection.
• Partial least squares regression (PLSR) and LASSO regression were used to develop calibration models.

Main Results:

• NIRS demonstrated fair predictive accuracy for dry matter (DM) (R² = 0.78, RPD = 2.15) using cohort-tuned models.
• NMR spectroscopy provided superior accuracy for chemical properties: TN (R² = 0.89, RPD = 1.74), NH4-N (R² = 0.97, RPD = 5.70), and TP (R² = 0.95, RPD = 2.64).
• Two- and three-band indices significantly improved NIRS performance, increasing R² by 34% for DM, 57% for TN, 25% for NH4-N, and 33% for TP.
• Feature selection effectively reduced spectral dimensionality without compromising prediction accuracy.

Conclusions:

• NIRS shows promise as a portable tool for on-site characterization of liquid manure, particularly for dry matter.
• NMR spectroscopy remains the benchmark for accurate laboratory validation of chemical manure properties.
• Combining NIRS with advanced pre-processing techniques enhances its utility for nutrient management applications.