Shirazi et al. (2026) Estimation of onion crop evapotranspiration and crop coefficients using weighing lysimeters and machine learning models in semi-arid region

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-04-09
• Authors: Saba Hashempour Motlagh Shirazi, Fatemeh Razzaghi, Ali Reza Sepaskhah
• DOI: 10.1038/s41598-026-43887-w

Research Groups

• Water Engineering Department, School of Agriculture, Shiraz University, Shiraz, Iran
• Drought Research Center, Shiraz University, Shiraz, Iran

Short Summary

This study measured onion crop evapotranspiration (ETc) and crop coefficients using weighing lysimeters over two years in a semi-arid region of Iran, and developed highly accurate machine learning models to predict ETc for improved irrigation management.

Objective

• To measure onion crop evapotranspiration (ETc) and crop coefficients using weighing lysimeters in a semi-arid region, and to develop and evaluate machine learning models for predicting ETc using easily accessible parameters.

Study Configuration

• Spatial Scale: Kooshkak Agricultural Research Station, Shiraz University, Iran (semi-arid region).
• Temporal Scale: Two-year field experiment.

Methodology and Data

• Models used: Artificial Neural Network (ANN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Lasso Regression (LASSO).
• Data sources:
  • Measured ETc using digital weighing lysimeters based on the water balance method.
  • Meteorological variables (temperature, relative humidity, wind speed, net radiation).
  • Crop parameters (leaf area index, plant height).

Main Results

• The total onion ETc was 447.1 mm in the first year and 432.2 mm in the second year.
• Soil evaporation constituted 36.6% and 32.8% of the total ETc in the first and second years, respectively.
• Average single crop coefficient values across both years were 0.41 (initial stage), 0.68 (mid-stage), and 0.51 (late stage).
• Average basal crop coefficient values across both years were 0.10 (initial stage), 0.51 (mid-stage), and 0.37 (late stage).
• Random Forest (RF) and Decision Tree (DT) models achieved the highest predictive accuracy (R² = 0.98, NRMSE = 0.04) for ETc estimation.
• Artificial Neural Network (ANN) and Support Vector Regression (SVR) models also performed well (ANN: R² = 0.97, NRMSE = 0.07; SVR: R² = 0.97, NRMSE = 0.08).
• Lasso Regression (LASSO) showed lower accuracy (R² = 0.85, NRMSE = 0.18) compared to other ML models.

Contributions

• Provides standardized and reliable data on onion ETc and crop coefficients for semi-arid conditions, addressing a critical data gap for precise irrigation management.
• Develops and validates highly accurate machine learning models for ETc estimation using lysimeter measurements as a robust reference.
• Offers a dependable framework for optimizing irrigation scheduling and enhancing water-use efficiency in onion cultivation under water-scarce, semi-arid conditions.

Funding

• Shiraz University (Grant no. 2GCB1M222407)
• Shiraz University Research Council
• Drought Research Center
• Center of Excellence for On-Farm Water Management

Citation

@article{Shirazi2026Estimation,
  author = {Shirazi, Saba Hashempour Motlagh and Razzaghi, Fatemeh and Sepaskhah, Ali Reza},
  title = {Estimation of onion crop evapotranspiration and crop coefficients using weighing lysimeters and machine learning models in semi-arid region},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-43887-w},
  url = {https://doi.org/10.1038/s41598-026-43887-w}
}