葉の膨圧ダイナミクスと機械学習を用いた温室点滴灌漑トマトおよびセロリの水状態診断

科学分野:

• Agricultural Engineering
• Plant Physiology
• Machine Learning

背景:

• Accurate crop water status monitoring is vital for efficient irrigation in controlled environments.; Traditional methods for assessing plant water status are often destructive or provide only intermittent data, limiting real-time management.; Developing non-invasive techniques is crucial for advancing precision agriculture.

研究 の 目的:

• To investigate the efficacy of the non-invasive leaf patch clamp pressure (LPCP) probe for evaluating the water status of drip-irrigated tomato and celery.; To characterize the LPCP probe's output parameter (Pp) and its relationship with environmental factors.; To develop predictive machine learning models for real-time water status assessment and irrigation optimization.

主な方法:

• Utilized the leaf patch clamp pressure (LPCP) probe to measure leaf turgor dynamics in tomato and celery under drip irrigation.; Analyzed diurnal patterns of the LPCP probe's output parameter (Pp) and identified distinct states corresponding to different moisture conditions.; Developed and validated machine learning models, including random forest, integrating soil water content (SWC) and environmental factors for water status prediction.

主要な成果:

• Diurnal Pp patterns exhibited two states: State I (unimodal) for mild stress and State II (troughed) for severe stress, with defined SWC thresholds for each.; Pp showed varying associations with environmental drivers (solar radiation, SWC, wind speed) depending on the stress state.; A random forest model integrating SWC and Pp substate prediction achieved high accuracy (R² = 0.995) in predicting crop water status.

結論:

• The LPCP probe offers a non-invasive method for real-time monitoring of crop water status in drip-irrigated systems.; Machine learning models, particularly random forest, can effectively predict water stress levels using Pp dynamics and SWC.; Findings support the optimization of precision irrigation strategies for greenhouse vegetables.