Mitigating saturation effects in rice nitrogen estimation using Dualex measurements and machine learning.

Area of Science:

• Agricultural Science
• Plant Physiology
• Remote Sensing

Background:

• Nitrogen is crucial for rice, but traditional assessment methods are labor-intensive and unreliable at high levels.
• Chlorophyll-based measurements saturate, limiting accurate nitrogen status assessment in rice under high nitrogen conditions.

Purpose of the Study:

• To evaluate flavonoid content (Flav) and Nitrogen Balance Index (NBI) using a Dualex sensor for precise rice nitrogen status estimation.
• To overcome saturation limitations of chlorophyll measurements in high nitrogen environments.

Main Methods:

• Field experiments with 15 rice varieties under varied nitrogen levels.
• Dualex sensor measurements of chlorophyll (Chl), Flav, and NBI from top leaves.
• Machine learning models (random forest, XGBoost) for nitrogen concentration prediction.

Main Results:

• Chlorophyll measurements showed saturation effects at high nitrogen levels.
• Flav and NBI remained sensitive across all nitrogen levels, accurately reflecting nitrogen status.
• Machine learning models achieved high prediction accuracy (R² > 0.82) for leaf and plant nitrogen concentrations.
• SHAP analysis identified NBI and Flav from top two leaves as key predictors.

Conclusions:

• Flavonoid content and NBI measurements effectively overcome chlorophyll saturation limitations.
• Combining Flav and NBI with machine learning enables precise nitrogen estimation in rice.
• This approach offers practical solutions for improved nitrogen management in rice cultivation.