Strengthening food security through rice crop area monitoring in India using EOS-04 (RISAT-1A) SAR data

This is the authors' abstract. We don't add key points for this paper.

Timely and accurate estimation of rice crop area and production is crucial for effective crop management and strengthening food security initiatives. Unlike optical sensors, Synthetic Aperture Radar (SAR) provides a significant advantage by penetrating cloud cover and enabling uninterrupted monitoring of croplands, even under persistent cloudy conditions. This study presents a decision rule-based rice crop mapping approach using temporal HH and HV dual-polarization data acquired in Medium Resolution ScanSAR (MRS) mode from Earth Observation Satellite-04 (EOS-04), the successor to Radar Imaging Satellite-1 (RISAT-1). Distinct backscatter responses observed at different growth stages of rice were used to establish decision rules, enabling reliable classification of rice versus non-rice areas. To ensure operational scalability, an automated framework was developed that directly retrieves satellite data from archives, integrates with the Data Processing (DP) Chain at the National Remote Sensing Centre (NRSC), and applies decision rules within a streamlined workflow for large-scale mapping. The methodology was first validated across eight test sites in diverse regions of India, achieving classification accuracies between 81% and 96% depending on data continuity. It was subsequently deployed to over 340 major rice-growing districts during the Kharif 2024 season, consistently achieving accuracies above 80%. Comparison with official government statistics further confirmed its robustness, with rice area estimates showing strong agreement (R2 > 0.9) across 343 districts. These results demonstrate the reliability and scalability of the EOS-04 dual-polarized SAR-based methodology, paving the way for operational, nationwide rice crop monitoring to support agricultural planning and food security.