High-Resolution Mapping of Paddy Rice Extent and Growth Stages across Peninsular Malaysia Using a Fusion of Sentinel-1 and 2 Time Series Data in Google Earth Engine

Rice is the staple crop for more than half world’s population, but there a lack of high-resolution maps outlining rice areas and their growth stages. Most remote sensing studies map extent; however, in tropical regions, grown throughout year with variable planting dates cropping frequency. Thus, mapping stages useful only extent. This study addressed this challenge by developing phenology-based method. The hypothesis was that unsupervised classification (k-means clustering) Sentinel-1 2 time-series data could identify fields stages, because (1) presence flooding during transplanting can be identified VH backscatter; (2) changes canopy (vegetative, generative, ripening phases) up to point harvesting Normalized Difference Vegetation Index (NDVI) time series. Using proposed method, mapped field extent calendars across Peninsular Malaysia (131,598 km2) on Google Earth Engine (GEE) platform. monthly series from January 2019 December 2020 were classified using k-means clustering similar phenological patterns. approach resulted 10-meter resolution extent, intensity, calendars. Validation very street view images showed predicted had an overall accuracy 95.95%, kappa coefficient 0.92. In addition, agreed well local government’s granary data. results show method cost-effective accurately over large areas. information will helpful measuring achievement self-sufficiency production estimates methane emissions cultivation.