Forecasting rainfed rice yield with biomass of early phenophases, peak intercepted PAR and ground based remotely sensed vegetation indices

Abstract

Rice is the main staple food of the country but crop productivity in some years declines due to erratic monsoon and non-uniformity in spatial and temporal distribution of rainfall. Hence, assessing productivity of the rice crop in advance using meteorological and plant physiological attributes will be helpful for planners to take decision on contingency measures. In this investigation, dry biomass of early phenophases (active tillering, panicle initiation, boot leaf stages, flowering),peak intercepted photosynthetically active radiation (IPAR), peak spectral reflectance based vegetation indices of 3 rice varieties under 3 nitrogen levels (50, 100 and 130 kg ha-1) were made correlated with grain yield.Based on interrelationship it was found that biomass of flowering period was better correlated with grain yield with the R2 value of 0.75.Inter-relationship between peak IPAR(%), remotely sensed peak simple ratio index (IR/R) and normalized difference vegetation index (NDVI)with the rice yield were also established. Multiple regression modelwas developed by interrelating yield as dependant variable with dry biomass of flowering stage, peak IPAR(%) and peak IR/R and NDVI as independent variables which may be used as an effective tool for early prediction of rice yield, at least 30-40 days in advance. The grain yield was also estimated through developed algorithm using MODIS satellite derived NDVI and compared with that of actual yield.