Identification of rice crop phenology using Scatsat-1 Ku-band scatterometer in Punjab and Haryana

Abstract

The present study was carried out to analyse rice crop phenology using SCATSAT-1 Scatterometer data. Owing to its higher frequency, Ku-band responds to the dormant elements of surface, corresponding to top of crop canopy. The Ku-band backscatter behaviour is influenced by the structure, moisture content and the roughness of the top canopy which varies with different crop stages. Rice growth undergoes three main phenological stages viz, maximum tillering giving a scattered rough surface (first peak); maximum vegetative stage with a smooth top appearance (first minima after puddling); and the panicle development stage giving more scattering due to the spikelets (second peak). This study aimed at deriving these three stages of rice crop. Rice crop map generated, under FASAL project, using Sentinel-1 SAR data has been utilized and further the sub-pixel area within the SCATSAT-1 pixel. The daily Sigma-0 product was utilised for estimation of dates of major phenological stages of rice crop. A software tool was developed to generate the three phenological stages at 2 km pixel level using the time series backscatter coefficient in VV polarization over rice pixels using polynomial curve fitting. It was observed that peak tillering stage coincides with the first maxima, first minima coincides with the maximum vegetative stage and second maxima coincides with the peak heading or panicle formation stage of the crop. The validation was carried out using ground information from crop cutting experiments and ground truth data. Validation result showed 4 to 12 days difference in maximum vegetative stage.