Hyperspectral data and vegetative indices for paddy: A case study in Kerala, India

Abstract

This paper reports the result of first attempt to study the variation in spectral signature of rice grown at different altitude and nutrient management practices using spectroradiometer in Kerala, India. Multiple experiments were conducted in Wayanad, Kozhikode, and Alappuzha districts, capturing essential spectral and agronomic data. The spectral data from Wayanad and Alappuzha plots at different growth stages revealed significant influences due to soil type and altitudinal variations. Furthermore, distinct spectral signatures emerged during different rice growth stages. The calculation of 15 indices, encompassing vegetative and water indices, yielded noteworthy variations at various physiological stages. Notably, the Leaf Area Index in Wayanad surpassed that in Alappuzha, maintaining this trend from panicle initiation to harvest stage. The spectral signatures also provided clear indications of water stress in the rice fields, with severity and mechanical processes were apparent through vegetative and water indices. Validation through conventional analysis of relative water content and membrane leakage corroborated these findings. Field experiment results confirmed the significant differences in spectral reflectance based on planting methods and nutrient management treatments. The spectral signature exhibited a robust correlation with rice yield (r = 0.94), capturing variations in management practices statistically. Regression analysis revealed a strong correlation between spectral and lab-analyzed values for different physiological and nutrient parameters. In conclusion, the developed spectral library for the rice cultivar Uma, along with the derived indices, proves instrumental in comprehending crop growth stages, identifying water or nutrient stress, and predicting the paddy yield during growing seasons, which will help for sustaining the rice production.