Jadhav et al. (2025) Quantification of Sugarcane Crop Water Footprint Using Remote Sensing and MachineLearning Techniques: Case Study of Kolhapur District, Maharashtra, India

Identification

• Journal: ISPRS annals of the photogrammetry, remote sensing and spatial information sciences
• Year: 2025
• Date: 2025-12-19
• Authors: Shrinivas Jadhav, Sahil K. Shah, Vidya Kumbhar, T. P. Singh
• DOI: 10.5194/isprs-annals-x-5-w2-2025-249-2025

Research Groups

• Symbiosis Institute of Geoinformatics (SIG), Symbiosis International (Deemed University), Pune, India.

Short Summary

This study integrates Sentinel-2 remote sensing imagery with machine learning algorithms to classify sugarcane crops and quantify their Blue and Green Water Footprints (WF) in the Kolhapur district of India. The research demonstrates that Random Forest models provide the highest precision for both crop identification and the prediction of water consumption patterns.

Objective

• To develop a machine learning-based framework for the spatial identification of sugarcane crops and the quantification of their Blue Water Footprint (BWF) and Green Water Footprint (GWF) to support sustainable irrigation management.

Study Configuration

• Spatial Scale: Regional (Kolhapur District, Maharashtra, India; approximately 7,685 $km^2$).
• Temporal Scale: 2017–2023.

Methodology and Data

• Models used:
  • Classification: Random Forest (RF), Support Vector Machines (SVM), and Logistic Regression (LR).
  • Regression/Prediction: Random Forest (RF), Support Vector Regression (SVR), Gradient Boosting Regression (GBR), and Artificial Neural Networks (ANN).
• Data sources:
  • Satellite Imagery: Sentinel-2 (10 m spatial resolution) for crop masking.
  • Evapotranspiration (ET): MODIS MOD16A2 (8-day composites).
  • Precipitation: CHIRPS gridded rainfall dataset via Google Earth Engine (GEE).
  • Temperature: MODIS MOD11A1 (Land Surface Temperature and Emissivity).
  • Ground Truth: Department of Agriculture and National Remote Sensing Centre (NRSC), India.
  • Yield Data: Directorate of Economics and Statistics, Kolhapur.

Main Results

• Crop Classification: The Random Forest (RF) model outperformed others with a classification accuracy of 99%, effectively utilizing the Near-Infrared (NIR) band to distinguish sugarcane.
• Water Footprint Quantification:
  • Green Water Footprint (GWF): Ranged from 50 to 350 $m^3/t$. Higher values were observed in southern talukas (Chandgad, Ajra) due to better rainwater retention.
  • Blue Water Footprint (BWF): Values expanded from a range of 68–365 mm in 2017 to 85–670 mm in 2022, indicating intensified irrigation requirements.
• Model Performance: RF was the most effective regressor for predicting WF with an $R^2$ of 0.92, followed by GBR ($R^2 = 0.88$). ANN also showed high reliability, while SVR performed poorly on high-variance data.
• Spatial Trends: Northern and eastern talukas (Shirol, Hatkanangale) showed lower WF values, while southern talukas exhibited higher water demands and potential water stress.

Contributions

• Provides a hybrid framework that replaces labor-intensive field surveys with automated satellite-based machine learning for crop-specific water footprinting.
• Validates the use of ensemble learning (RF and GBR) for predicting agricultural water consumption based on climatic variables (ET, precipitation, and temperature).
• Offers a scalable methodology for precision agriculture and regional water resource management in semi-arid, water-stressed environments.

Funding

• Not explicitly mentioned in the provided text.

Citation

@article{Jadhav2025Quantification,
  author = {Jadhav, Shrinivas and Shah, Sahil K. and Kumbhar, Vidya and Singh, T. P.},
  title = {Quantification of Sugarcane Crop Water Footprint Using Remote Sensing and MachineLearning Techniques: Case Study of Kolhapur District, Maharashtra, India},
  journal = {ISPRS annals of the photogrammetry, remote sensing and spatial information sciences},
  year = {2025},
  doi = {10.5194/isprs-annals-x-5-w2-2025-249-2025},
  url = {https://doi.org/10.5194/isprs-annals-x-5-w2-2025-249-2025}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)