Dubinin et al. (2026) Linking plant water status dynamics to yield and fruit cracking in citrus orchards using UAV multi-sensor data and machine learning

Identification

• Journal: Agricultural Water Management
• Year: 2026
• Date: 2026-07-17
• Authors: Vladislav Moshe Dubinin, Michael Morozov, Ohaliav Keisar, Guy Lidor, Victor Alchanatis B, Avi Sadka, Tarin Paz‐Kagan
• DOI: 10.1016/j.agwat.2026.110634

Research Groups

• French Associates Institute for Agriculture and Biotechnology of Dryland, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus
• Institute of Plant Sciences, Volcani Agricultural Research Institute
• Institute of Agricultural and Biosystems Engineering, Volcani Agricultural Research Institute

Short Summary

This study introduces a dynamic, physiology-based framework that links temporal plant water status (PWS) trajectories during key phenological stages to fruit-cracking risk at the individual-tree scale. The framework integrates UAV-derived multispectral, thermal, and LiDAR observations with field physiological measurements and machine-learning models.

Objective

• To estimate key canopy-scale PWS indicators using UAV-derived multispectral, thermal, LiDAR, and environmental variables under different irrigation applications.
• To predict yield and fruit-cracking incidence using temporally resolved PWS indicators and identify the dominant predictors and their timing during the growing season.
• To quantify the effects of irrigation application and climatic variability on the temporal evolution of PWS and associated yield and cracking responses across seasons.

Study Configuration

• Spatial Scale: Individual-tree scale, with a focus on the 'Ori' mandarin orchard in Israel.
• Temporal Scale: Three growing seasons (2023–2025), with monthly UAV flights and field measurements.

Methodology and Data

• Models used: Random Forest, Extreme Gradient Boosting, Artificial Neural Networks, Support Vector Regression.
• Data sources: UAV-derived multispectral, thermal, LiDAR data; meteorological data from the Kfar Chabad station; soil moisture data from 12 sensors installed at depths of 15, 30, 45, and 60 cm.

Main Results

• Temporally resolved PWS indicators (TG, SWP, SC, PAI) were estimated using UAV-derived multispectral, thermal, LiDAR data and machine-learning models.
• The framework successfully predicted yield and fruit-cracking incidence at the individual-tree scale, with high accuracy (R² = 0.896 for yield prediction).
• The study found that mid-season SWP and SC were strongly associated with yield formation and cracking patterns.

Contributions

• This study provides a novel framework for linking temporal PWS dynamics to fruit-cracking risk at the individual-tree scale.
• The framework integrates UAV-derived multispectral, thermal, LiDAR observations with field physiological measurements and machine-learning models.
• The study highlights the importance of temporally resolved PWS trajectories in predicting yield and fruit-cracking incidence.

Funding

• This research was funded by the French Associates Institute for Agriculture and Biotechnology of Dryland, Ben-Gurion University of the Negev.

Citation

@article{Dubinin2026Linking,
  author = {Dubinin, Vladislav Moshe and Morozov, Michael and Keisar, Ohaliav and Lidor, Guy and B, Victor Alchanatis and Sadka, Avi and Paz‐Kagan, Tarin},
  title = {Linking plant water status dynamics to yield and fruit cracking in citrus orchards using UAV multi-sensor data and machine learning},
  journal = {Agricultural Water Management},
  year = {2026},
  doi = {10.1016/j.agwat.2026.110634},
  url = {https://doi.org/10.1016/j.agwat.2026.110634}
}