Chandrasan et al. (2026) Towards an Irrigation Decision Support System for Ragi in Karnataka: A Systematic Review of Satellite-Based Root-Zone Soil Moisture Estimation

Identification

• Journal: Current Agriculture Research Journal
• Year: 2026
• Date: 2026-09-10
• Authors: Bhavya Chandrasan, Solomon Jebaraj
• DOI: 10.12944/carj.14.2.4

Research Groups

• Dept. of CSIT, Centre for Research Excellence and Innovation, JAIN (Deemed-to-be University), Bengaluru, Karnataka, India
• Dept. of CSIT, JAIN (Deemed-to-be University), Bengaluru, Karnataka, India

Short Summary

This systematic review consolidates satellite-based and machine learning techniques for Root-Zone Soil Moisture (RZSM) estimation, evaluating their applicability for ragi irrigation in Karnataka, India. It identifies five critical research gaps that currently prevent the development of a free-access Decision Support System (DSS) for ragi farmers, despite the high accuracy (R² > 0.85) of ensemble machine learning models using multi-source satellite data.

Objective

• Consolidate remote sensing and machine learning approaches relevant for RZSM assessment.
• Evaluate their appropriateness for application on Karnataka’s laterite soils and monsoon climate conditions.
• Pinpoint important research challenges and gaps.
• Propose a research framework for an open-data-based ragi irrigation DSS.

Study Configuration

• Spatial Scale: Focus on ragi-growing districts in Karnataka, India (e.g., Hassan, Tumkur, Kolar, Mandya). Discusses data resolutions ranging from 10 meters (Sentinel-1/2) to 9 kilometers (SMAP L4, ERA5-Land), 5.5 kilometers (CHIRPS), and 250 meters (SoilGrids). Ragi rooting depth is considered 20 to 60 centimeters.
• Temporal Scale: Systematic literature review covers articles published between January 2018 and December 2025. Data products discussed have temporal resolutions such as 3-day (SMAP), 6-day (Sentinel-1), 5-day (Sentinel-2), hourly (ERA5-Land), and daily (CHIRPS).

Methodology and Data

• Models used: Systematic literature review following PRISMA guidelines. Discusses various machine learning models (Random Forest, eXtreme Gradient Boosting, LightGBM, SVM, GBM, KNN, CNN, ConvLSTM, Deep Neural Networks, Stacking), semi-empirical SAR models (Oh, Dubois, Water Cloud), physics-based models (Richards equation, Hydrus-1D), and depth extrapolation techniques (Exponential Filter).
• Data sources: Satellite (Soil Moisture Active Passive (SMAP) Level-4, Sentinel-1 C-Band Synthetic Aperture Radar (SAR), Sentinel-2 multi-spectral imagery), Reanalysis (ERA5-Land), Precipitation (Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS)), Soil properties (SoilGrids).

Main Results

• Ensemble machine learning models (Random Forest, eXtreme Gradient Boosting, LightGBM) trained on fused multi-source satellite inputs (SMAP, Sentinel-1/2, ERA5-Land, CHIRPS, SoilGrids) reliably predict soil moisture with R² values exceeding 0.85.
• Multi-resolution downscaling of SMAP soil moisture data (from 9 kilometers) using high-resolution active remote sensing (Sentinel-1) combined with depth extrapolation techniques (e.g., Exponential Filter) is a successful strategy for linking satellite-based surface soil moisture and RZSM estimates.
• Five interlinked research gaps were identified that hinder the development of an open-data RZSM-based irrigation DSS for ragi in Karnataka:
  1. Absence of a ragi-specific operational RZSM-DSS.
  2. Limitations in root zone depth estimation and unvalidated extrapolation for Karnataka’s lateritic Alfisols.
  3. Challenges in model transferability without costly in-situ calibration.
  4. Inconsistent validation frameworks across different studies.
  5. Lack of integration of irrigation forecasts into current models.

Contributions

• Provides the first systematic review (following PRISMA guidelines) specifically focused on satellite-based RZSM estimation and machine learning for ragi irrigation in Karnataka, India.
• Consolidates and evaluates the appropriateness of remote sensing and machine learning approaches for RZSM assessment in Karnataka's semi-arid agro-climatic conditions and lateritic soils.
• Identifies and articulates five critical, interlinked research gaps that currently impede the development of an operational, free-access RZSM-based Decision Support System for ragi farmers.
• Proposes a comprehensive research framework to address these identified gaps, aiming to bridge the research-to-practice divide for ragi irrigation advice using open remote sensing data.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Citation

@article{Chandrasan2026Towards,
  author = {Chandrasan, Bhavya and Jebaraj, Solomon},
  title = {Towards an Irrigation Decision Support System for Ragi in Karnataka: A Systematic Review of Satellite-Based Root-Zone Soil Moisture Estimation},
  journal = {Current Agriculture Research Journal},
  year = {2026},
  doi = {10.12944/carj.14.2.4},
  url = {https://doi.org/10.12944/carj.14.2.4}
}