A Remote Sensing–Based Methodology for Assessing Alternate Wetting and Drying Practices in Punjab
Subject Tags
- Data Science and Artificial Intelligence
- Land management
- Regenerative food systems
Abstract
Alternate Wetting and Drying (AWD) can reduce the water required for rice cultivation while lowering methane emissions. However, monitoring its adoption across individual fields and large agricultural areas is difficult through field visits alone. This technical note presents a remote sensing methodology developed to detect and monitor AWD practices across rice-growing areas in Punjab.
The study combines L-band Synthetic Aperture Radar satellite data with field observations and machine learning. Researchers collected 4,352 field observations during the 2024 Kharif season across Patiala, Firozpur and Sangrur, classifying rice fields as flooded, wet or dry. Five machine learning models were trained and evaluated using the combined satellite and field data.
The Random Forest model delivered the strongest overall performance, achieving approximately 73% accuracy in distinguishing among the three field conditions. The findings demonstrate the potential of satellite-based monitoring to assess AWD adoption across large agricultural landscapes. The methodology could support water conservation initiatives, climate-smart agriculture and results-based incentive programs, while further improvements in observation frequency, field validation and radar data processing could increase accuracy.
TNC Authors
-
Rajveer Singh
Senior Agronomist
Nature Conservancy India Solutions Private Limited
Email: rajveer.singh@tnc.org -
Fateh Guram
Assistant Manager – Communications
Nature Conservancy India Solutions Private Limited
Email: fateh.guram@tnc.org -
Laia Domenech
Deputy Director Conservation
Nature Conservancy India Solutions Private Limited
Email: laia.domenech@tnc.org