Comprehensive nutrient analysis in agricultural organic amendments through non-destructive assays using machine...

Area of Science:

• Analytical Chemistry
• Soil Science
• Environmental Science

Background:

• Organic amendments are crucial for soil health and nutrient management.
• Accurate and rapid characterization of organic amendments is essential for quality control.
• Existing methods for elemental analysis can be time-consuming and costly.

Purpose of the Study:

• To assess portable X-ray fluorescence (pXRF) and DRIFT-MIR spectroscopy for analyzing organic amendments.
• To develop machine learning models for quantifying nutrient elements (C, N) and other elements.
• To propose a system for quality assessment of organic amendments using these techniques.

Main Methods:

• Utilized portable X-ray fluorescence (pXRF) and DRIFT-MIR spectroscopy for elemental analysis.
• Developed and applied machine learning algorithms (forest regression, extreme gradient boosting).
• Employed cross-validation to evaluate model generalizability across instruments.

Main Results:

• Both pXRF and DRIFT-MIR spectroscopy, with machine learning, showed strong capabilities for nutrient estimation.
• pXRF proved effective for quantifying nutrients and identifying contaminants.
• The developed system enables rapid, comprehensive nutrient analysis and quality assessment.

Conclusions:

• Portable spectrometry coupled with machine learning provides a scalable and cost-effective solution for organic amendment analysis.
• This integrated approach facilitates improved nutrient management and quality control in agricultural and environmental applications.