Monitoring and Risk Prediction of Low-Temperature Stress in Strawberries through Fusion of Multisource Phenotypic...

Area of Science:

• Agricultural Science
• Plant Physiology
• Smart Agriculture Technologies

Background:

• Crop phenotyping is crucial for smart agriculture, but spatial information in imaging is underutilized.
• Understanding crop responses to environmental stressors like cold is vital for agricultural resilience.

Purpose of the Study:

• To evaluate the feasibility of monitoring crop cold stress using phenotypic spatial variability.
• To analyze the relationship between temperature-time interactions and crop physiological/phenotypic traits.

Main Methods:

• Controlled experiments on 'Toyonoka' strawberry plants under varying cooling gradients and stress durations.
• Analysis of photosynthetic physiology and phenotypic traits, focusing on mutual information.
• Development of an XGBoost model integrating Photosynthetic Physiological Potential Index (PPPI) and relative negative accumulated temperature (RNAT) to predict cold damage risk (CDRI).

Main Results:

• Highest mutual information was found between NPQ/1D-Parallel/TENT, Y(NO)/2D-Region/INEM, qP/1D-Parallel/TENT and key physiological parameters (Pmax, REC, Chl a+b).
• A novel Cold Damage Risk Index (CDRI) was calculated using PPPI and RNAT.
• The XGBoost model achieved high accuracy (R²=0.98, RMSE=0.337, accuracy=92.13%, Kappa=0.904), outperforming other methods, with qP/1D-Parallel/TENT being the most influential feature.

Conclusions:

• Phenotypic spatial variability holds significant potential for crop cold stress monitoring.
• The developed XGBoost model provides a robust tool for assessing cold damage risk in strawberries.
• This research offers a scientific foundation for advanced phenotypic information mining and agro-meteorological disaster monitoring.