Karunakaran et al. (2026) Field Spectroscopy-Assisted Machine Learning Framework for Assessing Agroclimatic Effects on Paddy

Identification

• Journal: Journal of Agrometeorology
• Year: 2026
• Date: 2026-09-07
• Authors: Karthik Karunakaran, Karuppasamy Sudalaimuthu
• DOI: 10.54386/jam.v28i3.3399

Research Groups

• Department of Civil Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, Kattankulathur Campus, Chengalpattu, Tamil Nadu, India.
• Association of Agrometeorologists.

Short Summary

This study presents a machine learning-based framework for assessing agroclimatic effects on paddy crops. The framework combines field spectroradiometer-calibrated vegetation indices and agrometeorological variables to estimate stage-specific yield-limiting stress in Kharif paddy in Thanjavur Delta.

Objective

• To develop a geospatial machine learning model that can estimate the impact of agroclimatic factors on paddy crop growth and yield variability.
• To identify the key predictors influencing paddy yield at different phenological stages.

Study Configuration

• Spatial Scale: Local scale, focusing on Thanjavur Delta region in Tamil Nadu, India.
• Temporal Scale: Kharif season (June-September) of 2023.

Methodology and Data

• Models used: XGBoost, Random Forest, Multiple Linear Regression, and K-Nearest Neighbors.
• Data sources: Satellite data from Landsat 8/9, field spectroradiometer measurements, agrometeorological variables (temperature, humidity, rainfall, wind speed), and soil parameters.

Main Results

• The study found that the XGBoost model performed best in estimating stage-specific yield-limiting stress.
• The key predictors influencing paddy yield at different phenological stages were identified as humidity, rainfall, and soil moisture.
• The study also found that the reproductive stage was the most sensitive to climatic changes.

Contributions

• This study provides a novel framework for assessing agroclimatic effects on paddy crops using machine learning techniques.
• The framework can help agricultural planners, policymakers, and farmers detect crop stress conditions on time and make better decisions concerning crop management.

Funding

• Not specified in the paper.

Citation

@article{Karunakaran2026Field,
  author = {Karunakaran, Karthik and Sudalaimuthu, Karuppasamy},
  title = {Field Spectroscopy-Assisted Machine Learning Framework for Assessing Agroclimatic Effects on Paddy},
  journal = {Journal of Agrometeorology},
  year = {2026},
  doi = {10.54386/jam.v28i3.3399},
  url = {https://doi.org/10.54386/jam.v28i3.3399}
}